NANR172 Human response to vibration in residential...

www.defra.gov.uk

www.defra.gov.uk

Human response to vibration in residential environments March 2007

Department for Environment, Food and Rural Affairs Nobel House 17 Smith Square London SW1P 3JR Tel: 020 7238 6000 Website: www.defra.gov.uk © Queen's Printer and Controller of HMSO 2007 This publication is value added. If you wish to re-use this material, please apply for a Click-Use Licence for value added material at http://www.opsi.gov.uk/click-use/value-added-licence-information/index.htm.

Office of Public Sector Information Information Policy Team St Clements House 2-16 Colegate Norwich NR3 1BQ

Fax: 01603 723000 Email: [email protected]

Information about this publication and further copies are available from: Local Environment Protection Defra Ashdown House 123 Victoria Street London SW1E 6DE This document is also available on the Defra website and has been prepared by Temple Group Ltd and Arup Acoustics. Published by the Department for Environment, Food and Rural Affairs

Defra NANR172 Human Response to Vibration in Residential EnvironmentsFinal Report

M:\119573-00\OUTGOING DOCUMENTS\FINAL REPORT\FINAL ISSUE\FINAL REPORT-ISSUE2.DOC AAc/119573-00/FINAL REPORT-ISSUE 2

TRL Temple ISVR Arup AcousticsIssue2 21 March 2007

Contents

Page Executive Summary i 1 Introduction 1

1.1 Background 1 1.2 Objectives 2 1.3 Definitions 3 1.4 Report Structure 3

2 Project Brief 4 2.1 Aims 4 2.2 Tasks and Objectives 4

3 Development of the Questionnaire, Measurement Protocol and Pilot Study. 5 3.1 Development Of The Measurement Protocol 5 3.2 Development of the Questionnaire 16 3.3 Development of the Pilot Study 19

4 The Pilot Study 27 4.1 Objectives (success criteria) 27 4.2 Questionnaire Summary 28 4.3 Measurement Protocol 31 4.4 Undertaking the Pilot Study 32 4.5 Analysis and Results 49

5 Recommendations 67 5.1 Social Survey 67 5.2 Measurement Protocol 67 5.3 Analysis Methodology 69

6 Conclusions 70 6.1 Measurement Location Selection 72 6.2 Questionnaire 72 6.3 Measurement Protocol 72 6.4 Combined Analysis 72

References

Appendices Appendix A Literature review Appendix B Equipment Appendix C




DATS Software output Appendix D Questionnaire Appendix E Letter to residents Appendix F Draft measurement protocol Appendix G Measurement location details Appendix H Measurement proforma Appendix I DATS Software Appendix J Analysis of vibration data



Page i TRL Temple ISVR Arup AcousticsIssue2 21 March 2007

Executive Summary This report describes research undertaken to develop a method by which human exposure to vibration in residential environments can be assessed. The work was carried out on behalf of the Government’s Department for Environment, Food and Rural Affairs (Defra) by two contractors working in close collaboration: Arup Acoustics supported by the Transport Research Laboratory and Temple Group Limited supported by the Institute of Sound and Vibration Research. The principal aims of the project were to:

• develop a measurement protocol that would enable the exposure to vibration to be quantified;

• develop a questionnaire that would solicit residents’ reaction to vibration within their homes caused by sources outside their property;

• conduct a pilot study to test and in so far as is possible, validate the questionnaire and measurement protocol; and hence

• provide advice on how best to progress towards conducting a vibration exposure-response study

The results of the project provide Defra with a methodology for undertaking a subsequent much wider study that would yield a robust relationship between exposure and response. This should provide a reliable basis for developing standards and guidance for the assessment of vibration in residential buildings.

Scope

The vibration sources considered are those affecting residents which are outside their control, e.g. road, rail, industry, construction and same building sources (e.g. neighbouring gym, heating system, washing machine, etc but not domestic sources within the same residence).

Measurement Protocol

A measurement protocol has been developed that manages the practical problems associated with both internal and external measurements. The protocol will also enable the follow on community dose response study to determine whether evaluative criteria can be defined (for external sources) based on an external assessment location. This would make the assessment of environmental noise and vibration more consistent which may be an advantage for Defra’s policy development.

The protocol is based on ‘off the shelf’ measurement equipment and analysis software.

Questionnaire

The social survey questionnaire was developed to provide a survey methodology and questionnaire that would yield robust information with maximum efficiency based on best practice in social survey and questionnaire design and interviews undertaken face to face with adults.

Pilot Study

The pilot study trialled the measurement protocol and the questionnaire and has generated over 100 case studies (questionnaire and measurement pairings) from three measurement locations adjacent to the busy East Coast Main Line (railway) north of London and at a construction site in south east England.

The pilot study was designed to test the measurement protocol and questionnaire, identify any limitations with either and establish how they might be optimised for a subsequent exposure-response study.

The data from the pilot study has been analysed to determine whether any ordinal relationship between vibration exposure and comment could be established, as a means of validating whether the larger study is likely to be able to identify and hence define a community dose-response relationship for vibration.



Page ii TRL Temple ISVR Arup AcousticsIssue2 21 March 2007

Principal Conclusions

The pilot study demonstrated that the measurement protocol, questionnaire and post processing / analysis methodologies were successful. Nevertheless, a number of areas have been identified where the approaches should be modified or investigated further. However, the scale of the modifications and investigations is such that they could be accommodated in the early stages of the exposure-response study.

In more detail, the principal conclusions were as follows.

Measurement location selection

Maps and aerial photographs available from the internet were valuable in undertaking a first sift for possible sites. Visits to site were necessary to refine the search before proceeding with the surveys.

A sifting process identified sites where a range of different levels of vibration was anticipated; where there were sufficient dwellings at different distances from the vibration source (especially close to the source); and where vibration from only one source could be measured without contamination from other sources.

Ideal measurement locations for carrying out the pilot study were not as numerous as might be supposed. This would be expected to be the case also for the future exposure-response study.

Measurement protocol

The proportion of occupants who were willing to allow access for internal measurements (average permission rate of 44%) was higher than expected which has benefits for data gathering and is particularly encouraging for the application of the approaches to internal sources of vibration. However, some people who initially responded positively were not able to provide access at times convenient for an economic survey programme.

The measurement protocol was found to be practicable in its application and data were reliably recorded for subsequent post processing.

The post processing software also proved reliable with event and period vibration indicators being successfully extracted (from long records) for all measurement positions at all locations. Valuable lessons have been learned that can be copied across the future full study.

Recommendations are given in the report for improvements in a number of other areas that should be included in the early stages of a full community dose response study.

Questionnaire

Responses indicated that the questionnaire enabled separate ratings for primary and secondary effects of vibration to be obtained. Thus, the results showed that respondents understood the distinction between terminology such as vibration and noise as used in the questionnaire.

Data obtained from questions on the neighbourhood and on annoyance from noise and vibration was suitable for the required factor analysis.

Significant correlations between annoyance from primary and secondary vibration effects and between annoyance from noise and vibration from related exposures indicated response consistency and reliability.

Analysis showed that the railway and construction site questionnaires were successful in extracting differences in responses to different vibration and noise exposures from a range of sources in close proximity to one or more of the sources of building vibration.

Combined analysis

For data from the railway sites, the presence of an ordinal relationship between vibration magnitude and annoyance gives a positive indication that the application of the pilot study methodologies in a



Page iii TRL Temple ISVR Arup AcousticsIssue2 21 March 2007

larger study may result in the required distribution of responses for an investigation into dose-response relationships. Significant correlations between annoyance and three measures of vibration magnitude provide some confidence that residents’ annoyance generally increased with increasing vibration magnitude.

The ordinal relationship was most significant between primary vibration measures and annoyance from the primary effect of feeling train vibration. The correlation coefficient was lower and less significant for correlations involving the secondary effects of annoyance from hearing and seeing vibration than the primary effect of feeling vibration. This finding suggests that the questionnaire extracted differences in response to primary and secondary vibration effects.

Highly significant correlations between annoyance ratings for feeling vibration from trains and both internal and external vibration measures indicate that both the internal and the external vibration measures may provide the ordinal relationship between vibration magnitude and annoyance required for the determination of a dose-response relationship in an extensive study.

The analysis demonstrates the success of the questionnaire and measurement protocol in generating the information required and also provides confidence that a full national study should identify a dose response relationship, for vibration at least, that could be used for future standard and policy development.

Recommendations

• The work undertaken in this project provides a tested and validated methodology for undertaking a national study of the community dose-response relationship to vibration in residential environments that could now be taken forward based on the findings of this work.

• A number of improvements to the measurement protocol and the questionnaire are set out in the report. These are of a scale that will enable them to be undertaken as an early stage of the national study.



Page 1 TRL Temple ISVR Arup AcousticsIssue2 21 March 2007

1 Introduction This report presents the results of a research study into the human response to vibration in residential environments undertaken on behalf of Defra (the Government’s Department for Environment, Food and Rural Affairs) by two contractors working in close collaboration: Arup Acoustics supported by the Transport Research Laboratory and Temple Group Limited supported by the Institute of Sound and Vibration Research. The principal aims of the project were to:

• develop a measurement protocol that would enable the exposure to vibration to be quantified;

• develop a questionnaire that would solicit residents’ reaction to vibration within their homes caused by sources outside their property;

• conduct a pilot study to test the questionnaire and measurement protocol; and hence

• provide advice on how best to progress towards conducting a vibration exposure-response study

The report is presented chronologically, recording how the project evolved through discussion between the project contractors, Defra and Defra’s Project Board, a team of recognized industry experts acting as advisors to Defra. Defra and the Project Board are referred to hereinafter simply as Defra. The original client brief is provided, followed by sections setting out the Definitive Specification document, Interim Report and how the project was trialed through a pilot study. Section 1.4 sets out the structure of the report in more detail.

1.1 Background

Vibration is experienced by many fewer people than noise. However where significant vibration occurs, it (or its secondary effects such as groundborne/structureborne noise, rattling fixtures and fittings) can be:

• a material planning consideration (either for new receptors e.g. housing; or new sources e.g. new railways);

• a cause of nuisance (or disturbance, or complaint); and/or

• a cause of health effects (e.g. sleep disturbance).

Current policy is based predominantly around British Standard BS 6472:1992 “Assessment of Human Exposure to Vibration in Buildings”. There are a number of challenges associated with using this standard which are currently being addressed by the British Standards Institution. There is also a lack of confidence in some sectors of industry in the use of the vibration indicator used in the standard and the associated rating values. This concern relates to the laboratory origin of the indicator and rating values and the need to consider better the secondary effects of vibration.

These challenges and difficulties affect policy and standard development in this field and also affect the consistent application of current policy and standards. This project aimed to develop robust survey tools that are an important precursor to carrying out a wide ranging study to establish how people in residential environments respond to vibration from external sources.

The vibration sources considered are those affecting residents which are outside their control, e.g. road, rail, industrial, construction. A measurement protocol and questionnaire have been developed and trialled through a pilot study of 100 case studies. The pilot study




aimed to test the proposed methods, confirm the quality of the data derived and validate them as far as possible within the constraints of a relatively small sample size.

The value of the project is the development of tools that could be applied in a future project, encompassing a much larger study area (in terms of geography, demography, etc), to evaluate the community exposure-response relationships for vibration emitted from different sources. This information would aim to inform future policy and standard development in this area without undue constraint on, or cost to, industry.

1.2 Objectives

The principal objectives of the project were to: • Develop a robust and efficient measurement protocol;

• Develop a social survey methodology and questionnaire that will yield reliable information based on best practice and using face to face interviews;

• Develop and undertake a pilot study to validate the proposed approaches; and

• Acquire data from the pilot study enabling investigation at an ordinal level of the relationship between vibration exposure and response in order to validate the proposed approaches.

Specific sub-objectives for the measurement protocol were as follows:

• To specify a method to measure vibration levels within a specified level of precision and be capable of capturing vibration events within the lower quartile of human perceptibility.

• To measure vibration simultaneously in the z, x and y axes in the time domain (i.e. as ‘raw’ time histories) so that all of the parameters and frequency weightings of interest could, if required, be analysed, at a later date, to a sufficient degree of accuracy;

• To measure vibration at position(s) which are representative of the points of entry to the body inside dwellings;

• To be repeatable and reproducible;

• To be representative of all the temporal variations associated with each source included in the study;

• To be cost effective and efficient;

• To allow efficient post processing of the recorded vibration signal such that the pilot study is not artificially confined to any single vibration metric at this stage;

• To measure relevant noise indices simultaneously so that the extent to which different elements may be a confounding factor can be properly determined;

• To be able to measure vibration inside occupied dwellings with a minimum level of intrusion, but in such a way that extraneous events e.g. footfalls, can be filtered out;

• To be capable of measuring vibration from a range of sources of vibration and over a range of vibration amplitudes including the lower quartile of perceptibility;

• To specify the sensitivity, performance characteristics and precision of the measurement and analysis equipment; and

• To specify the calibration of equipment traceable to standards and guidance.

The specific objective for the social survey questionnaire would be to provide an instrument that is psychometrically valid and reliable and which provides a sufficient measurement of human response to vibration. The questionnaire was also required to assess the impact of




confounding factors, such as secondary effects of vibration, e.g. rattling of crockery, groundborne noise, etc.

1.3 Definitions

The following terminology is used in this document: • Case study: a pairing of a questionnaire and associated measurements at one

property.

• Measurement location: a geographical area around a vibration source within which a number of case studies will be undertaken.

• Measurement position: a position either within or external to a building at which vibration (and in some cases) noise is measured.

1.4 Report Structure

This report is structured as follows.

Section 2 provides the original project brief provided by Defra.

Section 3 describes the development of the measurement protocol, questionnaire and the pilot study methodology. Initially this was through the Definitive Specification document that defined the way in which the project would be delivered.

An Interim Report was then developed. This document provided the outcome of the literature review, described the rationale behind and development of the measurement protocol and questionnaire and provided the questionnaire and measurement protocol to be implemented in the pilot study. The pilot study methodology and success criteria were defined.

Both documents were initially developed by the project contractors and sub-contractors and then subjected to review by Defra. Review meetings were then held with Defra to finalise the methodologies before the final versions of the documents were submitted and approved.

Section 4 describes how the pilot study was actually undertaken, how the data was analysed and the conclusions reached.

Section 5 provides recommendations on how the measurement protocol, the questionnaire and the data analysis should be applied in an exposure-response study. This includes areas that have been identified through the pilot study as needing additional development.

Section 6 presents the conclusions reached from the project.




2 Project Brief This section sets out the brief for the project as issued by Defra at the time of tender, which was the scope of works undertaken in the study.

2.1 Aims

The aims of this project are to produce a robust methodology for acquiring the vibration data and social study information to be able to determine a dose-response relationship.

Unweighted vibration data in the x, y and z-axis should be collated in the form of a database, from which a further investigation for an index that relates vibration to levels of annoyance could be undertaken.

2.2 Tasks and Objectives

The initial contract will involve the development of a suitable measurement protocol, social study questionnaire, and a pilot study to trial them on a selected, adequate number of case studies to validate the protocol and questionnaire.

The department is keen to encourage the most efficient and robust methodology, with innovative ideas as appropriate. Consideration should be given in the proposal to the following issues; including justifications and clarifications as appropriate (this list is not exhaustive):

1. Sources of Vibration Vibration sources to be considered are those impacting residents outside their control, e.g. road, rail (both above and below ground level), industrial, construction, and plant items affecting homes such as external plant, boilers, etc (but in separate locations for source and receiver).

2. Measurement Protocol techniques Consider how and where measurements will be taken, and the practical problems associated with both internal and external measurements.

3. Pilot Study Delivery An understanding of the inter-relationship between the measurements and the social study should be demonstrated.

4. Social Survey An understanding of best practice in social survey and questionnaire design should be demonstrated, along with consideration of the most appropriate approach for the pilot study.

5. Pilot Study Variables Whilst the investigation for the pilot study should be driven by the case studies identified, please state in the proposal the variables to be covered, such as location, time of day, vibration source, propagation paths, numbers of complainants, age, gender, and any other factors that you consider to be relevant.

6. Number of Cases In order to facilitate equality in the bids, please base the pilot studies on 100 completed cases of questionnaire and accompanying measurements, covering as wide a range of samples as possible. Interviews will be undertaken face to face on adults.

A Final Report was required outlining the results of the pilot study, the final measurement protocol and social study questionnaire on which to base further work, and incorporating Defra’s suggestions.




3 Development of the Questionnaire, Measurement Protocol and Pilot Study. This Section describes the rationale behind and the evolution of the measurement protocol, the questionnaire and pilot study through initial development by the Project Partners, review and comment by Defra, followed by further development and a second review. The Project Partners’ initial proposals were submitted in the Definitive Specification document. The final iteration was described in the Interim Report.

3.1 Development Of The Measurement Protocol

This section describes the evolution of the methods and choices of equipment that were used during the pilot study. The objectives of the project were set out in Section 2.

3.1.1 Literature Review A literature review was undertaken to advise the development of the measurement protocol. This drew on publications and experience from investigations undertaken by the Project Partners as well as studies reported elsewhere. The principle published works considered are summarised in Appendix A.

3.1.2 Considerations in Development of Measurement Protocol In undertaking the pilot study it would be necessary to determine the vibration exposure, either directly or indirectly, at the property of each successful interview. The acquisition of suitable data for evaluation in the pilot study gave rise to a number of issues, the principal issues being identified as:

• Respondent – source separation: evidence available suggests that there can be a community response to vibration inside residential environments arising from external sources at levels of exposure that are significantly lower than the exposure to other ambient vibration arising from domestic activity within the respondent’s own dwelling environment. This makes it impossible to base a measurement protocol on any simple quantification of the overall ambient vibration exposure inside dwellings.

• Need to minimise measurement uncertainty: The questionnaire and measurement protocol were developed to enable (in the future) the community dose response relationship to vibration to be evaluated. Evaluating dose-response relationships is always challenging because there are so many sources of statistical variability (not least of which is variability in human response to a single level of vibration exposure). Thus it is essential that the dose-response evaluation methodology minimises measurement uncertainty. To do otherwise runs the very significant risk that the combined variability and measurement uncertainty will render it impossible to identify any underlying community dose-response relationship. Reducing measurement uncertainty has two main elements (albeit that both present major challenges):

(i) In positional terms, measure respondent exposure as closely as possible to the point of entry; and

(ii) In vibration event terms, be able to quantify, as accurately as possible, the contribution made by the events of interest (i.e. external source such as railway, construction activity etc).

• The need to maximise case-study (questionnaire and vibration measurement pairing) sample size: the justification for this is fundamentally the same as above – the greater the sample size the greater the ability to evaluate any underlying dose-response relationship.

• Need for pilot study to measure inside and outside properties: the Project Partners suggested that there would be significant practical advantages in applying the




measurement protocol in the future if an approach could be developed and verified based only on measurements made outside properties. A key issue that therefore needed to be resolved was whether it would be sufficient to measure vibration outside properties without taking measurements indoors. However, to rely only on external measurements during the pilot stage would run a number of potentially significant risks:

(i) Current standards (especially BS 6472) are predicated on measurements on the floors of buildings as the closest practical means of quantifying vibration at the point of entry to the respondent. Thus the conclusions of the pilot study would be vulnerable to challenge unless it could be clearly demonstrated that there is no significant difference between response to vibration exposure inside properties calculated from scaled (as necessary) external measurements and measurements directly inside the properties, at locations currently accepted as providing a reasonable basis for quantifying exposure in residential environments.

(ii) It is possible that variability in ground to floor transfer functions is so large that it would not be possible to evaluate the vibration exposure of the subject on the basis of external measurements. Therefore, capturing as much data as possible inside properties would enable assessment of the advantages, or otherwise, of carrying out internal measurements.

Capturing data indoors would also enable the protocol to be based on and validated against internal measurements if, in the event, the pilot study were to show that it is necessary to measure vibration indoors. Furthermore, simultaneously made measurements inside and outside properties would provide another data source for evaluating building transfer functions. Applied to the future larger scale study, this data set would provide very valuable information.

Another clear disadvantage associated with an approach based only on external measurements was that the methodology would not be applicable to environmental vibration generated by sources inside buildings (but not within the domicile of the respondent). Whilst this type of environmental vibration is far less widespread than external sources of vibration, it is a type of vibration that would fall within the scope of the follow on larger study and is an issue for local authorities from time to time and hence for Defra.

It was decided therefore that the pilot study would aim to test the validity of using external measurements. Internal data would be gathered as part of the measurement protocol (pilot study and future full study), and this information would be gathered simultaneously (phase locked where possible) with vibration data from outside each building, in order to maximise the quantity and depth of data acquired and hence to test the suitability of dose quantification based on external data.

• Tension between the need to minimise measurement uncertainty and maximising sample size to increase the statistical validity of the study. These tensions arise principally from the following:

(i) With current commercially available equipment, minimising measurement uncertainty inside a property requires an ‘invasive’ approach (a separate measurement visit with either complex equipment, possibly cabled outside the property, or attended monitoring) which would significantly reduce the number of internal measurements permitted by questionnaire respondents; and

(ii) The ability of current commercially available equipment / systems operating unattended to evaluate reliably vibration indicators for events of interest (e.g. railway vibration occurring indoors distinct from the other, probably greater amplitude, ambient vibration events) without recourse to labour intensive post




processing that will be impracticable in programme and cost terms for the eventual full dose response survey.

3.1.3 Approach to Development of the Measurement Protocol Two fundamentally different approaches to delivery of the project could have been used. These are summarised in Figure 3.1 and set out below.

FIGURE 3.1: Summary of approaches to delivery of the measurement protocol

Approach 1: use equipment currently available ‘off the shelf’ (either available in-house to the Project Partners or newly procured) and adapt the measurement protocol to fit. This would mean that some aspects of the protocol would be compromised, although these would be managed to yield the best result possible.

Approach 2: develop a measurement system (using off the shelf components where possible) that is as close as is practicable to the specification of the idealised monitoring equipment described in Appendix B and is also economically viable. This would provide the only means of achieving both minimised measurement uncertainty and maximised sample size in a reasonably economic way (there would also be significant legacy benefits). Critically, the equipment inside properties would be left by the questionnaire surveyors, as this would provide the most effective means of ensuring uptake and hence maximise sample size.

Within Approach 1, two broad methodologies could be adopted, which dictate the nature and quantity of the data that would be acquired:

Approach 1a: Minimise measurement uncertainty: Acquire resolved, short-term data either by attended monitoring or through cabled multi-channel data acquisition triggered by external source ’event’ information; or

Approach 1b: Maximise measurement sample size: This approach would be based on small, self contained data loggers or recorders acquiring data over 24hr periods at all positions or alternatively for internal measurements combined with multi-channel data acquisition used externally. Critically the equipment inside properties would be left by the questionnaire surveyors, as this will the most effective means of ensuring uptake and hence sample size.

Approach 1a would be expected to yield a smaller volume of data than 1b, since measurements are short-term and the approach is significantly more intrusive to the householder and hence take up would be expected to be lower. Also measurement positions would be constrained due to cable runs. However, while the volume of data may be smaller, the degree of uncertainty in the evaluated dose would be minimised. Approach 1b would be much less intrusive, more likely to be acceptable to householders (yielding a greater volume of data) and would not constrain measurement positions.

Ideal equipment

Optimum measurement protocol

Time, budget, practicability and delivery uncertain

Available equipment

Measurement protocol to be optimised within the constraints of the equipment

Time, budget, practicability and delivery more predictable

Approach 1

Approach 2

Approach 1a

Approach 1b

Lower volume, higher quality data

Higher volume, more uncertain data

Optimised data setIdeal equipment


Time, budget, practicability and delivery uncertainTime, budget, practicability and delivery uncertain

Available equipment


Time, budget, practicability and delivery more predictableTime, budget, practicability and delivery more predictable

Approach 1

Approach 2

Approach 1aApproach 1a

Approach 1bApproach 1b

Lower volume, higher quality data

Higher volume, more uncertain data

Optimised data set




However, using economic automatic event identification (either in the field or post processing) would result in a high level of uncertainty associated with the measured dose (e.g. due to uncertainty over the raw data analysed - source / cause of vibration events). Whilst this could be partially resolved by attended monitoring or post-processing, it could never be totally resolved and such labour intensive solutions would be expensive when applied to a full community dose response relationship survey.

It was therefore the view of the Project Partners that Approach 1a would provide a statistically more robust basis for evaluating the community dose response relationship to vibration in residential environments, albeit that the approach needed to be extended such that more data would be acquired (e.g. over a 24hr period) at one measurement position, at least, in a manner that would enable the greater event information to be scaled to the other short term measurement positions.

At the beginning of the study, a significant amount of work was carried out to explore the possibility of developing equipment that would ideally meet the requirements of the dose-response study. Appendix B provides a description of this work. The aspiration under Approach 2 (see Figure 3.1) was to identify (or develop) a noise and vibration monitoring device that would be self-contained and sufficiently small that:

• it could be left in peoples’ homes without being unduly intrusive;

• it could record all the required ‘raw’ noise and vibration data over a 24 hour period;

• it could be remotely triggered by a device near to the vibration source of interest;

• it would analyse the required noise and vibration indicators for the ‘triggered’ events; and

• it would be simple to use and therefore could be installed by the questionnaire survey team to maximise take up by interviewees.

However, Approach 2 was effectively discounted following discussions with Defra. In essence, it may have been practical to have developed ideal equipment for fairly steady and repetitive sources of vibration (e.g. railway vibration) but significant development work would be necessary to develop ideal equipment that could be applied to all sources of vibration of interest to the study. The development of the measurement protocol was therefore based up on Approach 1a. Nevertheless, the outline specification for an ‘ideal vibration meter’ provided a useful reference point for the development of the actual measurement protocol. The outline specification also provides a useful starting point for further development of the ‘ideal meter’ in the future.

3.1.4 Methodological Issues Associated with Measuring Source Specific Vibration

It was essential that the measurement protocol could separate source specific vibration from “false triggers” from internal sources (due to, for example, footfalls, door slams, etc) and that records can be correlated with events due to the vibration source being studied. False triggering was a particular concern for unattended measurement options.

One of the most significant challenges to the project has been in identifying equipment that will achieve source separation and perform all the necessary data acquisition functions. Several possible options were explored and discounted.

The use of remote triggering devices to trigger recording equipment set up inside properties was investigated. Remote triggering would ensure that only events from the source of interest were recorded and spurious events from other (internal and external) sources would be largely avoided. However, this would not obviate the risk of other sources affecting the measurement during the triggered recording period. If raw time history data were recorded, any such events could be identified by visual inspection of the traces and discounted. An




alternative approach, which could be applied either during data acquisition or through post processing, would be the use of event recognition technology to remove spurious events.

Remote triggering was not progressed as a methodology because currently there are no ‘off-the-shelf’ systems available. Cabled solutions are too restrictive on site and whilst wireless technologies exist none have been developed or proven for this type of use.

It was found that, using equipment off-the-shelf, it should be possible to determine source specific vibration levels using unattended measurements using software routines to recognise discreet events from continuous recordings of the time-history. This approach has the added merit, for linear sources like railways, of managing the time delay between event measurements at different measurement positions (that will be different positions along the length of the line source).

It was acknowledged that applying such techniques to unattended measurement of specific vibration levels from sources with random time-histories and with highly variable amplitudes, such as that generated by construction activities, would be more problematic.

Internal source specific vibration levels could be determined for random and variable sources using unattended measurements, but this would require an array of transducers with complex, phase locked, trigger functions. This was considered to be impractical within the context of this study.

3.1.5 Evolution of the Measurement Protocol Attended monitoring provides a reliable means of achieving source separation and overcomes the limitation of using unattended measurements described above. However, attended monitoring is resource intensive and intrusive when carried out indoors. A number of options were explored in an attempt to develop an optimal solution which struck an appropriate balance between the various, and often competing, factors such as:

• cost,

• efficiency,

• accuracy,

• acceptability (privacy and inconvenience to the respondent); and

• sufficiency.

The Project Partners spent considerable time evaluating the strengths and weaknesses of different options, which attempted to strike this balance.

The preferred methodology was developed analogous to the approach used for many environmental noise dose-response studies. The principal components of the approach can be summarised as:

• continuous unattended monitoring at one position (probably close to the vibration source of interest);

• short term attended monitoring outside of the majority of properties where questionnaires have been completed (interpolated between measurement positions to provide data at remaining properties);

• short term attended monitoring inside as many properties as possible where questionnaires have been completed (data to be phase locked with data acquired simultaneously outside the same property).

Environmental vibration is rarely experienced in the absence of noise and earlier studies have demonstrated that response to vibration can be significantly affected by the presence of noise and other stimuli. It was decided therefore that the protocol must address the possible combined effects of noise and vibration, both in terms of noise associated with the




vibration (e.g. groundborne noise, airborne noise directly from the vibration source, rattling of items within the property) and the effects of other (separate) noise sources on the perception of vibration.

Internal noise (groundborne or secondary effects such as rattling) would be measured at each study location, as would overall external ambient and background noise levels. In addition, a pro forma would be developed for recording observations by the field measurement and questionnaire survey teams in a systematic way to supplement the measurement data. These would provide additional information in support of the social survey analysis.

3.1.5.1 Key Tests for the Measurement Protocol The decision was taken to aim to base the measurement protocol on external measurements because of:

• The limitations imposed by equipment currently available ‘off the shelf’,

• The difficulties in gaining access to measure vibration inside properties, and

• The potential for internal measurements to be affected by domestic activity and other non-source specific vibration.

However, it is widely known that vibration levels can vary significantly between buildings even when they are relatively close and of similar construction. Consequently, there is a risk that the use of external measurements may not be adequate or representative, which dictated that it was important for the pilot study and future main study to test:

• Is it necessary to measure vibration inside dwellings?

• If it is necessary to measure inside homes, is it sufficient to measure vibration on a structural surface at a single measurement location?

The following other objectives were therefore identified for the measurement protocol:

• Measure vibration inside and outside as many properties at possible; and

• Demonstrate whether a single defined measurement location inside each property is appropriate.

3.1.5.2 Measurement Methodology for Pilot Study The methodology adopted for the pilot study was developed on the basis that it would not be economic to base the overall approach on having to gather internal measurements at every property where a questionnaire was completed. The methodology is therefore intended to allow correlation either directly between internally measured vibration and response (where data is available – i.e. a subset of all questionnaire locations) or between externally measured vibration and response. Correlation with external vibration would be via a transfer function, to account for building response.

The general approach adopted for the pilot study was therefore:

• Identify the required sample size (defined by the project brief as 100 case studies);

• Set up a questionnaire implementation strategy focused on obtaining completed questionnaires to meet the required sample size (including a representative sample of the community);

• Measure vibration directly outside respondents’ properties at a sufficient number of locations at a standardised position outside each property. These would then be converted to estimates of internal vibration levels using scaling factors / transfer functions;




• Measure vibration inside at points corresponding to defined external measurement locations where respondents agreed to permit access. Where possible, internal measurements would be taken simultaneously and phase locked with those outside the property; and

• Estimate the vibration outside and inside any remaining respondent’s properties based on interpolation / extrapolation of other measurement data (converted to internal vibration exposure using defined scaling factors / transfer functions as necessary).

It was considered that the determination of transfer functions should be treated as an ongoing process undertaken throughout the pilot study and follow-on main study at all properties where it is possible to measure vibration both inside and outside the building.

The transfer functions would be used as necessary to calculate the internal vibration levels on the basis of external vibration measurements at properties from which a questionnaire response was obtained but the respondent would not allow internal vibration monitoring. Where it was not possible to gather relevant transfer functions on site, transfer functions gathered during the literature review were to be used.

During meetings with Defra held during the development process, it was noted that measurement of vibration externally is only appropriate to sources of vibration located outside properties; vibration assessment methods may also be required for sources within buildings (i.e. within the same structure but not the same domicile as a resident). Hence it was agreed that internal data should be gathered as part of the measurement protocol, and that, where possible, this information would be gathered simultaneously (phase locked) with the vibration outside each building. This maximises the quantity and depth of data acquired and hence enables testing of the suitability of dose quantification based on external data.

The measurement approach that was adopted for trial in the pilot study is detailed below and is shown schematically in Figure 3.2.

Figure 3.2: Schematic of measurement protocol

• Source of vibration

C+ 24 hr control monitoring position

x Interpolated data

x Short term attended monitoring

Questionnaire response

x

x x

x

C+x

x

x

• Source of vibration

C+ 24 hr control monitoring position

x Interpolated data

x Short term attended monitoring

Questionnaire response

x

x x

x

C+x

x

x




As noted earlier, the principal components of the approach can be summarised as:

• continuous unattended monitoring at one position (close to the vibration source of interest) – the control monitoring position;

• short term attended monitoring outside the majority of properties where questionnaires were completed (interpolate between measurement positions to provide data at remaining properties); and

• short term attended monitoring inside as many properties as possible where questionnaires have been completed (data to be phase locked with data acquires simultaneously outside the same property).

The specification and more detailed nature of the measurements obtained at each type of situation are described in the following subsections. Vibration Measurements The measurement surveys would follow as soon as practicable after the questionnaire had been conducted. This was particularly important for the construction site (or other transient situations), where the vibration may vary significantly from one day to the next. The speed with which the measurements followed on would be based upon feedback from the questionnaire team indicating how transient or variable was the vibration exposure.

The survey period would aim to cover specific times of the day / night where feedback from the questionnaire team noted the most significant vibration reaction from the respondents.

All measurements were to be conducted by vibration and noise specialists, thereby ensuring robust data of a high quality was acquired.

A ‘control’ measurement position would be used to record continuously raw vibration time history data close to the vibration source in order to aid identification of events of interest and to enable quantification of the overall vibration exposure. The control position would be located as close as practicable to the vibration source.

At properties from which a questionnaire response was received, ‘snapshot’ attended noise (see next subsection) and vibration measurements (synchronised with the ‘control’ by aligning internal clocks) would be made outside and at a standardised distance of 2m from the closest façade to the vibration source, where practicable.

Transducer mountings appropriate to the surface would be used, taking into consideration guidance on mounting strategies given by the Association of Noise Consultants (2001).

If a large number of responses were received from a particular measurement location (i.e. more properties than it was practicable to measure at within the available timeframe), then the measurement positions would be chosen at distances appropriate for determination of a ‘site law’, enabling interpolation of data to other properties.

Where responses were received from adjoining properties or from properties in close proximity to each other, which were of similar construction, one measurement position was considered representative of the group. This would be determined from experience and available guidance, and would depend on the distance from the vibration source, the distance between the properties, and the orientation of the properties in relation to the source, etc.

Total exposure to vibration at each property would be subsequently calculated from the ‘control’ data, scaled as required to each ‘questionnaire property’ by a correction factor (or factors) based on the ratio of levels during the snapshot recording and the levels recorded simultaneously at the control position. The external snapshot data were also to be scaled (such that it is an estimation of the vibration inside the property) using an appropriate building response function determined from either:




• the simultaneous (phase locked, where possible) internal and external measurements made at the ‘questionnaire property’ or the equivalent data derived from the another property at the location of a similar construction, or;

• transfer functions available from other sources (e.g. public domain literature).

Where the building response transfer functions used in estimating internal vibration were not those derived from the survey work at the same location, this was to be clearly identified in the results database, enabling subsequent elimination from analysis, should any consistent variation from other data be identified.

For vibration measurements inside properties, the following additional issues were considered:

• All internal measurements would be attended, ensuring that the recorded data were due to sources external to the domicile and not due to internal activities.

• A standardised measurement position was required, as close as practicable to the centre of the room in which the respondent considered the vibration to be most significant (worst case).

• Measurements would be acquired simultaneously inside and outside the property to facilitate consequent analysis of transfer functions for each building.

• As far as practicable, measurements would be made over a period of time sufficient to achieve a representative sample of data. Where this presented difficulties due to the need for attended monitoring over an extended period, acquisition of a small sample of data over a relatively short period of time was considered preferable to obtaining no internal data at all.

Noise Measurements Environmental vibration is rarely experienced in the absence of noise and earlier studies have demonstrated that response to vibration can be significantly affected by the presence of noise and other stimuli. It was therefore essential that the measurement protocol could account for the possible combined effects of noise and vibration, both in terms of noise associated with the vibration (e.g. groundborne noise, airborne noise directly from the vibration source, rattling of items within the property) and the effects of other (separate) noise sources on the perception of vibration.

Noise was to be assessed in the following ways:

• Where ‘snapshot’ vibration measurements were undertaken, noise measurements would be attended throughout and the audio output signal from sound level meters recorded onto the data acquisition unit. This would ensure that the noise source was known and that any spurious (internally generated) events could be identified and eliminated. The spectral content of the recorded audio output signal from sound level meters would, however, be frequency limited to the selected vibration sampling frequency (2.56 kHz).

• Subjective observations of noise within properties would be recorded by the interviewers and measurement team.

• Summary information would be recorded from the sound level meter.

Groundborne noise levels were to be calculated subsequently using proven methods (refer to Association of Noise Consultants, 2001) based on 1/3 octave band vibration data and measured where possible, enabling an assessment of the contribution of the groundborne noise component to the overall degree of disturbance. Secondary effects (rattling crockery, etc) were to be assessed through the questionnaire and observations made by the interviewers.




Duration of Measurement Period The duration of measurements required depended upon the nature of the source. Measurements would be made over a representative period (including any key ‘disturbance periods’ identified by the questionnaire) that, combined with information about the source (e.g. train timetable or construction programme), was to be used to evaluate exposure for the relevant daytime and night-time periods. A priority action was therefore placed upon the interviewers to inform the vibration survey team of any times identified specifically by the interviewees as being particularly significant in terms of vibration.

For repetitive / consistent sources, determining a representative measurement period was relatively straightforward; for sources that vary greatly over time or are of a temporary nature (e.g. construction works) this issue is more complex.

For the railway and construction sites, the majority of events would be expected to occur during the daytime period, although freight train movements would also be expected at night. Short term ‘snapshot’ measurements would therefore be undertaken during this period in order to obtain an event sample size that was representative of the overall exposure.

The control position would be monitored continuously for a minimum 24 hour period (or throughout the whole duration of a day’s activity on the construction site) which, when considered in conjunction with the ‘snapshot’ data, would enable the assessment of diurnal vibration at the respondent location. The control position would operate over any period in which ‘snapshot’ measurements were being undertaken. Equipment Requirements The interdependence of the equipment requirements and the survey methodology were such that it was necessary to consider the two aspects in parallel. The following equipment requirements were considered as central to the adequate fulfilment of the study objectives:

• Simultaneous measurement of raw time history data for tri-axial vibration and noise;

• Suitable mounting of transducers to provide for various possible environments;

• Frequency response from 1Hz to 500 Hz for calculation of groundborne noise as well as assessment of vibration against BS 6472 (1Hz to 80Hz);

• Adequate dynamic range to provide sufficient information from vibration levels an order of magnitude or more below accepted thresholds of perception (to enable calculation of groundborne noise) to the highest levels that may be anticipated (potentially to levels just below those that could cause building damage);

• Ability to store all relevant data for 24 hours monitoring at the control monitoring position (or for a sufficient period to make occasional download and re-start practicable);

• Storage of raw time history data and, if possible, ‘on-the-fly’ calculation of required noise and vibration indicators;

• Conditioning of data – amplification and anti alias filtering;

• Power source sufficient to ensure uninterrupted operation for 24 hours for control monitoring position (or for a sufficient period to make occasional download and re-start practicable); and

• A weatherproof, secure and tamper proof method of installation, particularly for the control monitor.

Preferred Equipment Options The following equipment was selected for use during the pilot study. It should be noted that other options may be available for conducting the follow up exposure-response relationship study, but at the time of the pilot study the equipment set out below was considered to




provide the best combination of functionality and cost against the required specification. Selection of the equipment took into account the scale of a full exposure-response study.

Data Acquisition Unit The selected data acquisition device was the Rion DA20 4-channel hard disk data recorder (Figure 3.3). The specification is provided in Table 3.1. This instrument provides suitable functionality for use at the control position, at external ‘snapshot’ positions and for the internal measurements. Three channels on each unit would record raw vibration time history data, whilst the fourth channel would record the audio output from the sound level meter (at internal snapshot positions). While the audio signal would be frequency limited by the selected sampling frequency (2.56 kHz) of the acquisition unit, it was anticipated that sufficient information would be stored in respect of groundborne / structureborne noise to enable scrutiny of data if subsequently required.

Figure 3.3: Rion DA20 hard disc data recorder

Frequency Range DC – 20 kHz Dynamic Range 80 dB Input Range ±10V Anti-aliasing filters Signal Conditioning ICP Capacity 2 GB (26 hrs, 4-channels @ 2kHz) Battery Life (Internal battery) 9.5 hrs Battery Life (External 3.2 A/h battery) 24+ hrs

Table 3.1: Specification for Rion DA20




Where there was the opportunity to obtain simultaneous internal and external measurements it was intended that the pilot study would make use of the 16 channel data acquisition unit available to the Project Partners. This would allow simultaneous recording inside and outside the buildings. However, to reduce the invasiveness of the monitoring it was also decided to try ‘daisy-chaining’ together Rion DA20 units. This is discussed further later in this report.

Transducers Transducers proposed for the pilot study were PCB 393B12 uni-axial accelerometers assembled into a tri-axial array and PCB 356B18 tri-axial accelerometers. Their specifications are provided in Table 3.2.

Sound Level Meters There are many options for sound level meters available which provide the required ‘on-the-fly’ analysis to give appropriate noise indices; the meters must also have the ability to provide an output to the data recorder. The Bruel and Kjaer 2260 and the Norsonic 118 were the instruments of choice for this study.

PCB 393B12 PCB 356B18 Sensitivity (±10%) 10 V/g 1 V/g Range 0.5 g pk 5 g pk Frequency Range (±5%) 0.15 Hz – 1 kHz 0.5 Hz – 3 kHz Resolution (1 Hz to 10 kHz) 8 × 10-6 g 5 x 10-5 g

Table 3.2: Specification for Accelerometers

3.1.5.3 Data Analysis Data from the acquisition unit require post-processing to yield the required indicators for the recorded events. The Rion DA-20 stores data in .WAV format. Due to limitations with the software supplied by Rion, alternatives were investigated. It was concluded that data analysis would be carried out using Prosig DATS Professional software.

This software offers advanced signal identification and triggering options and has the ability to calculate all of the vibration indicators currently used in UK and European vibration assessment.

For discrete events such as train pass bys, determination of the required parameters is relatively straightforward. Within the analysis software, threshold levels can be specified to define the start and end points of events supported as necessary by synchronised clock information. This enables the automated extraction and analysis of the required data. For continuous sources such as construction or industrial processes where the vibration level may continue above ambient for long periods, albeit at a variable level, the approach proposed was to subdivide the signal into contiguous sections, in order to extract a set of numerical values that can be analysed. Contiguous one minute samples were to be used for evaluation of data obtained during the pilot study.

3.1.5.4 Site Observations A proforma was developed to collect and record site observations; this included detailed observations of the construction of the buildings which may affect building response.

3.2 Development of the Questionnaire

This Section describes the review of previous social surveys and field and laboratory studies of vibration and noise and the development of the questionnaire for trial in the pilot study.

3.2.1 Literature Review A review of previous studies has been conducted as part of the process of developing the questionnaire. The review included an evaluation of previous field and laboratory studies of the effects of noise and vibration in buildings in the UK, Sweden and Norway and studies of




social survey methods and publications concerned with methods of evaluating the severity of vibration and noise. The findings and methods of several of the studies reviewed are considered to be valuable and were taken into account in the development of the questionnaire of the present study. Of particular value are the national Noise Attitude Survey (NAS) conducted by Grimwood et al (2002), the “Standardised Interview to Assess Domestic Noise Complaints and their Effects” (SIANCE) conducted by Stansfeld et al (2000), the survey of the effect of railway-induced building vibration on the community (Woodroof and Griffin, 1987) and the field study of Fields and Walker (1977). The process of developing the questionnaire was conducted by devising new questions for aspects not covered in earlier investigations and by the adaptation and revision of some sections of questionnaires employed in previous studies to meet the requirements of the present study.

3.2.2 The Development Process The development of the questionnaire involved obtaining feedback and information from a social survey expert and Defra. At different stages in the development process feedback was sought from Stephen Stansfeld, Professor of Psychiatry at the Queen Mary’s School of Medicine and Dentistry, University of London and author of the SIANCE study.

The process of developing the social survey questionnaire involved consideration of the key objective of the questionnaire which is to provide an instrument for obtaining community response data which are psychometrically valid and reliable and which provides an accurate measurement of human response to vibration and confounding factors. The social survey questionnaire was developed to elicit responses of residents to vibration and noise from a range of sources and to identify factors influencing acceptability, adverse comment or annoyance.

The questionnaire has been designed to:

• ensure reliable interpretation of terminology such as vibration and noise;

• ensure elicited responses produce data suitable for statistical analysis and to ensure that responses are unaffected by misinterpretation, influence or suggestion;

• be suitable for administration by formal interviewing techniques, in which all respondents are asked the same questions in the same order;

• obtain data which are reliable, repeatable and comparable;

• elicit responses of residents to vibration and noise from a range of sources in close proximity to one or more of the sources of building vibration that are of interest to Defra;

• minimise response errors in the resulting data so as to provide an accurate record of the true response of residents;

• provide data suitable for determining dose-response relationships.

3.2.3 Questionnaire Design and Factors Affecting Response

3.2.3.1 Response Terms Where questions employ a rating scale to obtain an indication of the degree of subjective response to vibration and noise, respondents are asked to rate their degree of bother, annoyance or disturbance. Previous studies of noise and vibration effects have used various combinations of these terms. For example, Klæboea, Turunen-Riseb, Hårvikc. and Madshusc (2003) used the response terms “annoy” and “notice” for ratings of vibration. Studies by Grimwood et al (2002) and Stansfeld (2000) used “bother, annoyance or disturbance” for noise rating. Öhstrtröm and Skånberg (1996) employed the terms “observe” and “annoy” for noise and vibration ratings. Woodroof and Griffin (1987) asked residents to rate their annoyance from railway-induced building vibration. In a study by Watts (1984) residents were asked to provide ratings of bother caused by noise and vibration from road




traffic. Fields and Walker (1977) asked residents to rate their bother and annoyance from railway noise and vibration.

It is considered that the use of response terms bother, annoyance and disturbance is the most appropriate in this study because the three terms encompass most likely reactions to building vibration and noise. Use of response terms bother, annoyance and disturbance is also consistent with the 1999 NAS survey of Grimwood et al (2002) and the SIANCE survey of Stansfeld et al (2000) which provided psychometrically valid and reliable data. Both studies resulted in an appropriate distribution of results, with the former study finding that 40% of respondents were bothered, annoyed or disturbed to some extent by road traffic noise.

3.2.3.2 Rating Scales The questionnaire employs both numerical and semantic rating scales. A 7-point numerical rating scale is used to indicate how much the area and home is liked; to rate bother, annoyance or disturbance from vibration (feeling, hearing or seeing) and noise from all sources; to rate the combined response to vibration and noise; for the rating of response to secondary effects of vibration (rattling of windows or objects or the swaying of pendulum lights); and for sensitivity to noise and vibration.

A 5-point semantic scale is employed for ratings of bother, annoyance or disturbance from feeling vibration and hearing noise from each source and for rating response due to the secondary effects of vibration (rattling of windows or objects or the swaying of pendulum lights).

The same 5-point semantic scale is employed in the 1999 NAS survey of Grimwood (2002) and the 7-point numeric scale is similar to the 7-point numeric scale also used in the NAS survey. Use of similar scales will enable a comparison of the results of the present study with those of previous studies. Use of two scales enables the results to be tested for consistency and reliability by correlating the responses obtained with the 5-point and 7-point scales. Woodroof and Griffin (1987) and Watts (1984) employed both numerical (7-point) and semantic (4-point) rating scales. Use of the 7-point rating scale (not at all to extremely) will enable comparison of response due to vibration from the present study to the findings of the vibration field study of Woodroof and Griffin in which the same scale was employed.

Stansfeld et al (2000) employed a 10-point numerical rating scale for bother, annoyance and disturbance from noise, a 5-point rating scale for frequency of different emotions produced by noise and a 7-point numerical scale for sensitivity to noise and for rating how much they like the area. In a health section, 4- and 5-point scales are employed. The types of rating scale were limited to two in the present study. Simplifying the task of the respondent by limiting the number of different scales used in the questionnaire may increase the reliability of the results by making the task easier.

3.2.3.3 The Use of Filter Questions The questionnaire of the present study makes minimal use of filter questions. Filtering, such as asking about a specific source of noise only if they reported hearing that noise, has been shown to bias the results of social surveys. Grimwood et al (2002) compared the results of two NAS surveys and found evidence that the use of a filter question which asks which noises are heard will introduce errors since “heard” may be interpreted as “notice”, “affected” or “bothered”. If questions about a specific noise source are then asked only if they reported hearing that noise, the results will be biased, with a greater proportion indicating that they are affected since the question will not be asked of those who are unaffected.

In the present study, bias of the results caused by filter questions is unlikely to be significant since the number of filter questions have been minimised. In addition, filter questions are not




employed in key questions which will be used to test the validity and reliability of the questionnaire.

3.2.3.4 Secondary Effects The secondary effects of vibration, which include the rattling of objects or swaying of pendulum lights, are addressed. Respondents are asked whether they ever hear or see windows or objects rattle or pendulum lights sway in the home, what they think this is caused by, whether it interferes with aspects of their home life and the degree of bother, annoyance or disturbance. In questions on secondary effects, “hear or see” is underlined to ensure that the respondent distinguishes between these and the primary effects of vibration of the floor, chair or bed; in questions on primary effects of vibration, “feel” is underlined.

3.2.3.5 Order of Questions It is suggested in the NAS study (Grimwood et al, 2002) that the order of questions affects response. For example, a general noise question may elicit greater response when preceded by questions on specific noise sources which act as a reminder to different aspects of noise. In the present questionnaire, a general question on the degree of bother, annoyance or disturbance from vibration is placed after questions on specific aspects of vibration. The specific questions may act as a prompt or reminder about past experiences of vibration and therefore may elicit a more accurate response. Similarly, a general question on the degree of bother, annoyance or disturbance from noise is placed after questions on specific aspects of noise.

Questions on sensitivity to noise and vibration are placed after questions on ratings of vibration and noise so as to avoid influence of general sensitivity questions on ratings of specific noise and vibration effects.

3.2.3.6 Conclusions The questionnaire development process involved a review of previous social surveys and field and laboratory studies. The key objectives of the pilot study were of primary consideration in the questionnaire development. The key objectives are to provide an instrument for obtaining community response data which are psychometrically valid and reliable and to provide an accurate measurement of human response to vibration and confounding factors. Possible effects of the use of different response terms, ratings scales, filter questions and order of questions were taken into account. The requirements for validation and statistical analyses of the questionnaire data and of combined questionnaire and vibration data have been taken into account in the questionnaire development.

The developed questionnaire is modular in form. The majority of questions are applicable in all situations. However, a central ‘modular’ element is selected depending on the nature of the main source of vibration (e.g. rail, construction, etc).

3.3 Development of the Pilot Study

This section sets out how the pilot study evolved and the thought processes and reasoning behind its development.

3.3.1 Methodology and validation considerations The Specification of Requirements asks for “the development of a suitable measurement protocol, social survey questionnaire, and a pilot study to trial them on a selected, adequate number of case studies to validate the protocol and questionnaire”.

Establishing dose-response relationships for vibration is beset with difficulties because of the complexity of the methodological issues associated with measuring vibration and human response to vibration under field conditions. For example, environmental vibration situations tend to be far more complex than noise situations. So, given the complexity of the problem, how can the measurement protocol and questionnaire be tested and validated? This question can only be answered by breaking it down into a series of questions or hypotheses




which can be tested to provide confidence in the answers obtained. It may be necessary to test a number of critical questions or hypotheses in order to gain confidence in the ability of the questionnaire and measurement protocol to determine dose-response relationships for different sources of vibration.

In the context of the pilot study, it was agreed that validation should be taken to mean a demonstration that the questionnaire and measurement protocol are capable of measuring a community dose response relationship (should there be one). It was also acknowledged that, within the limitations of the pilot study, it would not be possible to fully validate the questionnaire and protocol. It was agreed, therefore, that the pilot study should test key issues. Some of these critical issues are addressed below.

What is a dose-response and is there a community dose response relationship for vibration exposure inside buildings? Numerous social surveys have been carried out to determine the correlation between a noise index and annoyance and have been used to derive dose-response relationship such as that derived by Schultz. Most of these studies have been cross-sectional studies, where the correlation between annoyance and noise exposure is determined for a sample population at a single point in time. These are referred to as steady-state dose-response relationships. These studies have found that annoyance generally increases with noise level, as measured using a noise index, but that community response to noise at the same level can vary enormously.

It is worth noting that in the UK, dose-response relationships have been determined for road traffic, aircraft and railway noise. Large social surveys have not been carried out in the UK for construction noise or industrial noise. The absence of such studies may be explained, in part, by the greater methodological issues associated with determining a dose-response relationship for construction noise and industrial noise.

There is no strong evidence obtained from field studies conducted so far in the UK that show a general increase in annoyance with vibration dose. It is conceivable that the human response to vibration is independent of vibration dose above certain thresholds i.e. perceptible thresholds. This meant that either:

• a dose-response relationship does not exist i.e. annoyance does not increase with vibration magnitude, or

• the power of the earlier studies and the methods used to measure vibration dose and response was not sufficient to elicit a dose-response relationship.

Is there a single dose response relationship for different types of vibration? The specified aim of the future larger study is to determine a dose-response relationship for vibration sources impacting on residents outside their control e.g. road, rail (both above and below ground level), industrial, construction, and intra-development activities (e.g. vibration generated by a gym / aerobics floor in adjacent residential property or vibration from plant items such as external plant, boilers etc).

Different sources of vibration generate vibration with markedly different characteristics, which may provoke differences in response. To illustrate this point, it is worth noting that BS6472 suggests that people in residential building are relatively tolerant of construction vibration. Clause 4.1 of the standard states that “situations exist where motion magnitudes above those generally corresponding to minimal adverse comment level can be tolerated, particularly for temporary disturbances and infrequent events of short term duration. An example is a construction or excavation project.”

It is likely therefore that the future larger study may not be able to determine a universal dose-response relationship, which covers all sources of interest. Specific studies may be necessary to determine source specific dose-response relationships and it may be




necessary to apply specific techniques to sources with unique characteristics such as construction vibration.

Can a sufficient sample size be achieved? It is likely that a very extensive study would be required to determine dose-response relationships for several different sources. Even with such an extensive study, there is a significant risk that it may not be possible to determine dose-response relationships for specific sources. It should be possible to identify a large number of people exposed to railway vibration. Even so, there is still some risk that it will not be possible to find a sample population exposed to a sufficiently large range of exposure levels to be able to determine a dose response.

Achieving a sufficient sample size will be more difficult for road traffic vibration and construction vibration (smaller number of people exposed). In addition, there is a significant risk that it may not be possible to find a sufficient number of respondents who are exposed to industrial vibration over a sufficient range of vibration dose and response so as to be able to derive a dose-response relationship for industrial sources.

These issues are relevant to the future study to derive a dose response relationship, or relationships, but they are even more relevant to the smaller pilot study.

Are we concerned just with whole-body vibration? Standards and guidance on the perception of vibration inside buildings are primarily concerned with the physical sensation of structural vibration when a human body is in contact with a vibrating structural surface such as a floor, chair or bed. However, structural vibration is often accompanied by audible and visual stimuli. When people are asked to describe vibration they often associate noise and visual sensations with vibration as well as physical sensations. It is likely that a person’s overall response will be determined by an aggregate of responses to different stimuli. Consequently, it is important to assess whether people can disassociate their response to structural vibration from their overall response and whether the questionnaire can reliably measure discreet response to whole body feelable vibration and aggregated response to noise and vibration stimuli arising from the same source.

With respect to vibration dose, people inside residential buildings are exposed to a range of internal and external sources of vibration. The magnitude of vibration generated inside people’s dwellings is often higher than that from external sources of environmental vibration or from other parts of the same building. There is some evidence to suggest that there can be a community response to vibration inside residential environments arising from external sources at levels of exposure that are significantly lower than the exposure to other ambient vibration arising from domestic activity within the respondent’s own dwelling environment.

Can the questionnaire and measurement protocol elicit a sufficient range of responses so as to be able to derive a dose-response relationship? The field study conducted by the ISVR (Woodroof and Griffin) measured vibration at properties within 100m of a railway. (Note that Fields and Walker found that 45% of residents living within 100m of the railway line did not feel vibration). Properties were selected in three distance bands: <33m, 34-66m and 67-100m, but there was no analysis to check that this successfully provided a range of vibration magnitudes or a range of responses. The range of responses in the study was not sufficient to derive any correlation between vibration and response. There is a significant risk that a similar study may yield similar results i.e. it will not find a dose response relationship. Consequently, one of the key objectives of the pilot study should be to check whether the questionnaire and measurement protocol will yield a sufficient range of response i.e. high exposure causes significantly higher response.




Can individual response to whole-body vibration be measured reliably? There is a significant risk that the questionnaire will not sufficiently discern human response to structural vibration from responses to other visible and/or audible stimuli. Vibration effects may be more subtle and less noticeable than noise and people may be less able to make evaluative judgements about vibration than they are about noise where noise dominates their overall response. Testing the validity of the questionnaire and its ability to scale individual responses to perceptible vibration and other stimuli is therefore a key pilot study objective.

Can individual response to whole-body vibration be measured reliably when the vibration dose varies considerably over time? It is possible to determine vibration levels over a sufficient length of time so as to capture and represent variations in exposure to vibration. The key question here, of course, is what is a representative time period? This may be a difficult question to answer for highly variable sources such as construction vibration but not an insurmountable one. A far more difficult challenge is to measure typical response to vibration when interviews are typically administered at single point in time. Consequently, measuring human response to vibration from sources which vary considerably with time, such as that typically found with construction vibration, is far more problematic and difficult than is measuring response to vibration from sources such as roads and railways. Where vibration varies considerably from one day to the next, there is a real risk that the response measured on a particular day will not be representative. In other words, the response measured on a particular day may vary considerably from the response measured on any another day. It is not known whether a person’s response to an interview will be strongly influenced by immediate events, which occurred at or close to the time of the interview, or whether their response will be affected by the whole vibration dose, with a tendency to ‘smooth over’ peaks and troughs in vibration exposure.

It is essential therefore to understand how response to vibration varies over time when people are exposed to considerable fluctuations in vibration level over the duration of the construction works. It will not be possible to use the questionnaire to measure human response to vibration with confidence until this question has been adequately answered.

The pilot study will provide an opportunity to test some of the ‘mechanics’ associated with applying the questionnaire and measurement protocol to a location affected by construction vibration. Of particular relevance will be the feedback from the questionnaire team to the measurement protocol team to confirm how and when vibration measurements should be undertaken to reflect the measured response to a transient source. However, as described above, it was anticipated that some additional work would be required to ensure that the questionnaire is optimised before it is rolled out on a wider basis to evaluate the dose –response relationship for construction vibration. It would be possible to investigate this issue by conducting a longitudinal study of construction vibration, where vibration and response is measured progressively over time. However, such a study is outside the scope of the current pilot study.

3.3.2 Pilot Study Design The above discussion (Section 3.3.1) about the methodological issues demonstrates that there are significant risks associated with the validation of the instruments being developed and that:

i) the questionnaire and measurement protocol may not be able to determine a dose-response relationship at all; and

ii) equally, there may not be a dose response relationship.

The aim of the future main study would be to determine a dose-response relationship for vibration sources impacting on residents outside their control e.g. road, rail (both above and




below ground level), industrial, construction, and plant items affecting homes such as external plant, boilers etc. Clearly, this means that the questionnaire and the measurement protocol must cover different sources of vibration. However, it was recognised that within the constraints of the pilot study it would not be possible to validate the questionnaire and measurement protocol for all of the different sources.

Consequently, the pilot study was designed to focus on the key issues and to maximise the ability of the instruments to identify a dose response if there is one (i.e. reducing risk i) above). In particular, the design of the Pilot Study was guided by the findings of the review of earlier studies, especially the field study conducted by the ISVR (Woodroof and Griffin). In that study, the range of responses was not sufficient to derive any correlation between vibration and response. Consequently, there was a significant risk that a similar study may yield similar results i.e. it will not find a dose response relationship. Consequently, one of the key objectives would be to check whether the questionnaire and measurement protocol would yield a sufficient range of exposure levels and a sufficient potential range of response i.e. high exposure causes significantly higher response.

Further significant risks were that the questionnaire may not provide an adequate description and quantification of response to vibration; and the questionnaire may not be capable of adequately discerning response to vibration from responses to other visible and/or audible stimuli. Testing the validity of the questionnaire and its ability to scale individual responses to perceptible vibration and other stimuli was another key objective.

It was considered necessary to be confident that the measurement protocol would be capable of reliably measuring vibration from each of the sources to be included in the main study. Measuring vibration from different sources presents different challenges. However, a number of the issues are essentially the same for each type of source e.g. filtering vibration from the source of interest from total ambient vibration. The risk was identified that the measurement protocol may not be sufficiently robust to measure vibration from all sources of vibration if it was not tested on sources with different characteristics.

3.3.3 Pilot Study Options Having established that the pilot study would not be large enough to address all of the key issues, the Project Partners explored two basic options in collaboration with Defra. These are illustrated in Figure 3.4 and are summarised as follows:

Aim 1: Test whether the measurement protocol and questionnaire would be likely to yield a methodology that will be applicable to all vibration sources

Aim 2: Test whether the measurement protocol and questionnaire would be likely to provide sufficient confidence that they would yield a dose-response relationship if used as a basis for a more extensive study.




FIGURE 3.4: Principal aims of the Pilot Study

A number of possible approaches for the pilot study were identified and considered, which could have fulfilled these aims. These approaches are summarised in Table 3.3.

Approach Description Purpose

1 Sites selected at different levels of exposure to vibration, ranging from high to low exposure, from a single source (railway)

By controlling the number of variables i.e. using a single source, it is possible to explore whether the higher levels of vibration lead to higher levels of annoyance, many other parameters remaining constant. It is possible that an adequate sample size can be achieved using a sufficiently large source e.g. a major railway line.

2 Select sites to cover different sources and different exposure conditions e.g. high and low vibration levels, different combinations of noise and vibration

This option would test whether the protocol and the questionnaire can be successfully applied to a range of sources and vibration exposure conditions, but would not allow investigation of an ordinal dose-response relationship.

3 A hybrid between 1 and 2, which would study most (5) sites from a single source (railway) and the remainder at other source types

This would be expected to allow some investigation of a dose-response relationship, from the single source data, although confidence in the conclusions would be reduced by the smaller quantity of data. The applicability of the protocol to different sources would be investigated in a limited manner.

TABLE 3.3: Possible approaches to the pilot study

Having considered the strengths and weaknesses of the different approaches it was decided, in agreement with Defra, to design the pilot study to focus on Aim 2 (Figure 3.4).and Approach 1 (Table 3.3).

This option was chosen because it provided the best approach to test whether the questionnaire and measurement protocol could yield a sufficient range of responses. A

Aim 1

Aim 2

Develop a protocol giving confidence of determining a dose-response relationship

Pilot Study focussed on one type of vibration source

…requires…

Develop a protocol applicable to different types of source

Pilot Study undertaken at different types of vibration source

…requires…

Aim 1

Aim 2

Develop a protocol giving confidence of determining a dose-response relationship

Pilot Study focussed on one type of vibration source

…requires…

Develop a protocol applicable to different types of source

Pilot Study undertaken at different types of vibration source

…requires…




number of primary tests and secondary tests were then formulated around this approach to evaluate objectively the success of the questionnaire and protocol.

3.3.4 Pilot Study Objectives In order to maximise the likelihood that an effective and valid outcome would be achieved from the pilot study, two primary objectives were agreed, namely:

• To demonstrate that the outputs from the questionnaire and measurement protocol are appropriate for the analysis of dose response relationships based on the current best practice; and

• Test whether the questionnaire and protocol can elicit an ordinal relationship between response and dose over a sufficient range at reasonable extremes of vibration exposure (a positive indicator that a dose response relationship may exist).

In order to meet these primary objectives it was decided to validate the questionnaire and protocol initially for a single type of source. In addition, it was decided that the pilot study should focus on and maximise the volume of data for repeatable and consistent vibration sources where it is possible to ‘control’ the vibration exposure by simple means, such as varying the distance from the source. A single railway (different locations along the railway) was therefore selected as the main source to be considered in the pilot study.

In order to maximise the range of exposure to vibration from the single railway source it was decided to:

• select pilot study measurement positions close to the source, where there would be a relatively high exposure to vibration; and

• select other positions at greater distances from the railway, where it could be predicted that there would be a low, but still measurable and perceptible vibration exposure (i.e. locations at or around the threshold of perception).

Given the obvious limitations of the single source approach, it was also agreed to test whether the questionnaire and measurement protocol would be reliable at a location with a very different source. Consequently, a construction site location was to be included in the pilot study at which an impulsive source of vibration would be used. The most common option for such a source would be impact driven piling.

3.3.5 Site Selection Pilot study measurement locations (i.e. residential areas where the questionnaire and protocol would be applied) would be selected to maximise the range of exposures to vibration where there would be a high density of housing (to maximise the potential number of potential respondents). Where possible, the locations were to be selected where residential property exists at a range of distances from the site. Properties close to and distant from the railway could then be targeted to ensure high and low dose measurement and questionnaire pairings.

It was decided to target the railway vibration site selection in the vicinity of the East Coast Main Line (ECML) between Peterborough and London. This line was chosen due to the large volume of rail traffic, density of population and its accessibility to the survey teams. Potential measurement locations would be selected on the basis of the following:

• They needed to be where there was known (or could be confidently predicted) to be significant levels of ground vibration and a high probability of perceptible vibration from the specific source;




• There needed to be a sufficient number of respondents in the locality (to ensure that sufficient residents could be identified who would be prepared to cooperate with the study);

• The location needed to be easily accessible and exposed to vibration from a single source, without interference from other sources of vibration;

• The vibration source needed to be identifiable and repeatable; and

• They needed to be in areas that would not risk compromising the health, safety and welfare of the survey teams.

Potential railway locations were identified at Peterborough; between St Neots and Huntingdon, around Potters Bar and Hadley Wood and on the London approach. If practicable, at least one of the locations would be selected over a tunnel to test the questionnaire and measurement protocol in a situation where there is significant groundborne noise in addition to feelable vibration, but where airborne noise would be largely eliminated.

It was intended that three high exposure sites would be selected close to the railway line, ideally within about 20 metres, to maximise exposure levels. Three low exposure sites would be chosen at locations where vibration would be expected to be perceptible at low levels (anticipated to be about 100 metres from the railway line). An intermediate site would be selected where there were sufficient properties at an intermediate distance between the high and low exposure sites.

3.3.6 Data Analysis As noted in Section 3.1.5.3, the raw data was analysed using Prosig DATS Professional software, since this provided all the required functionality, including the ability to extract individual events from long continuous time histories and to determine any parameters that may be required for analysis.

For the pilot study, vibration data would only be required to be analysed to obtain the peak un-weighted acceleration and rms un-weighted acceleration, although the software does have the facility to derive all other metrics from the recorded data that may be required. Examples of some of the possible parameters available are illustrated in Appendix C.




4 The Pilot Study The preceding section of this report has described how the questionnaire, measurement protocol and pilot study methodology evolved through the project. This section of the report describes the work actually undertaken to validate the measurement protocol and questionnaire through the pilot study. The methods used for the fieldwork and for the subsequent data handling and analysis are described.

4.1 Objectives (success criteria)

A number of evaluation criteria were explicitly identified, prior to the start of the pilot study, so as to be able to assess objectively the success of the study. The table below is repeated in the conclusions section with additional information briefly describing the performance of the study against these criteria.

Evaluation criteria to assess the performance of the measurement protocol Reference Evaluation criteria 1 Raw vibration time history data can be reliably and continuously recorded at

each assessment location for at least 24hr (or full operating period, as appropriate to the source).

2 Vibration data can be gathered, using the same equipment specification, in a number of different positions both inside and outside properties at a range of distances from the source and over the full range of possible human exposure (from threshold of perception – noise and vibration – to very high exposure).

3 Noise and vibration (in 3 orthogonal axes) can be recorded simultaneously at each measurement position.

4 The method is minimally intrusive, maximising acceptability to residents and hence maximising permissions to undertake measurements inside properties

5 Phase locked noise and vibration can be recorded simultaneously inside and outside properties.

6 The method is efficient and straightforward on site, supporting high (measurement) production rate, making a full exposure-response relationship study viable.

7 Vibration events can be identified, extracted and analysed automatically from continuous records during post-processing.

8 Suitable for quantifying vibration exposure from different sources. 9 Post processing can reliably evaluate (automatically) all current vibration

indicators (e.g. weighted and unweighted acceleration, VDVg, VDVd, VDVb, PPV etc)

10 Feedback from the questionnaire survey team to the measurement team can be provided sufficiently quickly and accurately to allow exposure to transient sources to be captured consistently in the measurement protocol and the questionnaire.

12 The accuracy of the measurement protocol is fit for purpose. 13 Questionnaire is sufficient to gather information required but short enough to

deliver high rate of acceptance for completing the questionnaire. 14 Questionnaire responses suitable for statistical and factor analysis. 15 Responses to questionnaire differ significantly when respondent is subject to

different exposure. 16 Questionnaire response reliably differentiates between noise and vibration

exposure. 17 Questionnaire responses are consistent. 18 Respondents demonstrate reliable interpretation of terminology (e.g. noise

and vibration). 19 Questionnaire responses are reliable, repeatable and comparable. 20 Questionnaire and measurement protocol combined analysis demonstrate

that there is an ordinal relationship between response and exposure.




4.2 Questionnaire Summary

This section provides a brief description of each section of the questionnaire as taken forward to the pilot study. The questionnaire as trialled is included at Appendix D. Section 4.2.2 describes work undertaken to test the questionnaire before commencing the pilot study.

4.2.1 Final Version of the Questionnaire: Section Summaries The questionnaire was developed to be modular, so that it could be easily adapted for use for different dominant sources of vibration. In the pilot study, two different modules for the questionnaire were required: the majority of the sites were railway sites. For the construction site, modifications were required to Sections F, H and I (see below). The questionnaire is provided in Appendix D and summarised below.

Section A: call, interview, dwelling and location information Section A is completed by the interviewer without input from the respondent. The interviewer specifies the address, date and start time and end time of the interview; the dwelling type; whether the residence is close to specific road types or railway lines; and whether it is in a village, town or city. In larger studies, the information in this and other sections may be employed to provide a broad picture of responses to building vibration in different types of residences and areas.

Section B: Personal characteristics of the respondents The interview with the respondent starts with Section B, which consists of questions concerning duration and hours of occupancy; and number of occupants. Duration of occupancy may influence response since there may be a habituation effect whereby respondents residing for longer durations experience lower levels of bother or annoyance from building vibration and noise. Section C: Satisfaction with neighbourhood and Section D: Satisfaction with home Sections C and D consist of questions on likes and dislikes about aspects of the home and neighbourhood. The purpose of these questions is to provide a general indication of the level of satisfaction with the neighbourhood and to determine whether vibration and noise are major contributory factors to general satisfaction. A 7-point scale is employed to provide an indication of general satisfaction. The responses may be compared with ratings of vibration and noise from questions in other sections to show the relative contribution to overall satisfaction. Section E: Vibration questions In Section E, questions are focussed first on the primary effects of vibration, i.e. feeling vibration of the floor, chair or bed. These are followed by questions on the secondary effects of hearing or seeing windows or objects rattle and pendulum lights sway. In questions on primary effects, it is emphasised that the questions are concerned only with the feeling of vibration by use of bold and underlined text for the term “feel”. In questions on secondary effects, it is emphasised that the questions are concerned only with the hearing or seeing of vibration by use of bold and underlined text for the terms “hear or see”. The aim of emphasising the text is to ensure the interviewee is responding to the primary or secondary effect where appropriate.

Respondents are asked for ratings of annoyance, bother or disturbance caused by feeling vibration of the floor, chair or bed using a 5-point semantic scale (Q27). If the response is “not at all”, “don’t know” or “don’t feel”, then further questions follow to ascertain whether vibration is ever felt. A “don’t feel” option is selected if the response to the additional questions is that the respondent never feels vibration. Hence, the additional questions ensure that the ratings are obtained only for the perception of vibration. This avoids distortion of the results caused by the inclusion of the lowest rating when vibration is not perceived.




For primary and secondary effects of vibration, questions aimed at obtaining ratings of annoyance, bother or disturbance caused (Q27 and Q31) are preceded by questions on the interference with activities (Q26 and Q30) so that respondents are reminded of the effects of feeling, seeing or hearing vibration. Such ordering of the questions may elicit more accurate response and higher ratings than the placing of activity disturbance questions after rating questions.

A 7-point numeric semantic scale is employed for the rating of combined response to feeling, seeing and hearing vibration from all sources (Q32). By comparison with the ratings of feeling vibration (Q39) and hearing and seeing vibration (Q31) caused by each source, the relative contribution of the primary and secondary effects from each source to the overall effect may be examined in an extensive study.

Question 33 asks whether the respondent is concerned that the dwelling may be damaged by vibration. In an extensive study, comparison of ratings of bother, annoyance and disturbance from residents who fear building damage with those who do not fear building damage may provide an indication of whether fear of building damage increases ratings.

Section F: Railway or construction vibration questions Ratings of bother, annoyance and disturbance caused by feeling, seeing or hearing vibration caused by the trains or by construction machinery are obtained from Question 35 using a 7-point scale.

The time of day when vibration from the trains or from construction machinery causes bother, annoyance or disturbance is obtained from Questions 36 and 37. Details of the type of train or construction machinery causing most bother, annoyance or disturbance from vibration are obtained from Question 38. In a larger study, the responses obtained may be correlated with measurements of building vibration to provide information on vibration characteristics resulting in adverse comment.

Section G: Noise questions Section G consists of questions on responses to noise which are of similar format to questions on vibration in Section E. The questions aim to determine sources of noise, interference with activities and ratings of bother, annoyance and disturbance caused by noise from each source.

Ratings of noise from each source obtained from Question 41 may be compared with ratings of primary and secondary effects of vibration from each source obtained from Questions 27 and 31 which employ the same 5-point scale. Such a comparison provides an indication of which aspect of the source, noise or primary or secondary vibration is the most annoying or disturbing.

A 7-point rating scale is employed to obtain ratings of bother, annoyance or disturbance from noise caused by all sources (Q42). In an extensive study, the results of Question 42 may be compared with ratings from Question 32, which employs the same scale for the degree of bother, annoyance or disturbance from vibration, to indicate relative importance of noise and vibration.

Section H: Railway and construction machinery noise questions Section H consists of questions on responses to railway or construction machinery noise which are of similar format to questions on railway or construction machinery vibration in Section F. The time of day when noise from the trains or construction machinery causes bother, annoyance or disturbance is obtained from Questions 45 and 46. Details of the type of train or construction machinery causing most bother, annoyance or disturbance from noise are obtained from Questions 47. In an extensive study, the responses may be correlated with




measurements of noise to provide information on noise characteristics resulting in adverse comment.

Section I: Relative and total response to noise and vibration In Question 48 of Section I, the respondent is asked to provide ratings of bother, annoyance or disturbance caused by noise and vibration from all sources using a 7-point numerical scale. In an extensive field study, ratings of the combined effect of noise and vibration may be correlated with ratings of the effect of vibration (Q32) and of the effect of noise (Q42) to indicate how ratings of individual stimuli may be combined to predict overall response. Ratings of the combined effects of noise and vibration from a field study involving vibration and noise measurements would allow comparison with the results of a laboratory study conducted by Howarth and Griffin (1990) in which a relationship for the combined effects of noise and vibration was determined.

In Question 49, the respondent is asked to indicate preference for the reduction of either noise or vibration from all sources. The question is then repeated for noise and vibration from trains (Q50). Responses to Question 50 from an extensive field study also involving measurements of noise and vibration would enable a comparison with the findings of the laboratory study of relative importance of noise and vibration from trains conducted by Howarth and Griffin (1990).

Section J: Interviewer assessment of vibration Section J was completed by the interviewer. The interviewer indicates whether they felt vibration of the house, floor or chair or whether they saw or heard rattling of windows or objects or swaying of pendulum lights. They are also asked to indicate possible sources of vibration effects. The section was completed at the end of the interview so that focussing the interviewer on their own responses does not lead to the interviewer unconsciously influencing the responses of the resident. Interviewer assessment of the vibration provides an indication of whether vibration effects occurred in the dwelling during the interview. The responses may be of value since they are independent and not affected by influences such as the fear of damage of the property, expectation of compensation or concern about the effect of vibration on property value.

4.2.2 Questionnaire Pre-pilot Study Verification and Training The completed questionnaire went through several stages of development before being finally used in the pilot study. These are described in this section.

An initial draft of the questionnaire was given to two interviewers for comment. Following this, the questionnaire was revised and provided to the interview team for a briefing session.

Both TRL and ISVR senior staff attended the briefing session and explained the questionnaire development and aims of the survey. The interview team read through the questionnaire item by item to ensure the team found each question comprehensible and unambiguous. They were asked to comment on the clarity of the questionnaire during use and to highlight any specific difficulties that might arise in administering the questionnaire. A mock interview was completed between two interviewers to check the time required for completion of the questionnaire.

The briefing also included guidance on:

• how to probe further for responses without influencing the content;

• avoiding influencing the respondent with leading questions;

• what information about the property and its furnishing should be recorded;

• the use of the vibration monitoring equipment that would be used by the ‘follow on’ measurement team and how to request, log and communicate to the measurement team permission to undertake measurements;




• their own observations of noise and vibration in the property to be recorded after the interview; and

• their personal security when visiting respondents’ homes.

As a result of the discussion at the briefing session, changes were made to the questionnaire in line with interviewer comments. In particular Q27 was altered to make it easier to administer. The layout of the questionnaire was changed to place the interviewer observation questions at the end of the survey.

Following these changes, a pre-survey was conducted at five properties in the local village. Various questions arose from this regarding the specifics of what should be recorded, and regarding changes to the format of the questionnaire to make it easier to complete in the field.

Further changes were then made to the questionnaire in line with the issues raised in the pre-survey, and a draft copy was prepared ready for use in the field. From these trials it was assessed that the questionnaire would take between 35 and 45 minutes to complete.

At the end of each interview, the respondent would be asked whether they would agree to vibration measurements being taken within their property by the measurement team who would contact them to make the necessary access arrangements. A letter was prepared that detailed that briefly outlined the measurement protocol (see Appendix E).

The questionnaire survey would be followed as soon as was practicable by the measurement survey. The social survey team therefore relayed to the measurement team the contact details of each resident who had agreed to the survey. This was done on the first working day following completion of the questionnaire survey.

4.3 Measurement Protocol

The protocol as initially trialed was the same as that presented in the Interim Report (see Section 3) and is included at Appendix F. The equipment and processes were trialed in advance of the pilot study, as described below.

4.3.1 Measurement Protocol Pre-pilot Study Verification and Training Before the pilot study commenced, the measurement equipment was used on a site, during a different project, to ensure that it performed reliably. This was also a railway vibration site where triaxial vibration was recorded. This also provided an opportunity for site staff to become familiar with the equipment in a ‘live’ situation.

The data from this site were sent to Prosig who wrote a prototype script and undertook prototype analysis to demonstrate their software’s capability.

An important aspect in ensuring a good take-up of the agreement to measure within people’s homes was the ability of the questionnaire survey team to describe in detail how the measurements would be conducted. In particular, it was considered that the size of the equipment and the impact on the respondents’ activities needed to be made clear. Close discussions between the questionnaire team and the measurement team and provision of images of the equipment were valuable in this respect.

The following section of the report describes how the pilot study was actually carried out. The first site in the pilot study was intended to be a ‘pre-pilot’ study, in which particular emphasis would be placed on verifying the processes. In the event, this turned out to be wholly successful and was progressed along with the rest of the pilot study. The only significant change made as a result of the pre-pilot study was in the method used for acquiring phase-locked internal and external vibration measurements, as described in Section 4.4.4.1.




4.4 Undertaking the Pilot Study

4.4.1 Objectives The fundamental objective of the pilot study was to test and validate, as far as possible within the constraints of a relatively small sample size, the proposed questionnaire and measurement protocol. This would aim to demonstrate that the approach would be appropriate for a full scale exposure-response relationship study and, if necessary, identify any limitations or difficulties that may need to be addressed within the full scale study. This fundamental objective for the pilot study is distinct from the objective of the full scale study, which would be to apply the questionnaire and measurement protocol in such a way as to be able to attempt to determine an exposure-response relationship. Section 4.1 detailed the objectives (success criteria) that were set for the pilot study.

4.4.2 Fieldwork Methodology This section of the report sets out the procedures used during the pilot study. Details of each measurement location and the work undertaken at each are provided in Appendix G.

4.4.2.1 Measurement Location Selection It was concluded in the Interim Report (see Section 3 of this report) that there were significant advantages in undertaking the whole pilot study around railway lines, principally for the following reasons:

• Railways are the most common and widespread source of perceptible environmental vibration;

• Focussing on one source would minimise the number of variables, enabling investigation of whether an ordinal response between vibration magnitude and subjective response;

• Railways provide a source of vibration that is essentially the same day-to-day, as opposed to construction sites, where the source may be variable in terms of both location and magnitude;

It was also considered that there would be value in trialling the questionnaire and measurement protocol at a site which had an impulsive vibration source in order to test the proposed approaches at a site with different characteristics, both in terms of the nature of the vibration and the variable position of the vibration source. This would enable the flexibility of the approaches to be evaluated. The most common source of vibration of this nature is hammer driven piling. This section of the report describes how the railway and piling sites were selected.

Railway Measurement Locations Pilot study measurement locations were chosen to maximise the range of exposures to vibration and to maximise the potential number of respondents, by selecting locations with a high density of housing at a range of distances from the vibration source. Properties as close to the railway as possible and at distances predicted to be towards the limit of perceptibility were targeted to ensure high and low vibration measurement-questionnaire pairings. Locations were also chosen, where possible, on the basis of a range of train types (to provide information for different temporal and frequency characteristics).

Potential locations were selected on the basis of the following:

• There was expected to be a significant level of ground vibration propagating away from the source, with a high probability that the level would be perceptible;

• A sufficient number of respondents was required in the locality exposed to both ‘high’ and ‘low’ vibration exposure (to allow validation of the combined questionnaire and measurement protocol) to ensure that sufficient residents could be approached who would cooperate with the study;




• The locations needed to be easily accessible;

• As far as possible, residences were required that would be exposed only to a single source of vibration – there would be no interference from other sources; and

• They needed to be in areas which would not risk compromising the health, safety and welfare of the survey teams.

It was decided to concentrate the search for potential locations alongside the East Coast Main Line (ECML) between Peterborough and London. The following process was used to identify sites:

• 1:5000 OS maps of the ECML from Stevenage to Alexandra Palace were inspected;

• 15 possible locations were identified on the basis of vicinity of houses to ECML; and

• Sites were rated in order of “density of population around railway” and “density of population around railway tunnel”.

Inspection of maps and photographs (i.e. not house counts) showed, as expected, population around ECML increases as London is approached, but in all locations there were few residential properties situated less than 30m from the railway.

A shortlist of possible measurement locations was generated from desk studies, followed by a visit to all shortlisted locations to assess their suitability. Many were rejected, largely on the basis of the housing density. There was found to be few locations where there were sufficient properties sufficiently close to the railway to provide ‘high’ exposure, although the total target number of properties could have been acquired from more distant, lower exposure, properties.

Three suitable measurement locations were chosen at which to conduct the surveys, one of which was at a location over a tunnel. This was intended to test the questionnaire and measurement protocol in a situation where there was likely to be significant groundborne noise in addition to perceptible vibration and where the source was neither visible nor creating appreciable airborne noise.

For each measurement location a map was produced indicating properties that fell into estimated zones of high and low vibration exposure (see Figures 4.1 to 4.3). These were provided to the questionnaire survey team to enable them to target a similar number of properties in both zones at each site.

Construction Site Measurement Location Selection The principle criteria for selection were as follows:

• A hammer driven piling site;

• Piling to be in a residential area and required as close as possible to the closest residential buildings;

• Start date and duration of the piling contract to be suitable to fit in with the research programme and to enable both the questionnaire and measurement surveys to be undertaken; and

• A location in the south east of England was preferred, since this is where the majority of the survey staff was based.




Figure 4.1: Target properties identified at Measurement Location A

Figure 4.2: Target properties identified at Measurement Location B

Figure 4.3: Target properties identified at Measurement Location C

Maps produced by MAGIC on 30 August 06. Copyright resides with the data suppliers and the map must not be reproduced without their permission. Some information in MAGIC is a snapshot of information being maintained or continually updated by the originating organisation. Please refer to the documentation for details, as information may be illustrative or representative rather than definitive at this stage




A suitable site was located following discussions with piling contractors. The Federation of Piling Specialists website and contacts within the Steel Construction Institute were also of assistance.

A number of contacts were approached. Dawson Construction Plant Limited offered a site but the closest houses were some 100 m away. Aarsleff Piling was able to offer a more suitable site. This was at a location operated by Kier Partnership Homes Limited, where concrete precast piles were being driven as foundations for residential buildings. Properties had already been constructed on the adjacent site and were occupied. The scheduled start date was suitable and the expected duration of the works was 7 to 8 working days, which provided just sufficient time for the works to start, the questionnaire survey to be undertaken, responses fed back to the measurement team and a monitoring programme to be planned and executed.

4.4.3 Undertaking the Questionnaire Survey The measurement team selected the measurement locations for the pilot study, so detailed maps were provided by the measurement team close to the time of the survey. These maps highlighted properties at particular distances from the railway line and were used by the interviewers when selecting properties to approach. The interviewers commented that the clarity of the mapping could have been improved to reduce the interviewer judgement required in selection.

Interviewers were provided with a quota of properties at which to achieve successful interviews. The contact method used by the interviewers was “cold-calling”, which involved a personal call at the dwelling rather than a letter drop. Respondents were approached to determine their willingness to take part in the survey and an appointment made to interview them face to face, if it was not convenient at the time of initial contact. The interview was conducted before the vibration survey was discussed, in order not to bias the respondents’ answers. In order to obtain the greatest chance of finding an occupant at home, questionnaire surveys were arranged for Fridays, Saturdays and Sundays.

Local authorities and Police were notified that the interviewers would be surveying in the area. One issue was that some properties were in flats where access was restricted by entry phone, and it was not possible for interviewers to “cold-call” at these properties. In some cases it was possible to use a successful interview as an opportunity to ask neighbours of the respondent to take part.

Once an appointment for an interview had been completed, the occupant was given contact details of the survey manager so that any questions may be answered.

Following a successful interview, the respondent was asked whether they were willing to allow measurements to be taken in their home. If they agreed, they were given a letter outlining the procedure that would be followed (see Appendix D).

Immediately following the interview, the interviewer was required to answer questions relating to their own observations of noise or the effects of vibration during the interview.

Tape recordings were made of some interviews, with the agreement of the respondent, for quality control purposes.

4.4.4 Undertaking the Exposure Survey using the Measurement Protocol

4.4.4.1 Measurement Equipment The measurement protocol required that a combination of vibration and noise measurements were made at control, satellite and internal positions during the pilot study. This section describes the equipment used at each position.

It was established early in the project that there existed no off-the-shelf equipment that would provided all functions ideally required for this project. Development of new equipment




was beyond the scope of the project, so the best approach using available equipment was developed. Equipment available to the Project Partners was used for the entirety of the pilot study. No equipment was hired or borrowed, but the Rion DA20 and Prosig software were procured for this project.

Data Acquisition

All vibration and noise data at control, satellite and internal locations was recorded using the hand held Rion DA20 4-channel hard disk data recorder. Data is saved by the DA20 as raw time history on 2Gb flash disks in wave (.WAV) format.

Transducer sensitivity in volts per unit can be assigned to individual channels. This information is saved in the wave file information and allows the signal, calibrated by sensitivity, to be viewed directly in the RION viewer software. However, the sensitivity is used to calibrate the signal within the RION software only and the saved wave file remains uncalibrated.

Sensitivities in the range 0.1mV/ms-2 to 99.99mV/ms-2 can be assigned to the acceleration channels. Vibration transducers used during the study had approximate sensitivities of either 1V/ms-2 or 0.1V/ms-2. Therefore corrections to the calibration were required prior to processing the data.

Noise data was recorded as a voltage i.e. raw data voltage output and a separate calibration tone recorded to enable subsequent calibration.

All data (noise and vibration) was sampled at a frequency of 2.56 kHz per channel

Phase-locked external and internal measurements

Prior to the pilot study it was intended that, where possible, simultaneous external and internal measurements would be phase-locked and recorded using a single acquisition device. This would be realised using the 16-channel CED vibration analyser.

Due to the lengthy set-up times of this equipment and intrusiveness of multiple long cable runs it was stated in the Interim Report that this method would only be used in suitable circumstances, i.e. if the respondent was happy with cabling run through their property and with the survey team being inside the property for what would become an extended measurement period.

Experience from the first set of measurements made on the first site showed that the set up times for the CED equipment were excessive in relation to the total measurement period required. Therefore, phase-locked measurements were undertaken using two simultaneously triggered RION DA-20s. A simple external gate trigger was used to synchronise the start and end of recording. Whilst data obtained thus are not truly phase locked, it was anticipated that the phase lag would be no more than 1/4 of a wavelength at the upper frequency limit. This approach allowed all simultaneous external and internal measurements to be phase-locked, since the set-up times were greatly reduced. Only a single cable was required to link internal and external measurements; this was acceptable to all property occupants.

Vibration Transducers

At all control, satellite and internal positions vibration was measured in the three orthogonal axes using a tri-axial transducer array. The arrays consisted of either a PCB 356B18 triaxial accelerometer or three PCB 393B12 uniaxial accelerometers assembled into a triaxial array using an orthogonal mounting block.

Both types of accelerometer require a constant current source, which was provided by the RION DA-20 via the input connectors.




Each axis of the tri-axial PCB356B18 has an approximate sensitivity of 0.1V/ms-2. The uni-axial PCB 393B12 accelerometers have an approximate sensitivity of 1V/ms-2.

The sensitivities of both transducers were found to be sufficient in all situations, however the higher signal to noise ratio (SNR) achieved with the higher sensitivity PCB 393B12 transducers was noticeable during the post-processing of the data and lead to more consistent event identification.

High levels of very low frequency noise (<1Hz) were seen using the PCB 256B18 tri-axial accelerometer. It is not known if this is a function of the transducer, DA-20 or a fault in either of the devices. Signals recorded using this set-up were filtered below 1Hz during the post-processing. This cut-off frequency allowed all data within the range covered by the British Standard BS6472 to be retained and analysed.

Transducer Mountings

To ensure good coupling between the accelerometer arrays and the measurement surface for a variety of measurement surfaces, three methods of mounting were required.

For control and external satellite positions, the measurement surface was either soil or a hard surface such as concrete or tarmac. The position was limited by the ground outside the property, especially if measurements were being made without permission of the property owner. For soil measurement surfaces a steel ground spike was driven into the ground. The accelerometer arrays were then mounted on the ground spikes using magnetic mounting blocks or bees’ wax. It is possible that the different mounting materials may account for the range of calculated transfer functions.

For hard surfaces, a steel fixing plate was bonded to the measurement surface using quick drying epoxy-based resin. The accelerometer arrays were then mounted on the plates using either a magnetic mounting block or bees’ wax. In some situations the use of epoxy-based resin as an adhesive was inappropriate, for example on an ornamentally paved driveway. In these situations the ability to adhere the PCB 356B18 to the measurement surface using bees’ wax was an advantage.

For internal measurements, surfaces encountered included carpet, wood laminate, synthetic laminate, linoleum and tiles. In all cases a steel measurement plate was used. The plate has been constructed to be of high mass to simulate the surface loading a person applies to the measurement surface. It also has a high bending stiffness so that mounting resonances are outside the range of measurements. Spiked ‘feet’ ensured good contact with the measurement structure on carpeted floor.

Noise measurements

Noise was measured at most of the external satellite and internal locations. Two types of sound level meters were used during the pilot study, a Bruel and Kjaer 2260 and a Norsonic 118. In each case the sound level meters were set up to record continuously and the output was fed into the RION DA-20 where it was recorded as a raw voltage time history. At each different measurement location, a calibration tone was recorded for use in the post processing of the noise data.

In one situation the property owner expressed a wish not to have noise measured inside his property. Adverse weather conditions such as high winds or rain prevented the measurement of noise in some cases.




Figure 4.4 - Noise and vibration measurement within a property. Note PCB 393B12 accelerometers mounted on plate with spiked feet to penetrate the carpet

4.4.4.2 Selection of Measurement Positions The selection of measurement positions was largely dependant on the number and location of questionnaire respondents. Having undertaken the questionnaire survey and sought agreement to monitor within the properties, it was then possible to plan an optimised monitoring strategy. It was therefore advantageous to wait for the majority of the information to be collated and passed on from the interviewers before selecting measurement positions.

The number and distribution of respondents also dictated the required time on site, as discussed below. On some sites, the response rate was considerably better than had been expected. This provided a degree of flexibility to allow the measurement programme to be optimised.

This section describes how the control measurement position and the internal and external satellite monitoring positions were selected.

Control Position

It was important to identify a suitable position for the control monitor before selecting satellite and internal measurement positions. The function of the control monitor is twofold:

• to log the vibration exposure of a site over a 24 hour (or other appropriate) period; and

• to provide a basis for scaling the shorter term external satellite measurements to account for all events during the 24 hour (or other appropriate) period.




Figure 4.5: Control monitoring position at the construction measurement location

The simultaneous recording of events at the control and satellite positions is required for the latter. Therefore satellite measurements are unusable in the analysis if there were no control monitor in place.

The main requirements for selecting the control position were:

• A position close enough to the source of vibration so that source events of all magnitudes were recorded;

• A secure position where equipment could be left unattended for up to 24 hours and for the entirety of the site survey; and

• A position where the survey staff could gain daily access to the monitor, on sites where several days monitoring was required.

For long-term unattended measurement the most secure option is on private property. For transport sources this will most likely be in the garden of a property close to the source of vibration. For construction sources, such as piling, this will most likely be within the boundaries of the construction site.

For transport sources, a shortlist of suitable locations was selected from the list of questionnaire respondents who had expressed willingness for the surveyors to measure inside their properties. The respondents were then contacted and through discussion it was determined whether their property fitted the above criteria. The most suitable position was then chosen.

At some sites it was found difficult to arrange access to a respondent’s property on more than one occasion to place and maintain the control position. Where access was granted on more than one occasion, it was not always possible to arrange it at a time or period convenient for surveying staff to perform satellite measurements around the site. This




highlights the reliance of the entire survey on good selection of the control location, since it needs to be easily and regularly accessible.

For the construction site, the site owners were contacted to assess the security of the site and to gain permission for placing the monitor. This was successful on this occasion.

A fallback option that was considered but was not needed for the pilot study was for the situation where there was no alternative but to locate the logger on publicly accessible land. In this case, it was proposed that a security company would be employed to guard the equipment continuously.

Satellite (External) Measurement Positions

Selection of satellite measurement positions is again dependant on the volume and distribution of questionnaire responses at each site. Once this information had been obtained from the interview team, the locations of questionnaire respondents were plotted on a site map. Using this map, a strategy was developed to maximise the number of questionnaire-measurement data pairs and optimise the number of satellite measurement locations required. For the pilot study, it was also important to ensure that a balance of questionnaire-data pairs from low and high exposure properties was obtained, to enable an ordinal relation between vibration exposure and level of comment to be investigated.

To maximise the efficiency of the data acquisition work, measurement positions were sought where interpolation between measurements or site laws could be used to assess levels of vibration at more properties than were measured directly outside. This helped maximise the questionnaire-measurement data pairs while retaining a practicable extent and duration of the vibration survey. Examples of where this approach was applied included:

• Terraced houses from which more than one response was received. Here, a satellite position was chosen at each end of the terrace and the data used to estimate levels in the intervening property or properties.

• Where more than one respondent property was located within close proximity to another, such as two adjacent semi-detached dwellings, or three adjacent dwellings in a terrace, a single satellite position was considered to represent each of them.

Further site-specific details of these approaches are given in later subsections of this report.

Where it was not possible to represent multiple respondent properties with a single satellite position, one satellite position per respondent property was used.

The draft measurement protocol had proposed that external vibration measurements were made at a standardised distance of 2 m from the façade of the building closest to the vibration source. In many cases this proved to be impracticable for the following reasons:

• Closest façade to the vibration source was not publicly accessible (e.g. houses where the back garden extends to the railway land);

• Measurement at 2 m would have required access to land for which permission had not been given (i.e. where no access had been granted to do external measurements on the property);

• No suitable surface on which to mount the transducers at the required distance.

In all cases, a measurement location as close as practicable to the façade facing the vibration source was chosen.




Figure 4.6 Example of external noise and vibration measurement position

At properties where internal measurements were made, access had been granted to the respondent’s property so a measurement location close to the appropriate façade of the property could be used, subject to there being suitable ground. Where access permission had not been given, measurements were made as close to the façade of the property as possible, such as at the end of the front garden.

Contact details of respondents who were not willing to let measurement teams inside their properties were not obtained during the social survey; therefore it was not possible to arrange access to measure externally within the grounds of the majority of respondents’ properties.

Internal Measurement Positions

The choice of internal measurement locations was not only limited by the number of questionnaire respondents, but also the number of respondents willing to let survey teams into their properties to measure and the availability of these respondents at times appropriate for the surveys.

The proportions of respondents willing to have internal vibration measurements carried out were:

• Measurement Location A: 26 out of 41 (63%)

• Measurement Location B: 22 out of 51 (43%)

• Measurement Location C: 5 out of 34 (15%)

• Construction Measurement Location: 13 out of 24 (54%)

• Overall 66 out of 150 (44%)




The choice of internal measurement locations at Location C in particular was therefore very limited. Although the take up rate at the other locations was high, of those willing, many were only available outside office hours and many were not contactable after the social survey teams had visited. The figures above therefore do not accurately reflect the number of respondents who were actually available to take part in the survey.

Additionally, time constraints meant it was not practicable to measure internally at all willing respondent properties. However, having some additional ‘spare’ willing respondents provided some flexibility in planning and programming the measurements.

For the purpose of the pilot study the survey teams aimed to measure internally at six properties at each measurement location. Obviously this was not possible at measurement location C.

Before commencing internal measurements, vibration teams asked respondents in which room most vibration was perceived and permission was asked to measure in this room. In many cases, survey teams were directed to a different room instead. The internal measurement location should be clarified before conducting a full exposure-response survey.

Once a suitable room had been identified a vibration measurement location was chosen as close to the centre of the room as possible, within the constraints of furniture layout, etc. Measurements were made at a single location in each room. A noise measurement location was chosen just off centre of the same room.

4.4.4.3 Satellite Measurement Duration For the railway sites, measurement durations were chosen to be long enough to record a minimum of 10 clear train events. This was dependent on the frequency of railway traffic and the presence of extraneous sources of vibration.

This made the typical measurement duration between 40 and 60 minutes long at one measurement location, but at others a higher frequency of train events meant the typical measurement duration was reduced to 30 minutes.

In cases where the magnitude of vibration was low or extraneous sources of vibration existed, such as internal vibration due to footfall etc, or road traffic near external locations, measurement durations were increased until vibration from a sufficient number of uncontaminated train events were recorded.

4.4.4.4 Measurement Record Pro-forma A measurement pro-forma was developed for the pilot study (see Appendix H) on which details of each measurement were recorded, including:

• Address of property;

• Control, external or internal measurement location;

• Transducer serial numbers and sensitivities;

• Data recorder serial number;

• Sound level meter calibration;

• Recording file references;

• Record start and stop time;

• Transducer mountings;

• Description of measurement location/property; and

• Observations, including clock times of noise/vibration events.




4.4.5 Fieldwork Lessons Learned The pilot study proved that the measurement protocol and questionnaire were very successful in obtaining the required data. The pilot study also aimed to identify potential areas for improvement which would benefit a future exposure-response study. The following sets out lessons learned at various stages in carrying out the pilot study that should be considered when taking this project to the next phase.

4.4.5.1 Site Selection Target properties were identified by the measurement team and marked on maps to inform the questionnaire surveyors which properties to approach. This was successful and enabled the vibration specialists to identify properties covering the required range of vibration levels. In some instances there was difficulty on the part of the questionnaire team in interpreting these maps and it was considered that it would be advantageous if precise details of the properties to be included in the survey could be provided.

The most successful route to identifying suitable construction sites was found to be speaking directly to the contractors or subcontractors to determine what work they have programmed. Generally they appear to have work on several sites spanning a few months ahead, so that they can allocate appropriate resources. This would allow a construction site monitoring programme to be developed to mirror the construction works. This may however require a greater degree of flexibility and mobility of questionnaire and measurement survey teams that was employed for the pilot study, to cover a wider geographical area.

4.4.5.2 Survey Planning and Programme It was found to be most efficient to wait until information from all respondents was received from the site survey team before attempting to organise internal measurements. This potentially leads to some short delay between social survey and measurements, but enabled the logistics to be optimised resulting in a more efficient measurement schedule.

A large proportion of respondents were only available after office hours, which made them difficult to contact initially and often unable to accommodate the measurement team. This has a significant effect on the total number of respondents who are both willing and able to participate in the measurement survey.

Prior to the pilot study, there was concern that the duration of the social survey could affect the responses received: if the time spent on site by interviewers was spread over a long period, or if multiple visits were made more than a few days apart, then news of the survey may pass through the neighbourhood. However, the interview surveys at measurement locations A and B were conducted over two consecutive weekends. Although the interview team reported that a small number of respondents had known that the survey was being conducted, it did not appear to have been generally discussed. The advantage of attending over two weekends was that it was possible to contact a larger number of residents who were able to take part at their convenience, as more time slots were available for the interviews.

4.4.5.3 Questionnaire Survey Due to the short time available for organising surveys at the construction measurement location, detailed maps of the survey buildings were not provided to the questionnaire interviewers. The interviewers therefore were required to use their judgement to decide which properties to approach. This location was successful in this respect, but it is better if the vibration experts advise the questionnaire survey team before they arrive on site. This will ensure that the correct properties are surveyed and will ensure time on site is maximised for obtaining interviews.

The response rate achieved in the pilot study was very encouraging. However, increasingly, neighbourhoods are posting notices rejecting “cold-calling” and this may have an effect on




future surveys. A letter drop serves to introduce the survey to the respondent and may lead to a higher response rate being achieved.

The interviewers and respondents found the questionnaire too long and repetitive. However the questionnaire was designed to be suitable for vibration from all possible sources, as requested by Defra. This meant that questions were included on sources of vibration not near to the residences. The inclusion of all sources in the questionnaire enabled a questionnaire suitable for all sources of vibration to be tested. Now that the pilot study has shown that the questionnaire was successful in extracting significant differences in response to different sources, the questionnaire can easily be tailored and reduced in length for future surveys on specific sources of vibration.

Question 9 asked for the number of floors or storeys of the property to be recorded. The interviewers found this confusing to complete, but the confusion arose from the inclusion of “Ground floor” in the specified responses which was included in error and should be removed in future questionnaires.

Questions 16, 17, 18, 20, 21 and 22 were difficult to complete and caused confusion for the interviewers, with each interpreting it in a different way. It would be better to make these open questions in future questionnaires and code them during the data entry phase.

Question 27 was thought to be very repetitive by both interviewers and respondents. In a future study, which might be focussed on a few sources of vibration, this question can be shortened by the removal of parts on sources not within a specified distance of the residences.

4.4.5.4 Measurement Survey The success rate in gaining agreement to access property for internal measurements was generally good and significantly higher than expected during the early stages of the project, indicating that the carefully planned approach taken by the questionnaire team was appropriate and successful. However, a number of respondents who had agreed by telephone to the survey forgot about appointments made to measure internally or to provide access for the control location. It may be helpful for future surveys to send a confirmation letter, email or text message to the occupier confirming arrangements, if time allows. Even so, making provision to approach other respondents at short notice to fill in time gaps may be advantageous. Alternatively, having “either/or” times with some householders, where convenient for them, may increase efficiency.

The measurement survey followed as soon as practicable after the questionnaire. For railway sites, the time between questionnaire and vibration surveys was not critical because of the unchanging nature of the source. For construction sites, this was more important, so that measured vibration was representative of that felt by respondents at the time that the questionnaire was conducted. The vibration team was able to be on site on the Monday morning following the weekend during which the questionnaire survey had been undertaken. Nevertheless, monitoring the vibration that was actually experienced from a piling site is difficult. Many householders reported to the measurement team that the vibration had been worse earlier in the works than it was at the time of the measurements. This is consistent with the piling programme, which started close to occupied property and moved away. Potential alternatives by which this might be improved are as follows:

• Further reduction in the time between the questionnaire survey and the measurements. The measurement team could be on site the same day as the questionnaire surveyors, at least for some properties, although this may present the risk of raising awareness in later interviewees, which may affect their answers to the questionnaire. Alternatively, it may be possible to train the questionnaire team to install a logger, where possible, immediately after completing the survey. This would, however, present the risks associated with unattended measurements.




• Calculation of previous actual levels based on the distances at earlier (worse) times. This could be done by determining a site law from data gathered at various distances from the piling to show how vibration attenuates with distance and then extrapolating to the closest position of the rig from the respondents’ properties. Caution would be required in extrapolating the data and it is suggested that this should only be done over a short distance relative to the range of distances from which data are acquired.

• Monitoring the whole construction process. Through liaison with the construction site operator or contractor, it may be possible to install a logger at a suitable position on the site boundary to record vibration for the entire duration of the works, or at least up until the time of the survey. Selection of the monitoring position(s) would need to consider the variable position of the source and the distribution of residential properties.

4.4.5.5 Equipment The equipment used for the survey was considered to be the best available off-the-shelf option and it proved very successful in this application. In particular, the ability to record a continuous time history at the required sampling frequency over 24 hours was essential. Furthermore, the small size and ease of use in the field made the Rion DA-20 ideal for the satellite measurements, both internally and externally.

Overall the equipment used provided the performance required. However, there are a number of limitations of the equipment, which are summarised as follows:

• Transducer sensitivities input to the DA-20 are only used for calibration within the DA-20 software; they are not saved to the raw data. Hence, calibration is still required during post-processing.

• Several flash card errors occurred during surveys, resulting in four flash cards becoming corrupt. Data recovery software was available but of the four measurements lost, only one could be recovered.

• If power to the DA-20 is lost while recording, all of the current recording is lost. This presents a particular risk for the control position logger.

• It was found that the 16 channel kit that was originally intended to be used for phase locked internal and external measurements took a disproportionate time to set up. An alternative using a simultaneous trigger to two Rion DA-20 loggers was used instead.

• Low frequency noise occurred on signals recorded using the PCB 356B18 accelerometers. The cause of this has not been established.

• The PCB 393B12 accelerometers require a shaker table for calibration, being too large for a portable calibrator. It is therefore not possible to calibrate the instruments on site; calibration of the instrumentation chain before going to site is therefore necessary.

• Sensitivity of transducers and the dynamic range of the instrumentation chain needs careful and site-specific consideration. At the low exposure railway sites, the PCB 356B18, which have a sensitivity of 0.1V/ms-2 were towards the limit of their useable range. Conversely, the more sensitive PCB 393B12 accelerometers, which have a sensitivity of 1V/ms-2 were almost too sensitive for use at the control position at the piling site and the output came close to limiting on the Rion DA-20.

4.4.5.6 Measurement Positions A standardised measurement distance of 2 m from the façade of the building closest to the vibration source was proposed for external vibration measurements. While this approach would be reliable and repeatable under some circumstances, in many cases it proved to be impracticable for reasons set out in Section 4.4.4.2.




Permission was not sought to perform external measurements. It is recommended that this should be requested even if the householder declined access for internal measurements. Without this permission it was difficult to get close to the façade of many properties, resulting in a limited number of positions to measure.

As far as possible, measurements were made at positions and on surfaces in accordance with the ANC’s Guidelines document. However, this needed to be balanced against the need to measure as close as possible to the intended standardised distance externally and, internally, the room requirement dictated the floor finish to some degree. Consequently, a variety of both internal and external measurement surfaces and mountings was encountered, which may account for some of the observed scatter in transfer functions. The measurement team need to be fully aware of this and seek to minimise the variation as much as practicable.

It had been intended to monitor in the centre of the habitable ground floor room that was closest to the vibration source. While this was the objective, the choice of room was not always at the behest of the survey team; the decision was often dictated by the property owner. This aspect should be discussed in advance with the householder, either by the questionnaire survey team when they first sought agreement, or when subsequent arrangements were being made by telephone. This would be expected to provide a greater consistency of approach, with consequential benefits to the data consistency.

The simultaneous use of two Rion DA-20 loggers enabled all internal and external vibration measurements to be used to derive external to internal transfer functions. This therefore provided transfer functions for a variety of different property types, but it was not possible to represent all different property types at each measurement location, particularly at measurement location B. It was therefore necessary to assume that transfer functions derived for one type of property were applicable to a range of properties. This is likely to be a sufficiently accurate approach where properties with similar floor slabs and assumed foundations could be identified and appropriate transfer functions from similar properties used.

Extrapolation between measurement positions and the use of single points to represent exposure at more than one property were useful efficiency tools. These approaches significantly reduced the time required on each site, which would be particularly beneficial should a larger follow up study be undertaken. Unfortunately, given the constraints imposed by budget and time, it was not possible to quantify the accuracy of this approach and to evaluate the degree of any errors that this might introduce. Further research on this would be a useful interim step towards the follow up study.

4.4.5.7 Quality of Data All instrumentation and procedures followed current best practice for noise and vibration measurement, within the constraints of each site. All equipment had been calibrated and had appropriate calibration certificates. Furthermore, measurement locations were carefully selected to minimise the risk that data may be corrupted by other sources of vibration than the principal source being studied. The quality of the data is therefore considered appropriate for the study and for an extension of the study.

Nonetheless, the following issues were identified that may affect the quality of the measurement data acquired and which therefore need to be considered when undertaking measurements of this nature:

• Unforeseen local vibration sources. Locations were carefully chosen to have a single dominant source of vibration affecting properties at all distances from that source. However, unexpectedly high volumes of traffic were sometimes experienced on roads near satellite positions (e.g. due to a road leading to a railway station or school), which




affected the local vibration environment. This was particularly a problem for properties at a long distance from the study vibration source.

• Internal vibration created by property owners. Residents were encouraged to limit their movement around their property during the survey but this was not always successful. Two internal surveys had to be written off due to this.

• Low levels of vibration. Levels of vibration were very low at measurement location C and at other sites at large distances from source. Although the instrumentation was capable of monitoring these low levels, the events were more susceptible to masking from other sources. The presence of other vibration sources therefore becomes increasingly important further from the principal vibration source.

• Accurate manual documentation of events during attended surveys was very important for use in post-processing. Where monitoring positions were not in line-of-sight and/or earshot of the vibration source (e.g. at measurement location C and other sites when property or topography shielded view of the vibration source) it was difficult to document events accurately.

4.4.6 Data Processing This section describes how the questionnaire and raw vibration data were processed to extract the information required for the subsequent analyses.

The analysis of the pilot study data was undertaken solely for the purposes of proving whether the acquisition process was successful and whether there may be any ordinal relationship between vibration magnitude and response. It was not an objective of this study to determine any exposure-response relationship any further than ascertaining whether the approaches used could demonstrate with any confidence that high exposure causes a greater response than low exposure.

The ordinal analyses used the mean unweighted peak and rms event acceleration levels and the overall period rms acceleration level. Equally, any other parameter might have been selected, such as velocity. Acceleration was chosen since it minimised the processing time as raw acceleration signals were recorded. Ideally, the Project Partners would have liked to analyse the data to obtain the full range of parameters including those currently used to assess vibration dose. Invariably, like other aspects of this study, limits had to be drawn and compromises had to be made. All efforts were made to focus on the critical elements of the study to validate the measurement protocol and questionnaire so that they could be used with confidence in the full dose-response study. The recorded data collected during the pilot study could be analysed to yield valuable information about the correlation of different parameters. Although it would have been nice to assess correlations with different parameters, including VDV, this assessment is not essential for the purposes of the pilot study and the critical tests set out in Section 4.1.

4.4.6.1 Questionnaire Data Responses from the questionnaires were encoded into SPSS as a spreadsheet to be used in statistical analysis. A 10% check on the accuracy of data entry was made, and corrections made where appropriate.

Using SPSS software, basic summary statistics including frequency tables and ordinal analyses would be undertaken on the social survey data. The sample population were described in terms of age, gender, occupation, property type, and distance from source of vibration.

Frequency tables were provided for all the questionnaire items to show proportions of residents experiencing, for example degree of annoyance, sleep disturbance, and concern over building damage.




For this survey, qualitative data, such as that from open questions, was analysed simply and commented on. Further analysis may take place in future surveys.

Some of the variables, such as questions 17 and 21 were time consuming to analyse because of the way they had been completed by the interviewers, which may be a significant issue for a large dataset as may be required by a follow up study. However, as has been suggested above, changing these to open questions may help the time taken for coding.

4.4.6.2 Measurement Data For the combined analysis of social survey and measured vibration data the measured data were required to be in a suitable format. As defined in the Interim Report the combined analysis was performed on both overall and event based noise and vibration indicators.

All recorded vibration measurement data were stored in raw form in wave (.WAV) format and no on-the-fly processing was performed. All overall and event based indictors were obtained through post-processing of the raw data.

External satellite position vibration data were gathered by taking snapshot measurements outside properties from where a questionnaire response had been obtained. By interpolating between measurement positions, or by using a single satellite position to be representative of more than one property, the number of case studies was increased. The total (i.e. 24 hours for railway sites; daily working hours for construction site) external vibration exposure for each property was then obtained by scaling the satellite data using transfer functions from control to external positions. The internal vibration exposure was then calculated using a transfer function for external to internal vibration. An analysis stage was therefore required to convert post-processed overall and event data, to vibration exposure experienced by respondents.

Post-processing Before processing the measured data, each recording was individually checked for quality using the RION DA-20 signal display software. Additionally all the appropriate transducers’ sensitivities and gains, stored with each measurement file, were checked and recorded for input to the post-processing software, Prosig DATS.

The noise recordings were also calibrated using the RION DA-20 software.

All measurement data from control, external satellite and internal measurement positions has been post-processed using the Prosig DATS analysis software. A description of the functionality and application of this software to this project are provided in Appendix I.

Review of post-processing approach The Rion DA-20 stored a continuous time history of all data at each measurement position. From this it was necessary to extract the required parameters for each event of interest, for which an automated process for identifying events was required. For discrete events such as train pass bys, this determination was relatively straightforward. Within the Prosig DATS analysis software, threshold levels can be specified to define the start and end points of events supported as necessary by synchronised clock information. This enabled the automated extraction and analysis of the required data. For continuous sources such as construction or industrial processes where the vibration level may continue above ambient for long periods, albeit at a variable level, the signal needs to be subdivided into contiguous sections in order to extract a set of numerical values that can be analysed. Contiguous one minute samples were used for evaluation of data obtained during the pilot study.

The combination of the Rion DA-20 and the Prosig DATS software worked very well for this project. Although a number of issues arose through processing the data using the DATS software, it is considered that most or all of these could be resolved with further development of the scripts. The software was acquired specifically for use on this project




and is much more powerful than it has been possible to realise through the Project Partners’ limited experience of it. Issues identified as requiring further attention are as follows:

• The volume of data generated may present a problem with storage for a larger study.

This is due to the way in which the DATS software works: reference signals are created from the raw data file to perform triggering, which is used to extract each event from the continuously recorded time history. This processing causes an approximately 10 fold increase in the file size. For 24 hour signals from the control position, the 1.2Gb file therefore results in 16Gb of processed data. This can be optimized with further development of post-processing, however temporary files of this size still must be created during processing.

• For the pilot study, internal, external and control measurements were processed separately. Although the data files were time synchronised, the trigger start and stop points will be different in all three cases for the same event, which may have some implication for the calculated transfer functions. It is recommended that simultaneous analysis of internal and external measurements is undertaken, since these were phase locked. However simultaneous analysis of external (or internal) and control position data is not possible as it is not possible to phase lock the data.

• High frequency of vibration events (e.g. two trains close together) results in merging of events after post-processing, making them difficult to extract automatically.

• Quality assurance of data is an issue, due to the automated extraction and analysis of the data. This is particularly a concern where low levels of vibration are measured, since there is a greater risk of the event being contaminated by extraneous sources or false triggers having occurred. This has consequential implications on the validity of transfer functions. It is necessary to check individual events, which is very time consuming: for example, there were more than 400 events in measurement location B in the 24 hour period.

DATS is very powerful software, but limited time restricted the ability to adapt it for the purposes of this study. Further development is recommended.

4.5 Analysis and Results

This section of the report describes the analyses of the data undertaken with the objective of verifying the questionnaire and measurement protocol. The results of the analyses are not intended to infer any generally applicable relationships between any of the variables, except insofar as they provide the validation for the methods used. For brevity, the raw results from the pilot study are not included in this report; a separate document has been prepared containing summaries of the results of the questionnaire and the measurement surveys.

It must be borne in mind that the key issues being tested by the pilot study related to the veracity of the questionnaire and the measurement protocol, The combined analysis (presented in Section 4.5.3), for which the data presented in this section was required, was a means to validation of the pilot study. The data presented in this section should therefore not be considered as generally representative of interrelationships or dependencies, as the sample is too small to draw any robust conclusions about relationships between the variables presented.

4.5.1 Analysis of Social Survey Data The aim of the questionnaire data analysis was to determine whether implementation of the questionnaire in an extensive study would produce data that would be reliable and suitable for determining vibration exposure-response relationships.

The data were examined to determine whether the questionnaire:




• resulted in reliable interpretation by the interviewees of terminology such as vibration and noise;

• elicited responses that produce data suitable for statistical analysis;

• obtained data which are reliable, repeatable and comparable;

• elicited responses of residents to vibration and noise from a range of sources in close proximity to one or more of the sources of building vibration;

• minimised response errors in the resulting data so as to provide an accurate record of the true response of residents; and

• provided data suitable for determining dose-response relationships.

4.5.1.1 Statistical Methods

Factor analysis Factor analysis may be employed to assist in the identification of underlying variables, or factors, that explain the pattern of correlations within a set of observed variables. Factor analysis is in data reduction to identify a small number of factors that explain most of the variance observed in a larger number of variables and can be used to generate hypotheses regarding causal mechanisms for subsequent analysis.

The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy tests whether the partial correlations among variables are small. Bartlett's test of sphericity tests whether the correlation matrix is an identity matrix, which would indicate that the factor model is inappropriate. The KMO should be greater than 0.5 for a satisfactory factor analysis to proceed. If the Bartlett's test of sphericity is significant this means the correlation matrix is not an identity matrix and therefore factor analysis is appropriate.

KMO and Bartlett’s tests were applied to responses from questions for which factor analysis may be useful in a larger study to group variables.

Internal consistency: Spearman’s correlation coefficient for agreement between annoyance scales. Spearman’s rank correlation coefficient was applied to test consistency between ratings of bother, annoyance or disturbance in response to vibration or noise from the same source. Spearman’s rank correlation coefficient was also determined to test the consistency between responses obtained using different scales to obtain ratings of annoyance from the same source.

Internal consistency: agreement between perception and annoyance. Consistency between responses to questions on perception of vibration and annoyance/disturbance were tested. Checks were made to determine whether those respondents who indicated that they did not feel vibration and that they did not hear or see vibration gave low ratings of annoyance from vibration.

To test for differences between annoyance ratings The Wilcoxon matched-pairs signed ranks test was employed to test for differences between annoyance ratings for noise and vibration from the same source. The Wilcoxon test was also applied to test for differences between ratings in response to two questions on annoyance caused by the same source of noise.

To determine whether questions on annoyance from vibration and noise resulted in differences in responses for different sources of vibration and noise, the Friedman test was applied




4.5.1.2 Analysis of Railway Sites Social Survey Results

Satisfaction with neighbourhood

Factor analysis

The KMO and Bartlett tests were applied to the data from Q15 which obtains ratings for various aspects of the neighbourhood to determine whether the data are suitable for factor analysis. The Keiser-Meyer-Olkin (KMO) measure of sampling adequacy was 0.320 and the Bartlett’s test of sphericity was not significant (p = 0.312) indicating that the data from Q15 are unsuitable for factor analysis. Forty-five of the respondents were unable to provide ratings for the standards of the local schools. When all ratings for the standards of the local schools were excluded from the analysis, the results of the KMO and Bartlett tests indicated that the remaining data are suitable for factor analysis (KMO = 0.734, Bartlett’s test of sphericity: p < 0.001).

Principal components factor analysis with varimax rotation was conducted on the data from Q15 with the extraction of one component having eigen values greater than unity. The variance accounted for by the extracted component was 42% of the total variance. The extraction of only one component with an eigen value greater than unity suggests that there was little difference between the ratings of the seven different aspects of the neighbourhood and that the seven parts to Q15 might be reduced to one general question on neighbourhood satisfaction.

The results of the KMO and Bartlett tests show that Q15 yields data suitable for factor analysis. The results of the factor analysis show that it may be possible to reduce the number of parts in the question with little loss of information.

Vibration questions

Factor analysis

The KMO and Bartlett tests were applied to annoyance ratings caused by feeling vibration from various sources in response to Q27 to determine whether the data are suitable for factor analysis. For analysis which included ratings of ‘1’ (not at all) for which respondents indicated that they did not feel the vibration, the Keiser-Meyer-Olkin (KMO) measure of sampling adequacy was 0.547 and the Bartlett’s test of sphericity was significant at p < 0.001, indicating that the data from Q27 are suitable for factor analysis.

Principal components factor analysis with varimax rotation was conducted on the data from Q27 with the extraction of three components having eigen values greater than unity. Responses from Q27E (annoyance from quarrying) was removed from the analysis because the variance of the data was zero. The variance accounted for by the three extracted components was 60% of the total variance.

The first factor appears to relate to vibration from road vehicles, road works and construction. The second component relates to vibration from aircraft and trains (over-ground and under-ground) and the third component relates to vibration from footfalls, door slamming and domestic appliances (inside the home and in neighbouring homes).

Factor analysis on data in which ratings of ‘1’ (not at all) were only included if respondents said they felt vibration, required the inclusion of data only from questions on cars, overground trains and on footfalls, slamming doors, domestic appliances in their own and neighbouring homes. Data from questions on vibration caused by all other sources were removed from the analysis as the variance of each question response was zero. The Keiser-Meyer-Olkin (KMO) measure of sampling adequacy was 7.49 and the Bartlett’s test of sphericity was not significant (p = 0.278), indicating that the data from Q27 are unsuitable for factor analysis if the “not felt” responses were removed.




The results of the KMO and Bartlett tests show that Q27 yields data suitable for factor analysis. In this study, factor analysis required the inclusion of the lowest rating from respondents who did not feel vibration from some sources. The results of the factor analysis suggest that it may be possible to reduce the number of parts to the question with little loss of information by grouping some question parts together.

Internal consistency

Consistency between responses to questions on perception of vibration and annoyance/disturbance was investigated by examining whether those respondents who indicated that they did not feel vibration in Questions 23, gave low ratings of the effect of vibration in Q27 and that they did not hear or see vibration in Questions 28, gave low ratings of the effect of vibration in Question 31.

Forty-two respondents (33%) said they did not feel vibration of the floor, chair, bed, or other in Q23. Of those, 24 gave the lowest annoyance rating, “not at all” annoyed, in Q27 and 18 gave ratings of either “a little” or “moderately” annoyed from feeling vibration caused by specific sources.

Fifty-six percent of respondents said they did not hear or see vibration in response to Q28, of which half of those gave ratings in Q31 of between “a little” and “extremely” annoyed for hearing or seeing vibration from specific sources.

The results show that some respondents who gave negative responses to general questions on feeling vibration gave positive responses to questions on specific sources. Questions on specific sources may assist respondents in recalling events. This finding highlights the importance of avoiding the use of general questions to filter out respondents from further more specific questions. Such filters may result in the loss of useful information and inaccurate conclusions on the level of annoyance caused by specific events.

Test of a correlation between annoyance ratings from different vibration questions

Ratings of annoyance caused by feeling vibration from overground trains (Q27c, 5-point scale) were highly correlated with annoyance ratings caused by feeling, seeing and hearing train vibration (Q35, 7-point scale); Spearman’s rank correlation coefficient rsp = 0.792, p < 0.001, N = 123. Ratings of annoyance caused by feeling vibration from overground trains were also highly correlated with annoyance ratings caused by feeling, seeing and hearing train vibration for a correlation conducted after excluding subjects who indicated that they did not feel vibration in Q27c; rsp = 0.757, p < 0.001, N = 81.

Ratings of annoyance caused by feeling vibration from overground trains (Q27c) were highly correlated with annoyance ratings from hearing or seeing vibrating, rattling, shaking or swaying caused by trains (Q31c); rsp = 0.709, p < 0.001 for a correlation conducted after the exclusion of eleven subjects who indicated that they did not hear or see vibration in response to Q31c. Ratings of annoyance caused by feeling vibration from overground trains (Q27c) were also highly correlated with annoyance ratings caused by seeing and hearing train vibration (Q31c) for a correlation conducted after the exclusion of subjects who indicated that they did not feel vibration or did not hear or see vibration; rsp = 0.717, p < 0.001, N = 81.

The results indicate that increased levels of annoyances from feeling train vibration were associated with increased annoyance levels from hearing and seeing vibration. Exposure to these primary and secondary effects is expected to be correlated since increased vibration of the floor is likely to be accompanied by, for example, increased vibration of windows.

Correlation between annoyance from these primary and secondary is therefore expected and demonstrates consistency in response.




Test of differences between annoyance ratings from vibration questions

There was no significant difference between the annoyance ratings of feeling vibration from trains (Q27c) and seeing and hearing vibration from trains (Q31c); p = 0.384 (Wilcoxon matched-pairs signed ranks test). Eleven subjects who indicated that they did not hear or see vibration in response to Q31c were excluded from the analysis (N = 114). However, when ratings of ‘1’ (not at all) from subjects who indicated they did not feel vibration from overground trains (Q27c) were excluded from the analysis, the difference was significant (p = 0.015, N = 81). The mean annoyance rating of those feeling vibration from overground trains was 2.67 on a 5-point scale, and for those hearing or seeing vibration from trains was 2.12. The design of the questionnaire was successful in eliciting significant differences in annoyance from the primary effects of feeling vibration and from the secondary effects of hearing or seeing things vibrate, rattle, shake, or sway.

Vibration source Median annoyance rating

Mean annoyance rating

Cars, lorries, buses or other road vehicles 1 1.4 Aircraft 1 1.1 Surface trains 2 2.1 Subsurface trains 1 1.1 Quarrying or mining 1 1.0 Construction 1 1.1 Roadworks 1 1.2 Footfalls, slamming doors, domestic appliances inside your home 1 1.3

Footfalls, slamming doors, domestic appliances in neighbouring homes 1 1.4

Table 4.1 - Median and mean annoyance caused by feeling vibration from different sources (Q27): pilot study railway sites.

Table 4.1 shows the median and mean annoyance ratings from feeling vibration from each source. Vibration from overground trains caused the most annoyance with median and mean ratings of 2 and 2.1 respectively (2 = “A little“ annoyed). Median and Mean ratings of annoyance from all other sources of vibration were about ‘1’ (“Not at all“ annoyed). The difference between the annoyance ratings caused by feeling vibration was highly significant between the nine different sources of vibration; p < 0.001, χ2 = 229 (Friedman). Ratings of ‘1’ (not at all) from respondents not feeling vibration were included in the analysis. Exclusion of respondents not feeling vibration resulted in too few cases for analysis. The results suggest that the survey was successful in showing differences in responses to different vibration exposures.

Noise questions Factor analysis

The KMO and Bartlett tests were applied to annoyance ratings caused by noise from various sources in response to Q41 to determine whether the data are suitable for factor analysis. Responses from Q41e (annoyance from quarrying) were removed from the analysis as the variance of the data was zero. The Keiser-Meyer-Olkin (KMO) measure of sampling adequacy was 0.587 and the Bartlett’s test of sphericity was significant at p < 0.001, indicating that the data from Q41 are suitable for factor analysis.

Principal components factor analysis with varimax rotation was conducted on the data from Q41 with the extraction of two components having eigen values greater than unity. The variance accounted for by the three extracted components was 48% of the total variance.

The first factor appears to relate to noise from road vehicles, aircraft, overground trains, construction, road works and footfalls, door slamming and domestic appliances (inside the




home and in neighbouring homes). The second component seems to relate to noise from aircraft and trains in tunnels.

The results of the KMO and Bartlett tests show that Q41 yields data suitable for factor analysis. The results of the factor analysis suggest that it may be possible to reduce the number of parts to the question with little loss of information by grouping some question parts together.

Test of a correlation between ratings from different noise questions

Ratings of annoyance caused by noise from overground trains (Q41c, 5-point scale) were highly correlated with annoyance ratings caused by noise from passing trains (Q44a, 7-point scale); Spearman’s correlation coefficient rsp = 0.726, p < 0.001. Two respondents were excluded from the analysis as they indicated that they did not hear noise from trains in response to Q41c.

Test of differences between annoyance ratings from noise questions

Table 4.2 shows the median and mean annoyance ratings from noise for each source in response to Q41. Noise from overground trains caused the most annoyance with median and mean ratings of 2 and 2.1 respectively (2 = “A little” annoyed). Median and Mean ratings of annoyance from all other sources of vibration were about 1 (“Not at all” annoyed). The difference between the annoyance ratings caused by noise was highly significant between the nine different sources of vibration; p < 0.001, χ2 = 229 (Friedman). Twenty-four respondents were excluded from the analysis as they indicated that they did not hear noise from one or more sources in response to Q41. The results suggest that the survey was successful in eliciting differences in responses to different noise exposures.

Comparison of noise ratings for passing trains from different questions

Respondents were asked to provide ratings of annoyance caused by noise from trains in two questions, Q41c and Q44a, using the same 7-point scale. The difference between annoyance ratings provided in response to the two questions was highly significant; p < 0.001 (Wilcoxon). The median response for both questions was ‘2’ (“a little”). The mean rating for Q41c (How bothered, annoyed or disturbed are you by noise from overground trains?) was 2.16. The mean rating for Q44a (How bothered, annoyed or disturbed are you by noise from the passage of nearby trains?) was 2.55. The most likely explanation for the higher mean rating in response to Q44a is that respondents were influenced by the



Cars, lorries, buses or other road vehicles 2 1.8

Aircraft 1 1.5 Overground trains 2 2.1 Trains in tunnels 1 1.1 Quarrying or mining 1 1.0 Construction 1 1.2 Roadworks 1 1.1 Footfalls, slamming doors, domestic appliances inside your home 1 1.2


Table 4.2 - Median and mean annoyance caused by noise from different sources (Q41): pilot study railway sites.




preceding question (Q43) in which respondents were asked how sensitive they are to noise. Sensitivity ratings on a 7-point scale ranged from 1 to 7 with a mean rating of 3.5. Twenty-nine respondents (23%) rated their sensitivity as 5, 6, or 7. It is possible that those who rated their sensitivity to noise at the higher end of the scale may then have responded to the following question on annoyance from noise (Q44a) with a higher rating than if they had not been asked about their sensitivity to noise. The placing of questions on sensitivity at the end of the questionnaire would eliminate any influence of the sensitivity question on responses to other questions.

Comparison of noise and vibration ratings

There was no significant difference in the annoyance ratings from Q27c (annoyance from feeling vibration from overground trains) and Q41c (annoyance from noise from overground trains) using a 5-point rating scale; p = 0.48, N = 122 (Wilcoxon), excluding responses from those not hearing noise from trains. The mean annoyance from vibration (Q27c) was 2.16 and from noise (Q41c) was 2.10. There was also no significant difference in the annoyance ratings from feeling vibration from overground trains and from noise from overground trains when responses were excluded from those not feeling vibration; p = 0.14, N = 82.

There was no significant difference in the annoyance ratings from Q35 (annoyance from feeling, hearing or seeing vibration from the passage of nearby trains) and Q44a (annoyance from noise produced by the passage of nearby trains) using a 7-point rating scale; p = 0.31 (Wilcoxon). The mean annoyance from Q35 was 2.60 and from Q44a was 2.55. The results suggest that respondents experienced similar levels of annoyance caused by vibration and noise from trains.

Summary of analysis of railway sites social survey data The questionnaire data from the railway sites were shown to be suitable for the required analysis. Results of KMO and Bartlett’s tests indicate that ratings of aspects of the neighbourhood were of the appropriate form for satisfactory factor analysis to proceed. Annoyance ratings for vibration and noise from various sources were also shown to be suitable for factor analysis. Questions on annoyance from vibration and noise yielded ratings on which correlation coefficients may be determined. Significant correlations between annoyance from primary and secondary vibration effects indicated response consistency. Annoyance ratings in response to questions on vibration and noise from different sources were suitable for the application of the Friedman test and showed that the railway questionnaire was successful in extracting differences in responses to different vibration and noise exposures from a range of sources in close proximity to one or more of the sources of building vibration. Application of Wilcoxon Signed Ranks test showed significant differences in annoyance from primary and secondary effects of vibration, indicating that the questionnaire enabled separate ratings of the two effects to be obtained. Thus, the results showed that respondents understood the distinction between terminology such as vibration and noise. Questions on sensitivity to noise and vibration should be placed at the end of the questionnaire to avoid influence on responses to following questions.

4.5.1.3 Analysis of Construction Site Social Survey Results

Vibration questions Factor analysis

The KMO and Bartlett tests were applied to annoyance ratings caused by feeling vibration from various sources in response to Q27 to determine whether the data are suitable for factor analysis. After exclusion of responses from questions on quarrying, aircraft, over-ground and under-ground trains which had zero variance, the Keiser-Meyer-Olkin (KMO) and the Bartlett’s test of sphericity indicated that the data from Q27 are suitable for factor analysis (KMO = 0.505, p = 0.022. The analysis included ratings of ‘1’ for which respondents indicated that they did not feel the vibration.





The first factor appears to relate to vibration from footfalls, slamming doors and domestic appliances (in own and neighbouring homes). The second component relates to vibration from construction, cars and road works.

Exclusion of those who did not feel vibration from the analysis yielded insufficient cases for factor analysis.

Internal consistency

Consistency between responses to questions on perception of vibration and annoyance/disturbance was investigated by examining whether those respondents who indicated that they did not feel vibration in Questions 23, gave low ratings of the effect of vibration in Q27 and that they did not hear or see vibration in Questions 28, gave low ratings of the effect of vibration in Question 31.

Thirteen respondents (54%) said they did not feel vibration of the floor, chair, bed, or other in Q23. Of those, four gave the lowest annoyance rating in Q27 (“not at all” annoyed) and nine respondents gave ratings from “a little” to “extremely” annoyed from feeling vibration caused by specific sources.

Sixteen respondents (67%) said they did not hear or see vibration in response to Q28, of which eleven gave ratings in Q31 of between “a little” and “extremely” annoyed for hearing or seeing vibration from specific sources.

The results show that some respondents who gave negative responses to general questions on feeling vibration gave positive responses to questions on specific sources, highlighting the importance of avoiding the use of general questions to filter out respondents from further more specific questions.

Test of a correlation between annoyance ratings from different vibration questions

Ratings of annoyance caused by feeling vibration from construction (Q27f, 5-point scale) were highly correlated with annoyance ratings caused by feeling, seeing and hearing vibration from nearby heavy construction machinery (Q35, 7-point scale); Spearman’s rank correlation coefficient rsp = 0.555, p = 0.005, N = 24. However, ratings of annoyance caused by feeling vibration from construction were not significantly correlated with annoyance ratings caused by feeling, seeing and hearing construction vibration for a correlation conducted after excluding subjects who indicated that they did not feel vibration in Q27f; rsp = 0.135, p < 0.618, N = 16.

Ratings of annoyance caused by feeling vibration from construction (Q27f) were highly correlated with annoyance ratings from hearing or seeing vibrating, rattling, shaking or swaying caused by construction (Q31f); rsp = 0.675, p < 0.001, N = 24. Correlation of ratings of annoyance caused by feeling vibration from construction (Q27f) with annoyance ratings caused by seeing and hearing construction vibration (Q31c) after the exclusion of subjects who indicated that they did not feel vibration was marginally significant; rsp = 0.480, p = 0.060, N = 16.

The results indicate that increased levels of annoyances from feeling construction vibration were associated with increased annoyance levels from hearing and seeing vibration. Exposure to these primary and secondary effects is expected to be correlated since increased vibration of the floor during the construction activities is likely to be accompanied by, for example, increased vibration of windows and demonstrates consistency in response.

Test of differences between annoyance ratings from vibration questions




There was no significant difference between the annoyance ratings of feeling vibration from construction (Q27f) and seeing and hearing vibration from construction (Q31f); p = 0.159, N = 24 (Wilcoxon matched-pairs signed ranks test). When ratings of ‘1’ (not at all) from subjects who indicated they did not feel vibration from overground trains (Q27f) were excluded from the analysis, the difference was also not significant (p = 0.098, N = 16). A more extensive study of annoyance caused by vibration from construction activities, involving a larger number of respondents, is required to determine whether the design of the questionnaire may elicit significant differences in annoyance from primary and secondary effects.

Table 4.3 shows the median and mean annoyance ratings from feeling vibration from each source. Vibration from construction caused the most annoyance with median and mean ratings of 3 and 2.4 respectively (2 = ‘A little’, 3 = ‘moderately’ annoyed). The difference between the annoyance ratings caused by feeling vibration was highly significant between the nine different sources of vibration; p < 0.001, χ2 = 61 (Friedman). Ratings of ‘1’ (not at all) from respondents not feeling vibration were included in the analysis. Exclusion of respondents not feeling vibration resulted in too few cases for analysis. The results suggest that the survey was successful in showing differences in responses to different vibration exposures.

Noise questions Factor analysis

The KMO and Bartlett tests were applied to annoyance ratings caused by noise from various sources in response to Q41 to determine whether the data are suitable for factor analysis. Responses from Q41c, d and e (annoyance from over-ground and under-ground trains and quarrying) were removed from the analysis as the variance of the data for each was zero. The Keiser-Meyer-Olkin (KMO) measure of sampling adequacy was 0.625 and the Bartlett’s test of sphericity was significant at p = 0.001, indicating that the data from Q41 are suitable for factor analysis.


The first factor appears to relate to noise from road vehicles and footfalls, door slamming and domestic appliances (inside the home and in neighbouring homes). The second component seems to relate to noise from aircraft and construction activities.




Aircraft 1 1.0 Overground trains 1 1.0 Trains in tunnels 1 1.0 Quarrying or mining 1 1.0 Construction 3 2.4 Roadworks 1 1.2 Footfalls, slamming doors, domestic appliances inside your home

1 1.8

Footfalls, slamming doors, domestic appliances in neighbouring homes

1 1.5

Table 4.3 - Median and mean annoyance caused by feeling vibration from different sources (Q27): pilot study construction site.




The results of the KMO and Bartlett tests show that Q41 yields data suitable for factor analysis. The results of the factor analysis suggest that it may be possible to reduce the number of parts to the question with little loss of information by grouping some question parts together.

Test of a correlation between ratings from different noise questions

Ratings of annoyance caused by noise from construction (Q41f, 5-point scale) were highly correlated with annoyance ratings caused by noise from piling or pounding machines (Q44a, 7-point scale); Spearman’s correlation coefficient rsp = 0.728, p < 0.001.

Test of differences between annoyance ratings from noise questions

Table 4.4 shows the median and mean annoyance ratings from noise for each source in response to Q41. Noise from construction caused the most annoyance with median and

mean ratings of 3 and 2.4 respectively (2 = “A little”, 3 = “moderately” annoyed). The difference between the annoyance ratings caused by noise was highly significant between the nine different sources of vibration; p < 0.001, χ2 = 80 (Friedman). The results suggest that the survey was successful in eliciting differences in responses to different noise exposures.

Comparison of noise ratings for construction from different questions

Respondents were asked to provide ratings of annoyance caused by construction in two questions, Q41f and Q44a, using the same 7-point scale. The difference between annoyance ratings provided in response to the two questions was highly significant; p < 0.001 (Wilcoxon). The median annoyance ratings for Q41f was ‘3’ (“moderately”) and for Q44a was ‘4’ (“very”). The mean rating for Q41f (How bothered, annoyed or disturbed are you by noise from construction?) was 4.33. The mean rating for Q44a (How bothered, annoyed or disturbed are you by noise from pounding or piling machines?) was 2.75. The most likely explanation for the higher mean rating in response to Q44a is that respondents were influenced by the preceding question (Q43) in which respondents were asked how sensitive they are to noise. Sensitivity ratings on a 7-point scale ranged from 1 to 7 with a mean rating of 3.7. Eight respondents (30%) rated their sensitivity as 5, 6, or 7. It is possible that those who rated their sensitivity to noise at the higher end of the scale may then have responded to the following question on annoyance from noise (Q44a) with a higher rating than if they had not been asked about their sensitivity to noise. The placing of questions on




Aircraft 1 1.2 Overground trains 1 1.0 Trains in tunnels 1 1.0 Quarrying or mining 1 1.0 Construction 3 2.8 Roadworks 1 1.5 Footfalls, slamming doors, domestic appliances inside your home 1 1.8


Table 4.4 - Median and mean annoyance caused by noise from different sources (Q41): pilot study construction site.




sensitivity at the end of the questionnaire would eliminate any influence of the sensitivity question on responses to other questions.

Comparison of noise and vibration ratings

There was no significant difference in the annoyance ratings from Q27f (annoyance from feeling vibration from construction) and Q41f (annoyance from construction noise) using a 5-point rating scale; p = 0.23, N = 24 (Wilcoxon). The mean annoyance from vibration (Q27f) was 2.42 and from noise (Q41c) was 2.75. There was also no significant difference in the annoyance ratings from feeling construction vibration and from construction noise when responses were excluded from those not feeling vibration; p = 0.29, N = 16.

The difference between ratings in response to Q35 (annoyance from feeling, hearing or seeing vibration from heavy construction machinery) and Q44a (annoyance from noise produced by piling or pounding machines) was highly significant; p = 0.002 (Wilcoxon). The mean annoyance from vibration (Q35) was 3.13 and from noise (Q44a) was 4.33. The results suggest that respondents were more annoyed by noise than by vibration caused by construction.

Summary of analysis of construction site social survey data The questionnaire data from the construction site were shown to be suitable for the required analysis. Results of KMO and Bartlett’s tests indicate that annoyance ratings for vibration and noise from various sources were of the appropriate form for satisfactory factor analysis to proceed. Questions on annoyance from vibration and noise yielded ratings on which correlation coefficients may be determined. Significant correlations between annoyance from primary and secondary vibration effects indicated response consistency. Annoyance ratings in response to questions on vibration and noise from different sources were suitable for the application of the Friedman test and showed that the construction questionnaire was successful in extracting differences in responses to different vibration and noise exposures from a range of sources in close proximity to one or more of the sources of building vibration. Application of Wilcoxon Signed Ranks test showed no significant difference in annoyance from primary and secondary effects of vibration. A more extensive study of annoyance caused by vibration from construction activities, involving a larger number of respondents, is required to determine whether the design of the questionnaire may elicit significant differences in annoyance from the primary effects of feeling vibration and from the secondary effects of hearing or seeing things vibrate, rattle, shake, or sway

4.5.1.4 Pilot Study Social Survey Analysis Conclusions The aims of the questionnaire analysis, as outlined in Section 4.5.1 above, are addressed below. These findings represent some of the principle conclusions drawn from the pilot study and are key to the validation of the approaches taken to the questionnaire and the measurement protocol.

Reliable interpretation of terminology such as vibration and noise

Significant differences in annoyance from the primary effects of feeling vibration and the secondary effects of seeing and hearing vibration indicated that the questionnaire enabled separate ratings of the two effects to be obtained. Thus, the results showed that respondents understood the distinction between terminology such as vibration and noise as used in the questionnaire.

Responses produce data suitable for statistical analysis

Data obtained from questions on the neighbourhood and on annoyance from noise and vibration were suitable for the required factor analysis. Questions on annoyance from vibration and noise yielded ratings on which Spearman’s rank correlation coefficients may be determined. Annoyance ratings in response to questions on vibration and noise from




different sources were suitable for the application of the Friedman and Wilcoxon Signed Ranks test.

Data are reliable, repeatable and comparable

Significant correlations between annoyance from primary and secondary vibration effects and between annoyance from noise and vibration from related exposures indicated response consistency and reliability. Questions on sensitivity to noise and vibration should be placed at the end of the questionnaire to avoid influence on responses to following questions.

Responses are obtained to vibration and noise from a range of sources in close proximity to one or more of the sources of building vibration

Analysis of ratings in response to questions on vibration and noise from different sources were suitable showed that the railway and construction site questionnaires were successful in extracting differences in responses to different vibration and noise exposures from a range of sources in close proximity to one or more of the sources of building vibration.

Data suitable for determining exposure-response relationships.

The suitability of the data for determining exposure-response relationships is addressed in Section 4.5.3 below.

4.5.2 Analysis of Vibration Measurement Data During the post-processing phase, vibration indicators were extracted from the raw acceleration data for each measurement location. Further interpolation and extrapolation of the data was required to predict exposure to vibration for every respondent’s dwelling.

The mechanics by which data were extracted is described in Appendix J. The software has the ability also to extract other parameters as may be required by the full exposure-response study, but which were not required for the pilot study. Examples are provided in Appendix C.

This section of the report describes how the analysis processes used have demonstrated the ‘fitness for purpose’ of the measurement protocol.

4.5.2.1 Control - External Transfer Functions External measurements at satellite positions were made over a ‘snapshot’ period that was simultaneous with the control position measurements. This enabled a transfer function between the two positions to be calculated for each event over that period. The transfer functions were then used to calculate the total exposure at the satellite position based on scaling the total exposure measured at the control location.

The pilot study verified that this approach provided a practicable and time efficient way of assessing total exposure outside each property. For the purposes of the pilot study, time domain transfer functions were calculated for each event in each of the vibration axes and for rms and peak acceleration. An average of the transfer functions was then taken to obtain x, y and z-axis vibration transfer functions for both rms and peak acceleration. While the use of single parameter transfer functions was acceptable for the purposes of the pilot study, it may be possible to reduce the scatter in the calculated control-external transfer functions by adopting a more complex approach, such as through calculation in third octave bands, or through the use of other vibration parameters.

The external transfer functions have been calculated by processing the recordings at the control position and the satellite position separately. This means that the start point, end point and duration (based on the time between evaluated dB down points) of events differ between the two measurements. This is an accepted way of determining transfer functions and is reliable for discrete events (such as trains) or events that are continuous over a long




period (such as piling). However, such an approach can introduce a source of variability that might otherwise be avoided if the recordings were analysed simultaneously. In the larger dose-response study it is recommended that the analysis methods are developed further so that the recordings can be analysed simultaneously in an efficient manner.

4.5.2.2 External - Internal Transfer Functions To predict the vibration within properties and to evaluate the hypothesis that response to vibration inside a property can be evaluated from the vibration measured outside, simultaneously recorded internal and external data were used to calculate transfer functions. These transfer functions were calculated in the same way as for the control-external transfer functions.

This is an established technique and was found to be reliable, provided that neither the internal or external signal was corrupted by vibration from extraneous sources. In some cases, especially at the low exposure sites, the quality of internal data resulted in few events remaining in the analysis. Increasing the measurement duration at each satellite position would reduce the risk of having insufficient data for analysis.

There was a large scatter of results for the transfer functions. This scatter was similar for both the railway and constructions sites. This scatter may be reduced by extracting external and internal events simultaneously, so that they were effectively phase-locked transfer functions.

4.5.2.3 Interpolation and Extrapolation of Data As described above, vibration measurements were not made at all properties from which a questionnaire response had been obtained. To increase the case studies available for analysis, and to minimise the number of measurement positions, the measurement protocol allows for interpolation and extrapolation of the data, where:

• A contiguous line of properties exists and external vibration measurements have been made at each end of the line; or

• Measured data for a nearby property exists where similar vibration conditions can be reasonably expected (i.e. from an adjacent property)

The measurement surveys were carried out shortly after the questionnaire surveys were completed. This allowed the measurement surveys to be planned in such a way as to maximise efficiency. Representative sample locations were chosen for lines of properties where there was a cluster of respondents. The external vibration at properties along the line was predicted using a site law. A number of site laws for each survey location were developed from the transfer functions measured at each end of a line of properties and interpolating to find the transfer functions for properties within the line. This proved to be an effective way of maximising the amount of data available for analysis without requiring an excessive amount of site work. This will be especially important for ensuring the cost efficiency of the follow on exposure-response study.

Similarly, a judgement was made whether properties adjacent or close to external satellite measurements positions were exposed to similar levels of vibration. If so the measured control-external transfer function is assumed to apply for both properties. This assumption is only valid for cases where the distance between the two properties is much less than the distance from the source of vibration.

4.5.2.4 Calculation of Total Exposure The 24 hour external vibration exposure at each case study property was calculated by scaling the event rms acceleration indicator in each of the three axes of vibration from the control position according to the calculated control-external transfer functions. Calculation of the 24 hour internal vibration exposure was achieved by scaling the predicted 24 hour external measurements using the measured external-internal transfer functions.




To obtain predictions for the internal vibration at case study properties where no internal measurements were made a representative external-internal transfer function was chosen, according to property type, from the measured transfer function from the appropriate study location.

For each case study in the pilot study, only the z-axis event rms acceleration, averaged over every event during the 24 hour period, was tabulated together with the predicted overall z-axis rms vibration for use in the combined analysis. For the follow on study, it will be necessary to calculate other indices from the raw vibration data. A small number of events were analysed using the Prosig software to obtain VDV and other indices to demonstrate that such parameters can be obtained from the recordings (see Appendix C).

4.5.2.5 Review of Analysis Method While the pilot study successfully demonstrated the suitability of the measurement protocol, the following aspects were identified that would benefit from further investigation or development early in the programme of a full scale exposure-response study:

• There were large variations in the transfer functions calculated for internal-external positions and control position-external satellite positions. Likely sources of this scatter are thought to be variability of the source, measurement position, measurement surface and post-processing. Given the number of potential sources of uncertainty there was no rigorous way of removing extraneous results. Measurement methods that may yield a greater degree of standardisation should be considered, which would then provide benefit to the investigation of alternative calculation of transfer functions.

• The pilot study sample size of internal measurements was not large enough to allow investigation of the variation in transfer functions or to obtain enough representative transfer functions, particularly at sites where there was a large variation in properties. Assumptions therefore had to be made that transfer functions from one property type were appropriate for another. This approach would be expected to be reliable, within the bounds of the inherent variability of measurements of this nature, for properties of similar foundation and floor constructions, even if floor plans differ in detail.

• The analysis process requires more automation before it could be applied to a more extensive study. This is possible within the available software, and without it, the follow on study would be expected to generate an unmanageably large volume of data.

• To improve efficiency in data manipulation, a standardised set of spreadsheets is required to calculate internal-external transfer functions, control-external transfer functions and site laws.

4.5.3 Combined Analysis One of the principal aims of the pilot study was to test the application of the questionnaire and measurement protocol in terms of their success in eliciting responses which have an ordinal relationship with vibration magnitude, i.e. that low magnitudes of vibration generally correspond to lower levels of annoyance and higher magnitudes of vibration generally give higher annoyance levels. The presence of an ordinal relationship between vibration magnitude and annoyance suggests that the application of the pilot study methodologies in a larger study may result in the required distribution of responses for an investigation into dose-response relationships.

Key questions which provide responses to test whether the questionnaire achieves this aim include Questions 27 and 31, which provide ratings of bother, annoyance and disturbance caused by the primary effects (feeling vibration) and secondary effects (hearing or seeing vibration) respectively; Question 35 which obtains ratings of annoyance from primary and secondary effects of vibration (feeling, hearing or seeing vibration) from trains; and Question 32 which provides annoyance ratings for feeling, hearing or seeing vibration from all sources.




The responses to the key questions were examined to determine whether suitable distributions of responses were obtained with a range of vibration exposures so that the required statistical analysis may be conducted on application of the survey to a larger population. The distributions of responses were examined for vibration exposures quantified in terms of measured z-axis internal and external vibration.

All the external vibration data used for the combined analysis were scaled from the control position data using the calculated control-external transfer function for each measurement position. In turn, all internal vibration data were scaled from external data using the calculated external-internal transfer functions.

Spearman’s rank correlation coefficient was determined between the three evaluation measures of vibration and annoyance ratings in response to the key questions. Spearman’s rank correlation coefficient is a non-parametric measure of correlation; it is a measure of the association between rank orders. Thus, a significant positive correlation between vibration magnitude and annoyance ratings indicates that there is an agreement between the rank orders of the two variables, i.e. that if the vibration magnitude and annoyance ratings are each placed in rank order of size, then there is an agreement between the ranks of the paired data so that low vibration magnitudes correspond to low annoyance ratings and similarly high vibration magnitudes correspond to high annoyance ratings. Unlike the Pearson product-moment correlation coefficient, Spearman’s rank correlation coefficient does not require the assumption that the relationship between the variables is linear, nor does it require the variables to be measured on interval scales; it can be used for variables measured at the ordinal or ranking level.

Railway sites

Spearman’s rank correlation coefficient was determined between the vibration measured at 76 properties and annoyance ratings from the residents of these properties. The correlation was conducted for annoyance ratings in response to Questions 27c, which asked for ratings for feeling vibration from over-ground trains (with and without respondents not feeling the vibration); Q31 which asked for ratings for hearing and seeing vibration from over-ground trains; Q35: feeling, hearing or seeing vibration from trains; and Q32: feeling, hearing or seeing vibration from all sources. Tables 4.5 and 4.6 give levels of significance (p) for the correlation between annoyance ratings from each question and three evaluation measures of vibration for internal and external vibration respectively.




Question providing annoyance ratings

Internal z-axis vibration acceleration Average event r.m.s. acceleration over 24h

rsp 0.378 Q27c: feeling vibration from trains p 0.001**

rsp 0.329 Q27c: feeling vibration from trains, excluding respondents not feeling vibration p 0.017*

rsp 0.161 Q31c: hearing and seeing vibration from trains

p 0.179

rsp 0.205 Q35: feeling, hearing and seeing vibration from trains

p 0.079† rsp 0.268 Q32: feeling, hearing and

seeing vibration from all sources p 0.020*

Table 4.5 Spearman’s rank correlation coefficient (rsp) – to test for an ordinal relationship - and significance levels (p) for correlations between annoyance ratings and internal vibration measures; pilot study railway sites. († marginally significant (p < 0.1), * significant (p < 0.05) , **highly significant (p < 0.01))


External z-axis vibration acceleration Average event r.m.s. acceleration over 24h

rsp 0.406 Q27c: feeling vibration from trains p <0.001**

rsp 0.422 Q27c: feeling vibration from trains, excluding respondents not feeling vibration p 0.002**

rsp 0.276 Q31c: hearing and seeing vibration from trains

p 0.020*

rsp 0.2485 Q35: feeling, hearing and seeing vibration from trains

p 0.033* rsp 0.291 Q32: feeling, hearing and

seeing vibration from all sources p 0.011* Table 4.6 Spearman’s rank correlation coefficient (rsp) – to test for an ordinal relationship - and significance levels (p) for correlations between annoyance ratings and external vibration measures; pilot study railway sites. († marginally significant (p < 0.1), * significant (p < 0.05) , **highly significant (p < 0.01))

Construction sites

Spearman’s rank correlation coefficient was determined between the vibration measured at 21 properties and annoyance ratings from the residents of these properties. The correlation was conducted for annoyance ratings in response to Questions 27f, which asked for ratings for feeling vibration from construction (with and without respondents not feeling the vibration); Q31f which asked for ratings for hearing and seeing vibration from construction; Q35: feeling, hearing or seeing vibration from the use of nearby heavy construction machinery; and Q32: feeling, hearing or seeing vibration from all sources. Tables 4.7 and 4.8 give levels of significance (p) for the correlation between annoyance ratings from each question and three evaluation measures of vibration for internal and external vibration respectively.





Internal z-axis vibration acceleration Average event r.m.s. acceleration over 6h

rsp 0.056 Q27f: feeling vibration from construction p 0.811

rsp -0.083 Q27f: feeling vibration from construction, excluding respondents not feeling vibration p 0.769

rsp 0.305 Q31f: hearing and seeing vibration from construction

p 0.179

rsp 0.553 Q35: feeling, hearing and seeing vibration from construction

p 0.009** rsp 0.399

Q32: feeling, hearing and seeing vibration from all sources

p 0.073† Table 4.7 Spearman’s rank correlation coefficient (rsp) – to test for an ordinal relationship - and significance levels (p) for correlations between annoyance ratings and internal vibration measures; pilot study construction site. († marginally significant (p < 0.1), * significant (p < 0.05) , **highly significant (p < 0.01))


External z-axis vibration acceleration Average event r.m.s. acceleration over 6h

rsp 0.144 Q27f: feeling vibration from construction p 0.533

rsp -0.008 Q27f: feeling vibration from construction, excluding respondents not feeling vibration p 0.977

rsp 0.302 Q31f: hearing and seeing vibration from construction

p 0.183

rsp 0.595 Q35: feeling, hearing and seeing vibration from construction

p 0.004** rsp 0.305

Q32: feeling, hearing and seeing vibration from all sources

p 0.180 Table 4.8 Spearman’s rank correlation coefficient (rsp) – to test for an ordinal relationship - and significance levels (p) for correlations between annoyance ratings and internal vibration measures; construction site, pilot study. († marginally significant (p < 0.1), * significant (p < 0.05) , **highly significant (p < 0.01))

Combined analysis conclusions

For data from the railway sites, significant Spearman’s rank correlations indicate an ordinal relationship between the vibration measured and ratings in response to several questions on annoyance from vibration. The presence of an ordinal relationship between vibration magnitude and annoyance suggests that the application of the pilot study methodologies in a larger study may result in the required distribution of responses for an investigation into dose-response relationships.

Highly significant correlations between annoyance ratings for feeling vibration from trains and both internal and external vibration measures indicate that there were ordinal relationships of similar significance between the level of annoyance from feeling vibration and internal and external vibration measures. This suggests that both the internal and the external vibration measures may provide the ordinal relationship between vibration magnitude and annoyance required for the determination of a dose-response relationship in an extensive study.




For data from the construction site, Spearman’s rank correlations were generally insignificant between the vibration evaluation measures and annoyance ratings. Further work, with a larger sample of vibration measurements and interviews, is required to investigate whether the methodologies employed in this study may be suitable for the investigation of a dose-response relationship for vibration from piling.

Implementation of a more extensive study to examine further the relationship between annoyance and vibration exposure must consider the influence of vibration frequency, direction and duration.




5 Recommendations The work undertaken in this project provides a tested and validated methodology for undertaking a national study of the community dose-response relationship to vibration in residential environments that could now be taken forward based on the findings of this work.

At the outset of the study a fundamental risk was identified, namely that the measurement protocol and questionnaire may not be able to yield a classic dose-response relationship where annoyance generally increases with vibration dose. The findings of the pilot study have helped to overcome this risk so that Defra can proceed with the full dose-response study.

Some further work needs to be carried out to validate fully the questionnaire and measurement protocol so that it can be applied to all sources of vibration which are of interest to Defra. This is particularly the case for mobile or temporary sources (such as construction sites) where vibration dose may change significantly from one day to the next.

The scale of the further development / validation work is such that it could be undertaken as an initial first stage of a national vibration incidence and attitude study.

The pilot study also demonstrated that members of the public were more willing to allow measurements to be made within their property than had been anticipated. The approaches have therefore provided useful information on the assessment of vibration sources within the resident’s building (but not within their domicile).

The improvements to the social survey and to the measurement protocol identified as well as potentially valuable parallel research opportunities are set out in the following subsections.

5.1 Social Survey

While the pilot study yielded a good response rate from residents, it is considered that a letter drop ahead of the questionnaire survey may further improve the response rate.

The questionnaire may be shortened by removal of questions on sources of vibration not present within a specified distance of the social survey site.

Questions on sensitivity to noise and vibration should be placed at the end of the questionnaire to avoid influence on responses to following questions.

Modules to enable the questionnaire to be applied to road traffic, industrial and internal vibration sources need to be developed.


The pilot study showed that the measurement protocol can yield vibration data which can be used to derive dose-response relationships. However, it is recommended that further work is necessary to investigate the causes of the variability of the transfer functions and to quantify the accuracy of different sampling strategies and assumptions which have been used to underpin the measurement protocol. The accuracy of the measurement protocol can be quantified using an intensive set of highly controlled vibration measurements taken over a relatively long period of time.

As expected before starting the pilot study, acquiring data from construction sites proved to be more difficult than acquiring representative vibration data from a railway. The pilot study confirmed that construction vibration can change significantly from one day to the next. Consequently, vibration measurements taken on one day may not be representative of the vibration exposure that occurred at the time that the social survey was conducted if there is any delay between the administration of the questionnaire and the measurement period. For example, the construction case study included in pilot study featured piling vibration




where the piling rig was moved around the site from day to day. This made it virtually impossible to measure the worst exposure, since the questionnaire had to be undertaken before the measurements could start, by which time the vibration source had have moved further away. Further development of the approaches to address such situations is therefore necessary. As set out in earlier sections of this report, possible options are to follow the questionnaire with the measurements even more closely; install an unattended logging device as soon as the questionnaire is completed; calculate the overall exposure through extrapolation from measured data; or install a logger (or loggers) close to the site boundary as soon as the works start, to record the construction vibration for several days, if not weeks, before and after the social survey was conducted.

The intention to measure vibration externally at a standardised distance of 2m from the closest façade to the vibration source proved to be impractical in several cases. Where access to a property was granted for internal measurements, this was not a significant issue. The difficulties arose where internal access was not granted. In future, respondents who are not amenable to measurements being made within their property should be asked if they will allow external access. This should enable a greater success rate in monitoring at a standard distance. An alternative that could be investigated would be to monitor at a position an equal distance from the source as would be a point 2m from the façade, but not necessarily directly in front of the property. For example, a position on a neighbour’s driveway, a footpath or road may be suitable.

The Rion DA-20 proved to be fully capable for undertaking the requirements of this project. It is believed that this equipment is the best currently available off-the-shelf for this work. Two key features are its portability and ease and speed of setting up.

During the pilot study a wide range of vibration measurement surfaces and rooms were encountered inside properties. It is not clear what effect this has on the range of measured vibrations and also what effect different finishes have on human response to vibration. This could have implications for the full dose response study. It is recommended that the full dose-response study investigates this further using a number of controlled measurements. For example, phased locked measurements could be carried out simultaneously inside and outside different pairings, or even different groups, of properties to determine external to internal transfer functions for similar properties, but which have different floor finishes.

It is often not economic to incorporate high specification isolation methods, such as isolation bearings, into new residential development to new residential dwellings. However, it is possible that alternative floor constructions or floor finishes could help to reduce or minimise exposure to vibration inside new proposed dwellings. For example, it is possible that some of the design solutions which are applied to new dwellings in order to satisfy Part E of the Building Regulations are more effective at minimising vibration inside buildings than others. It is even conceivable that cost effective solutions can be found that can be used to alleviate or mitigate existing vibration problems. It is recommended that a search is carried out of different floor finishes and current practice within the residential development sector to identify different floor constructions and finishes, especially those which might be more effective at reducing exposure to vibration. Furthermore it is recommended that a range of different floor construction types and finishes are included in the full dose-response study so that the vibration isolation performance of different constructions and finishes are investigated.

One of the aspects considered by Defra but not able to be addressed directly within the scope of this report was how the effects and response to vibration generated by sources within buildings could be assessed, e.g. from a neighbour’s heating system or washing machine. It is considered that the measurement protocol and questionnaire could be easily adapted so that it can be applied to internal sources. However, the uncertainty associated with internal measurements, described above, would equally apply and need to be resolved.




5.3 Analysis Methodology

Further development of the vibration data analysis processing scripts is required. The Prosig DATS software provides all the functionality required for more detailed processing and analysis of the data but the way it is automated requires optimization. For example, the current processing method used by the software increases the file size by a factor of approximately 10, which creates data storage and handling difficulties.

It has been demonstrated that the measurement protocol can reliably acquire both the noise and vibration data required. Little use has been made in the current study of the noise data: this should be analysed in the exposure-response study to investigate how much influence the internal and external noise levels have on the response to vibration.

There was a large scatter in the calculated transfer functions (control to external positions and external to internal positions). Some of this scatter is likely to be due to variations in the frequency of vibration generated by different sources or events, which could be investigated further by frequency analysis of results for a larger sample size of events. However, it is thought that much of the scatter is due to the method of event extraction, where control, external and internal measurements are processed independently from one another. Transfer functions are therefore evaluated in the time domain for pairs of events with differing start points, end points and durations. Further work should be undertaken to ensure that the methodology is fully robust in all situations.

The combined analysis indicated that assessing human response to vibration in residential environments from external sources on the basis of measurements made externally may be a valid approach. If possible this would have enormous cost benefits for conducting a much more extensive survey. However, there are insufficient data to be wholly conclusive and it is recommended that this aspect is considered as part of the exposure-response study. If confidence can be gained in the use of external data only, there may be significant cost and programme benefits for the full study.

It was not the intention of the pilot study to assess the correlation between annoyance and different vibration dose parameters. That said, a large volume of data has been acquired for this project, but relatively little analysis has been undertaken to date. Undoubtedly, one of the objectives for the larger follow on study will be to find the best parameter for quantifying vibration dose (such as vibration dose value, weighted acceleration, velocity, etc) to determine dose-response relationships. The recordings obtained from the pilot study could be analysed further to provide data which could be analysed to provide an early guide to the best available vibration dose parameter(s). Indeed, a specific analysis of the pilot study recordings to examine the correlation between annoyance and a range of vibration parameters could provide valuable information about the suitability of different parameters to be included in a larger follow-on study. Any further extensive study to examine the relationship between annoyance and vibration exposure must consider the influence of vibration frequency, direction and duration.




6 Conclusions This report presents the findings of the work undertaken to develop and trial a social survey questionnaire and a measurement protocol that together aim to establish a methodology by which human exposure to vibration in residential environments may be assessed through a more extensive study. The data collected were analysed as a means of validating the approaches taken, by investigating whether an ordinal relationship between relatively simple vibration parameters and response could be identified. It was not the purpose of the pilot study to determine an exposure-response relationship. Neither was it the purpose to investigate correlations between measured response and different vibration dose parameters or indices such as VDV.

It is considered that the pilot study demonstrated that the approaches suggested and trialled were very successful and provide a robust basis for undertaking an exposure-response study. In particular, an ordinal relationship was found between vibration magnitude and annoyance over a sufficient range which suggests that the application of the pilot study methodologies in a larger study may result in the required distribution of responses for an investigation into dose-response relationships. This finding achieves the primary objective of the pilot study and means that Defra can now proceed with a larger dose-response study with far greater confidence than before the pilot study was carried out.

As is always the case with research studies, especially at this early stage, a number of recommendations have been made for further work. In addition, a number of specific recommendations have been made to improve the questionnaire and the measurement protocol.

The pilot study focussed on railways as the most common and widespread source of perceptible environmental vibration. The processes developed were also trialled on a construction site with an impulsive vibration source. While some development and adaptations may be required for them to be fully appropriate to other sources and situations, it is considered that the approaches developed could be easily applied more widely. In particular, the success in gaining access for internal measurements provides confidence that the approaches are appropriate for assessing response to vibration sources within buildings.

The following table briefly describes the performance of the study against the ‘success’ criteria defined for the pilot study.




Evaluation criteria to assess the performance of the measurement protocol Ref. Evaluation criteria Achieved 1 Raw vibration time history data can be reliably and

continuously recorded at each assessment location for at least 24hr (or full operating period, as appropriate to the source).

Yes

2 Vibration data can be gathered, using the same equipment specification, in a number of different positions both inside and outside properties at a range of distances from the source and over the full range of possible human exposure (from threshold of perception – noise and vibration – to very high exposure).

Yes

3 Noise and vibration (in 3 orthogonal axes) can be recorded simultaneously at each and every measurement position.

Yes

4 The method is minimally intrusive maximising acceptability to residents and hence maximising permissions to undertake measurements inside properties.

Yes

5 Phase locked noise and vibration can be recorded simultaneously inside and outside properties.

Yes

6 The method is efficient and straightforward on site, supporting high (measurement) production rate, making a full exposure-response relationship study viable.

Yes

7 Vibration events can be identified, extracted and analysed automatically from continuous records during post-processing.

Yes Further development of post processing software recommended

8 Suitable for quantifying vibration exposure from different sources.

Yes Modules for industrial, road and internal sources to be developed

9 Post processing can reliably evaluate (automatically) all current vibration indicators (e.g. weighted and unweighted acceleration, VDVg, VDVd, VDVb, PPV etc).

Yes

10 Feedback from the questionnaire survey team to the measurement team can be provided sufficiently quickly and accurately to allow exposure to transient sources to be captured consistently in the measurement protocol and the questionnaire.

Partially Methodological improvements required for time varying sources (e.g. construction)

12 The accuracy of the measurement protocol is fit for purpose.

Partially Further tests recommended

13 Questionnaire is sufficient to gather information required but short enough to deliver high rate of acceptance for completing the questionnaire.

Yes Further optimisation possible

14 Questionnaire responses suitable for statistical and factor analysis.

Yes

15 Responses to questionnaire differ significantly when respondent is subject to different exposure.

Yes

16 Questionnaire response reliably differentiates between noise and vibration exposure.

Yes

17 Questionnaire responses are consistent. Yes 18 Respondents demonstrate reliable interpretation of

terminology (e.g.: noise and vibration). Yes

19 Questionnaire responses are reliable, repeatable and comparable.

Yes

20 Questionnaire and measurement protocol combined analysis demonstrate that there is an ordinal relationship between response and exposure.

Yes




In detail, the conclusions reached from the research are as follows.

6.1 Measurement Location Selection

Ideal measurement locations for carrying out the pilot study were not as numerous as might be supposed and this would be expected to carry through to the exposure-response study. It was challenging to identify sufficient sites where a large range of different levels of vibration were anticipated; where there were sufficient dwellings at different distances from the vibration source (especially close to the source); and where vibration from only one source could be measured without contamination from other sources. This should be borne in mind when taking forward future survey work.

Maps and aerial photographs available from the internet were valuable in undertaking a first sift for possible sites. Visits to site were necessary to refine the search before proceeding with the surveys.

6.2 Questionnaire

Responses indicated that the questionnaire enabled separate ratings for primary and secondary effects of vibration to be obtained. Thus, the results showed that respondents understood the distinction between terminology such as vibration and noise as used in the questionnaire.

Data obtained from questions on the neighbourhood and on annoyance from noise and vibration were suitable for the required factor analysis.

Significant correlations between annoyance from primary and secondary vibration effects and between annoyance from noise and vibration from related exposures indicated response consistency and reliability.

Analysis showed that the questionnaire was successful in extracting differences in responses to different vibration and noise exposures from a range of sources in close proximity to one or more of the sources of building vibration.


The proportion of occupants who were willing to allow access for internal measurements (average permission rate of 44%) was higher than expected which has benefits for data gathering. However, some people who initially responded positively were not able to provide access at times convenient for an economic survey programme.

The measurement protocol was found to be practicable in its application and data for subsequent post processing were recorded reliably, with a few exceptions.

The post processing software also proved reliable with event and period vibration indicators being successfully extracted (from long records) for all measurement positions at all locations. Valuable lessons have been learned that can be copied across the future full study

Recommendations are given for improvements in a number of other areas that should be considered within the full community dose response study.

6.4 Combined Analysis

The presence of an ordinal relationship between vibration magnitude and annoyance suggests that the application of the pilot study methodologies in a larger study may result in the required distribution of responses for an investigation into dose-response relationships.

Significant correlations between annoyance and three measures of vibration magnitude provide some confidence that residents’ annoyance generally increased with increasing vibration acceleration.




The results show that correlations were most significant between the vibration measured and annoyance from the primary effect of feeling train vibration. The correlation coefficient was lower and less significant for correlations involving the secondary effects of annoyance from hearing and seeing vibration than the primary effect of feeling vibration. This finding suggests that the questionnaire was successful in extracting differences in response to primary and secondary vibration effects.

Significant ordinal correlations between annoyance ratings for feeling vibration from trains and both internal and external vibration measures indicate that there were ordinal relationships of similar significance between the level of annoyance from feeling vibration and internal and external vibration measures. This suggests that external vibration measures may be as good as internal measurements for describing vibration dose in order to derive a dose response relationship or relationships. This finding could provide significant benefits in terms of cost and access to premises if it can be confirmed. However, the pilot study was not sufficiently conclusive to enable the follow up exposure-response study to be based solely on external measurements, and further work involving a larger sample size will be necessary to validate all aspects of the measurement protocol and before Defra can be entirely confident that the full dose-response study can rely exclusively upon external measurements.

The analysis demonstrates the success of the questionnaire and measurement protocol in generating the information required and also provides confidence that a full national study should identify a dose response relationship, for vibration at least, that could be used for future standard and policy development.

REFERENCES




Ashdown Environmental Limited. (1995) Channel Tunnel Rail Link – Vibration and Groundborne noise calculation procedures for the Channel Tunnel Rail Link. Final report 1 of 2 January 1995 Association of Noise Consultants (2001) Guidelines Measurement & Assessment of Groundborne Noise & Vibration. ANC/Fresco, Uckfield British Standard 6472:1992 ‘Guide to evaluation of human exposure to vibration in buildings’. British Standards Institution, London British Standard 6841:1987 ‘Guide to Measurement and evaluation of human exposure to whole-body mechanical vibration and repeated shock’. British Standards Institution, London. British Standard 14837-1 (2005) Mechanical vibration – Ground-borne noise and vibration arising from rail systems – Part 1: General guidance. British Standards Institution, London. David Trevor-Jones Associates. IoA Round robin – VDV. 09 May 2005 DTJ536/R2/1 62 DIN4150-2 Structural vibration Part 2: Human exposure to vibration in buildings. June 1999. Deutsches Institut fur Normung, Berlin Fields, J.M. and Walker, J.G. (1977) The effects of railway and vibration on the community. Contract Report 77/18. Institute of Sound and Vibration Research, University of Southampton. Greer, R.J., (1993) AEL methodology for the prediction of re-radiated noise in residential buildings from trains travelling in tunnels. Proceedings of the 1993 International Conference on Noise Control Engineering Leuven, Belgium. Greer, R., Thornely-Taylor, R, Malam, D., Williams, P., Pollard, J., Brodowski, T. and Evans, P. (2005) ANC round robin VDV measurement exercise analysis of eVDV data. Acoustics Bulletin, 30, (2), Mar/April 2005, 20-23 Grimwood, C.J., Skinner, C.J and Raw, G.J. (2002) The UK National Noise Attitude Survey 1999/2000. Noise Forum Conference 20 May 2002. Howarth, H.V.C. and Griffin, M.J. (1990) The relative importance of noise and vibration from railways. Applied Ergonomics 21(2), 129-134. International Standard ISO2631-1 Mechanical vibration and shock – Evaluation of human exposure to whole-body vibration – Part 1: Vibration in buildings (1Hz to 80Hz). Second edition, 2003. International Organization for Standardization, Geneva International Standard ISO2631-2 Mechanical vibration and shock – Evaluation of human exposure to whole-body vibration – Part 2: Vibration in buildings (1Hz to 80Hz). Second edition, 2003. International Organization for Standardization, Geneva Klæboea, R. Turunen-Riseb, I.H. Hårvikc, L. and Madshusc, C. (2003) Vibration in dwellings from road and rail traffic- Part II: exposure-effect relationships based on ordinal logit and logistic regression models. Applied Acoustics 64, 89–109. NS8176 (1999) Vibration and shock – measurement of vibration in buildings from land based transport and guidance to evaluation of its effects on human beings. Norwegian Council for Building Standardization. Öhstrtröm, E. and Skånberg, A.-B. (1996) A field survey on effects of exposure to noise and vibration from railway traffic, Part 1: annoyance and activity disturbance effects. Journal of Sound and Vibration, 193(1), 38-47. Stansfeld, S., Brown, B., Haines, M. and Cobbing, C. (2000).The Development of a ‘Standardised Interview to Assess Domestic Noise Complaints and their Effects’ (SIANCE). Final Report, December 2000. Department of Psychiatry, St Bartholomew’s and the Royal London School of Medicine and Dentistry, Queen Mary and Westfield College. Trevor-Jones, D. (2002) Issues raised in Review of BS6472. Proc. Inst. Acoustics, 24. Turunen-Rise, I.H., Brekke, A., Hårvik, L., Madshus, C. and Klæboe, R. (2003) Vibration in dwellings from road and rail traffic – Part 1: a new Norwegian measurement standard and classification system. Applied Acoustics 64, (1), 71-87 Watts, G.R. (1984) Vibration nuisance from road traffic – results of a 50 site survey. TRRL Laboratory report 1119. Watts, G.R. (1987) Traffic-induced ground borne vibrations in dwellings. TRRL Research report 102. Woodroof, H.J. and Griffin, M.J. (1987) A survey of the effect of railway-induced building vibration on the community. Institute of Sound and Vibration Research Technical Report No. 160, University of Southampton.

Appendix A Literature review

Defra NANR172 Human Response to Vibration in Residential Environments Final Report

M:\119573-00\OUTGOING DOCUMENTS\FINAL REPORT\FINAL ISSUE\APPENDICES\APPENDIX_A_LITERATURE_REVIEW.DOC

TRL Temple ISVR Arup AcousticsISSUE 7 March 2007

Contents A1 Literature Review

A1.1 Development of the measurement protocol A1.2 Development of the Questionnaire.



Page A1 TRL Temple ISVR Arup AcousticsISSUE 7 March 2007

A1 Literature Review This Appendix sets out the findings of the literature review and, in particular, draws upon publications and experience from investigations undertaken by the Project Partners as well as studies reported elsewhere.

A1.1 Development of the measurement protocol

A1.1.1 Human perception of and response to vibration British Standard BS 6472 (1992) [1] provides advice on the measurement and assessment of vibration between 1Hz and 80Hz effecting humans in buildings. It defines three axes, x, y, and z, in which vibration can be measured and assessed and relates these to the orientation of the subject. Equal response frequency curves are also provided. The response curves are presented as base curves for r.m.s. acceleration and peak velocity for the x, y and z axes. Multiples of the base curves are also given with semantic descriptions of the vibration effect in each case to provide the assessment of continuous narrow band vibration signals. The base curves are the basis for the ‘w ’ weighting (for the z-axis) and ‘wd ’ weighting (for the x and y-axes) defined in BS 6841 [2].

With reference to the measurement of vibration, BS 6472 suggests that vibration should be measured at a point on the building structure representative of entry to the respondent. If this is not possible, the standard states that transfer functions, to obtain representative data from a point elsewhere on the structure, should be used and declared.

Measurements should be taken tri-axially and then weighted according to the known orientation of the occupant. If the occupant’s orientation is unknown or variable, the measurements should be weighted for all axes and the highest value used for assessment. Full time history vibration data should be recorded to allow for various subsequent assessment methods defined in the standard, such as the Vibration Dose Value (VDV).

The current version of BS 6472 is not straightforward to use. The Association of Noise Consultants (ANC) Guidelines, as discussed below, were therefore developed to provide an independent view on the most appropriate use of the standard in different circumstances and on good practice. In 2002 a BSi panel was convened to review BS 6472 with the aim of correcting, clarifying and updating the standard where necessary. At the time of writing, the updated document is under preparation.

International standard ISO 2631 Part 1 (1997) and Part 2 (2003) [4][5] are the international equivalent to BS 6841 and BS 6472 respectively. Definitions of the human axial system and frequency weightings definitions are the same as in BS 6472.

It is important to note that the latest revision of Part 2 has led to larger and more significant differences between the international standard and its British counterpart than was previously the case. The 2003 version of ISO 2631-2, for example, does not provide advice on assessment indicators or criteria values (these are left to national standards). Furthermore, the recommended frequency weighting has also been modified compared to the previous version.

Although a subject-orientated axial system is defined, measurements are required to be made with respect to the axes of the building rather than the subject and only a single frequency weighting wm is used for all axes. An ‘informative’ note in ISO standards identifies that the individual frequency weightings presented in ISO 2631-1 can be used where the orientation of the respondent is known.




Measurements are to be made solely based on the expected occupation of the property with respect to the tasks performed by the occupants. This means that the assessment position and, where practicable, the measurement position in a room must reflect the highest magnitude of vibration. This standard also recommends the assessment and recording of associated effects such as the level of structureborne and airborne noise, rattling and visual clues.

BS ISO 14837-1:2005 [6] provides general guidance on groundborne vibration generated by the operation of rail systems and the resultant groundborne noise in buildings. Reference is made to [4,5] when discussing human perception. The standard recommends measurement locations ‘generally’ in the centre of floor spans for assessing the effect of vibration on humans. Other recommendations for measurement are consistent with the ANC guidelines [7] (see below). In addition, guidance on predicting and calibrating models of groundborne noise and vibration is given.

DIN 4150-2 (1999) [8] is the German equivalent to BS 6472 and ISO 2631-2 and, similarly, applies to mechanical and structural vibration in the frequency range 1Hz to 80Hz. Unlike references [1] and [4] a different frequency weighting function is defined: the time-frequency-weighting KBF(t) . This weighting is specified for an unknown body posture and is derived from a combination of seated and standing weightings.

With regard to measurement, like ISO 2631-2, DIN 4150-2 defines measurements with respect to the axes of the building. Measurements should be made at the location of maximum vibration, where a different location for each axis can be specified. For example, the floor may be chosen as the location of maximum vertical vibration while a door or window may be chosen as the location of maximum lateral vibration. Based on experience, DIN 4150-2 recommends that the centre of the floor should be chosen in preference to a more representative point of entry of vibration to the body, as this provides some consistency between measurements. The period of measurement should be characteristic of exposure and can be shorter than the actual exposure duration.

NS 8176 (1999) [9] is a recently developed Scandinavian standard developed by Norwegian, Swedish and Finnish Research institutes. The aim of the standard is to [10]:

“to present an efficient measurement procedure to determine a sufficiently stable estimate, a single number quantity, that gives a representative description of the vibrations in a building caused by traffic passing by”.

The assumptions made to increase the efficiency of the method are that pass-by vibration can be treated as a statistical quantity and that a single indicator at a single point in the receiving building can adequately predict the response of occupants. The single indicator adopted is the 95th percentile of the maximum weighted vibration measured at the location of highest response in the floor or structural frame of the building. The adopted weighting is the ISO combined whole body rating.

Guidance on measurement protocol defines tri-axial measurement locations to be on the floor of a habitable room at the point where maximum vibration occurs. It is suggested that this will occur at the mid-span of floors or beams. More than one location can be selected if necessary. Some guidance on mounting transducers is given, such as the use of a mounting tripod plate for measuring on carpeted floors.

In 2001 the Association of Noise Consultants (ANC) published a set of guidelines for the measurement and assessment of groundborne noise and vibration [7], intended for use by experienced practitioners. The need for a set of guidelines arose due to:

• difficulties in the application of BS6472:1992 and similar standards;

• a lack of suitable measurement equipment;




• large variation in measurements between different organisations;

• the adoption of different criteria by consultants and local authorities; and

• contemporary involvement in major projects involving groundborne noise and vibration.

The ANC Guidelines provide a clarification of the implementation of the BS 6472 and provide advice on good industry practice.

The ANC Guidelines clarify that for most vibration sources, BS 6472 recommends that the VDV indicator should be used. Process, or ‘application’, flow diagrams are presented to make clear how BS 6472 should be applied to different vibration sources.

The guidelines identify that, like DIN 4150-2, in practice, measurements and assessments are made at the centre of floor spans (rather than specifically at the point of entry) as these locations are generally worst case and provide consistency between assessments.

Advice on transducers, mounting and types of recording equipment is also discussed. The importance of specifications including frequency range, dynamic range and sensitivity are all discussed.

With regard to measurement scope, if a particular source is being investigated, it is recommended that ambient measurements should be made when this source is not present. The presence of internal sources, such as slamming doors, and a requirement to flag these sources is also identified.

The ANC Guidelines also provide guidance on the prediction of internal vibration from free-field ground vibration measurements with the use of transfer functions.

Studies supporting these guidelines have been performed [11,12], in which users and manufacturers of VDV equipment were invited to participate in a comparison of a wide range of equipment and software for measuring VDV under controlled conditions. Apart from a few exceptions, the small variations seen were encouraging from the point of view of a user of the equipment.

Woodroof and Griffin [13] present an early vibration dose study of 720 respondents at 24 sites within 100m of a railway. As well as a social questionnaire, 24 hour vibration monitoring was undertaken. The measurement technique followed the then early ISO 2631. Measurements were made in a habitable room in which the magnitude of vibration was perceived to be the highest. Where possible, a measurement location was chosen and only vertical x-axis vibration was measured.

An important finding of the survey was that the location of the measurement within the building affected the frequency and magnitude of the vibration. The frequency response was more determined by the building response than by the source characteristics.

A1.1.2 Transfer functions The calculation procedure for the predictions of groundborne noise for the Channel Tunnel Rail Link Project [14] presents a method for predicting the groundborne vibration and groundborne noise inside properties from measurements of free-field ground vibration which may be of value for the current project. This is achieved with the use of separate building response functions for perceptible vibration and for airborne noise at the centre of rooms.

The building response functions in [14] have been derived from a measurement study of external vibration and internal vibration and noise for properties located over the London Underground Central Line [15]. The resulting response functions were subsequently




validated, as part of a complete groundborne noise prediction model, against measured data from tunnels in France and Germany.

With regard to the measurement protocol for the Defra research project, it is useful to consider the above prediction methods, not as a reason not to measure inside dwellings but as a useful parallel study to internal measurement. If a relationship can be found between ground vibration measured at a building façade and the corresponding community response to internal vibration, this could lead to a more efficient protocol in the future. This would require a large database of simultaneously measured façade level vibration and internal groundborne noise and vibration. In addition to published information, a large resource of previous project experience and data is available to the Project Partners, from which transfer functions could be validated and developed.

A1.1.3 References [1] British Standard 6472:1992 ‘Guide to evaluation of human exposure to vibration in buildings’. British Standards Institution, London

[2] British Standard 6841:1987 ‘Guide to Measurement and evaluation of human exposure to whole-body mechanical vibration and repeated shock’. British Standards Institution, London.

[3] Trevor-Jones, D. (2002) Issues raised in Review of BS6472. Proc. Inst. Acoustics, 24,(),

[4] International Standard ISO2631-1 Mechanical vibration and shock – Evaluation of human exposure to whole-body vibration – Part 1: Vibration in buildings (1Hz to 80Hz). Second edition, 2003. International Organization for Standardization, Geneva

[5] International Standard ISO2631-2 Mechanical vibration and shock – Evaluation of human exposure to whole-body vibration – Part 2: Vibration in buildings (1Hz to 80Hz). Second edition, 2003. International Organization for Standardization, Geneva

[6] British Standard 14837-1 (2005) Mechanical vibration – Ground-borne noise and vibration arising from rail systems – Part 1: General guidance. British Standards Institution, London.

[7] The Association of Noise Consultants (2001) Guidelines Measurement & Assessment of Groundborne Noise & Vibration. ANC/Fresco, Uckfield.

[8] DIN4150-2 Structural vibration Part 2: Human exposure to vibration in buildings. June 1999. Deutsches Institut fur Normung, Berlin

[9] NS8176 (1999) Vibration and shock – measurement of vibration in buildings from land based transport and guidance to evaluation of its effects on human beings. Norwegian Council for Building Standardization.

[10] Turunen-Rise, I.H., Brekke, A., Hårvik, L., Madshus, C. and Klæboe, R. (2003) Vibration in dwellings from road and rail traffic – Part 1: a new Norwegian measurement standard and classification system. Applied Acoustics 64, (1), 71-87

[11] (2003) IoA Round robin – VDV. David Trevor-Jones associates 09 May 2005 DTJ536/R2/1 62

[12] Greer, R., Thornely-Taylor, R, Malam, D., Williams, P., Pollard, J., Brodowski, T. and Evans, P. (2005) ANC round robin VDV measurement exercise analysis of eVDV data. Acoustics Bulletin, 30, (2), Mar/April 2005, 20-23

[13] Woodroof, H.J., Griffin, M.J. (1987) A survey of the effect of railway-induced building vibration on the community. ISVR Technical Report 160. Institute of Sound and Vibration Research, University of Southampton.




[14] Ashdown Environmental Limited. (1995) Channel Tunnel Rail Link – Vibration and Groundborne noise calculation procedures for the Channel Tunnel Rail Link. Final report 1 of 2 January 1995

[15] Greer, R.J., (1993) AEL methodology for the prediction of re-radiated noise in residential buildings from trains travelling in tunnels. Proceedings of the 1993 International Conference on Noise Control Engineering Leuven, Belgium.

A1.2 Development of the Questionnaire.

A review of previous studies has been conducted as part of the process of developing the questionnaire. The review included an evaluation of previous field and laboratory studies of the effects of noise and vibration in buildings in the UK, Sweden and Norway and studies of social survey methods and publications concerned with methods of evaluating the severity of vibration and noise. A summary of the studies included in the review and an evaluation with respect to the value to the present study is given below.

A1.2.1 Field Studies Klæboea, R. Turunen-Riseb, I.H. Hårvikc, L. and Madshusc, C. (2003) Vibration in dwellings from road and rail traffic- Part II: exposure-effect relationships based on ordinal logit and logistic regression models. Applied Acoustics 64, 89–109.

This Norwegian socio-vibration survey, with 1503 respondents from 14 study areas, was undertaken in the autumns of 1997 and 1998. The main purpose of the study was to provide input for establishing comfort criteria for dwellings exposed to vibration. Telephone interviews were used for obtaining people’s reactions to vibrations in dwellings. In-dwelling vibration values were calculated for 1427 of these respondents. Vibration severity was evaluated by determination of the 95-percentile of maximum weighted vibration velocities (or acceleration), vw,95. The logarithmic transformation of vw,95, as defined in Norwegian Standard NS 8176, was selected as the exposure measure. Full details of the questionnaire are not provided. However, examples of some questions are given. A 5 point semantic annoyance scale was employed to obtain subjective reaction: ‘‘highly annoying, somewhat annoying, a little annoying, not annoying, does not notice”. The questionnaire contained questions on how often people notice different effects of the vibrations (that the house shakes or vibrates, rattling of furniture and household items, that things move about and/or that they can bodily feel the vibrations). By coding ‘‘often’’ as unity and other responses as zero, the responses were dichotomized. Logistic regression analysis was then conducted for the exposure-effect relationships for each effect with a dichotomous dependent variable. The authors recognise that a disadvantage of the logistic regression is the elimination of information on the transitions between the other response categories as a function of the vibration measure. Therefore, ordinal logic analysis was also undertaken for the degree of annoyance as a function of the vibration exposure measure.

Systematic information on the number and duration of vibration events associated with the different rail and road stretches for the different study areas was not available. Logistic regression was conducted to determine the growth of annoyance as a function of the vibration exposure in terms of the statistical maximum weighted vibration value vw,95. The authors suggest that number and duration of events were sufficiently controlled between study areas to be insignificant factors.

The study is of limited value to the present work since the complete questionnaire is not presented and only a small number of the questions are reproduced. The analysis methods do not consider the duration and number of events and are therefore considered




unsatisfactory for the present study. A logistic regression analysis with a dichotomous dependent variable is considered an unsatisfactory method for determination of a dose-response due to loss of response information and therefore will not be incorporated in the present study.

Grimwood, C.J., Skinner, C.J and Raw, G.J. (2002) The UK National Noise Attitude Survey 1999/2000. Noise Forum Conference 20 May 2002.

In 1991, BRE undertook a National Noise Attitude Survey (NAS) for Defra. The aim of the survey was to sample the population of England and Wales to assess attitudes to environmental noise in the home. The survey was part of a long-term project to produce a ‘general noise annoyance model’. A further National Noise Attitude Survey and research project was conducted in 1999/2000 with the following objectives:

• To undertake a survey of attitudes to environmental noise in England and Wales and to track any changes in attitudes to noise between 1991 and 1999.

• To undertake a UK wide survey of attitudes to environmental noise utilising a revised questionnaire in order to provide an estimate of current attitudes to noise.

• To investigate the importance of questionnaire design in noise attitude surveys.

In 1999/2000, two sample groups, each approximately equivalent in size to that used in 1991, were interviewed in England and Wales; the first with the 1991 questionnaire, and the second with a new modular questionnaire (1999 NAS questionnaire). During 2000, the survey using the new modular questionnaire was extended to include Scotland and Northern Ireland in order to enable UK estimates of attitudes to environmental noise to be made.

The 1999 questionnaire was designed to be modular in nature, to enable individual sections (supplementary questionnaires) to be used independently of each other. It was designed so that it was possible to select a supplementary questionnaire to use for specific surveys. For example, the main section could be used independently of supplementary questionnaires to establish a broad response of a population to environmental noise, whilst just a supplementary questionnaire could be used on a targeted basis to establish response to a particular noise source.

A comparison of the results of the two questionnaires show that different phrasing on the same line of questioning may yield different results and that responses obtained may differ significantly owing to a variety of other factors within the questionnaire and its administration. Factors effecting response include (i) routing within the questionnaires and the use of filter questions; (ii) question wording and the options given for responses; (iii) interviewer coding instructions; (iv) use of show cards; (v) focus of questions on specific noise sources or general categories of noise; (vi) interview technique; (vii) questionnaire structure and the order of questions.

The findings of the surveys and the comparison of the two questionnaires provide invaluable information that has been carefully considered in the design of the questionnaire for the present study.

A number of findings of the study are worthy of particular consideration in the design of the present social survey questionnaire. The authors note that a disadvantage of subdividing noise into types is that there will be fewer samples in the groups. However, a focus group of respondents in the 1991 study yielded the comment that specific noise source descriptors are preferable to general categories as they help understanding and recall. Where questions were asked about specific noise sources, the proportions reporting being bothered, annoyed or disturbed by one or more specific noise sources in a category is significantly higher than the proportion that report the same for the general category of environmental noise.




It is noted that nearly all respondents of the focus group assumed that “don’t hear” means “don’t notice” or “not affected by, not bothered by”. All participants were in favour of filter questions but many did not think it should be based on “hear” stating that “you get used to a noise and then don’t hear it any more”.

The 1991 questionnaire made extensive use of filter questions. Respondents were questioned in detail about a specific source of noise only if they reported hearing that noise, and further questions were only asked of those respondents who reported being adversely affected by the specific source. A comparison of the 1991 and 1999 surveys showed that this filtering biased the results when compared to asking the same question to all respondents. It was concluded that filtering should be kept to a minimum, and the consequences of any such routing should be considered carefully.

Two stages of questioning were employed in the 1999 survey; the first to establish whether the extent to which they were bothered, annoyed or disturbed. A second-stage involved the ranking of the sources to determine which supplementary questionnaire sections should be completed. If a respondent reported “don’t know”, a check was made to see if this meant that they never heard the noise or if they did mean “don’t know”. If they rated the noise as bothering, annoying or disturbing them “not at all”, they were asked if this was because they never heard the noise. In either case, if they said that they never heard the noise then this was coded as “don’t hear (rather than “not at all” or “don’t know”). The questioning method is described as indirect and was adopted because it was believed that it was preferable not to ask a respondent explicitly whether they heard a noise source and that by obtaining the response indirectly a more accurate measure would be obtained of the number of respondents who actually heard the noise sources. This indirect method resulted in a very much higher proportion of respondents indicating that they hear noise from specific sources than the direct method of the 1991 questionnaire which simply asked “do you hear any of the following noises”. This agrees with the findings from the focus groups which concluded that respondents would assume that the term “don’t hear” meant “don’t notice” or “not affected by” or “not bothered by”. The comparison of the 1991 and 1999 questionnaires show that the 1999 questionnaire was more effective in obtaining responses that reflect the true reaction of the respondents to specific noise sources and therefore the questionnaire of the present study employs a similar line of questioning to elicit responses without the use of filters and by indirect questioning on the extent of bother, annoyance and disturbance and the hearing and feeling of noise and vibration for specific sources.

In addition to a 5-point semantic scale, the supplementary questionnaire for railway noise of the 1999 study employed a 7-point numeric scale for the rating of bother, annoyance or disturbance. Correlation between the responses obtained with the two scales enables the results to be tested for consistency and reliability. The questionnaire for the present study makes use of a semantic and a numeric rating scale which are similar to those used in the 1999 NAS study.

The NAS 1999 survey made extensive uses of show cards. The primary purpose of the show cards was to remind the respondent of possible appropriate responses and to inform the respondents of specific sources of noise that may be included in a given category. The authors state that the main effect of the use of show cards is likely to be an improvement in the consistency of the resulting responses, as all respondents should be answering the question on the same basis. Show cards will be used extensively in the social survey questionnaire of the present study.

The authors highlight the importance of interview technique. One aspect raised is the effect of emphasising specific words in the question. For example, evidence was found that when respondents were asked about aspects they liked in the home, quietness was less likely to be mentioned if the word “home” was emphasised. It is suggested that this may be due to




quietness being considered an attribute of the neighbourhood and not the home. In the questionnaire of the present study, use is made of bold and underlined text where appropriate to indicate emphasis on specific words to focus the respondent on specific aspects and to improve interpretation and consistency.

It is suggested in the NAS study that the order of questions affects response. For example, a general noise question may elicit greater response when preceded by questions on specific noise sources which as a reminder to different aspects of noise. In the present questionnaire, general questions on vibration and noise are placed after questions on specific aspects of vibration or noise. Questions on sensitivity to noise and vibration are placed after ratings of noise and vibration so as to avoid influence of sensitivity questions on noise and vibration rating.

The 1999 NAS questionnaire has a section for the interviewer to complete of their own observations of noise. A similar section is included in the questionnaire of the present study. Other topics covered in the 1999 NAS and the present survey include dwelling and location information; satisfaction with the neighbourhood; satisfaction with the home; sensitivity; activity disturbance and degree of bother, annoyance or disturbance.

Stansfeld, S., Brown, B., Haines, M. and Cobbing, C. (2000).The Development of a ‘Standardised Interview to Assess Domestic Noise Complaints and their Effects’ (SIANCE). Final Report, December 2000. Department of Psychiatry, St Bartholomew’s and the Royal London School of Medicine and Dentistry, Queen Mary and Westfield College.

The Department of Health and the Department of Environment, Transport and the Regions funded a study to develop a “Standardised Interview to Assess Domestic Noise Complaints and their Effects”. The aim of the project was to develop a standardised questionnaire that provides a tool for Local Authority officers to use to assess the effects of domestic noise and that can be used to research the effects of domestic noise. The questionnaire includes sections on information on the noise; health and behavioural effects (annoyance, sleep and activity disturbance); characteristics of complainant; the environment (housing details); and noise assessment.

The questionnaire was field tested on a sample of noise complainants from London Boroughs. The results of the field study suggest that the Standardised Interview to Assess Domestic Noise Complaints and their Effects (SIANCE) is a psychometrically reliable and valid measure. Several different scales are used to obtain measures of response to noise. Internal consistency was good for all sub-scales.

The questionnaire design and the methods of assessing the validity of the questionnaire have been considered in the development of the present social survey. The wording of questions on degree of bother, annoyance and disturbance, noise sensitivity and satisfaction with neighbourhood are similar in the SIANCE and present surveys.

The statistical analysis methods employed in the SIANCE study is of interest in the consideration of analysis to be conducted on questionnaire data obtained in the present study. Similar methods of testing the reliability and validity of the data will be employed. In the SIANCE survey, the intra-class correlation coefficient was determined to test the inter-rater reliability and Spearmen’s rank correlation coefficient to test the test-retest reliability. To test scale reliability, Cronbach’s alpha correlation coefficient was determined. Pearson’s correlation coefficient was determined to test agreement between results using different scales. Factor analyses were conducted to identify underlying variables that determine the pattern of correlations within the set of variables for activity disturbance, emotional response and for coping with the noise. The questionnaire of the present study has been designed to provide data suitable for the statistical analyses conducted in the SIANCE study.




Öhstrtröm, E. and Skånberg, A.-B. (1996) A field survey on effects of exposure to noise and vibration from railway traffic, Part 1: annoyance and activity disturbance effects. Journal of Sound and Vibration, 193(1), 38-47.

This Swedish study presents some of the results from investigations of the effects of exposure to noise and vibration from railway traffic. Effects on annoyance, sleep disturbances and psycho-social well-being, as well as disturbance of different activities were evaluated by a postal questionnaire. Fifteen different sites located near railway lines in Sweden were investigated.

The study covered areas with different frequencies of passing trains and with three categories of railway-induced building vibration magnitude: where the velocity exceeded 1 mms-1, where the velocity was less than 1 mms-1 and in areas without vibration. The results for general annoyance and activity disturbances in the study show that train noise was more annoying in areas in which there was simultaneous exposure to vibration from trains. The authors suggest that this result may have been due to difficulties people have in differentiating between noise and vibration. For respondents residing within 200m of the railway line, the average annoyance was at least as high for vibration as for noise.

The effects were evaluated by a postal questionnaire. The main questionnaire contained questions about the dwelling and the neighbourhood, annoyance from different sources in the neighbourhood (noise, dust, fumes, vibration etc.), work environment and sleep and sleep disturbances as well as questions on health and general well-being. General annoyance was evaluated by a 5-point verbal category scale: do not observe (0), observe but not annoyed (1), not very annoyed (2), rather annoyed (3), very annoyed (4). The questionnaire also contained questions on different parameters thought to affect annoyance, such as noise sensitivity, the position of bedroom windows and the construction of the house. Those who stated that they were “rather” or “very” annoyed by noise or vibration from railway track received a second questionnaire with specific questions on disturbance of different activities.

While the findings of the study are of interest in highlighting the significant contribution to annoyance of vibration from nearby trains, the work reported is of limited value in the design of the social survey for the present study as the questionnaire is not reproduced in the paper.

Woodroof, H.J. and Griffin, M.J. (1987) A survey of the effect of railway-induced building vibration on the community. Institute of Sound and Vibration Research Technical Report No. 160, University of Southampton.

A survey was conducted of annoyance caused by railway-induced building vibration in Scotland. A questionnaire was designed to determine the number of residents who perceive railway-induced building vibration in their home and to determine the degree of annoyance caused. Vibration was recorded continuously for 24 hours in 52 dwellings occupied by respondents who noticed vibration. The severity of the vibration was assessed by 90 alternative objective measures. The number of trains to pass the respondent’s dwelling in 24 hours provided the highest correlation with annoyance. The response was not significantly affected by any of the characteristics of the vibration. The results suggest that subjective assessments of vibration annoyance were not solely influenced by the perception of the vibration. It was concluded that the absence of a significant correlation between subjective response and measures of vibration severity occurred because railway-induced building vibration did not cause significant annoyance in the respondents, though the results suggest that 35% of residents within 100 m of the railway notice the vibration.

The questionnaire has questions on general satisfaction with the area, feeling vibration of the house/floor/chair or bed from trains, hearing rattling of objects, seeing movements of




objects and sleep disturbance, fear of building damage. The study provides valuable information on the design of the questionnaire which may provide data from which a correlation between subjective and objective measures can be obtained in locations where a significant degree of annoyance occurs.

The statistical analysis conducted to test correlations between different measures of the severity of the vibration and subjective response included determination of Kendall’s tau with corresponding z-scores to indicate the significance level. Partial correlation was also conducted; Pearson’s correlation coefficient was computed to investigate the contribution of specific characteristics of vibration to the overall correlation. The questionnaire of the present study has been developed to obtain data from which such analysis may be conducted.

Correlation was conducted between measures of vibration magnitude and annoyance ratings obtained from the question, “How annoyed are you by the vibration here from passenger, goods and other trains? Which rating most closely corresponds to your own feeling: 0 (Not annoyed) – 6 (extremely annoyed)?” Correlation was also conducted with ratings obtained from the question, “Does the vibration from the trains annoy you: not at all / a little / moderately / very much?” The author states that since the response to the annoyance question was influenced by the number of trains, it is likely to be the hearing of the trains that made them aware of the number of trains passing. It might be concluded that the question designed to elicit judgements of annoyance from the vibration in fact provided judgements of the noise. The survey includes questions on feeling vibration of the house, floor, chair or bed from trains, hearing rattling of objects and seeing movements of objects. The respondents are they asked, “How annoyed are you by the vibration here from passenger, goods and other trains? Which rating most closely corresponds to your own feeling: 0 (Not annoyed) – 6 (extremely annoyed)? However, it is not clear which vibration effect (feeling, hearing or seeing vibration) the respondents are being asked to rate. The questions in the present social survey have been carefully worded to differentiate between ratings of bother, annoyance and disturbance from feeling vibration, hearing noise, seeing or hearing objects rattle.

Watts, G.R. (1987) Traffic-induced ground borne vibrations in dwellings. TRRL Research report 102.

A study was conducted to assess the effect of road traffic-induced ground-borne vibration on vibration disturbance response. Vibration measurements were made at 50 UK sites inside and outside dwellings located close to large road surface irregularities during the passage of light and heavy vehicles. Annoyance scores from an earlier study at the same sites (see below) were obtained using a 7-point scale employed by residents in homes estimated to be exposed to perceptible vibration to provide an overall assessment of vibration nuisance. The study is of limited value in the design of the questionnaire for the present study as the relation between subjective response and objective vibration measures was not explored.

Watts, G.R. (1984) Vibration nuisance from road traffic – results of a 50 site survey. TRRL Laboratory report 1119.

A 50-site survey was conducted to investigate the vibration nuisance in residential properties caused by road traffic. The study involved conducting interviews with residents and measurements of noise and window vibration and traffic flow parameters. Noise exposure levels measured at the façade of dwellings and window pane vibration were correlated with ratings of vibration nuisance. Two rating scales were employed by residents to indicate the degree of bother from vibration caused by road traffic: a 7-point rating scale ranging from 0 (not at all bothered) to 6 (extremely bothered) and a 4-point semantic scale (not at all / not very much / quite a lot / very much). The respondents asked whether they




notice different vibration effects caused by road traffic, e.g. windows rattling, floors shaking. They are then asked to rate the bother from the vibration. However, the question posed to the residents does not make clear which vibration effect is to be rated. Eighteen-hour noise exposure measures were found to be more closely associated with vibration and noise nuisance ratings than window vibration and traffic flow levels. The questionnaire of this study was considered in the design of questions in the present study.

Fields, J.M. and Walker, J.G. (1977) The effects of railway and vibration on the community. Contract Report 77/18. Institute of Sound and Vibration Research, University of Southampton.

The effect of railway noise on residents near railway routes was investigated with a combined social survey and noise measurement programme conducted along railway routes in the UK. The objectives of the study were to obtain an estimation of the number of residents in the UK near railway lines who are exposed to railway noise and to investigate correlations between noise measures and annoyance. It was found that the 24-h LAeq was more highly correlated with annoyance than any other noise measure. The 45-minute questionnaire explored annoyance caused by different types of railway noise. Other aspects of the railways causing annoyance were also examined. The questionnaire has sections on likes and dislikes about the area, satisfaction with the area, rating of aspects about the area, sources of noises heard in the home, ratings of satisfaction with noise different types of noise, noise sensitivity, time of day when train noise causes more annoyance, trains types causing most annoyance; rating of bother and annoyance caused by specific sources of noise from railways. The survey also includes questions on whether the respondent notices vibration of the house, floor, bed or chair or of objects caused by trains and a section on actions taken due to railway noise. The study provides valuable information on the design of questions concerned with annoyance caused by railways which was considered in the design of the questionnaire for the present study.

A1.2.2 Laboratory Studies Howarth, H.V.C. and Griffin, M.J. (1990) The relative importance of noise and vibration from railways. Applied Ergonomics 21(2), 129-134.

Howarth and Griffin (1987) reported a laboratory experiment in which seated subjects judged laboratory simulations of railway noise and railway-induced building vibration. The study involved the presentation of six magnitudes of vibration and six levels of noise in all 36 possible paired combinations. Subjects were asked which of the two stimuli (noise and vibration) they would prefer to be reduced. A subjective equivalence contour was determined from the levels at which 50% preferred the reduction of noise and 50% referred the reduction of vibration. The contour may be described by the relation LAE = 29.3 log10 VDV + 89.2 where LAE is the sound exposure level and VDV is the vibration dose value. The relation may be used to determine which stimulus it would be more beneficial to reduce in noise and vibration environments. The survey of the present study includes questions on the preference for the reduction of noise or vibration. The results of an extensive study involving application of the questionnaire together with vibration measurements may be compared with the findings of the laboratory study to determine whether the response of residents may be predicted from the results of laboratory studies.

Howarth, H.V.C. and Griffin, M.J. (1990) Subjective response to combined noise and vibration: summation and interaction effects. Journal of Sound and Vibration, 143(3), 443-454.

The influence of noise on judgements of vibration, the influence of vibration on the judgements of noise, and a composite reaction to combined noise and vibration were




investigated in a laboratory study by Howarth and Griffin (1990) involving simulations of noise and vibration from railways. It was concluded that, although vibration had little effect on the judgement of noise, the assessments of vibration could be increased or decreased by noise, depending on the relative magnitudes of the two stimuli. A composite measure was evolved to predict annoyance caused by a combination of noise and vibration from a summation of the individual effects. The questionnaire of the present study includes questions on ratings of the combined effects of noise and vibration, the responses of which may be employed with measurements of noise and vibration to provide a comparison with the findings of the laboratory study.

A1.2.3 Social Surveys and Evaluation Methodologies Preliminary work for Defra on this research programme was undertaken by a separate contactor. This work was reviewed and used to inform the work described in this report. In addition, questionnaires employed in past UK surveys, summarised herein, have been used to advise development of the questionnaire.

Turunen-Risea, I.H., Brekkeb, A., Harvikc, L., Madshusc, C. Klæboed R. (2003) Vibration in dwellings from road and rail traffic - Part I: a new Norwegian measurement standard and classification system. Applied Acoustics, 64, 71–87.

An objective method of assessing the severity of vibration is described. The objective method involves the determination of the 95-percentile of maximum weighted vibration velocities (or acceleration) from a minimum number of representative road and rail traffic events, vw,95. The relation between different values of the objective measure, vw,95 and the strength of people’s reactions was investigated. The objective method of assessing the severity of the vibration may be considered at the combined analysis stage of the project. The method was evaluated by means of a Round Robin social survey study (Klæboea et al (2003), see below).

Klæboea, R., Ohrstromb, E., Turunen-Risec, I.H., Bendtsend, H. and Nykanen, H. (2003) Vibration in dwellings from road and rail traffic - Part III: towards a common methodology for socio-vibrational surveys. Applied Acoustics 64, 111-120.

Drawing on the results of a Norwegian socio-vibration survey (Klæboea, Turunen-Riseb, Hårvikc and Madshusc (2003), see above) and a Swedish socio-acoustic survey supplemented with vibration measures (Öhstrtröm and Skånberg (1996), see above), a method of measuring annoyance caused by vibration in dwellings from road and rail traffic is proposed in this study. The method, entitled the Nordtest method, proposes use of a 5-point categorical annoyance scale and an 11-point numerical annoyance scale, both with lower anchoring point ‘‘do not notice’’. The study includes some discussion of the advantages and disadvantages of filter questions. Advantages given of filter questions include the reduction in length of the interview. It is claimed that a simple comparison of the data obtained using a verbal vibration annoyance question in a Swedish survey which did not have such a filter question with the Norwegian which had a filter question showed no evidence of people misinterpreting a question about whether they noticed the vibrations or not. However, it is stated that an ISO working group raised concerns about the use of filter questions in conjunction with the perception of sound/noise. The study raises the issue of whether questions on assessment of vibration and noise should include a specific time frame (the vibration annoyance question in the study specifies “in the last 12 months or so”) and whether the questions should specify “when you are at home” or “when you are indoors”. It is suggested that “when you are at home” is appropriate for noise while “when you are indoors” is appropriate for vibration. In the survey for the present study it is specified that noise/vibration “in your home” and over a time period of 12 months is to be considered for questions on the rating of bother, annoyance or disturbance from vibration and noise.




The study discusses the benefits of including a semantic response scale in addition to a numeric rating scale to indicate the degree of annoyance. The authors state that an argument for using two annoyance questions is the increased reliability of combining information from two questions. It is stated that providing one annoyance question with semantic categories and one with a numerical rating, thus having different answer formats and directionality, might enhance the validity of a combined reaction measure. However, the Nordtest Method does not contain procedures or instructions on how to combine results from the analysis of semantic and numerical annoyance questions. The questionnaire of the present study employs both numerical scales and semantic category scales.

A numerical rating using whole numbers from 0 to 10 is employed in the study. The authors state that this allows a finer gradation of answers than a scale having a smaller range of numbers. The highest rating (10) is anchored at ‘‘extremely annoyed’’. The lowest rating (0) is to be used when the vibrations are not noticeable. The questionnaire of the present study employs a 7-point numeric rating scale, as employed by Watts (1984), Woodroof and Griffin (1987) and Grimwood et al (2002).

Appendix B Equipment

Defra NANR172 Human Response to Vibration in Residential Environments Final report

M:\119573-00\OUTGOING DOCUMENTS\FINAL REPORT\FINAL ISSUE\APPENDICES\APPENDIX_B_EQUIPMENT.DOC

TRL Temple ISVR Arup Acoustics ISSUE 7 March 2007

Contents B1 INTRODUCTION B2 THE IDEAL INTERNAL VIBRATION ACQUISITION AND ANALYSIS DEVICE

B2.1 Measurement parameters B2.2 2.2 Transducers B2.3 Data acquisition and storage B2.4 Analysis requirements B2.5 Noise Filtering B2.6 Installation B2.7 Casing B2.8 Size and weight B2.9 Peripheral and Supporting Devices

B3 PROCURING THE IDEAL INTERNAL VIBRATION ACQUISITION AND ANALYSIS DEVICE B3.1 Investigation into availability of ‘ideal’ components



Page B1 TRL Temple ISVR Arup AcousticsISSUE 7 March 2007

B1 INTRODUCTION This appendix describes the attributes of vibration measurement equipment that were considered to be required to deliver the measurement protocol. Consideration was given to the conflicting issues of the need for intensive and precise monitoring within properties and a need for an adequate uptake from property owners.

There requirement for 24 hour logging within occupied properties was considered. This would limit the scope for using cabling to external logging devices. The size and amount of equipment required may also be a factor in whether a survey would be acceptable to a property owner.

A further key issue was the need to record ‘raw’ vibration time histories (for potential further analysis) and ‘on the fly’ calculation of vibration and noise indicators.

Ideally, a single piece of equipment would exist that could meet all of the (sometimes conflicting) requirements given below.

The ideal requirements were compared with equipment available to the project partners ‘off the shelf’. This allowed an assessment of whether any current equipment could deliver the greater part of what was required or whether bespoke equipment would need be assembled.

B2 THE IDEAL INTERNAL VIBRATION ACQUISITION AND ANALYSIS DEVICE To be able to provide all of the functions set out in the DEFRA project specification, the Ideal Internal vibration Acquisition and Analysis device would require an assemblage of components each providing an important function for the device as a complete system. The type of components required is given in Table B1 and schematic of how these components would need to be assembled is shown in Figure B1.

In this section the required specification of each component is discussed and where ‘off the shelf’ components exist, examples of the choices available are discussed and critically compared.

Component Function

1. Vibration transducer/tranducers

To measure ground borne vibration.

2. Sound pressure transducer To measure ground borne noise.

3. Recording/Storage device To acquire and store measured vibration/noise signals.

4. Processor To analyse stored data.

5. Mounting device To provide good coupling between transducers and vibrating surface.

6. Power source To provide uninterrupted power to the system.

7. Casing To house system and provide tamper proofing.

8. Trigger To notify the system of external events.

Table B1. The components required for the Ideal Internal Vibration Acquisition and Analysis device.




3.

4.

5.

8.

2.

7.

1.

6.

Figure B1. A schematic of the Ideal Internal Vibration Acquisition and Analysis device

B2.1 Measurement parameters

The measurement protocol requires vibration measurement simultaneously in the z, x and y axes and down to levels within the lower quartile of human perceptibility. This would require either three vibration transducers or a transducer with tri-axial capability.

The effect of what ISO 2631 terms “associated phenomena” (such as airborne noise, structure-borne or groundborne noise from walls or the ‘rattling’ of objects within the property) on a person’s perception of vibration were also to be monitored. A noise transducer was therefore to be included.

B2.2 2.2 Transducers

Three tri-axially mounted transducers were required to measure vibration in the frequency range 1 Hz to 80 Hz. The interquartile range of an individual’s ability to perceive vibration may extend from 0.01 m.s-2 to 0.02 m.s-2 (BS 6841:1987). Although not perceptible to human beings as whole body vibration, significant human response due to groundborne noise can occur at levels of 35dB LAmax,S which is approximately equivalent to 5 × 10-3 mm.s-1 r.m.s. in the frequency range 50 Hz to 250 Hz. The peak vibration due to internal sources, such as footsteps, may be as much as 1 mm.s-1 .

It is important that transducer resonances do not occur in the 1 Hz to 80 Hz frequency range. If piezoelectric type acceleration transducers are used it is unlikely that this will occur. However, velocity type transducers, such as geophones, have a low internal natural frequency. The effect of this can be corrected either pre or post acquisition if only the amplitude of the vibration is of interest. Correction of the phase distortion caused by the low frequency resonance in the transducer is less straight-forward.

A standard calibrated Sound Level Meter type microphone and pre-amp is sufficient for recording audible groundborne noise.

B2.2.1 Accelerometers To measure vibration within the lower quartile of human perceptibility a transducer with a noise floor of 0.001 m.s-2 is required. This is within the range of many acceleration




transducers on the market. However, if it is required that the system is to measure all non-perceivable vibration, a transducer noise floor of approximately 3 × 10-6 m.s-2 is needed for a signal-to-noise-ratio of 20 dB to be achieved. A number of charge-type accelerometers are available for this application, such as the B&K type 4321 and ENDEVCO 2230 tri-axial accelerometers. These transducers have a relatively low mass and only a single transducer is required to measure all three axes of vibration. However, a three channel signal conditioner is required.

If low impedance line-driven transducers are used then no charge amplifiers are required, as signal conditioning is applied directly from most recording devices. If no amplification is applied to the acceleration signal a transducer of sensitivity 10 V/g would be required for sufficient resolution on a typical 10 V recording device. 10 V/g is the limit to the sensitivity of commercially available accelerometers and can be resolved with large, typically high mass, seismic accelerometers such as the ENDEVCO type 86 accelerometer. Furthermore tri-axial transducers with a sensitivity of more than 1 V/g are uncommon, meaning that 3 high mass accelerometers would be required. The only relatively low mass 10 V/g accelerometer that has been found in this initial study is the PCB type 393B05 accelerometer.

The PCB type 3703G2FD3G accelerometer is a tri-axial capacitance accelerometer with a sensitivity of 1 V/g. This transducer is lightweight and would have the required sensitivity if amplification is used, though the fragility of capacitance transducers must be considered.

Typically the low frequency residual noise on line-drive accelerometers is of the order of 1×10-6 m.s-2 . This brings in to question the feasibility of measuring such low accelerations using accelerometers.

B2.2.2 Geophones Geophones are a cheap and robust alternative to acceleration transducers. Sensitivity and noise are not an issue with this type of transducer. It was mentioned above that geophones have an inherent natural frequency, typically occurring in the 1 Hz to 10 Hz frequency range. This natural frequency is a function of the moving mass and suspension stiffness of the transducer. At this natural frequency there is a peak in the sensitivity of the response. This can be corrected by damping this resonance, which causes a phase lag in the system. For the pilot study, only amplitude data is required and therefore this is not an issue. However, below the transducer resonance, even after damping has been applied, the sensitivity rapidly rolls off with decreasing frequency. For a flat transducer response in the 1 Hz to 80 Hz, a geophone with natural frequency of less than 1 Hz is required. Although such transducers exist, they are costly and less robust than lower specification geophones.

B2.3 Data acquisition and storage

In order to record tri-axial vibration and background noise simultaneously, a stable, error free 4-channel recording device is required, such as a HDD recorder. A 16-bit device would provide an adequate dynamic range.

If line-drive transducers are used, constant current analogue inputs on each channel would be valuable as the lack of these inputs will result in a need for separate signal conditioning units. It is also probable that the acquisition device would need some form of trigger input or the ability to trigger from a measurement channel. (See below).

While real-time determination of measurement indices would greatly reduce the time required for post processing, it would also increase the likelihood of ‘void’ measurements, as no checking could take place to make sure only external sources have been included in the analysis. For this reason it was important to record raw vibration time history data so that anomalies could be checked and further post-processing performed if required.




Furthermore, measurements when external sources are not present would provide useful background noise data.

Privacy concerns denoted that raw sound time history data should not be recorded if unattended monitoring were to be used. While such data would be useful, this will reduce the required data storage space of the device if full audio frequency recordings were to be made.

The space required for data storage will be set by the raw vibration data. For example, a sampling frequency of 1024 For 3 vibration channels recording for a maximum 24-hour period, storage space of approximately 700 Mb is required. Alternatively, the acquisition device should have the capability of sending measured data wirelessly to a separate device for storage.

Many options for acquisition are available of the shelf. Each has its own advantages and limitations. Some examples of standard equipment that was considered for the pilot study are given below.

B2.3.1 Sound Devices 744T This is the lowest cost device of those researched (around £3000). It is a 4-channel digital HDD recording device developed for the broadcasting industry. The strength of this device is that the storage capacity can be up to 40 GB depending on the internal HDD chosen. However it only has two non-microphone line-inputs with no ICP/IEPE conditioning and the low frequency limit of the device is 20 Hz. This product is therefore not suitable for this application.

B2.3.2 Squadriga 1369 The 1369 is a compact 4-channel device costing around £4000. This device has ICP/IEPE capability and it is possible to trigger the device remotely. A disadvantage is that it has insufficient battery life for this project. The low frequency limit of the device is insufficient at 4 Hz.

B2.3.3 TEAC ES8 The ES8 is the most compact device of those researched. It is an 8-channel device with a frequency range of 0 Hz to 2 kHz. It has triggering capability and internal memory will be sufficient if correct flash memory is chosen. No ICP/IEPE is available for the inputs. Therefore a separate signal conditioning device would be required.

B2.3.4 ROGA DAQ This is a pc based acquisition system allowing 16 ICP/IEPE inputs. The frequency range is 1 Hz to 20 kHz. As this is pc based system, it has no internal recording capability. This would be achieved with a notebook pc running in parallel to the device.

B2.3.5 National Instruments USB 9233 This is another pc-based acquisition system that is connected to a pc via a USB connection. It has 4 ICP/IEPE enabled inputs and a lower frequency limit of 0.5 TEAC GX-1

This is the most expensive of the acquisition devices investigated (>£10000). However it is the only device that meets all aspects of the required specification.

B2.3.6 RION DA-20 The Rion DA-20 is a compact handheld 4-channel hard disk data recorder. Data is saved by the DA20 as raw time history on 2Gb flash cards in wave (.wav) format. Each of the 4 inputs are compatible with a variety of transducers and ICP/IEPE can be selected on a




channel-by-channel basis if line-drive transducers are to be used. The low frequency limit of the device is specified as <1Hz. The cost of each RION DA-20 unit is approximately £3000.

B2.4 Analysis requirements

The ability to download vibration and noise measurement indices directly from the acquisition device at the end of each survey would minimise post-processing. For this, the device would need on-board processing capabilities and/or software. The system would also require event identification.

Important vibration indicators that must be processed by the device are basic parameters such as rms and peak particle velocity (PPV), duration of vibration exposure and crest factor. More complex statistical quantities are the vibration dose value (VDV with a range of different frequency weightings) and the frequency spectra and total levels of the signals. Similar statistical noise indicators are also required such as Leq , Lmax and LA10 .

If some signal recognition were necessary then additional on-board real-time signal processing is required.

It is possible that real-time processing could be achieved with a palm-top type system running standard matrix algebra software such as MATLAB. With such a system, code could easily be implemented to perform the statistical calculations in both the time and frequency domain. Furthermore the size of most palm-top systems is in keeping with the desire to minimise the size of the complete device.

B2.5 Noise Filtering

A significant challenge for developing the measurement protocol is that it is necessary to perform vibration and noise surveys within occupied properties. In practice it is possible that the internal vibration created by residents will be higher in level and longer in duration than that due to external sources. This creates a problem of event identification and signal masking.

Internal sources of vibration, such as those created by or under the control of the residents needed to be excluded from measurement. Vibration due to these sources therefore needed to be not recorded by the device or was to be rejected before analysis of the data. The following options were been identified to address this issue.

B2.5.1 Triggering It would be possible to trigger the recording device within the property using a transducer located in another location, ideally as close as possible to the source to reduce the chances of false triggering.

It was conceivable that the measurement location could be up to 300 m from the source. The use of long cables between the trigger and equipment inside a property would therefore not be practicable. A wireless system would therefore be required to perform the triggering. Alternatively, time synchronisation and post processing of the data could be used, although this would increase the storage capacity of the equipment, so that it could accommodate valid and spurious events.

Methods of wireless communication that could be used to implement a remote trigger are Bluetooth, WiFi, GSM or RF communication. It is unlikely that Bluetooth or WiFi will provide the required range for this application. In theory, the triggering range that could be achieved with a GSM technique is limitless. However the method is not reliable in terms of delivery time of the triggering signal, which is essential in this case.




The most promising option appeared to be ‘Radio Telemetry’. The technique transmits analogue or digital signals at radio frequencies in the 100 MHz to 1000 MHz frequency band and is already well proven in many monitoring applications, such as temperature monitoring or large scale fleet monitoring. The approach considered for vibration measurement protocol was that a transducer mounted near the source would act as a trigger sending a signal would via radio telemetry transmitter. The signal received at the telemetry receiver attached to the acquisition device inside the residential property would then trigger a recording.

This method could reliably transmit to an acquisition device inside a building. It would not be possible to transmit from inside tunnels so the trigger would need to be located on the surface in the case of an underground source. Over a distance of 300 m the transmission delay of the trigger signal would be around 2 s. This would not cause a problem if the recording system was constantly buffering for approximately 10 s for example.

Preliminary studies indicated that a complete radio telemetry system would cost around £1000.

B2.5.2 Signal rejection As an alternative to triggering, it is possible to identify unwanted events and reject them from the analysis. As with the triggering approach, a signal rejection approach could be applied in a number of different ways.

Firstly, triggers could be set up within the property to identify internal events, such as doors slamming or household appliances operating. However, this method is likely to be impracticable due to the potentially unlimited number of internal events occurring within the property.

Another approach would be to rely entirely on analysis of each event. If software could be implemented that could analyse each event and identify a signal with similar time and frequency characteristics to those of the external source, internal events could be rejected. This approach would ideally be applied real-time to save on post-processing. However, each device would have to be ‘trained’ to identify the characteristics of external events particular to each property and the statistical robustness of, or the development time involved with creating such software are areas of high uncertainty.

If a solely triggering or rejection approach were adopted there would be no means to prevent the recording of simultaneous internal and external events. This suggests that some aspects of both approaches would be required.

Additionally, the statistical likelihood of all useful data being masked by internal events was considered, as this would fundamentally make the required measurement protocol impossible.

Triggering and signal rejection issues can be minimised by carrying out fully attended monitoring to apply these functions manually.

B2.6 Installation

B2.6.1 Power source The use of power cables running from the mains supply within the room to the centre of the room may be considered intrusive. Furthermore, householders may be unhappy about the cost to themselves of the electricity required, particularly if equipment were to be installed for a long period. The equipment should therefore include an internal battery supply that will power the device for up to 24 hours. For reasons of practicality it would also be useful to be able to power the device directly from the mains if required.




B2.6.2 Transducer mounting The transducers must be mounted directly to the vibrating floor, either rigidly or sufficiently held down by its own weight. Ideally the device would have ‘feet’ to allow for a large range of measurement surfaces such as carpet or wooden floors, without causing damage. Spiked feet would be required to make good contact with the floor through carpet while domed feet would be used on solid surfaces.

Consideration was given to the mounting resonance of the device mounting resonance of the device, which must be outside the 1 Hz to 80 Hz frequency range. Additionally the dynamics of the mounting must ensure that no whole body resonance of the device on the floor occur in this range.

B2.7 Casing

As the data acquisition system may have needed to be left unattended in occupied properties (if unattended monitoring were to be used), it must be tamperproof. This could be achieved with a lockable sealed case. To enable the device to be used to measure vibration outside, if required, or to be used as a trigger for a sister device, a weather proof case would be advantageous.

B2.8 Size and weight

The Association of Noise Consultants (ANC) guide to the measurement and assessment of groundborne noise and vibration recommends a measurement location close to the centre of occupied rooms. This highlights the requirement for a device of limited intrusiveness in terms of its size, particularly for unattended monitoring.

A practical limitation to the weight of the device is set by a need for a ‘surveyor’ to be able to easily transport, install and remove this device from a residential property, which may require the device to be carried up a flight of stairs. For this reason a weight of less than 10 kg (for any component) is desirable.

A lower limit on the weight of the device is set by a need for strong coupling between the device and the floor.

B2.9 Peripheral and Supporting Devices

B2.9.1 Trigger The difficulty of identifying external events has been discussed. Event triggering was identified as a method of increasing the likelihood of recording only external events. If triggering were used, a separate supporting device would be required. The triggering system would require 3 components:

• a transducer to measure a trigger signal when an event at the source occurs

• a signal conditioner to convert the signal to a communicable entity

• a means of transmitting the trigger signal to the vibration acquisition and analysis device

Options for transmitting the trigger signal to the vibration acquisition and analysis device have already been discussed. Therefore thought is confined to acquiring a trigger signal in this instance. Ideally the triggering device should be located as close to the source as possible, as this is where the source signal is likely to be of the highest magnitude. It would be possible in many cases to locate a transducer at the source i.e. a strain gauge mounted on the track if the source is a railway. However, this method would require access to the




source which may be inaccessible. Additionally, triggering from underground sources is limited by the need for a clear communication path to the primary device.

A way to ensure the triggering device could be used for many different source types is to allow triggering from a location as close to the source as possible. This could be achieved by measuring groundborne vibration a few metres from the source or on the surface in the case of an underground source. The choice of vibration transducers is as discussed for the Vibration Acquisition and Analysis device above and as the only function of this transducer is to identify an event there are no strict requirements on the specification (low frequency limit, noise floor, sensitivity. For this reason a low spec geophone will be adequate and inexpensive for this application.

Now considering a signal conditioning device, a simple ‘stop-go’ signal is all that is required. For this application a trip-amp could be used. This device, when connected to a transducer as an input, will output a ‘go’ signal when the input signal reaches a pre-defined level. A trip amp would cost around £150.

B2.9.2 External measurement A means to investigate the relationship between vibration within a property and vibration outside a property (transfer functions) was required, as this may lead to a method of assessing the human response to vibration within residential properties using only external measurements. As for the Ideal Internal Vibration Acquisition and Analysis device, an external device would require almost identical functions and specifications. This leads to a view that two identical devices could be used to simultaneously measure vibration and noise inside and outside properties. Such an approach would make this device more appealing commercially, had it been decided to assemble a bespoke piece of equipment for this application.

Some modifications and considerations must be made if the device were to be used externally. The casing would need to be weatherproof. The mounting system would require modification, for example the vibration transducers would need to be mounted on ground spikes that have been driven into the ground if the surface at the façade of the property is soil or grass.

B3 PROCURING THE IDEAL INTERNAL VIBRATION ACQUISITION AND ANALYSIS DEVICE In the previous section the required functions of the Ideal Internal Vibration Acquisition and Analysis device with respect to the project requirements has been discussed together with the type and specification of the components that would make up the system. The possibility that some elements of the specification may not be met with commercially available equipment has also been mentioned. Here, the ways in which this device can be procured and assembled in a practicable manner are discussed.

Investigation of commercially available devices that could perform the required functions up to the required specification has shown that there is currently no “off the shelf” solution available that is wholly suited to delivering what would be seen as the ideal measurement protocol. Two remaining approaches to delivery of the project are therefore available, which are summarised in Figure B2 and set out below.

Approach 1: use equipment currently available to the project partners (either available in house or newly procured) and adapt the measurement protocol to fit. This may mean that some aspects of the protocol are compromised, although these would be managed to yield the best result. For example, Arup Acoustics’ multi-channel equipment could be used to




measure vibration at up to 5 positions simultaneously, but requires cables between transducers and the data acquisition unit. This makes measurements inside property difficult, due to the need to pass cables through a window or door and trail them around in the volunteers’ living space. There also is a difficulty in crossing roads. On the positive side, this equipment would provide time synchronised simultaneous data and the ability to trigger from an instrument close to the vibration source. This could be used for external measurements whilst alternative (e.g. Svan 948) more portable equipment was used to log inside a property.

Approach 2: develop new equipment to a specification that is as close as is practicable and economically viable to the specification of the idealised monitoring equipment described above, a specification that has been developed to fulfil the detailed monitoring protocol.

The investigation of possible equipment options is on-going, and the requirement is dependent upon the final preferred and agreed measurement protocol, but it may prove necessary to consider commissioning some equipment or to revise the scope and approach of the methodology to suit existing equipment.

FIGURE B2. Summary of options to delivery of the measurement protocol

B3.1 Investigation into availability of ‘ideal’ components

An investigation of the available “of the shelf” components that could be assembled according to Figure 1 to form the Internal Vibration Acquisition and Analysis device was conducted. Each component was assessed according to its ability to meet the ‘ideal’ specifications and if the equipment is already owned by each of the partners involved with the project.

The results of this investigation are shown in Tables B2 to B4.

B3.1.1 Vibration transducers In Table B2 commercially available vibration transducers have been assessed according to their low frequency (LF) limit, sensitivity, noise floor, whether or not it is a tri-axial transducer (as this will reduce required number of units) and portability.

In general the Geophones provide a low cost solution, however it is only the LF-24 that has the required low frequency limit. This is the most expensive and least portable of all of the geophones investigated, particularly as 3 units would be required for a single device.

Ideal equipmentOptimum measurement protocol

Time, budget, practicability and delivery uncertain

Available equipment



Approach 1

Approach 2

Ideal equipmentOptimum measurement protocol

Time, budget, practicability and delivery uncertainIdeal equipment


Time, budget, practicability and delivery uncertainTime, budget, practicability and delivery uncertain

Available equipment



Available equipment


Time, budget, practicability and delivery more predictableTime, budget, practicability and delivery more predictable

Approach 1

Approach 2




Although the low cost geophones show poor LF performance, these transducers will be adequate for use as part of a triggering device.

Many types of accelerometer have been found which have a sufficient LF limit. There is little difference in the total cost of the required accelerometers. Charge-type transducers are in inherently less portable due to a need for an accompanying signal conditioner, while the more portable line-drive accelerometers do not have the required noise floor. It must be noted that the line drive transducers have a low enough noise floor to measure vibration in the lower quartile of human perceptibility, but not to measure non-perceivable vibration.

This investigation showed that it would not be possible to meet ‘ideal’ specification with commercially available transducers. To what level the choice of transducer would compromise the ‘ideal’ specification will be down to the low frequency limit vs. portability, if geophones were chosen, and portability vs. noise floor if accelerometers were chosen.

Considering the low frequency limit and portability as essential elements of the specification, the highest performing transducer is the PCB 3703G2FD3G tri-axial capacitance accelerometer.

B3.1.2 Data acquisition The data acquisition devices have been assessed according to their LF limit, dynamic range, ICP/IEPE capability (as this without a signal conditioner is required), the ability to store raw time history data, battery life and portability.

Two mid-priced complete vibration monitoring devices have been investigated, the Instantel Minimate and the Profound VIBRA. The advantage of these systems is that they are “off the shelf” systems that perform all of the functions shown in Table 1 with no modification to the device. They are also the only systems which have the required battery life. However these systems have no capability of recording raw vibration time history data and the geophones included with these devices (Table B2) do not meet the ‘ideal’ specification. Therefore the use of these devices would compromise the ‘ideal’ specification in many areas.

The remaining systems shown are either stand-alone recording systems or pc-based systems. None of these systems possess the required battery life. Therefore a separate power source is required. Of the stand-alone systems, only the TEAC GX-1 and ES8 have the required low frequency limit and are the most expensive systems with the latter also having no ICP/IEPE capability. If the ideal low frequency limit was compromised to 4 Hz the Squadriga would provide a high performance compact stand-alone system.

The remaining systems are each PC-based. No compromise would need to be made to the ‘ideal’ device assuming a separate power supply could be found. The preferred system would be the NI USB 9233 system as it is compact and could easily be connected to an accompanying pc via a USB connection. The need for accompanying control system is the main disadvantage of these devices. However this will have a limited effect on the portability of the system and, in any case, a processing device, such as a notebook pc, is required as part of the Vibration Acquisition and Analysis device.

B3.1.3 Sound pressure transducers The requirements of the ‘ideal’ sound pressure transducer a less demanding as this is a more typical, well practiced method, additionally sound pressure is a secondary measurement to accompany the vibration data. Many adequate systems exist. A few examples are presented in Table B4.

The low-cost, portable solution would be to use a standard microphone and pre-amp and to connect this to a pc-based acquisition system or stand-alone HDD recorder. Such a setup




would not require a separate microphone power supply if the TEAC GX-1, Squadriga, NI USB 9233 or ROGA systems.

A more high-cost solution would be to use a standard sound level meter (SLM), such as a B&K 2260 which include all inbuilt software to perform real-time analysis of the noise data. A number of SLM devices are already available to the partners, making cost less of an issue. Although marked as ‘not portable’ in Table 4, newer models of SLM are relatively compact.

M:\119573-00\OUTGOING DOCUMENTS\FINAL REPORT\FINAL ISSUE\APPENDICES\APPENDIX_B_EQUIPMENT.DOC Page B12 TRL Temple ISVR Arup Acoustics 7 March 2007


Type Manufacturer Model Procure? LF limit (Hz)

Sensitivity Noise Tri-axial

Portability Cost (£)

Notes

Tri-axial capacitance ICP accelerometer

PCB 3703G2FD3G 0 1255 Noise = 1.5 × 10-4 m/s² @ 1 Hz

Seismic ICP accelerometer

PCB 393B05 0.7 1470 Noise = 1.3 × 10-5 m/s² @ 1 Hz. Poor portability as 3 units required. Price for 3 units.

Seismic ICP accelerometer

PCB 393B12 3 in Aac 0.15 Noise = 1.3 × 10-5 m/s² @ 1 Hz. Poor portability as 3 units required.

Seismic IEPE accelerometer

B & K Endevco 86 0.5 Poor portability as 3 units required

Tri-axial charge-type accelerometer

B & K Endevco 2230

1 2842 Poor portability as 3-channel charge amp required

Tri-axial charge-type accelerometer

B & K 4321 0.1 2613 Poor portability as 3-channel charge amp required

Geophone I-O SM-6 4.5 450 price for 3 units

Geophone I-O LF-24 1 2025 3 large units required, price for 3 units.

Minimate geophones

Instantel 2 2160 price for 3 units

Geophone Profound VIBRA 8 Poor LF limit, however tri-axial

Table B2. Summary of commercially available vibration transducers



Type Manufacturer Model Procure? LF

limit (Hz)

Dynamic range

ICP/IEPE? Time history

Battery Portability Cost (£) Notes

8-channel HDD recorder

TEAC GX-1 10 >10,000


Head Acoustics

Squadriga 4 4000

4-channel PC based USB DAQ system

National Instruments

USB-9233

0.5 2140 Cost includes £1000 for notebook PC

16-channel PC based DAQ system

Roga DAQ16 1 4495 Cost includes £1000 for notebook PC


TEAC ES8 0 6750

4-channel HDD Recorder

RION DA-20 1 3000

Vibration monitor

Instantel Minimate 2 N/A 2290

Vibration monitor

Profound Vibra 8 N/A

Table B3. Summary of commercially available data acquisition devices.



Type Manufacturer Model Procure? Freq

Range (Hz)

Dynamic range

Portability Cost (£)

Notes

Sound level meter B & K 2260 time history can be recorded via line out

Sound level meter B & K 2250 5000

Sound level meter B & K 2236

Microphone & preamp

B & K Many 1000 Wired directly into pc-based DAQ device

Table B4. Summary of some commercially available sound pressure transducers.

Appendix C DATS Software Output


M:\119573-00\OUTGOING DOCUMENTS\FINAL REPORT\FINAL ISSUE\APPENDICES\APPENDIX_C_DATS SOFTWARE OUTPUT.DOC

Page C1 TRL Temple ISVR Arup Acoustics 7 March 2007

Event Duration s Weighting Crest Factor RMS m/s2 VDV eVDV MTVV46 30.0 Wg 6.90130 0.00048 0.00204 0.00157 0.0013847 30.0 Wg 4.56312 0.00086 0.00324 0.00281 0.0018448 30.0 Wg 4.57192 0.00191 0.00656 0.00627 0.0036249 30.0 Wg 6.77224 0.00043 0.00173 0.00140 0.0010650 30.0 Wg 5.13552 0.00204 0.00756 0.00668 0.0039551 30.0 Wg 7.59098 0.00116 0.00528 0.00382 0.0036052 30.0 Wg 8.27697 0.00038 0.00170 0.00124 0.0012053 30.0 Wg 6.97672 0.00073 0.00303 0.00239 0.0018554 30.0 Wg 8.60995 0.00082 0.00375 0.00268 0.0026955 30.0 Wg 5.79094 0.00132 0.00496 0.00432 0.0032156 30.0 Wg 7.57425 0.00050 0.00214 0.00164 0.00137

Channel 1 - Vertical z - axis

0

20

40

60

80

100

120

3.15 4 5 6.3 8 10 12.5 16 20 25 31.5 40 50 63 80 100 125 160 200 250 315 400 500 630 800 1000

one-third octave band centre freq Hz

Velo

city

dB

re 1

e-9

ms-

1

St. Neots ExternalSt. Neots Internal

Barnet ExternalBarnet InternalWood Green ExternalWood Green Internal

Epsom ExternalEpsom Internal

C1 DATS Software Output For the Pilot Study, unweighted acceleration data were extracted from the raw time history files. The Prosig DATS software has the ability to extract other parameters that may be required for investigation of a dose-response relationship. Examples of some of the available output options are illustrated below.

Appendix D Questionnaire

Version 12/01/07 Page 1 of 33

Defra NANR172

Human Response to Vibration in Residential Environments

Social survey questionnaire



SECTION A: interview and dwelling information

1. Serial number

2. Address

Interview details

3. Date

4. Interview start time

5. Interview end time

6. Site number

Dwelling information

7. In which of the following is the dwelling situated? Mark one of the following Centre of a large city

1 Village or small town 4

Suburbs/outskirts of a large city 2 Countryside

5

Large town or small city 3 Other

6

(please specify) ……………………..

8. Type of dwelling

detached 1 apartment 5

Semi-detached 2 bed-sit 6

terraced 3 Other 7

End terrace 4 (Please specify) ……………………….……..

Ground floor 1 2 3 Other (specify) 9. a. Number of storeys if detached, semi

or terraced.

1 2 3 4 5 …………………

Ground floor 1st 2nd 3rd Other (specify) OR

b. Floor if apartment or bed-sit 1 2 3 4 5 ………………….



SECTION B: Occupancy information

Introduction to be read to interviewee

We are conducting a survey on behalf of the Department for the Environment, Food and rural Affairs (DEFRA) on the effects of local traffic on people in their homes. We are interested in your views about living in this area and how satisfied you are with different aspects of your neighbourhood and home. The survey will take 30- 40 minutes. If you are not able to take part now, is there a more convenient time that I could call back?

Appointment time …………………………………………

10. How long have you been living at this address? Select one. Show card 1

Less than 6 months 1

Please specify how long. … ………………………………………………..

(If less than 1 week, please thank and close)

6 months up to 1 year 2

Over 1 year, up to 2 years 3

Over 2 years, up to 5 years 4

Over 5 years 5

11. During a typical weekday (Monday to Friday), how many hours, both day and night, do you spend inside this house? Select one. Show card 2

Less than 10 hours 10 -14 hours 14-18 hours More than 18 hours

1 2 3 4

12. During a typical weekend day (Saturday or Sunday), how many hours, both day and night, do you spend inside this house? Select one. Show card 2

Less than 10 hours 10 -14 hours 14-18 hours More than 18 hours

1 2 3 4

13. How many adults and children live at this address?

Number of adults (18 years or over) Number of children (under 18 years)

1 2



SECTION C Satisfaction with neighbourhood

14. On the whole, how much do you like living in this neighbourhood? Choose a number to show your rating. Show card 3

don’t like at all 1 2 3 4 5 6

like very much 7

1 2 3 4 5 6 7

15. How would you rate this neighbourhood for each of the following? Show card 4

Very poor

1 2 3 4 5 6

Excellen

t 7

Don’t know

a. Being close to the shops 1 2 3 4 5 6 7 8

b. Standard of the local schools 1 2 3 4 5 6 7 8

c. Up-keep of roads 1 2 3 4 5 6 7 8

d. Appearance of buildings 1 2 3 4 5 6 7 8

e. Quietness/peacefulness of the area 1 2 3 4 5 6 7 8

f. Parks and open spaces 1 2 3 4 5 6 7 8

g. Public transport 1 2 3 4 5 6 7 8

16. Is there anything you particularly like about this neighbourhood? (Respondent may select more than one.) Please prompt with “anything else you particularly like?” to obtain several responses

a. Any mention of quietness/ peacefulness / freedom from noises or disturbances (specify)

………………………………………………………………………………………..….………… 1

b. Any mention of sounds / noises that are liked (specify) …………………………………….…… 1

c. Any mention of vibrations that are liked (specify) …………………………….…………..……. 1

d. Any mention of possible sources of noise that are liked (specify) …………………………….... 1

e. Any mention of possible sources of vibrations that are liked (specify) ………………………….. 1

f. Any mention of aspects of railway/road/traffic that are liked (specify) ………………..………… 1

g. Any other features that are liked (specify) …………………………………………….….....…… 1

h. Nothing liked 1



17. Is there anything you particularly dislike about this neighbourhood? (Respondent may select more than one.) Please prompt with “anything else you particularly dislike?” to obtain several responses

a. Any mention of noise or vibration/ noise or vibration effects / lack of peace and quiet /

disturbances (specify) ……………………………………………………………..………. 1

b. Any mention of noises that are disliked (specify) ……………………………….….…………… 1

c. Any mention of vibrations that are disliked (specify) ……………………………..…………..… 1

d. Any mention of possible sources of noise that are disliked (specify) …………….…..….……… 1

e. Any mention of possible sources of vibrations that are disliked (specify) ………….…….……... 1

f. Any mention of aspects of railway/road/traffic that are disliked (specify) …….………….…… 1

g. Any other features that are disliked (specify) ……………………….……..………………..…… 1

h. Nothing disliked (Go to Q19) 1

18. You mentioned that you dislike ……..…. why is that? (Record any effects on daily life for each item selected in Q17) (insert letter of selected items in Q17 in box above comments column)

a. Noise or vibration / noise or vibration effects / lack of piece and quiet / disturbances ……………… ……………… ………………

b. Noises that are disliked ……………… ……………… ………………

c. Vibrations that are disliked ……………… ……………… ………………

d. Sources of noise that are disliked ……………… ……………… ………………

e. Sources of vibrations that are disliked ……………… ……………… ………………

f. Aspects of railway/road/traffic that are disliked ……………… ……………… ………………

g. Other features that are disliked ……………… ……………… ………………



19. On the whole, how much do you like living in this home? Choose a number to show your rating.

Show card 3:

don’t like at all

1 2 3 4 5 6

like very much

7

1 2 3 4 5 6 7

20. What are the things you like about living in this home? (Respondent may select more than one.) Please prompt with “anything else you particularly like?” to obtain several responses

a. Any mention of quietness/ peacefulness / freedom from noises or disturbances (specify)

………………………………………………………………………………………………. 1

b. Any mention of sounds / noises that are liked (specify) ……………………………..…………. 1

c. Any mention of vibrations that are liked (specify) ………………………………….…………….. 1

d. Any mention of possible sources of noise that are liked (specify) ………………………….…… 1

e. Any mention of possible sources of vibrations that are liked (specify) …………………………. 1

f. Any other features that are liked (specify) ……………………………………………….….....… 1

g. Nothing liked 1

21. Is there anything you dislike about living in this home? (Respondent may select more than one.) Please prompt with “anything else you particularly dislike?” to obtain several responses

a. Any mention of noise or vibration/ noise or vibration effects / lack of piece and quiet /

disturbances (specify) ……………………………………………………………..………. 1

b. Any mention of noises that are disliked (specify) ……………………………….….…………… 1

c. Any mention of vibrations that are disliked (specify) ……………………………..…………..… 1

d. Any mention of possible sources of noise that are disliked (specify) …………….…..….……… 1

e. Any mention of possible sources of vibrations that are disliked (specify) ………….…….……... 1

f. Any mention of aspects of railway/road/traffic that are disliked (specify) …….………….…… 1

g. Any other features that are disliked (specify) ……………………….……..………………..…… 1

h. Nothing disliked (Go to Q23) 1

SECTION D: Satisfaction with home



22. You mentioned that you dislike ……… why is that? (Record any effects on daily life for each item selected in Q21) (insert letter of selected items in Q21 in box above comments column)

a. Noise or vibration / noise or vibration effects / lack of piece and quiet / disturbances ……………… ……………… ………………

b. Noises that are disliked ……………… ……………… ………………

c. Vibrations that are disliked ……………… ……………… ………………

d. Sources of noise that are disliked ……………… ……………… ………………

e. Sources of vibrations that are disliked ……………… ……………… ………………

f. Aspects of railway/road/traffic that are disliked ……………… ……………… ………………

g. Other features that are disliked ……………… ……………… ………………

SECTION E: vibration questions

23. Thinking about the last 12 months (or over the period of occupancy if less than 12 months), when you are in your home, do you ever feel vibration or shaking of the following? You may select more than one.

Show card 5:

a. The floor 1

b. A chair in which you are sitting 1

c. A bed in which you are lying 1

d. Other (specify) …………………………………………………….. 1

e. None felt (Go to Q26) 1



24. In which rooms do you feel this vibration or shaking? (specify)

…………………………………………………………………………………………………………………

25. What do you think this vibration or shaking that you feel is caused by? You may select more than one. Show card 6:

a. Cars, lorries, buses or other road vehicles 1

b. Aircraft 1

c. Overground trains 1

d. Underground trains 1

e. Quarrying or mining 1

f. Construction 1

g. Road works 1

h. Footfalls, slamming doors, domestic appliances inside your home 1

i. Footfalls, slamming doors, domestic appliances in neighbouring homes 1

j. Other (specify) ………………………………………………………….. 1

k. Don’t know 1

26. Does feeling vibration or shaking of the floor, chair or bed interfere with any of these aspects of your home life? You may select more than one. Show card 7:

a. Listening to TV, radio, music 1

b. Having a conversation (including on the telephone) 1

c. Reading, writing or other quiet activities 1

d. Concentrating 1

e. Resting 1

f. Sleeping 1

g. Other (specify) ……………………………………………………… 1

h. No 1



27. A. Thinking about the last 12 months (or over the period of occupancy if less than 12 months), when you are in your home; how bothered, annoyed or disturbed are you by feeling vibration or shaking of the floor, chair or bed caused by Cars, lorries, buses or other road vehicles? Show card 8:

Not at all 1

Please clarify, you never feel or sometimes feel this kind of vibration?

Never Go to Q27B

Sometimes Go to (i)

A little 2

Moderately 3 Go to Q27 B

Very 4

Extremely 5

Don’t feel 6


Never Go to Q27 B

Sometimes Go to (i)

Don’t know 7 Please clarify, you never feel or sometimes feel this kind of vibration?

Never Go to Q27 B

Sometimes Go to (i)

(i) So, can you confirm whether you are bothered, annoyed or disturbed by feeling vibration from Cars, lorries, buses or other road vehicles. Show card 8

not at all 1 a little 2 Moderately 3 Very 4 Extremely 5

27 B. Thinking about the last 12 months (or over the period of occupancy if less than 12 months), when you are in your home; how bothered, annoyed or disturbed are you by feeling vibration or shaking of the floor, chair or bed caused by Aircraft? Show card 8

Not at all 1


Never Go to Q27C

Sometimes Go to (i)

A little 2

Moderately 3 Go to Q27 C

Very 4

Extremely 5

Don’t feel 6


Never Go to Q27 C

Sometimes Go to (i)


Never Go to Q27 C

Sometimes Go to (i)

(i) So, can you confirm whether you are bothered, annoyed or disturbed by feeling vibration from Aircraft? Show card 8




27 C. How bothered, annoyed or disturbed are you by feeling vibration or shaking of the floor, chair or bed caused by Overground trains? Show card 8

Not at all 1


Never Go to Q27D

Sometimes Go to (i)

A little 2

Moderately 3 Go to Q27 D

Very 4

Extremely 5

Don’t feel 6


Never Go to Q27 D

Sometimes Go to (i)


Never Go to Q27 D

Sometimes Go to (i)

(i) So, can you confirm whether you are bothered, annoyed or disturbed by feeling vibration from Overground trains? Show card 8


27 D. How bothered, annoyed or disturbed are you by feeling vibration or shaking of the floor, chair or bed caused by Underground trains? Show card 8

Not at all 1


Never Go to Q27E

Sometimes Go to (i)

A little 2

Moderately 3 Go to Q27 E

Very 4

Extremely 5

Don’t feel 6


Never Go to Q27 E

Sometimes Go to (i)


Never Go to Q27 E

Sometimes Go to (i)

(i) So, can you confirm whether you are bothered, annoyed or disturbed by feeling vibration from Underground trains? Show card 8




27 E. How bothered, annoyed or disturbed are you by feeling vibration or shaking of the floor, chair or bed caused by Quarrying or mining? Show card 8

Not at all 1


Never Go to Q27F

Sometimes Go to (i)

A little 2

Moderately 3 Go to Q27 F

Very 4

Extremely 5

Don’t feel 6


Never Go to Q27 F

Sometimes Go to (i)


Never Go to Q27 F

Sometimes Go to (i)

(i) So, can you confirm whether you are bothered, annoyed or disturbed by feeling vibration from Quarrying or mining? Show card 8


27 F. How bothered, annoyed or disturbed are you by feeling vibration or shaking of the floor, chair or bed caused by Construction? Show card 8

Not at all 1


Never Go to Q27G

Sometimes Go to (i)

A little 2

Moderately 3 Go to Q27 G

Very 4

Extremely 5

Don’t feel 6


Never Go to Q27 G

Sometimes Go to (i)


Never Go to Q27 G

Sometimes Go to (i)

(i) So, can you confirm whether you are bothered, annoyed or disturbed by feeling vibration from Construction? Show card 8




27 G. How bothered, annoyed or disturbed are you by feeling vibration or shaking of the floor, chair or bed caused by Roadworks? Show card 8

Not at all 1


Never Go to Q27H

Sometimes Go to (i)

A little 2

Moderately 3 Go to Q27 H

Very 4

Extremely 5

Don’t feel 6


Never Go to Q27 H

Sometimes Go to (i)


Never Go to Q27 H

Sometimes Go to (i)

(i) So, can you confirm whether you are bothered, annoyed or disturbed by feeling vibration from Roadworks? Show card 8


27 H. How bothered, annoyed or disturbed are you by feeling vibration or shaking of the floor, chair or bed caused by Footfalls, slamming doors, domestic appliances inside your home? Show card 8

Not at all 1


Never Go to Q27I

Sometimes Go to (i)

A little 2

Moderately 3 Go to Q27 I

Very 4

Extremely 5

Don’t feel 6


Never Go to Q27 I

Sometimes Go to (i)


Never Go to Q27 I

Sometimes Go to (i)

(i) So, can you confirm whether you are bothered, annoyed or disturbed by feeling vibration from Footfalls, slamming doors, domestic appliances inside your home? Show card 8




27 I. How bothered, annoyed or disturbed are you by feeling vibration or shaking of the floor, chair or bed caused by Footfalls, slamming doors, domestic appliances in neighbouring homes? Show card 8

Not at all 1


Never Go to Q27J

Sometimes Go to (i)

A little 2

Moderately 3 Go to Q27 J

Very 4

Extremely 5

Don’t feel 6


Never Go to Q27 J

Sometimes Go to (i)


Never Go to Q27 J

Sometimes Go to (i)

(i) So, can you confirm whether you are bothered, annoyed or disturbed by feeling vibration from Footfalls, slamming doors, domestic appliances in neighbouring homes? Show card 8


27 J. How bothered, annoyed or disturbed are you by feeling vibration or shaking of the floor, chair or bed caused by Other things? (please specify……………………………………………….……….) Show card 8

Not at all 1


Never Go to Q28

Sometimes Go to (i)

A little 2

Moderately 3 Go to Q28

Very 4

Extremely 5

Don’t feel 6


Never Go to Q28

Sometimes Go to (i)


Never Go to Q28

Sometimes Go to (i)

(i) So, can you confirm whether you are bothered, annoyed or disturbed by feeling vibration from Other things? (please specify ……………………………………………………..) ? Show card 8




28. Do you ever hear things rattle or see things vibrate, shake or sway in your home? You may select more than one. Show card 9:

a. Windows rattle 1

b. Objects rattle 1

c. Lights sway 1

d. Other (specify)………………………………………………………… 1

e. No (Go to Q29) 1

29. What do you think this vibrating, rattling, shaking or swaying that you hear or see is caused by? You may select more than one. Show card 6:


b. Aircraft 1




f. Construction 1

g. Road works 1



j. Other (specify) ………………………………………………………… 1

k. Don’t know 1



30. Does hearing or seeing vibrating, rattling, shaking or swaying of things in your home interfere with any of these aspects of your home life? You may select more than one. Show car 7:




d. Concentrating 1

e. Resting 1

f. Sleeping 1

g. Other (specify) ……………………………………………………… 1

h. No 1 31. Thinking about the last 12 months (or over the period of occupancy if less than 12 months), how bothered,

annoyed or disturbed are you by hearing or seeing vibrating, rattling, shaking or swaying of things caused by……..? Show card 8

not at all A little Moderately very extremely don’t hear

or see

a. Cars, lorries, buses or other road vehicles 1 2 3 4 5 6

b. Aircraft 1 2 3 4 5 6

c. Overground trains 1 2 3 4 5 6

d. Underground trains 1 2 3 4 5 6

e. Quarrying or mining 1 2 3 4 5 6

f. Construction 1 2 3 4 5 6

g. Road works 1 2 3 4 5 6

h. Footfalls, slamming doors, domestic appliances inside your home

1 2 3 4 5 6

i. Footfalls, slamming doors, domestic appliances in neighbouring homes

1 2 3 4 5 6

j. Other (specify )

………………………………… 1 2 3 4 5 6



32. Thinking about the last 12 months (or over the period of occupancy if less than 12 months), how bothered, annoyed or disturbed are you by feeling, hearing or seeing vibration or shaking in your home?

Choose a number to show your rating. Show card 10:

Not at all 1 2 3 4 5 6

extremely 7

1 2 3 4 5 6 7

33. Are you concerned that this building may be damaged by vibration or shaking?

Yes No

1 2

34. On this scale, how sensitive would you say you are to vibration in general? Show card 11:

Not at all sensitive

1 2 3 4 5 6

Extremely sensitive

7

1 2 3 4 5 6 7

For railway sites only. SECTION F: railway vibration

35. Thinking about the last 12 months (or over the period of occupancy if less than 12 months), how bothered, annoyed or disturbed are you by feeling, hearing or seeing vibration or shaking in your home from the passage of nearby trains? Choose a number to show your rating.

Show card 9: Not at all

1

2 3 4 5 6 extremely

7 1

Go to Q39 2 3 4 5 6 7

36. Does the vibration or shaking in your home from the passage of trains bother, annoy or disturb you at these times during the week (Monday to Friday)? You may select more than one. Show card 12:

a. Day (07:00 - 19:00) 1

b. Evening (19:00 - 23:00) 1

c. Night (23:00 - 07:00) 1

d. None of these 1



37. Does the vibration or shaking in your home from the passage of trains bother, annoy or disturb you at these times during the weekend (Saturday and Sunday)? You may select more than one. Show card 12:

a. Day (07:00 - 19:00) 1

b. Evening (19:00 - 23:00) 1

c. Night (23:00 - 07:00) 1

d. None of these 1

38. Is there one type of train that causes the most annoying or disturbing vibration?

Yes No Don’t know

1 2 3 If yes:

(i) What type of train, passenger or freight? Passenger Freight Don’t know

1 2 3 (ii) How often does this type of train pass by?

(number per day/hour)

Per 24 hours Per hour Don’t know

1 2 3

(iii) At what time of day or night does this type of train pass by? Show card 12:

Day (07:00 - 19:00) specify time(s) if possible:

Evening (19:00 – 23:00) specify time(s) if possible:

Night (23:00 – 07:00) specify time(s) if possible:



37. Does the vibration or shaking in your home from heavy construction machinery bother, annoy or disturb you at these times during the weekend (Saturday and Sunday)? You may select more than one.Show card 12:

a. Day (07:00 - 19:00) 1

b. Evening (19:00 - 23:00) 1

c. Night (23:00 - 07:00) 1

d. None of these 1

For construction sites only. SECTION F: construction vibration

35. Thinking about the last 12 months (or over the period of occupancy if less than 12 months), how bothered, annoyed or disturbed are you by feeling, hearing or seeing vibration or shaking in your home from the use of nearby heavy construction machinery? Choose a number to show your rating.

Show card 10: Not at all

1

2 3 4 5 6 extremely

7 1

Go to Q39 2 3 4 5 6 7

36. Does the vibration or shaking in your home from heavy construction machinery bother, annoy or disturb you at these times during the week (Monday to Friday)? You may select more than one. Show card 12:

a. Day (07:00 - 19:00) 1

b. Evening (19:00 - 23:00) 1

c. Night (23:00 - 07:00) 1

d. None of these 1



38. Is there one type of heavy construction machinery that causes the most annoying or disturbing vibration?

Yes No Don’t know

1 2 3 If yes:

(i) What type of construction machinery? piling or pounding machines, or other machines or lorries?

piling or pounding machines

other machines/lorrie

s

Don’t know

1 2 3 (ii) How often does this type of disturbance occur?

(iii) At what time of day or night does this type of disturbance occur? Show card 12:






39. Thinking about the last 12 months (or over the period of occupancy if less than 12 months), when you are in your home, do you ever hear noise from the following? You may select more than one. Show card 6:


b. Aircraft 1




f. Construction 1

g. Road works 1



j. Other (specify) ………………………………………………………… 1

k. Don’t know 1 l. No 1

40. Does noise interfere with any of these aspects of your home life? You may select more than one. Show card 7:




d. Concentrating 1

e. Resting 1

f. Sleeping 1

g. Other (specify) ……………………………………………………… 1

h. No 1

SECTION G: Noise questions



41. Thinking about the last 12 months (or over the period of occupancy if less than 12 months), when you are in your home; how bothered, annoyed or disturbed are you by noise from …..? Show card 8

not at all A little Moderately very extremely don’t hear

or see

a. Cars, lorries, buses or other road vehicles 1 2 3 4 5 6

b.Aircraft 1 2 3 4 5 6

c. Overground trains 1 2 3 4 5 6

d. Underground trains 1 2 3 4 5 6

e. Quarrying or mining 1 2 3 4 5 6

f. Construction 1 2 3 4 5 6

g. Road works 1 2 3 4 5 6

h. Footfalls, slamming doors, domestic appliances inside your home

1 2 3 4 5 6

i. Footfalls, slamming doors, domestic appliances in neighbouring homes

1 2 3 4 5 6

j. Other (specify )

………………………………… 1 2 3 4 5 6

42. Thinking about the last 12 months (or over the period of occupancy if less than 12 months), how bothered, annoyed or disturbed are you by noise in your home from all sources, including air, rail, road, construction, and neighbours? Choose a number to show your rating. Show card 10:

Not at all

1 2 3 4 5 6 extremely

7

1 2 3 4 5 6 7

43. On this scale, how sensitive would you say you are to noise in general? Show card 11:

Not at all sensitive

1 2 3 4 5 6

Extremely sensitive

7

1 2 3 4 5 6 7



46. Does the noise in your home from the passage of trains bother, annoy or disturb you at these times during the weekend (Saturday and Sunday)? You may select more than one. Show card 12:

a. Day (07:00 - 19:00) 1

b. Evening (19:00 - 23:00) 1

c. Night (23:00 - 07:00) 1

d. None of these 1

For railway sites only. SECTION H: Railway noise

44. Thinking about the last 12 months (or over the period of occupancy if less than 12 months), when you are in your home; how bothered, annoyed or disturbed are you by hearing noise from …..? Show card 10:

Not at all


7 Go to Q48 Go to Q45

a. Passage of nearby trains 1 2 3 4 5 6 7

b. Train horns 1 2 3 4 5 6 7

c. Stations (loudspeakers, etc.) 1 2 3 4 5 6 7 d. Goods yards (shunting, freight handling, etc.) 1 2 3 4 5 6 7

e. Railway / track maintenance 1 2 3 4 5 6 7

45. Does the noise in your home from the passage of trains bother, annoy or disturb you at these times during the week (Monday to Friday)? You may select more than one. Show card 12:

a. Day (07:00 - 19:00) 1

b. Evening (19:00 - 23:00) 1

c. Night (23:00 - 07:00) 1

d. None of these 1



47. Is there one type of train that causes the most annoying or disturbing noise?

Yes No Don’t know

1 2 1

If yes:

(i) What type of train, passenger or freight? Passenger Freight Don’t know

1 2 3 (ii) How often does this type of train pass by?

(number per day/hour)

Per 24 hours Per hour Don’t know

1 2 3

(iii) At what time of day or night does this type of train pass by? Show card 12:






46. Does the noise in your home from the heavy construction machinery bother, annoy or disturb you at these times during the weekend (Saturday and Sunday)? You may select more than one. Show card 12:

a. Day (07:00 - 19:00) 1

b. Evening (19:00 - 23:00) 1

c. Night (23:00 - 07:00) 1

d. None of these 1

For construction sites only. SECTION H: construction noise

44. Thinking about the last 12 months (or over the period of occupancy if less than 12 months), when you are in your home; how bothered, annoyed or disturbed are you by hearing noise from …..? Show card 10:

Not at all


7 Go to Q48 Go to Q45

a. Piling or pounding machines 1 2 3 4 5 6 7

b. Lorries on construction sites 1 2 3 4 5 6 7

c. Other construction machines 1 2 3 4 5 6 7

45. Does the noise in your home from the heavy construction machinery bother, annoy or disturb you at these times during the week (Monday to Friday)? You may select more than one. Show card 12:

a. Day (07:00 - 19:00) 1

b. Evening (19:00 - 23:00) 1

c. Night (23:00 - 07:00) 1

d. None of these 1



47. Is there one type of heavy construction machinery that causes the most annoying or disturbing noise?

Yes No Don’t know

1 2 1

If yes:

(i) What type of construction machinery? piling or pounding machines, or other machines or lorries?

piling or pounding machines

other machines/lorrie

s

Don’t know

1 2 3 (ii) How often does this type of disturbance occur?

(iii) At what time of day or night does this type of disturbance occur? Show card 12:






SECTION I: Comparison of noise and vibration, glazing, traffic visibility, gender, age, ownership/tenancy

48. Thinking about the last 12 months or so, how bothered, annoyed or disturbed are you by vibration and noise in your home from all sources, including air, rail, road, construction, and neighbours?

Choose a number to show your rating. Show card 10:

Not at all 1 2 3 4 5 6

extremely 7

1 2 3 4 5 6 7

49. Considering vibration and noise in your home from all sources, including air, rail, road, construction, and neighbours, which would you prefer to be reduced, the vibration or the noise?

vibration noise

1 2

50. Railway sites only Considering vibration and noise in your home from the passage of nearby trains, which would you prefer to be reduced, the vibration or the noise?

vibration noise

1 2 50. Construction sites only

Considering vibration and noise in your home from nearby heavy construction machinery, which would you prefer to be reduced, the vibration or the noise?

vibration noise

1 2 51. Is there anything else you would like to say about vibration and noise in your home?



52. Thinking about the time while we have been speaking, would you say that you can usually feel more, less or the same vibration?

Usually more Usually less Usually same

1 2 3

53. Are all the windows in the dwelling double-glazed (full glazing), only some of them (partial glazing) or no windows double-glazed?

Full glazing Partial glazing No glazing

1 2 3

54. Are sources of noise or vibration visible? Specify (e.g. road or rail traffic, construction)

Road traffic Rail traffic Air traffic construction Other, specify

1 2 3 4 5 …………………………..

55. Is any of the following passing traffic visible from within your house from? You may mark more than one. Show card 13

a) Motorway 1

b) Other dual carriageway road 1

c) Single carriageway main road 1

d) Residential / estate road 1

e) Country lane 1

f) railway track 1

g) Other (specify) …………………………………………………… 1

h) None of these 1

56. Gender

Male Female

1 2



57. In which of these age categories are you? Show card 14:

16-18 19-24 25-34 35-44 45-54 55-64 Over 65

1 2 3 4 5 6 7

58. Which of these describes the ownership of your home? Select one.

Being bought on a mortgage 1

Owned outright by household 2

Rented from local authority 3

Rented from private landlord 4

Rented from registered social landlord (RSL) 5

Other………………………………………………… 6

End of Questionnaire Interview end time:



Ask all Respondents the following and, if they are willing to take part, hand them the explanatory letter

Interviewer: now complete Section J



SECTION J Interviewer assessment of vibration and noise

59. While in the dwelling, did you feel vibration or shaking of the following? You may mark more than one.

a. The floor 1

b. The chair in which you were sitting 2

c. Other (specify) ……………………………………………………… 3

d. None of these Go to Q61 4

60. What do think this is caused by? You may mark more than one.


b. Aircraft 1




f. Construction 1

g. Road works 1



j. Other (specify) ………………………………………………………… 1

k. Don’t know 1

61. In which rooms did you feel vibration or shaking? (specify)

………………………………………………………………………………………………………………

…



62. While in the dwelling, did you see or hear any of the following? You may mark more than one.

a. Rattling of the windows 1

b. Rattling of objects 1

c. Swaying of pendulum lights 1

d. Other (specify) …………………………………………………………… 1 e. None of these Go to Q64

1

63. What do think this is caused by? You may mark more than one. a. Cars, lorries, buses or other road vehicles 1

b. Aircraft 1




f. Construction 1

g. Road works 1

h. Footfalls, slamming doors, domestic appliances inside the home 1


j. Other (specify) ……………………………………………………………… 1

k. Don’t know 1



64. While in the dwelling, did you hear noise from the following? You may select more than one.


b. Aircraft 1




f. Construction 1

g. Road works 1

h. Footfalls, slamming doors, domestic appliances inside the home 1


j. Other 1

k. None of these 1 Location information

65. Residential area type. Mark one of the following.

Mostly residential/housing estate 1

Mixed residential/commercial (shops/offices) 2

Mixed residential/ industrial (factories) 3

Mostly industrial/ commercial, with some residential 4

Mixed residential/countryside 5

Mostly countryside 6

INTERVIEWER DECLARATION

I declare that this is a true record of an interview for this survey, and has been conducted within the MRS code of conduct. Interviewer Name

Signature Date



Interviewer, Please note below any difficulties encountered with completing questionnaire (for example questions the Respondent had difficulty in understanding, problems with the interview etc)

For office use only

66. Is any part of the dwelling located within 20 m of any of the following? You may mark more than one.

a) Motorway 1




e) Country lane 1

f) Railway track 1

g) Other (specify) …………………………………………………… 1

h) None of these 1

67. Is the dwelling located within 800 m (half a mile) of any of the following? You may mark more than one.

a) Motorway 1




e) Country lane 1

f) Railway track 1

g) Other (specify) …………………………………………………… 1

h) None of these 1

Appendix E Letter To Residents


M:\119573-00\OUTGOING DOCUMENTS\FINAL REPORT\FINAL ISSUE\APPENDICES\APPENDIX_E_LETTER TO RESIDENTS.DOC

Page E1 TRL Temple ISVR Arup Acoustics7 March 2007

E1 Letter to Residents The following is the main text from the letter to residents outlining the nature of the project and survey work.

Vibration Survey

Thank you for agreeing to help with this survey. This research has been commissioned by DEFRA (Department of the Environment, Food and Rural Affairs), and is being conducted by a partnership of Arup, Temple, TRL, and the University of Southampton.

As part of the research, the project team would like to measure vibration within as many residential properties as possible where the owner/occupier has already completed the questionnaire. At the same time, we would like to measure noise levels and the vibration occurring in the ground outside the property.

You have indicated that you would be willing to allow measurements to be made within your home. The following describes the procedure.

Measurement Procedure

The survey will be undertaken by a team of up to 2 people. Measurements will be made within one of the rooms in which vibration is felt, as indicated in responses to the questionnaire.

Vibration measuring equipment will be placed at the centre of the floor of the room. There will be no rigid connection between the transducers and the floor, meaning that floor surfaces (such as carpet or laminate) will not be damaged or tarnished. This equipment will be linked to a vibration analyser, which will also be placed in the room. The analyser is similar in size to a typical desktop computer. We would appreciate permission to plug our equipment into a mains electricity socket.

A Sound Level Meter will also be placed on a photographer’s tripod in the room to measure secondary noise effects.

In addition, cabling will be run from this analyser to equipment located outside of the property to link the two sets of measurements. The cabling with be run through existing openings in the building such as a letter boxes or windows.

Setting up the equipment will take approximately 1 hour. The surveys will take between 2 and 4 hours and will be attended by the measurement team at all times.

Removing the equipment will take approximately 45 minutes. The measurement team would need access to the measurement room during the measurement period (2 to 4 hours). It would also be very helpful if activity around the rest of the property was at a minimum during this period (to avoid interrupting the measurements).

Data Protection – Fair Collection Notice

With your agreement, your name, telephone number and address will be passed to members of the research team from Arup, Temple, and the University of Southampton to enable them to contact you to arrange a mutually acceptable time for the vibration measurements to be made.

The personal information you have provided to TRL will only be used for the purposes of this research.*


M:\119573-00\OUTGOING DOCUMENTS\FINAL REPORT\FINAL ISSUE\APPENDICES\APPENDIX_E_LETTER TO RESIDENTS.DOC

Page E2 TRL Temple ISVR Arup Acoustics7 March 2007

* The Data Protection Act 1998 gives you the right of access to your personal information held by TRL. An administrative charge of £10 (0% VAT) may be charged for such requests, and you will receive a response within 40 calendar days. Requests of this nature must be in writing, and you will be required to provide verification of your identity to authorise release of your information.

If you have concerns about the way TRL are using your personal information, contact the Data Protection Manager at the above address.

Appendix F Draft Measurement Protocol


M:\119573-00\OUTGOING DOCUMENTS\FINAL REPORT\FINAL ISSUE\APPENDICES\APPENDIX_F_DRAFT MEASUREMENT PROTOCOL.DOC


Contents F1 Measurement Protocol (draft)

F1.1 Definitions F1.2 Summary of the Measurement Protocol. F1.3 External Vibration Measurements F1.4 Internal Vibration Measurement F1.5 Noise Measurement F1.6 Measurement Duration (Control Position) F1.7 Equipment F1.8 Analysis F1.9 Reporting



Page F1 TRL Temple ISVR Arup AcousticsISSUE 7 March 2007

F1 Measurement Protocol (draft)

F1.1 Definitions

The following terminology is used in this document:

• Case study: a pairing of a questionnaire and associated measurements at one property.

• Measurement Location: a geographical area around a vibration source within which a number of case studies will be undertaken.

• Measurement Position: a position either within or external to a building at which vibration (and in some cases) noise is measured.

F1.2 Summary of the Measurement Protocol.

The following bullet points provide an outline of the approach to be taken. A more detailed description of the Protocol is given in the following subsections.

• Data shall be acquired, wherever possible, outside questionnaire respondents’ properties.

• The vibration survey shall follow as soon as practicable after the questionnaire has been conducted. Guidance on timing shall be sought from the questionnaire team on the day that the questionnaire is applied to a property.

• Suitably qualified and experience vibration and noise specialists shall conduct all measurements.

• Measurements shall be made continuously at a ‘control’ position to record all events from the source of interest. All vibration indicators of interest (see later) shall be evaluated from this ‘reference’ data.

• ‘Snapshot’ measurements shall be made outside properties for a period of time sufficient to get a representative sample of events. The acquisition of this data shall be synchronised with the control position.

• The difference between each equivalent event at the ‘snap shot’ and ‘control’ data shall be evaluated. The arithmetic average of the evaluated differences shall be used to scale the Reference indicator values to provide indicator values for each ‘snap shot’ location.

• Where permitted by residents, measurements shall be made inside properties. These shall be simultaneous with and synchronised with the external ‘snapshot measurements’.

• Where possible, internal and external measurements shall be recorded using a system that enables the two sets of data to be phase locked.

• Noise measurements shall be made simultaneously with the any internal vibration measurements.

• Raw time history data shall be obtained from each measurement position.




F1.3 External Vibration Measurements

External vibration measurements shall be undertaken as follows.

• A ‘control’ measurement position shall be established to record continuous raw, linear vibration acceleration time history data close to the vibration source, for a period to be specified (see below).

• Tri-axial measurements shall be acquired. Axes shall be vertical, parallel to the railway and perpendicular to the railway.

• For point or line sources, the control position shall be as close as reasonably practicable to the source to aid identification of events of interest. For large area sources (particularly construction sites) the positioning may be more complex and may require more than one control location. It is not necessary for measurements at the control location(s) to be attended throughout.

• Depending upon the site, it may be appropriate to use the ‘control’ location as one of the positions to characterise the vibration exposure at a property, for example by locating it at (the standardised distance from) a property which is close to a railway line.

• ‘Snapshot’ attended noise and vibration measurements shall be made outside properties from which questionnaire responses have been received

• Where reasonably practicable, Snapshot measurements shall be made at a standardised distance of 2m from the closest façade of the building to the vibration source.

• ‘Snapshot’ and ‘control’ monitors shall be synchronised. This may be achieved by ensuring that internal clocks are synchronised. Other synchronisation methods may be used.

• If a large number of questionnaire responses is received at a measurement location, snapshot measurement positions may be chosen at distances appropriate for determination of an attenuation ‘site law’ which may then be used for interpolation of the data to other properties. Extrapolation beyond the extremes of the measurement distances (to greater or smaller distances than those where measurements have been made) is not acceptable.

• If responses are received from several properties in close proximity, one measurement position may be considered representative of all of them. This will require expert consideration and depend on the distance from the vibration source, the distance between the properties, etc. Full justification for the decision to adopt this approach shall be provided in the survey report.

• Transducers shall be mounted using methods given in the Association of Noise Consultants’ Guidelines for Measurement and Assessment of Groundborne Noise and Vibration.

• Data shall be sampled at a frequency sufficient to record faithfully the vibration of interest. Anti-alias filtering shall be applied. Guidance provided in the Association of Noise Consultants’ Guidelines for Measurement and Assessment of Groundborne Noise and Vibration shall be followed.




F1.4 Internal Vibration Measurement

Vibration measurements within properties shall be undertaken as follows.

• Measurements of vibration shall be made inside properties from which questionnaire responses are received and whose occupants have agreed to such measurements.

• Continuous raw, linear vibration acceleration time history shall be recorded for a period to be specified (see below).

• Tri-axial measurements shall be acquired. Axes shall be vertical, parallel to the railway and perpendicular to the railway.

• Measurements shall be made (as a minimum) at the centre of at least one ground floor (habitable) room (that closest to the vibration source).

• Internal measurements shall be made simultaneously with external measurements taken outside the property (as described in Section 3).

• Internal and external measurements shall be time synchronised.

• Where permitted by the resident internal and external measurements shall be made using a system that will allow phase locked data to be acquired. It is acknowledged that this approach may not be acceptable to many residents.

• If a large number of respondents at a measurement location agree to internal measurements, it is acceptable not to measure inside all properties. Properties within which measurements are made shall be chosen at a range of distances from the vibration source. Estimates of the vibration exposure in other properties based on the attenuation ‘site law’ described in Section 3 shall be made.

• All internal measurements shall be attended, to ensure that the data recorded are due to sources external to the domicile, and not due to internal activities.

• Measurements shall be made at a standardised measurement position, which shall be as close as practicable to the centre of the room in which the respondent considers the vibration to be most significant.

• As far as practicable, measurements shall be made over a period of time sufficient to achieve a representative sample of data. However, this may be difficult due to the requirement to attend the monitoring and may need to be balanced against the need to ensure residents are amenable to a survey.

• Transducers shall be mounted using methods given in the Association of Noise Consultants’ Guidelines for Measurement and Assessment of Groundborne Noise and Vibration.

• Data shall be sampled at a frequency sufficient to faithfully record the vibration of interest. Anti-alias filtering shall be applied. Guidance provided in the Association of Noise Consultants’ Guidelines for Measurement and Assessment of Groundborne Noise and Vibration shall be followed.

F1.5 Noise Measurement

External noise shall be measured at the ‘snapshot’ vibration monitoring locations and internal noise shall be measured within properties where vibration is also measured. Noise shall be assessed in the following ways:

Continuous raw, linear sound pressure level shall be recorded for a period to be specified (see below).




Internal measurements shall be made (as a minimum) near the centre of at least one ground floor (habitable) room (that closest to the vibration source).

Where internal vibration measurements are undertaken noise measurements shall record raw sound pressure time histories and event indicators. Measurements shall include, but not necessarily be limited to the LAeq, LA,max,S and LA,max,F for each event.

It is not necessary for the internal noise raw time history to be sampled to provide full audio range recording. The sampling frequency shall be no lower than 2kHz.

Where external vibration measurements are undertaken, the external noise climate shall be measured during the vibration measurements in terms of standard indicator values only. Raw time history recordings are not required. Measurements shall include, but not necessarily be limited to SEL,LA,max,S and LA,max,F for each event and LA90, LA10, LAeq over 15 min contiguous periods over the total duration of the survey.

Subjective observations of noise shall be recorded by the surveyor, to include source of the noise; whether structureborne or airborne; contribution from any secondary effects.

F1.6 Measurement Duration (Control Position)

The duration of measurements at the control position shall be at least 24hrs and will depend upon the nature of the source. Measurements shall be made over a representative period and shall include any key ‘disturbance periods’ identified by the questionnaire.

The control position shall record for sufficient length of time to reflect the time varying nature of the sources.

Snapshot and internal measurements shall be for sufficient duration to provide confidence that a robust transfer function between the control and snapshot positions can be established.

F1.7 Equipment

All equipment shall have appropriate calibration certificates or records.

The equipment required for this work shall have the following minimum functionality:

• Simultaneous measurement of raw time history data for triaxial vibration and noise;

• Adequate dynamic range to provide sufficient information from vibration to cover, as a minimum, the range 0.001mm/s r.m.s to 7mm/s r.m.s (10mm/s PPV)

• Frequency response from 1Hz to 80Hz (for vibration);

• Sampling frequency minimum 2kHz per channel;

• Full audio frequency response for determination of (external) noise indices;

• Storage of raw time history data (it is also desirable, but not essential, to facilitate ‘on-the-fly’ calculation of noise and vibration indices);

• Ability to store all required raw time history data and indices, or the capacity to record for a sufficient period to make occasional download and re-start practicable;

• Anti alias filtering;

• Power source sufficient to minimise any need for interruption to continuous recording;

• A secure, weatherproof and tamperproof method of installation, particularly for the control monitor.




F1.8 Analysis

Data shall be analysed to yield the following parameters for each event (where discrete events occur such as train pass bys) or for each contiguous 1 minute sample (for continuous sources such as construction or industrial processes): peak unweighted acceleration; rms unweighted acceleration.

Total exposure to vibration at each property shall be calculated from the ‘control’ data scaled by a correction factor (or factors) based on the arithmetic average ratio of vibration magnitudes during the snapshot recording and the levels recorded simultaneously at the Control position and using an appropriate building response function.

Building response functions shall be determined from the arithmetic average ratio of vibration magnitudes during the snapshot recorded simultaneously internal and external to each property.

If it has not been possible to acquire site specific building response functions, then an approximation based on data from other sources shall be used. Full details of the assumptions made and the source of the response functions shall be included in the Survey Report.

Groundborne noise levels shall be measured where possible. Where direct measurement has not been possible, groundborne noise levels shall be calculated using proven methods based on 1/3 octave band vibration data. This data will enable an assessment of the contribution of the groundborne noise component to the overall degree of disturbance. Guidance provided in the Association of Noise Consultants’ Guidelines for Measurement and Assessment of Groundborne Noise and Vibration shall be followed.

F1.9 Reporting

For each measurement location, a Survey Report shall be prepared. This shall include, but not necessarily be limited to:

• A map or plan showing the location of the principal and any other vibration sources and the measurement positions;

• Details of the dates and times of day at which all measurements were made at each measurement position;

• A description of the principle vibration and noise sources, including details of the frequency of events and any variability through the time of the survey;

• For internal measurements, details of the property in which the measurements were made, to include the type and of building, opinion on probable foundation type, nature of the floor and covering in the room monitored, description of the position within the room where the measurements were made, any sources of secondary noise, record of any events generated internally that may have affected the measurements;

• Justification for the duration of the measurement periods, both for control and snapshot measurements;

• Details of the equipment used, including evidence of calibration;

• Description of building response functions used and any assumptions made in their determination;

• The weather conditions during the measurements; and

• Tabulated measurement data – this may be provided in electronic format.

Appendix G Site Details


M:\119573-00\OUTGOING DOCUMENTS\FINAL REPORT\FINAL ISSUE\APPENDICES\APPENDIX_G_SITE DETAILS.DOC

TRl Temple ISVR Arup Acoustics ISSUE 7 March 2007

Contents G1 Site Details

G1.1 Measurement Location A – Railway Site G1.2 Measurement Location B – Railway Site G1.3 Measurement Location C – Railway Site G1.4 Construction site



Page G1 TRL Temple ISVR Arup AcousticsISSUE 21 March 2007

G1 Site Details

G1.1 Measurement Location A – Railway Site

Measurement location A is a residential estate. The housing consists of brick built semi-detached and terraced housing situated around quiet estate roads with little variation in property type around the site. A map of the site is shown in Figure G1.

The East Coast Mainline (ECML) runs to the east of the site. The ECML is made up of 4 bidirectional tracks at this location. The closest properties are approximately 20m from the nearest track.

The topography of the location is generally flat. The ECML is elevated above the foundation height of properties by an average of 2m across the whole location.

The rolling stock along this section of the ECML consists of cross country Great North Eastern Railways (GNER) Diesel Motorised Units (DMUs), local Electric Motorised Units (EMUs) and freight.

At peak times there was a train event approximately once every 2.5 minutes and every 4 minutes at other times.

Figure G1 shows the locations of questionnaire responses, external satellite and internal-external measurement positions. The control monitor was located at the ECML end of one of the property’s garden.

The monitoring was undertaken by two (sometimes three) staff for a total of approximately five working days on site. Simultaneous Internal-external measurements were made at six respondent properties. An additional nine external satellite measurements were made. In all, 24 case studies were obtained from this site.

The total time spent on this site was longer than other sites, in part because it was the first site. Productivity increased as familiarity with the processes and equipment increased.




Figure G1 – Measurement Location A showing questionnaire responses, external satellite and internal measurement positions




G1.2 Measurement Location B – Railway Site

Measurement location B is a residential area where properties were situated on both sides of the ECML.

Figure G2 shows a map of the site. The ECML runs down the centre of the site and consists of four bidirectional tracks. The closest properties are approximately 15m from the nearest track.

For the majority of the location the topography is generally flat with building foundations and rail height at a similar level. This excludes some properties to the West of the ECML. Remote from the railway, some properties are approximately 5m below rail level whereas others are up to approximately 5m above rail level.

The traffic along this section of the ECML is similar to that at measurement location A. However a higher frequency of EMUs at B results in train movements of up to two per minute at peak times and one every two minutes at other times.

There was a large variation in housing type across the site, including circa 1950 semi-detached housing, more modern terraced and semi-detached housing and purpose-built three-story flats.

Also shown in Figure G2 are the location of questionnaire responses, external satellite and internal-external measurement positions. The control monitor was located at the ECML end of a garden.

Measurements were made by two staff, who spent approximately three working days on site. Simultaneous Internal-external measurements were made at seven properties. An additional 13 external satellite measurements were made across the location. In all, 31 case studies were obtained.




Figure G2 – Measurement location B showing questionnaire responses, external satellite and internal measurement positions




G1.3 Measurement Location C – Railway Site

This location consists of residential development situated above a tunnel (see Figure G3).

The measurement location is on a hill, with the highest point running directly along the line of the tunnel. The tunnel runs north west to south east across the site. No properties are situated directly above the tunnel. The perpendicular distance of the closest properties to the nearside track is approximately 20m.

This section of the ECML consists of four bi-directional tracks. Due to the presence of the tunnel, the trains could not be seen during the survey. However, it is reasonable to assume that traffic was similar to that at measurement location B. One train event each minute was recorded at peak times reducing to one every three minutes at other times.

Property type at this location was reasonably uniform. Properties were circa 1950 and 1960 semi-detached properties.

Also shown in Figure G3 are the location of questionnaire responses, external satellite and internal-external measurement positions. The control monitor was located at the end of a garden closest to ECML.

Two staff spent approximately two working days at this location. Simultaneous internal-external measurements were made at three properties. An additional 10 external satellite measurements were made. In all, 23 cases studies were obtained.




Figure G3 – Measurement location C showing questionnaire responses, external satellite and internal measurement positions




G1.4 Construction site

The piling site was a new residential development, for which an earlier phase of the development, adjacent to the construction site had already been constructed and was occupied. Piles were 10 m to 14 m long, 250 mm square section pre-cast concrete. They were driven into soft to firm, becoming firm to stiff clay using a Banut 500 rig (see Figure G4). Each pile took around 6 to 8 minutes to drive.

Figure G4 - Banut piling rig operating at the construction measurement location

The same equipment and essentially the same measurement approach were used on the construction site as were used on the railway sites.

The control monitor was positioned in a corner of the site where it would not interfere with or be damaged by site activities. Unlike the railway sites, it was not necessary for the logger to be left in position for a 24 hour period since the piling activity only took place during the daytime. The logger was used to record for the duration of the site working day.

The measurements required two staff for approximately two working days on site. Simultaneous internal and external measurements were made at five properties. Measurements were made at an additional six external satellite measurement locations. In all 13 case studies were obtained, covering a range of distances of approximately 50 m to 200 m from the piling rig. It is understood that the piling activity had been as close as approximately 20 m from the closest residences on some days before the vibration team was on site.

Measurements were taken inside and outside properties at the positions shown in Figure G5. Some measurements were taken on upper floors




Figure G5 - The construction measurement location, showing questionnaire responses, external satellite and internal measurement positions

Appendix H Measurement Proforma


M:\119573-00\OUTGOING DOCUMENTS\FINAL REPORT\FINAL ISSUE\APPENDICES\APPENDIX_H_MEASUREMENT PROFORMA.DOC

Page H1 TRL Temple ISVR Arup Acoustics ISSUE 7 March 2007

NANR 172 Survey Sheet Location:

Control / External / Internal Sheet of

Address:

Made by Date

Checked by Date

Recording ref: Start Stop

Equipment Accelerometer S/N & Sensitivity Ch1 (z-axis)

Data recorder ref:

Ch2 (x-axis)

Ch3 (y-axis)

Transducer fixing: Plate / spike / resin

Ch4 (noise)

SLM Ref: Cal Before Cal after

Observations Time Description

Description of measurement location and property if internal:


M:\119573-00\OUTGOING DOCUMENTS\FINAL REPORT\FINAL ISSUE\APPENDICES\APPENDIX_H_MEASUREMENT PROFORMA.DOC

Page H2 TRL Temple ISVR Arup Acoustics ISSUE 7 March 2007

NANR 172 Survey Sheet Location:

Control / External / Internal Sheet of

Address:

Made by Date

Checked by Date

Recording ref: Start Stop

Observations Time Description

Appendix I DATS Software


M:\119573-00\OUTGOING DOCUMENTS\FINAL REPORT\FINAL ISSUE\APPENDICES\APPENDIX_I_DATS SOFTWARE.DOC


Contents I1 DATS Software I2 General I3 Overload removal I4 Railway event extraction I5 Construction event extraction I6 Statistical analysis and output of results



Page I1 TRL Temple ISVR Arup Acoustics ISSUE 21 March 2007

I1 DATS Software I2 General

All measurement data from control, external satellite and internal measurement positions has been post-processed using the Prosig DATS analysis software. The software can be used for analysis and generation of a range of data and signal types. It can be used solely as a pc-based data analysis package, but has also been designed for use with a range of data acquisition hardware. For the pilot study only the pc-based analysis capability has been used.

The software contains an extensive toolbox of data manipulation, signal processing, generation and analysis functions.

The software may be operated in two distinct ways; interactively using the graphical worksheet interface or by using automated scripts, programmed in Visual Basic with selectable interaction. For this study the software was operated using automated scripts.

Due to the differences between data recorded at the control positions and that recorded at the satellite positions, separate scripts were written for each. Also, due to the different source characteristics between the railway and construction sites, separate scripts were written for each type of site. This is explained in more detail below.

Figure I1 shows how the post-processing is implemented in a number of phases using a DATS script file. The process is as follows:

Figure I1 Process diagram of the measurement data post-processing phase

Raw Data / Calibration information

Import to DATS

Noise removal

Event extraction

Calibration

Tabulated results

DATS analysis software

Statistical analysis




For this study each measurement (control, external, internal) has been analysed independently from the others. Therefore a script was executed and a results spreadsheet obtained for each individual measurement.

On execution of an analysis script file, the measurement file to be processed, transducer sensitivities, gains and signal units are input interactively via a call-out box (Figure I2). Each .wav file has four-channels; x-axis, y-axis, z-axis vibration and noise. Therefore calibration, gain and units are required for each. The file is then imported to the software, converted to a DATS dataset (.dac) and calibrated.

Not shown in Figure I1, the signals are then filtered using high pass filters with a cut off frequency of 1 Hz to remove all near DC components.

Figure I2 The data import call-out box used in the DATS script

Figure I3 Input of event removal parameters in DATS




A noise removal and event extraction process is then implemented using the software. The parameters for the event removal are input interactively using a second call out box (Figure I3). Here a reference signal is chosen from which to perform the event extraction for all four noise and vibration channels.

I3 Overload removal First, any overload from the four measurement channels is removed. Points above in the signal above a defined overload threshold (Figure I3) are removed and replaced with a zero level. A Haversine window replacement function ensures the replacement has minimal effect on the frequency spectrum of the signal.

Secondly any transients are removed from the data. The software inspects the first derivative of the measured signal normalised to the standard deviation signal. Points above a defined Transient Threshold (Figure I3) are replaced with a zero level using the Haversine replacement function.

For signals with significant levels of noise, this function was found to remove large amounts of relevant data. Therefore the transient threshold was set to high levels in practice.

Next vibration events are extracted from the signals. The method used is different for the railway and construction sites.

I4 Railway event extraction For the railway sites the event removal uses a single reference channel for the removal of events on each channel. For this study, the z-axis vibration channel was set as the reference channel in all cases.

The software calculates the normalised standard deviation of the reference signal. The first point above a defined ‘event threshold’ (Figure I3) is defined as the start of an event. The dB down level below the next peak in the reference signal is then found, defining the end of the event.

The dB down level is input as an event extraction parameter (‘dB down’ in Figure I3).

Pre and post trigger values were also set to ensure the entirety of the event is captured.

In the situation that the defined dB down level is not found, a search limit, in seconds, can be defined as the pseudo end point of the event.

The process is repeated until all events in the measurement signal are found and ‘event mask’ signal is created. The ‘event mask’ is then used to extract the events from each of the four vibration or noise channels. Each event for each channel is then saved as a separate dataset.

An example of a z-axis vibration railway event extracted using the DATS script for a control position is shown in Figure I4.

I5 Construction event extraction For the construction site, the source characteristics of the piling are very different to that of train pass-bys. For this reason it was decided split the construction site measurements into contiguous 60 second signals for analysis.

This was achieved using the same event extraction functions as for the railway site, but using a generated sine wave with the period of 60s as the reference signal. The event extraction function defines any point in the signal with a zero crossing and a positive differential as the start of an event.




Figure I4 An example of a train event extracted from a 24hr control recording

To ensure that the 60 second events from separate but simultaneous recordings could be compared, the reference signal was offset by a defined number of points so that extraction began at the nearest minute (real time).

An example of a 60s event extracted from the piling control position is shown in Figure I5.

I6 Statistical analysis and output of results After extracting all events in the measurement a statistical analysis function is used to extract the peak and rms acceleration or noise for each event dataset. At this stage it is possible to implement calculations for other noise and vibrations indicators, such as VDV or weight the data appropriately for human perception. For the purpose of the pilot study data was left un-weighted and only the basic indicators were extracted. For the pilot study, only the vibration data was extracted, since noise data analysis was not required as part of the pilot study.

The results for each event and each channel are then output to an excel spreadsheet for use in further analysis. Additional information included in the spreadsheet to aid in analysis is the real event time, event time in seconds from the start of the recording and the duration of the event. An example output of the DATS script is shown in Table I6.




Figure I5 An example of a 60s signal extracted from the control location at the construction location



Page I6 TRL Temple ISVR Arup AcousticsISSUE 21 March 2007

Data : \Defra Vibration Data\Site A\\Internal\D00002.WAV

Channel 1

0.91659 EU/V Channel 2

0.91912 EU/V Channel 3

0.3154 EU/V Channel 4 6.268 EU/V

Event Time Real Time Clock (s) Duration(s) Abs Max RMS Abs Max RMS Abs Max RMS Abs Max RMS 1 18/Sep/2006 11:07:01 256.2 11.7 0.04456 0.00924 0.02066 0.00477 0.01911 0.00489 0.92287 0.18709 2 18/Sep/2006 11:11:59 554.2 11.4 0.01840 0.00334 0.00825 0.00154 0.00709 0.00145 0.27324 0.05796 3 18/Sep/2006 11:23:30 1245.1 15.0 0.05045 0.00943 0.03591 0.00789 0.01852 0.00389 0.86017 0.13984 4 18/Sep/2006 11:29:42 1617.8 13.6 0.03929 0.00805 0.01244 0.00244 0.01114 0.00233 0.68006 0.14557 5 18/Sep/2006 11:36:37 2032.9 11.6 0.04364 0.00955 0.02746 0.00585 0.02621 0.00673 0.74230 0.18151 6 18/Sep/2006 11:41:14 2309.4 6.1 0.00636 0.00173 0.00438 0.00115 0.00444 0.00112 0.13800 0.03616 7 18/Sep/2006 11:45:49 2584.3 10.1 0.03898 0.00623 0.02835 0.00623 0.04240 0.00801 0.80760 0.18139 8 18/Sep/2006 11:46:13 2608.9 11.7 0.01745 0.00281 0.00673 0.00138 0.00476 0.00103 0.26910 0.06120 9 18/Sep/2006 11:46:47 2642.4 19.9 0.04124 0.00756 0.02116 0.00317 0.01544 0.00259 0.70589 0.10678

10 18/Sep/2006 11:52:04 2959.1 10.3 0.03712 0.00718 0.03109 0.00709 0.04108 0.00793 0.67729 0.16620 11 18/Sep/2006 12:05:29 3764.1 11.9 0.04009 0.00929 0.02251 0.00558 0.02810 0.00689 0.92706 0.15774 12 18/Sep/2006 12:08:10 3925.7 13.5 0.06960 0.01188 0.03352 0.00536 0.02047 0.00417 0.63621 0.14056 13 18/Sep/2006 12:13:39 4255.0 7.4 0.02467 0.00540 0.00885 0.00218 0.01109 0.00281 0.66921 0.15600 14 18/Sep/2006 12:27:27 5082.3 10.1 0.04021 0.00766 0.03403 0.00900 0.05174 0.01059 0.81071 0.17451

Table I6: Example output of the DATS script

Appendix J Analysis of Vibration Measurements


M:\119573-00\OUTGOING DOCUMENTS\FINAL REPORT\FINAL ISSUE\APPENDICES\APPENDIX J ANALYSIS OF VIBRATION MEASUREMENTS.DOC

TRL Temple ISVR Arup AcousticsISSUE 7 March 2007

Contents J1 Analysis of Vibration Measurements J2 Control-external transfer functions J3 External-internal transfer functions J4 Interpolation of data J5 Scaling of control data



Page J1 TRL Temple ISVR Arup AcousticsISSUE 7 March 2007

J1 Analysis of Vibration Measurements During the post-processing phase, event noise and vibration indicators for all control, external and internal measurements were extracted. Further data analysis and manipulation was required to provide predictions of the exposure to vibration in all cases study properties and to present this data in a format suitable for combined analysis.

This Appendix describes the analysis process undertaken and summarised in Figure J1.

Figure J1 Process diagram of measurement data analysis

J2 Control-external transfer functions The satellite measurements have been made over a ‘snapshot’ period simultaneously to the control measurements. The transfer functions are required to scale the 24hr exposure data, logged at the control position, to each satellite location.

The individual event indicators for vibration at each external location are compared to the events occurring during the same period at the control location. Each individual event is checked for quality and extraneous events are removed from the analysis.

For cases where external satellites were subject to additional sources of vibration, many events were removed from the analysis for quality reasons. This resulted in a small sample of events from which to evaluate transfer functions. Recommend increased measurement duration.

For the piling site, 60s measurements were compared. All events where piling was not happening were removed from the analysis.

Time domain transfers functions for each simultaneous control and external event in each of the vibration axis was calculated for both rms and peak acceleration. An average of the transfer functions was then taken to obtain x, y and z-axis vibration transfer functions for both rms and peak acceleration.

Control, external & satellite data

Control-external transfer functions

Interpolation of data

Scaling of control data

External-internal transfer functions

Tabulated case study data




The process was repeated for every external satellite location.

As the control and satellite transfer functions have been processed separately. (i.e. with independent reference signals) the start point, end point and duration of events differ between the two measurements. This will introduce a source of error to the transfer functions, particularly in terms of the rms acceleration.

The error could be reduced by processing control and external events using the same reference signal. This would ensure transfer functions were evaluated over identical durations, though the error would only be reduced as the measurements cannot be phase-locked.

Additionally, in the case of the railway sites, the control and satellite positions may be positioned some distance apart and at a similar distance from the source. Hence there is an inevitable delay between the event measured at source and at receiver, or vice verse (depending upon direction of traffic). Therefore evaluating the transfer function in the time domain between the two separately evaluated dB down points may introduce more error. Ideally the transfer function should be evaluated from the point the signal is first detected at the control to the point it is no longer detected at the external position, or vice verse.

J3 External-internal transfer functions To predict the vibration within properties and to validate the hypothesis that a person’s response to vibration inside a property can be evaluated from the vibration outside a property, internal and external measured data was used to calculate appropriate transfer functions.

These transfer functions were calculated in the same way as for the control-external transfer functions.

The event vibration and noise indicators from the post-processed simultaneous internal and external measurements were compared. Each individual event was inspected for quality and extraneous results removed.

In some cases, the quality of internal data resulted in few events remaining in the analysis. Recommend increasing measurement duration.

Again for the piling site, 60s events where no piling was occurring where removed from the analysis.

Transfer functions in the x, y and z axes where then calculated for both rms and peak acceleration. An average of each was then taken to represent the property.

The process was then repeated for all simultaneous internal and external measurements.

Again a large scatter of results was seen, similar for both the railway and constructions sites. It is thought that this scatter could be reduced by extracting external and internal events from the same reference signal. As the measurements were simultaneously recorded, this technique would result in phase-locked transfer functions.

J4 Interpolation of data As described above, external or internal vibration measurements were not made at all properties where a questionnaire response was obtained. To increase the case studies obtained, and minimise the number of measurement positions the protocol allows for predictions of the external and internal vibration to be made.

The external vibration exposure of properties where no measurements have been made can be reasonably predicted if:




A contiguous lines of properties exists where more than one questionnaire response is situated and external vibration measurements have been made at each end of the line; or

Measured data for a nearby property exists (i.e an adjacent property)

During the survey planning stages, contiguous lines of properties where multiple questionnaire responses where situated where identified and external satellite measurements positions were chosen accordingly.

The external vibration at properties along the line is predicted using a site law. A number of site laws for each survey location were developed from the transfer functions measured at each end of a line of properties and interpolating to find the transfer functions for properties within the line.

For adjacent or nearby properties where measurements external measurements had been made, a ‘nearest measurement’ law was applied where the measured control-external transfer function is assumed for both properties. This prediction is only valid for cases where the distance between the two properties is much less than the distance from the source of vibration.

J5 Scaling of control data To obtain a prediction of the 24hr external vibration exposure of all case study properties the event indicators in each of the three axes of vibration from the 24hr control at each study location have been scaled using the calculated control external transfer functions.

Prediction of the 24hr internal vibration exposure of case study properties where internal vibration has been measured was achieved by scaling the predicted 24hr external measurements with the measured external-internal transfer functions.

To obtain predictions for the internal vibration at case study properties where no internal measurements were made a representative external-internal transfer function was chosen, according to property type, from the measured transfer function from the appropriate study location.

Due to the limited number of simultaneous internal and external measurements made at each location, a representative external-internal transfer function was not always available. In this situation an average external-internal transfer function taken from the measured transfer functions of the study location was used to represent the case study property.

For each case study, the z-axis event peak and rms acceleration, averaged over every event during the 24hr period was tabulated together with the predicted overall z-axis rms vibration for use in the combined analysis.

NANR172 Human response to vibration in residential...

Documents

Transcript of NANR172 Human response to vibration in residential...