Manual of Petroleum Measurement Standards Chapter 19.3 ...ballots.api.org/copm/cele/ballots/docs/API...

This document is not an API Standard; it is under consideration within an API technical committee but has not received all approvals required to become an API Standard. It shall not be reproduced or circulated or quoted, in whole or in part, outside of API committee activities except with the approval of the Chairman of the committee having jurisdiction and API staff. Copyright API. All rights reserved.

Manual of Petroleum Measurement Standards

Chapter 19.3— Evaporative Loss Measurement

Part H—Specification for Establishing Evaporative Loss Factors for Floating-Roof Tank Devices

SECOND EDITION

PART H—SPECIFICATION FOR ESTABLISHING EVAPORATIVE LOSS FACTORS FOR FLOATING-ROOF TANK DEVICES This document is not an API Standard; it is under consideration within an API technical committee but has not received all approvals required to become an API Standard. It shall not be reproduced or circulated or quoted, in whole or in part, outside of API committee activities except with the approval of the Chairman of the committee having jurisdiction and API staff. Copyright API. All rights reserved.

FOREWORD

This edition of API MPMS Chapter 19.3 Part H incorporates API MPMS Chapter 19.3 Part F, 1st edition, Evaporative Loss Factor for Storage Tanks Certification Program, and API MPMS Chapter 19.3 Part G, 1st edition, Certified Loss Factor Testing Laboratory Registration, both of which are withdrawn.

PART H—SPECIFICATION FOR ESTABLISHING EVAPORATIVE LOSS FACTORS FOR FLOATING-ROOF TANK DEVICES 1 This document is not an API Standard; it is under consideration within an API technical committee but has not received all approvals required to become an API Standard. It shall not be reproduced or circulated or quoted, in whole or in part, outside of API committee activities except with the approval of the Chairman of the committee having jurisdiction and API staff. Copyright API. All rights reserved.

Chapter 19.3—Evaporative Loss Measurement

PART H—SPECIFICATION FOR ESTABLISHING EVAPORATIVE LOSS FACTORS FOR FLOATING-ROOF TANK DEVICES

1 Scope

The purpose of this standard is to specify requirements for the use of the test methods in API’s Manual of Petroleum Measurement Standards (MPMS), Chapter 19.3, Parts A through E, in order to develop evaporative loss factors for floating-roof rim seals, deck fittings, and deck seams (hereinafter referred to as floating-roof devices).

This standard illustrates how other standards of the API MPMS Chapter 19.3 series are integrated into an overall loss factor development program to enable the user to develop a loss factor for a given floating-roof device. This standard presents procedures for the evaluations to be performed under such a program, including preparation for protocol testing of individual devices, monitoring of the tests, and analysis and reporting of test results for the purposes of establishing evaporative loss factors.

It is not the purpose of this loss factor development program as given in the MPMS Chapter 19.3 series of standards to specify procedures to be used in the design, manufacture, or field installation of floating-roof devices. Furthermore, equipment can not necessarily be selected for use solely on the basis of evaporative-loss considerations. Many other factors—such as tank operation, maintenance, and safety—are important in designing and selecting tank equipment for a given application.

This standard may involve hazardous materials, operations, and equipment. This standard does not purport to address all of the safety problems associated with its use. It is the responsibility of the user of this standard to establish appropriate safety and health practices and determine the applicability of regulatory limitations prior to use.

2 Normative References

The following referenced documents are indispensable for the application of this document. For dated references, only the edition cited applies. For undated references, the latest edition of the referenced document (including any amendments) applies.

API Manual of Petroleum Measurement Standards (MPMS)

Chapter 15, Guidelines for the Use of the International System of Units (SI) in the Petroleum and Allied Industries

Chapter 19.2, Evaporative Loss from Floating-Roof Tanks, 3rd Edition, 2012

Chapter 19.3 Part A, Wind Tunnel Test Method for the Measurement of Deck-Fitting Loss Factors for External Floating-Roof Tanks, 1st Edition, 1997

Chapter 19.3 Part B, Air Concentration Test Method for the Measurement of Rim-Seal Loss Factors for Floating-Roof Tanks, 1st Edition, 1997

Chapter 19.3 Part C, Weight Loss Test Method for the Measurement of Rim-Seal Loss Factors for Internal Floating-Roof Tanks , 1st Edition, 1998

Chapter 19.3 Part D, Fugitive Emission Test Method for the Measurement of Deck-Seam Loss Factors for Internal Floating-Roof Tanks, 1st Edition, 2001

2 CHAPTER 19.3—EVAPORATIVE LOSS MEASUREMENT This document is not an API Standard; it is under consideration within an API technical committee but has not received all approvals required to become an API Standard. It shall not be reproduced or circulated or quoted, in whole or in part, outside of API committee activities except with the approval of the Chairman of the committee having jurisdiction and API staff. Copyright API. All rights reserved.

Chapter 19.3 Part E, Weight Loss Test Method for the Measurement of Deck-Fitting Loss Factors for Internal Floating-Roof Tanks , 1st Edition, 1997

Publ 2517D, “Documentation File for API Publication 2517—Evaporative Loss from External Floating-Roof Tanks”, 1st Edition, 1993

ASTM1 D323, Standard Test Method for Vapor Pressure of Petroleum Products (Reid Method)

ASTM1 D5191, Standard Test Method for Vapor Pressure of Petroleum Products (Mini Method)

3 Terminology

3.1 Definitions

For the purposes of this document, the following definitions apply.

standard.

3.1.1 data acquisition The process of receiving signals from the sensors, determining the values corresponding to the signals, and recording the results.

3.1.2 deck That part of a floating roof which provides buoyancy and structure, and which covers the majority of the liquid surface in a bulk liquid storage tank. The deck has an annular space around its perimeter to allow it to rise and descend (as the tank is filled and emptied) without binding against the tank shell. This annular space is closed by a flexible device called a rim seal. The deck may also have penetrations, closed by deck fittings, which accommodate some functional or operational feature of the tank.

3.1.3 deck fitting The device which substantially closes a penetration in the deck of a floating roof in a bulk liquid storage tank. Such penetrations are typically for the purpose of accommodating some functional or operational feature of the tank.

3.1.4 deck seam The joint attaching adjacent sheets or panels in the floating-roof deck.

3.1.5

floating-roof device A feature of a floating roof such as a deck fitting, rim seal, or deck seam where evaporative losses are possible.

3.1.6 floating roof.

A device that floats on the surface of the stored liquid in a floating-roof tank. A floating roof substantially covers the liquid product surface, thereby reducing its potential for exposure to evaporation. Floating roofs are comprised of a deck, a rim seal, and miscellaneous deck fittings

1American Society for Testing and Materials, 100 Barr Harbor Drive, West Conshohocken, Pennsylvania 19428.


3.1.7 instrument A device used in the measurement process to sense, transmit or record observations.

3.1.8 loss factor Factor that indicates the average amount of evaporation loss associated with a device. In order to obtain the total standing storage evaporative‐loss rate for a floating‐roof tank, the sum of the evaporative‐loss factors for each of the individual devices is modified by certain characteristics of both climate conditions and the stored liquid. The characteristics of the stored liquid are expressed as a vapor pressure function, the stock vapor molecular weight, and a product factor.

3.1.9 product factor A factor that describes the evaporative-loss characteristics of a given liquid. The product factor, the stock vapor molecular weight, and the vapor pressure function are multiplied by the sum of the floating-roof loss factors to determine the total standing storage evaporative-loss rate of a floating-roof tank

3.1.10 protocol test A test performed in accordance with the approved and published standards and procedures of the API MPMS, Chapter 19.3 series, as applicable.

3.1.11 rim seal A flexible device attached to the rim of a floating-roof deck that spans the annular space between the deck and the tank shell.

3.1.12 rim seal gap area The total cumulative horizontal area of all spaces or openings between the rim seal and the tank shell that provide an unobstructed path for a 0.125-in diameter probe to pass freely from a position above the rim seal down to the stored product.

3.1.13 floating roof standing loss Loss of stored liquid stock by evaporation past the floating roof during normal service conditions. This does not include evaporation of liquid that clings to the tank shell and is exposed to evaporation when the tank is being emptied (withdrawal loss), nor does it include vapor loss that may occur when the liquid level is sufficiently low so as to allow the floating roof to rest on its support legs. This does include, however, evaporative losses from the rim seal, deck seams, and deck fittings.

3.1.14 vapor pressure function A dimensionless factor used in the loss estimation procedure, that is a function of the ratio of the vapor pressure of the stored liquid to average atmospheric pressure at the storage location.

3.2 Units of Measurement

3.2.1 Basic Units

The unit of length is either the mile, mi, the foot, ft, or the inch, in. The unit of mass is the pound mass, pound or lb. The unit of force is the pound force, pound-force or lbf. The unit of time is either the hour, h, or the year, yr. The unit of temperature is the degree Fahrenheit, °F, or the degree Rankine, °R. The unit of electromotive force is the volt, v.

Formatted: Font: Arial, Bold

Formatted: Font: Bold


3.2.2 Loss Factors

The unit of reporting loss factors is the pound-mole per year, lb-mole/yr, or the pound-mole per foot year, lb-mole/ft yr.

The pound-mole units of the loss factor (Kf for deck fittings, Kr for rim seals, or Kd for deck seams) do not actually indicate pound-moles of vapor loss over time, but rather are units of a factor that have to be multiplied by certain coefficients (which are dimensionless) in order to determine actual pound-moles of evaporative loss over time for a given liquid product. To convert the pound-mole units of the loss factor to a loss rate in terms of actual pound-moles, the loss factor (Kf or Kr or Kd) is multiplied by the dimensionless coefficients P*, which is a function of the product vapor pressure, and Kc , the product factor.

A pound-mole is an amount of a substance the mass of which, when expressed in pounds, is equal to the numerical value of the molecular weight of the substance. To convert the actual pound-moles per year loss rate to pounds per year of a given liquid product, the loss rate (Kf or Kr or Kd)×(P* Kc) is multiplied by the molecular weight of the product in its vapor phase, Mv , with molecular weight having units of pounds per pound-mole. Additional information can be found in the API MPMS Chapter 19.2.

3.2.3 Pressure

The unit of pressure is the pound-force per square inch absolute, designated psia.

3.2.4 System of Units

This standard employs the inch-pound units or U.S. customary units. Values shall be referenced to the U.S. National Institute of Standards and Technology (NIST) values (formerly the U.S. National Bureau of Standards). The text of this standard does not include the equivalent International System of Units (SI) values, but guidance for conversion to SI and other metric units is provided in Annex I.

3.2.5 Velocity

The unit of velocity is the mile per hour, designated mi/h or mph.

3.3 NOMENCLATURE

Table 1 provides a description of the symbols and units that are used in this standard.

Table 1—Description of the Symbols and Units

Symbol Description Units

Ap Constant in the vapor pressure equation Dimensionless

Bp Constant in the vapor pressure equation °R

Kc Product factor Dimensionless

Kd Deck seam loss factor lb-mole/ft yr

Kf Deck fitting loss factor lb-mole/yr

Kr Rim seal loss factor lb-mole/ft yr

L Deck fitting loss rate for a single test lb/h

Mv Molecular weight of stock vapor lb/lb-mole


P True vapor pressure of the stock psia

Pa Atmospheric pressure psia

P* Vapor pressure function Dimensionless

T Stock liquid temperature °R or °F

V Wind Speed mi/h

Note: See 3.2 for definitions of abbreviations for the units.

4 Summary of the Loss Factor Development Program

This loss factor development program includes preparation for protocol testing, performance of the tests, and analysis of the results. Annex A illustrates the overall loss factor development program in the form of flow diagrams for the wind tunnel test method of API MPMS Chapter 19.3 Part A.

Preparation for a protocol test includes certain responsibilities of the testing laboratory, as specified in Section 5. Section 6 provides guidance for review of the documentation for a proposed protocol test.

The testing laboratory shall perform the protocol tests as specified in the test methods of the API MPMS Chapter 19.3 series of standards. Parts A and E of API MPMS Chapter 19.3 address test methods for deck fittings, Parts B and C contain rim seal test methods, and Part D contains a deck seam test method. General guidance for the performance of protocol testing is given in Section 7.

Test reports shall conform to the requirements of the testing protocol. Analyze the test data and assign loss factors to the tested devices as specified in Section 8 below.

It is the responsibility of the testing laboratory to document compliance with each applicable requirement of the API MPMS Chapter 19.3 series of standards. Documentation of compliance shall always be shown as a measured value compared to a specified requirement, rather than simply as a statement certifying compliance.

5 Testing Laboratory Preparation

The process of preparing a testing laboratory under this loss factor development program is outlined in Figure A.1. The requirements for a particular test facility are specified in the corresponding test method of API MPMS Chapter 19.3. Figures 1 and 2 illustrate the test facility requirements for the wind tunnel test method of Chapter 19.3, Part A.

The testing laboratory calibrates the test instruments in accordance with the appropriate test method of API MPMS Chapter 19.3. The test methods specify the maximum tolerable error, the maximum calibration interval, and the re-quired sensitivity for the test instruments. Table 2 illustrates the instrument requirements for the wind tunnel test method of Chapter 19.3 Part A, 1st edition.

The testing laboratory also programs the data acquisition system (DAS) as specified in the test methods of the API MPMS Chapter 19.3 series of standards. This programming includes corrections for instrument bias and temperature effects that may be applicable. The wind tunnel test method of API MPMS Chapter 19.3 Part A, for example, requires dead-weight testing of the load cells through a range of temperatures in order to establish a load cell temperature correction coefficient. Appendix A of API MPMS Chapter 19.3 Part A, 1st edition, describes two procedures for determining this correction coefficient. The procedure given in A.3 of API MPMS Chapter 19.3 Part A, 1st edition, shall be acceptable for assigning a temperature correction coefficient to a scale or load cell if the coefficient of determination, r2, is greater than or equal to 0.99 for the linear regression of measured weight loss, Wmi , on the temperature of the load cell, Tmi , resulting in the model:


This calculation is presented in more detail in B.3.5 of Annex B, which summarizes the statistical calculations required by API MPMS Chapter 19.3 Part A, 1st edition.

Upon completion of the instrument calibrations and DAS programming, the test facility shall be operated to demonstrate compliance with the allowable variations specified in the test methods. The allowable variations for the wind tunnel test method of API MPMS Chapter 19.3 Part A are summarized in Table B.2. API MPMS Chapter 19.3 Part A, further requires that a velocity profile survey be conducted for the wind tunnel. Annex C of this standard outlines a procedure for the wind tunnel velocity profile survey.

Having demonstrated that all instruments and equipment are functioning within acceptable limits, the testing laboratory shall perform tests of standard devices as required by Section 6.4. Annex D stipulates the standard devices for deck-fitting, rim seal, and deck seam testing, and specifies the method for determining whether the test results are acceptable.

Wmi a dTmi+=


Notes: 1. Citations are to API MPMS Chapter 19.3 Part A, 1st edition. 2. Perform velocity surveys at 5 mph, 10 mph, and 15 mph (see 7.3.4.1 of API MPMS Chapter 19.3 Part A, 1st edition). 3. Locate the wind speed sensors in the first or third measuring sections where the measured wind speed is within ±5 % of the average, but not within 6 in of the sides of the wind tunnel (see 10.5.1 of API MPMS Chapter 19.3 Part A, 1st edition). 4. Air flow straighteners shall be at least 2 ft long. The first air flow straighteners shall be located at least 6 in downstream of the air flow distribution mechanism (see 7.3.4.2 of API MPMS Chapter 19.3 Part A, 1st edition). 5. The 5-ft length shown for the measuring sections (measuring stations) is a minimum (see 7.3.2 of API MPMS Chapter 19.3 Part A, 1st edition). 6. When testing is performed with test items of different sizes in the wind tunnel, larger items are to be placed downstream of smaller ones (see 8.3 of API MPMSChapter 19.3 Part A, 1st edition).

Figure 1—Elevation View of a Typical Wind Tunnel Test Facility


Note: Citations are to the API MPMS Chapter 19.3 Part A, 1st edition.

Figure 2—Plan View of a Typical Wind Tunnel Test Facility

Table 2—Instrument Requirements for the API MPMS 19.3 Part A, 1st edition

Variable To Be Measured

Instrument Type Maximum

Tolerable Error Maximum

Calibration IntervalRequired

Sensitivity Reference in Part A

Weight of the test apparatus Scale ±0.1 % 3 months ±0.01 % of the test assembly wt.

10.3

Time of the observation Clock of the DAS ±0.1 % 6 months 1 s 10.2

Wind speed Vane anemometer ±5 % 6 months 0.1 mph 10.5

Pitot tube n/a n/a

Temperature of the air in the wind tunnel

Thermocouple ±0.5 °F 6 months ±0.2 °F 10.4

Average bulk temperature of the test liquid


Temperature of the air in the test room


Temperature of the scale or load cell


Voltage delivered by the power supply

Voltmeter of the DAS

±0.1 % 6 months

Atmospheric pressure Pressure transducer ±0.05 psia 6 months ±0.01 psia 10.6

Notes: 1. The first four columns are from API MPMS Chapter 19.3 Part A, 1st edition, Table 2. 2. The accuracy of the instruments shall be demonstrated using NIST-traceable standards, and shall be based on readings indicated by the DAS (API MPMS Chapter 19.3 Part A, 1st edition, 10.1). 3. The reference weights shall have certified accuracies of ±0.1 % (API MPMS Chapter 19.3 Part A, 1st edition, 10.3.2.1).


6 Preparation for Protocol Testing

6.1 General

The activities involved in the preparation for protocol testing of a particular floating-roof device are outlined in Figure A.2. The results from all tests of devices that are substantially the same in configuration, dimensions, and materials of construction shall be included in the determination of a loss factor. The combinations of configuration, dimensions, and materials to be included in the protocol testing shall be described in the test report in order to identify the entire range being claimed. Also determine, for the wind tunnel test method of API MPMS Chapter 19.3 Part A, whether any nonsymmetrical features of the device warrant testing it in multiple orientations. Annex E provides guidance for the review of preparation for protocol testing.

6.2 Physical Identification of Tested Device

The floating-roof device to be tested, and the recommended installation procedure for the device, shall be described in writing, including sketches and drawings, so that the test facility can properly mount the device in the test apparatus in a manner consistent with its intended use.

The description of the floating-roof device shall identify it in a manner that distinguishes it from other devices of similar type and construction, such that determination of whether a device installed in the field is substantially similar to the device tested can be ascertained by inspection of installed components in the field. All tolerances on dimensions, variations on configuration or substitution of material intended to be included in the description of the device shall be identified. These descriptions need not identify proprietary compositions of matter except in general terms, or internal mechanical details that have no impact on the evaporative loss performance. The material shall be sufficiently detailed to conclusively identify the device that was tested, distinguish the device from similar devices, and permit the reconstruction of the device should retesting be necessary.

Each floating-roof device prepared for testing shall be positively and permanently marked with a unique alphanumeric identifier. The identifier shall be included in all records of testing.

Accurate mechanical drawings shall be prepared of the device to be tested and its configuration during testing. These drawings shall show all details that are relevant to reproducing an identical device should retesting be necessary.

Each tested device shall be photographed in as much detail as necessary to show clearly the important features of the tested device and to allow identification of the device tested at a later date.

Each tested device, the identifying photographs and the mechanical drawings shall be retained for a period of 5 years by the test facility.

6.3 Data Collection and Integrity

The results of any testing conducted under these protocols shall be measured and recorded as specified herein.

All time-dependent loss measurements shall be recorded through the use of an electronic data collection system. The data files created by the system shall contain in each record a unique test identification code, a time stamp that can be reconciled against other standard time stamps, and the serial number of the sensor generating the signal.

The sensors shall not be recalibrated, respanned, or the signals adjusted during a test without clear documentation of the change made and the time stamp associated with the change.

The facility shall calibrate the equipment used for testing under this standard using the means specified in the appropriate test protocols. Equipment may be calibrated more often than the maximum time interval specified.


The mass and temperature calibration standards shall be traceable to the National Institute of Standards and Technology or a national standards agency.

6.4 Testing of Standard Devices

As part of the preparation of a testing laboratory for the testing of deck fittings, the facility shall perform testing of specific standard devices as specified in D.3 for calibration of the test facility.

The facility will rerun the testing of the specified standard devices, following modifications to the test facility, for the purpose of assuring that the modifications did not have an adverse impact on the quality of data produced by the test apparatus. If the results of the rerun tests are not comparable to historical results, the facility shall resolve the differences.

Furthermore, one standard device shall be included within each group of 20 test runs as specified in D.4. Wind tunnel tests at different wind speeds shall individually count as one run each.

For example, in a four-test-position wind tunnel it is expected that at least one standard device will be tested for every five times a wind tunnel test is performed.

7 Performing Protocol Testing

The activities involved in the protocol testing of a particular floating-roof device are outlined in Figure A.3. The test methods for this loss factor development program are published as separate standards as part of the API MPMS Chapter 19.3 series, as referenced previously. The application of these test methods is illustrated in this standard by reviewing the requirements for the wind tunnel test method in API MPMS Chapter 19.3 Part A, 1st edition.

Each demonstration of compliance with any requirement of the loss factor development program shall be documented in a quantified manner (i.e., by stating the value of each measurement or observation, and by stating the reference value or range to which it compares).

Each test report shall include documentation of compliance with the requirements of the test method itself, as well as documentation that demonstrates compliance with all other conditions of the loss factor development program. These other conditions include documentation that the testing laboratory was properly prepared to perform that particular test method, documentation of the required tests of standard devices, documentation that all instrument calibrations were current, and documentation of the properties of the test liquid. When n-hexane, technical grade or better, is used as the test liquid, the requirement of API MPMS Chapter 19.3 Part A, 1st edition, 7.6.2, is satisfied by monthly testing of the Reid vapor pressure in accordance with either ASTM D323 or ASTM D5191. The stated concern with respect to preferential evaporation of the lighter components is not relevant to a high-purity, single-component liquid. When multicomponent test liquids are used, however, the Reid vapor pressure shall be measured before and after each test.

API MPMS Chapter 19.3 Part A, 1st edition, sections 9.3 and 11.3 indicate that deck fittings having a relatively high rate of evaporative loss may initially experience unstable conditions during testing, due to the effect of evaporative cooling on the surface of the test liquid. A procedure for determining when steady-state conditions have been achieved is presented in Annex F.

Numerous statistical calculations are required in the determination of test results. These are summarized in Annex B for the wind tunnel test method in API MPMS Chapter 19.3 Part A, 1st edition, and include the requirement of B.3.7 concerning limits on the variability of the test results.

When evaluating loss factors to be obtained from the wind tunnel test method, the loss factor coefficients shall be determined as specified in API MPMS Chapter 19.3 Part A, 1st edition, Appendix C.


8 Reporting of Protocol Testing

Reports of protocol testing shall document that each step of the loss factor development program specified herein has been followed. The procedure for review of the loss factor development steps is outlined in Figure A.4. This includes reviewing the test preparation, and reviewing that all documentations, calibrations, and testing of standard devices were current and satisfactory.

Documentation of how the test assembly was installed in the test facility is compared to the installation drawings and procedure, in order to confirm compliance with the reference dimensions specified in the test methods. The test method drawing of the test assembly shall show the proper placement of the test assembly in the test facility. Figure 3 illustrates the requirements of the wind tunnel test method specified in API MPMS Chapter 19.3 Part A, 1st edition, for installing a deck-fitting test assembly.

Note: Citations are to the API MPMS Chapter 19.3 Part A, 1st edition.

Figure 3—Test Assembly

The test results are then reviewed for validity and consistency of the data, including checks of the statistical calculations summarized in Annex B. If no errors are found in the data or the statistical calculations, and the variability of the test results is within the limits of B.3.7, then review the determination of the loss factor coefficients. Test results showing very low loss rates shall be evaluated to determine whether they exceed the de minimis criteria of Annex G. Finally, evaluate the uncertainty in the loss factors as outlined in Annex H.


All tests of the device shall be reported. However, tests that were conducted for product development, at conditions not permitted by the test protocols, or on objectively different devices, should not be included in the report. It is the intent of this standard that the loss factors be based on the widest available pool of information.

The report shall include the loss factors for use in API evaporative loss calculation methods, the 95 % confidence limits for the factors, and a physical description of the device. The report shall include copies of all test measurements, drawings, specifications and reports required by the test protocols used, as well as any additional data material to the evaluation of the loss factor.

The testing laboratory shall assure that all statements, representations, data, and reports provided are accurate and complete to the best of their knowledge.

9 Specifications for Protocol Testing

9.1 General

This section specifies test conditions and number of tests to develop a loss factor for each type of device.

9.2 Deck Fittings

Deck fittings for use only with internal or covered floating roofs shall be tested at the nominal zero miles per hour wind speed, in accordance with the test method of either Part A or Part E of API MPMS Chapter 19.3. A minimum of three tests shall be performed.

The minimum test requirements for deck fittings for use with external floating roofs are summarized in Table 3. Deck fittings that are orientation-dependent shall be tested with the prominent feature oriented at 0°, 45°, and 90° to the direction of wind, for each of the nonzero wind speeds. That is, for orientation-dependent deck fittings, the number of tests shown in Table 3 at wind speeds other than zero shall be performed at each of the specified orientations.

Wind speeds of 4.3 mph, 8.5 mph, and 11.9 mph are specified to allow comparison to the test results for generic deck fittings which were tested at these wind speeds by API as described in API MPMS Chapter 19.2, 3rd edition, D.2. Since these are actual wind speeds at the deck fitting, they correspond to ambient wind speeds at the tank site of approximately 6 mph, 12 mph, and 17 mph, after applying the fitting wind-speed correction factor of 0.7 (as specified in API MPMS Chapter 19.2, 3rd edition, 4.2.2.3).

Table 3—Minimum Number of Tests for Deck Fittings

Wind Speed (mph) 01 4.32 8.52,3 11.92

Number of tests 3 1 3 1

Test method (API MPMS Chapter 19.3) Part A or Part E

Part A Part A Part A

Notes: 1. The nominal zero mph condition is any level of wind speed less than 0.5 mph. 2. The number of tests shown shall be performed at each orientation, for orientation-dependent deck fittings. The specified orientations are 0°, 45°, and 90°. 3. The nominal 10 mph wind speed shall be taken as a wind tunnel wind speed of 8.5 mph for consistency with previous API testing as described in API MPMS Chapter 19.2, 3rd edition, D.2.

The determination of acceptable variability of the test results, as specified in Annex B, shall be performed at 0 mph for all deck fittings, and additionally at 8.5 mph for deck fittings to be used with external floating roofs. The limits specified in Table B.2, shall apply for both wind speed conditions.


9.3 RIM SEALS

The rim-seal test method described in Chapter 19.3 Part B, 1st edition, section 11.2, specifies that the wind speed levels to be used for testing the evaporative loss rates for rim seals shall be 0 mph, 5 mph, 10 mph, and 15 mph. Rim seals to be used only with internal or covered floating roofs shall be tested at the nominal zero mph wind speed, in accordance with the test method of either Part B or Part C of API MPMS Chapter 19.3. A minimum of three tests shall be performed.

The minimum test requirements for rim seals to be used with external floating roofs are summarized in Table 4.

Floating-roof rim seals in actual practice sometimes have gaps between the rim seal and the shell of the tank. The cumulative area of all such gaps (rim-seal gap area) for an individual tank is expressed as the ratio of the total area of gaps divided by the diameter of the tank (in2/ft). Each rim-seal test result shall be determined as the weighted average of evaporative loss-rate measurements for various rim-seal gap areas, on the basis of an assumed distribution of rim-seal gap areas among the actual tank population for each type of rim seal. The rim-seal gap areas to be tested for the determination of rim-seal loss factors are summarized in Table 5, as well as the assumed distribution to be used in calculating a weighted-average test result.

Table 4—Minimum Number of Tests for Rim Seals

Wind Speed (mph) 01 5 10 15

Number of Tests2 3 1 3 1

Test Method (API MPMS Chapter 19.3) Part B or Part C

Part B Part B Part B

Notes: 1. The nominal zero mph condition is any level of wind speed less than or equal to 0.5 mph. 2. Each test result shall be determined as the weighted average of evaporative loss-rate measurements for various rim-seal gap areas, as shown in Table 5.

The determination of acceptable variability of the test results, as specified in B.3.7, shall be performed at 0 mph for all rim seals, and additionally at 10 mph for rim seals to be used with external floating roofs. The limits specified in Table B.2, shall apply for both wind speed conditions.

Table 5—Rim-Seal Gap Areas1 and Distributions for Average-Fitting Rim Seals

Type of Primary Rim Seal Primary Rim-Seal Gap

Area (in2/ft)

Secondary Seal (if present) Gap Area

(in2/ft)

Distribution for Weighted Average (%)

Mechanical shoe-seal, welded tank shell2 0 0 10

2.8 0 80

9.4 1.0 10

Mechanical shoe-seal, riveted tank shell3 0 0 5

2.8 0 60

9.4 1.0 35

Liquid-mounted seal2 0 0 65


1.3 0 25

2.6 1.0 10

Vapor-mounted seal2 0 0 65

1.0 0 25

1.0 1.0 10

Notes: 1. A method for the determination of rim-seal gap areas is specified in MPMS, Chapter 19.3 Part B, 1st edition, section 11. The measured rim-seal gap area for a given test shall be within ±10 % of the nominal value selected from this table, except for the zero rim-seal gap area, which shall be as described in Chapter 19.3 Part B, 1st edition, 11.1.4. 2. Loss factors for tight-fitting rim seals, as described in API MPMS Chapter 19.2, 3rd edition, 4.2.2.2, shall be based on only the 0 in2/ft rim-seal gap area. 3. The potential for rivet heads to hold the metallic shoe of a mechanical-shoe seal away from a riveted tank shell results in an assumption of larger gaps for this combination of primary-seal type and tank construction.

9.4 DECK SEAMS

The deck seam test method described in Chapter 19.3 Part D, 1st edition, section 11.1, specifies that the pressure differences to be used for testing the evaporative loss rates for deck seams shall be 0.02, 0.04, 0.06, 0.08 and 0.10 inches of water column for each test. A minimum of three tests shall be performed for each deck seam loss factor to be determined, with each test involving measurements at each of the specified pressure differences.

Floating-roof deck seams in actual practice have to sometimes accommodate gaps between adjoining sections of the floating roof deck. At least one of the tests for the determination of a deck seam loss factor shall be performed on a deck seam constructed with an intentional gap in the fit up of the deck seam. This gap shall be described in the test report to document the maximum gap between adjoining sections for which the reported deck seam loss factor is applicable.

Testing for the determination of a deck seam loss factor shall include at least three deck joints representative of where the deck seams intersect each other and at least three deck joints representative of where deck seams intersect the rim of the floating roof. The deck joint configurations tested shall represent the full range of angles at which the deck seam may intersect the rim of the floating roof. The range of angles tested shall be described in the test report to document the full range for which the reported deck seam loss factor is applicable.

Any use of elastomers such as gaskets or caulk in the deck seams or the deck joints shall be photographed and described in sufficient detail such that it can be ascertained as to whether the details in actual construction are reasonably similar to the details that were tested.

The deck seam loss factor shall be determined from the average values of the results from each test at a pressure difference of 0.05 in of water column. If the test assemblies include representative types and quantities of deck joints, then the deck seam loss factor shall be calculated as follows:

n

KK

n

i dd

1 05.0

4

where:

Kd is the deck seam loss factor to be used in API MPMS Chapter 19.2,


05.0dK is the deck seam loss factor determined from the test method in API MPMS Chapter 19.3 Part D, 1st

edition, for an individual test, i, at a pressure difference of 0.05 in of water column;

n is the number of tests.

If the quantities of deck joints in the test assemblies are not representative of the frequency of deck joints per foot of deck seam in actual construction, then the loss rate shall be determined for the deck joints separately from determining the loss rate for the deck seams, and the overall deck seam loss factor shall be calculated as follows:

rim

j

tee

j

n

i d

d L

K

L

K

n

KK rimtee1 05.0

4

where:

Kd is the deck seam loss factor to be used in API MPMS Chapter 19.2,

05.0dK is the deck seam loss factor determined from the test method in API MPMS Chapter 19.3, Part D, 1st

edition, for an individual test, i, at a pressure difference of 0.05 in of water column,

n is the number of tests,

teejK is the average loss rate for deck joints where the deck seams intersect each other, at a pressure

difference of 0.05 in of water column,

Ltee is the typical length of deck seams per joint where the deck seams intersect each other,

rimjK is the average loss rate for deck joints where the deck seams intersect the rim of the floating roof, at

a pressure difference of 0.05 in of water column;

Lrim is the typical length of deck seams per joint where the deck seams intersect the rim of the floating roof. A tank diameter of 90 ft may be assumed for purposes of determining Lrim.

Testing conducted for the development of the test method in API MPMS Chapter 19.3 Part D, 1st edition, demonstrated that significant variability can occur in the observed loss rate for a given type of deck seam, particularly in the event of an imperfection in the fit up of the deck seam or deck joint in the test assembly. In that fit up can also be an issue during construction of the floating roof deck in an actual tank, imperfections in the fit up of the deck seam or deck joint or imperfections in the installation of caulk or gasketing in the test assembly shall not be a basis for rejecting the results of a test.

Given the expected variability of the test results, there is no limit on the acceptable variability of the test results for deck seam testing, That is, three tests shall be deemed acceptable for determination of a deck seam loss factor, regardless of the variability of the test results, as long as the three tests include representative deck joints and at least one test accommodates the maximum gap for which the reported deck seam loss factor is to be applicable.


ANNEX A—FLOW DIAGRAM OF THE LOSS FACTOR DEVELOPMENT PROGRAM

A.1 General

This annex provides flow diagrams illustrating the steps of the loss factor development program for the wind tunnel test method of API MPMS Chapter 19.3 Part A. Figure A.1 presents the procedure by which a testing laboratory develops a loss factor for a given device. The responsibilities involved in preparing to perform protocol testing are listed in Figure A.2. Figure A.3 outlines the procedure for actually performing the test method, and Figure A.4 is a checklist of the steps involved in the review of the testing program. Citations are to this Part H of API MPMS Chapter 19.3, unless another part is indicated.


ANNEX B—STATISTICAL CALCULATIONS

B.1 General

This annex provides the equations for the statistical calculations required by the loss factor development program. While the calculations are illustrated by their application to the wind tunnel test method of API MPMS Chapter 19.3 Part A, they are also generally applicable to the other test methods.

B.2 Nomenclature

The symbols listed in Table B.1 are used in the statistical calculations in addition to the nomenclature defined in 3.3.

Table B.1—Nomenclature

Symbol Description and Units

A is a constant in the correlation of weight change to temperature (pounds).

D is a factor to correct for variations in the temperature of the scale load cell (pounds per °F).

EX is the per unit uncertainty for a variable X (dimensionless); .

PCI is the percent confidence interval, also known as the percent error, and is equal to the per unit uncertainty expressed as a percent, EX × 100.

r2 is the coefficient of determination (dimensionless); .

S

is the sample standard deviation (same units as X); .

Tmi is the measured temperature of the scale load cell at time tmi (°F).

Ta is the average temperature of the scale load cell during a test (°F); .

tmi is the time of reading i, (i = 1,2,…,n).

t(1–∞/2,n–1) is the (1–∞/2) percentile of the student’s t-distribution at (n–1) degrees of freedom. For a 95 % confidence interval, then, this would be t0.975 (or t0.025) at (n-1) degrees of freedom.

UX is an expression of uncertainty, where a two-sided confidence interval for a variable X is expressed as

.

Wai is the correlated (fitted) weight loss at time tmi (pounds), also known as .

Wci is the measured weight loss at time tmi after correcting for variations in the temperature of the scale load cell (pounds).

wmi is the measured (observed) weight at time tmi (pounds).

EX UX X=

r2

Xi X– Yi Y– i 1=

n

2

Xi X– 2

Yi Y– 2

i 1=

n

i 1=

n

=

S Xi X– 2

n 1– i 1=

n

0.5

=

Ta1n--- Tmi

i 1=

n

=

X UX

Wc i


Wmi is the measured weight loss at time tmi (pounds); Wmi = (w0 – wmi).

w0 is the initial weight, measured at the beginning of the test period (pounds).

Z has the same meaning as UX, and is used for the particular case of the uncertainty in the repeatability of multiple tests at a given level of wind speed (i.e., 0 mph and 10 mph).


B.3 Statistical Formulas

B.3.1 Overview

B.3.2 addresses the evaluation of test results for deck-fitting standard devices. Table B.3 then summarizes the data to be collected and the statistical calculations to be performed for the wind tunnel test method of API MPMS Chapter 19.3 Part A. The statistical calculations referenced in Table B.3 are described in B.3.3 through B.3.8 below.

B.3.2 Evaluation of Standard Device Test Results

The testing laboratory shall perform tests of one or more standard devices for each test method to be performed. The results of these tests shall be compared to the reference values for these standard devices. Selection of standard devices and procedures for evaluating standard device test results are given in Annex D.

B.3.3 Standard Deviations of Weight And Wind Speed

The sample standard deviation, S, for each reading of weight and wind speed shall be estimated. The measurements, Xi , used to determine S are each of the 30 observations of the parameter in question at a given hourly reading. This is summarized on the first page of Table B.3, in the fifth column, and is to be calculated and recorded automatically by the data acquisition system (DAS).

B.3.4 Uncertainty in the Mean of Measured Values

The uncertainty of a given variable, X, may be expressed as . Since hourly readings are recorded for the wind speed (V), atmospheric pressure (Pa), and test liquid temperature (T), the uncertainty for each of these parameters is

UX = (t(1–∞/2, n–1))S/ . The values for Xi are the hourly readings, because the variance in question is for the duration of the test, rather than for an individual reading.

The absolute uncertainty, UX , is then converted to a per unit uncertainty, EX , by the expression . This is

summarized for EV , , and ET on the top of the first page of Table B.3, in the last column. Note that the sample standard deviations recorded in B.3.3 do not enter into this calculation.

B.3.5 Uncertainty in the Loss Rate for a Single Test

In that the loss rate, L, for a given test is assumed to be linear, it is obtained as the slope of a linear regression of the measured weight loss, Wmi , on time, tmi . The measured weight loss, Wmi , is determined as the difference between the initial weight, w0 , and the measured weight, wmi , at time tmi . Readings from the load cell sensing the weight, however, are affected by variations in the temperature of the load cell. These variations are also assumed to be linear, but the slope of the temperature-dependent curve varies from load cell to load cell. The first step in determining the loss rate, then, is to determine the temperature correction factor, d, for the load cell to be used. This may be done by measuring

X UX

n

EX UX X=

EPa


a weight of known mass over a range of temperature levels, and then performing a linear regression of the weight change, Wmi, on load cell temperature, Tmi

where

a = a constant in the weight change correlation (pounds).

d = the slope of the weight versus temperature curve (pounds per °F).

Tmi = measured load cell temperature at time tmi (°F).

= the estimated (fitted) value of weight change at temperature Tmi (pounds).

When a value for d is determined by a separate dead-weight test, then the weight loss measurements from protocol testing of a device are corrected for variations in the load cell temperature as follows:

Wci = (Wmi) – d (Tmi – Ta)

where

Wci = corrected value of the measured weight loss at time tmi (pounds).

Ta = average load cell temperature during the test period (°F).

The loss rate, L, is finally determined by a linear regression of the corrected weight loss, Wci , on time, tmi , resulting in a correlated, or fitted, weight loss, Wai

where Wai represents the estimated (fitted) value of corrected weight loss, .

In order to determine the uncertainty of the loss rate, L, use the variance of the slope

where Xi = tmi.

The uncertainty of the loss rate, UL , is then

UL = (t(1–∞/2, n–2)) ,

and EL = UL/L.

This procedure is summarized as the first calculated value presented at the end of Table B.3.

Wmi a dTmi+=

Wmi

Wai Wci Ltmi= =

Wci

S2

Xi X– 2

i 1=

n

-----------------------------

S 1

Xi X– 2

i 1=

n

-----------------------------


The coefficient of determination, r2, is an indication of the proportion of variation in the data that is explained by the temperature of the load cell. As specified in Section 5, this separate regression of dead-weight test data to determine d shall only be used when the resulting r2 is greater than or equal to 0.99 (or some other level if so documented in the test report). In all other cases, values for d and L shall be determined from a simultaneous regression of weight loss on both temperature and time, from the protocol testing of a device. This method is presented in Section A.5 of Parts A, C and E of the API MPMS Chapter 19.3 series of standards.

B.3.6 Uncertainty in the Loss Factor for a Single Test

The loss factor for a given test is a normalized expression of the loss rate. The per unit uncertainty, , of the loss factor, Kf, is given in Appendix B of the API MPMS Chapter19.3 Part A, 1st edition, as

.

The per unit uncertainty for the loss rate EL is obtained as outlined in B.3.5 above. is determined as a function of

and ET , each of which are obtained from B.3.4 above, as well as and , each of which have assumed

values assigned in Table B.3. The formulas for calculating given , ET, , and are given in API MPMS

Chapter 19.3 Part A, 1st edition, B.4.1 and B.4.2. Finally, and are assigned assumed values in Table B.3 of this document.

B.3.7 Variability Among Test Results at a Given Wind Speed

In addition to estimating the uncertainty of an individual test result, Kf , the uncertainty for the average of several tests at a specified level of wind speed shall be estimated. The expression for estimating this uncertainty is

; which is the same as ,

where each Xi in the determination of S is a test result, Kf , at the selected level of wind speed. The nominal 10 mph wind speed condition shall be taken as an actual wind speed level of 8.5 mph for the wind tunnel test method of API MPMS Chapter 19.3 Part A, for consistency with previous test results. This procedure is summarized to the right of the calculation of the loss factor near the end of Table B.3.

Noting that Z as defined above has the same meaning as UX in B.3.4 above, the per unit uncertainty, EX, could be

expressed as . Expressing this term as a percent yields the percent error, , defined as the percent confidence interval (PCI).

Note that this estimate of uncertainty for the average value of the loss factor, Kf, at a given level of wind speed is simply a function of the repeatability of the tests at that level of wind speed, and is not dependent upon the per unit

uncertainty of the individual test results, , from B.3.6 above.

The test procedure shall be performed three times. The repetition of the test shall include removal of the tested device from the test apparatus, disassembly and reassembly or replacement with a new device, and retesting through the entire specified test sequence. In those situations where the design or number of the test equipment at a specific facility could permit it, simultaneous tests of multiple test devices can be used to satisfy this requirement.

The loss factor shall be the average of the three values obtained whereby the result of the three tests are summed and divided by three. In order to determine if these three values are within an acceptable level of variability the statistical test shown below shall be used.

EKf

EKfEL

2 EP2 EMv

2 EKc

2+ + + 0.5

=

EP

EPaEAp

EBp

EPEPa

EApEBp

EKcEMv

Z t0.025 n 1– S

n-------= t0.975 n 1– S

n-------

EX Z X = Z X 100

EKf


Step 1: Determine the sample standard deviation (S) of the sample using the following formula:

(1)

Where:

is the value of the loss factor at the specified wind speed for a given test.

is the average of the loss factors for a given set of tests.


Step 2: Determine the 95 % confidence (Z) of the sample using the following formula:

(2)

Where:

t0.025 is the Student’s t Test value for 95 % probability at the appropriate degrees of freedom. The degrees of freedom is n-1.

S is the sample standard deviation.


Step 3: Express the percent confidence interval (PCI) using the following formula:

(3)

Where:

PCI = 95 % confidence as a percent of the mean.

Z = the confidence interval (Student’s t Test).

= the average of the loss factors for a given set of tests.

Step 4: The 95 % confidence as a percentage of the mean (PCI) as calculated by Equation 3 should not exceed the limits for each device as shown in Table B.2.

Table B.2—Acceptable Levels of Variability in Testing Data

Device Type of Test

Maximum 95 % Confidence Limit as a Percent

of Mean (PCI) Wind Speed Used for Calculation (miles/h)

Slotted guide poles Wind tunnel 25 10

Roof legs Wind tunnel 35 10

S xi x– 2

n 1–----------------------=

xi

x

PCIZx--- 100 percent=

x

Z t0.025S

n-- - - -- - =


Gauge hatch/sample well Wind tunnel 75 10

Vacuum breaker Wind tunnel 75 10

Gauge-float well Wind tunnel 65 10

Access hatch Wind tunnel 50 10

Internal rim seals Weight loss test or

Air concentration test 10 Not applicable

External rim seals Air concentration test 5 10

Deck seams Weight loss test Not Applicable Not applicable

If the PCI as calculated by Equation 3 exceeds the number listed in Table B.2, the test shall be repeated three more times. The new average will be the sum of all loss factors divided by six.

B.3.8 Uncertainty in the Loss Factor for a Tested Device

The uncertainty in the predicted value of the loss factor at a given level of wind speed is complicated by the use of a log transformation of the data to yield a linear relationship. The resulting linear expression is {log(Kf – Kfa) = log(Kfb) + m log(V)}. An estimator for the unbiased variance of this expression is not readily available. In the alternative, Annex H presents a procedure for comparing the measured loss factor at a given level of wind speed to the reference value for the type of device in question.


Table B.3—Summary of Parameters and Statistical Calculations for the API MPMS Chapter 19.3 Part A, 1st edition

Raw Data from the Individual

Readings

Frequency & Determination of a

Reading

Allowable Variation

Purpose of Measuring This Parameter

Sample Standard Deviation, S, of Each Reading, Where Each

Xi is One of the 30 Observations

95 % Confidence Interval

, Where Each Xi is an Hourly Reading

is the Avg. of the Hourly Readings

Weight 11.2.3.2 hourly—

mean of 30 observations

Variable 12.2 and A.4

Loss rate correlation

12.2 Required for record

purposes only

Not required directly—see UL for uncertainty of the loss rate, L, on the next page

Wind speed 11.2.3.1 hourly—

continuous record avg over 30 seconds

7.4 and 11.1.1 ±10 %

12.7 & Appendix C Loss factor

determination & correlation

13.2 Required for record

purposes only

Table B-3 Required

EV = UV / V

Atmos. pressure 11.2.3.4 hourly—

read directly Variable

12.3 & B.4.2 To obtain P* for the

loss factor determination

Not required B.4.2

Required

Test liquid temp. 11.2.3.2 hourly—


Variable

12.3 & B.4.1 To obtain the TVP for


Not required B.4.1

Required ET = UT / T

Load cell temp. 11.2.3.2 hourly—


Variable 12.2, A.3 or A.5

To correct the weight measurements

Not required Not required

Wind tunnel air temperature

11.2.3.2 hourly—mean of 30

observations

7.2.4 ±10 °F of the test

room air Quality control only Not required Not required

Test room air temp.

11.2.3.2 hourly—mean of 30

observations

7.2 ±5 °F

Quality control only Not required Not required

Time

11.2.3.1, 3.2, 3.3, 3.4 hourly—read

simultaneously with other readings

Variable 12.2, A.4

Loss rate correlation Not required Not required

UX t0.975 n 1– S

n------- =

EX UX X=

X

EPaUPa

Pa=


Voltage 11.2.3.3 hourly—read

directly

10.6 ±1 % of the calibration

voltage

Quality control only Not required Not required

The data shown above constitute the raw data to be recorded for each reading. The DAS shall automatically calculate and record the sample standard deviation of each reading of weight and wind speed (10.2). Section 13.2 of API MPMS Chapter 19.3 Part A, 1st edition, requires that these data, along with the corrected weight measurements, shall be included with the documentation for the loss rate curve.

Note: All citations in this table are in reference to the wind tunnel test method for deck fittings of the API MPMS Chapter 19.3 Part A, 1st edition, unless noted otherwise.

Table B.3—Summary of Parameters and Statistical Calculations for the API MPMS Chapter 19.3, Part A (Continued)

Test Liquid Properties

Frequency & Determination of a

Reading

Allowable Variation

Purpose of Measuring This Parameter

is the Avg. of the Hourly Readings

Reid vapor pressure

7.6.2 beginning & end of each test, by ASTM

D323 or D5191

7.6.2 –5 %

Quality control only

B.4.1 EP is determined

from

The following are assumed values for n-hexane, technical grade or better:

Vapor pressure constant, Ap

13.824



Vapor pressure constant, Bp

6907.2 °R



Molecular weight of the test liquid vapor, Mv

86.18 lb/lb-mole 12.4 & B.4.3 Loss factor

determination

Product factor, Kc

1.0 12.4 & B.4.3 Loss factor

determination

EX UX X=

X

EAPEBP

ET

EAp1

3–10=

EBp1

3–10=

EMv1

3–10=

EKc0=


Calculated Values

Loss rate (L) for each test

10.3.2.2.1 & A.3 Estimate the temp. correction factor, d, by a linear regression of data from a separate dead weight test, where:

A.3 Obtain the corrected

weight, Wci , at each test reading from:

Wci = Wmi –d (Tmi–Ta)

A.4 Estimate the loss rate, L, by a linear

regression of:

Estimate the variance of

L from where Xi = tmi

Estimate the 95 % confidence interval of L from

where 95 % C.I. = L ±UL

Loss factor (Kf) for each test

12.4 Obtain the loss factor Kf for each test

from the loss rate, L, as follows:

B.4 Estimate the uncertainty

(95 % C.I.) of an individual test result Kf , given the values of EX

determined above

B.3.6 Estimate the 95 % C.I. of Kf at 0 mph and 8.5 mph from

/ where each Xi is a separate test result Kf at the given wind

speed

Loss factor equation

12.7 & C.3 Kfa + KfbVm

B.3.6 & C.3 Obtain Kfa as the

average Kf from at least 3 tests at 0 mph (i.e.,

< 0.5 mph)

C.3 On a log-log scale, using all test results Kf for wind speeds of V > 0.5 mph, obtain Kfb and m

from a linear regression of: log (Kf –Kfa) = log (Kfb) + m log (V)

Wmi a dTmi+= Wai Wci Ltmi= =

S2

Xi X– 2

i 1=

n

-------------------------------

UL t0.975 n 1– S 1

Xi X– 2

i 1=

n

-------------------------------=

KfL 24 hr/day 365.25 days/yr

PMvKc

-------------------------------------------------------------------------=

Z t0.975 n 1– S= n


ANNEX C—WIND TUNNEL VELOCITY PROFILE

C.1 General

This annex provides the procedure for performing a survey of the wind tunnel velocity profile, as required by API MPMS Chapter 19.3 Part A, 1st edition, 7.3.4.1. Surveys shall be conducted at nominal wind speeds of 5 mph, 10 mph, and 15 mph with an empty wind tunnel as part of the start-up documentation of a testing laboratory for performing the wind tunnel test method of API MPMS Chapter 19.3 Part A, 1st edition. Following the initial wind tunnel survey for this test method, a single velocity profile survey shall be conducted every six months at a nominal wind speed of 10 mph.

C.2 Nomenclature

The symbols listed in Table C.1 are used in this annex in addition to those defined in 3.3.

Table C.1—Nomenclature

Symbol Description and Units

Vi is the wind speed measured at profile location i in the cross section of the wind tunnel (mph).

is the mean of the wind speeds, Vi , measured at profile locations i = 1 through i = n (mph).

Vref is the wind speed at the geometric center of a cross section of the wind tunnel (mph).

is the mean of the reference wind speeds, Vref , measured with each profile wind speed, Vi (mph).

Vj is the wind speed at trial location j in the cross section of the wind tunnel (mph)

Vj (ref) is the reference wind speed measured with a trial location wind speed, Vj (mph).


C.3 Survey Procedure

C.3.1 Measurement Locations

The survey of the velocity profile shall be performed at a cross section of the wind tunnel midway along the length of a measuring station. The profile shall be obtained by measuring the wind speed, Vi, at each location i on a 6 in square grid. The boundary points of the measurement grid shall be located 3 in from the perimeter of the wind tunnel. Such a grid is shown in Figure C.1 for a wind tunnel having cross-sectional dimensions of 3 ft by 3 ft.

V

Vref


Figure C.1—Grid for Velocity Profile Measurements

During each measurement of wind speed for the velocity profile, also record a reference wind speed, Vref , which shall be measured at the geometric center of a cross section of the wind tunnel. The cross section for the reference wind speed shall be midway along the length of a measuring station other than where the profile is being obtained.

C.3.2 MEASUREMENT PROCEDURES

Each reading of wind speed shall be determined by continuous record averaging over a 30 s time period. Measurements of the velocity profile shall be obtained using a pitot tube, meeting the requirements of API MPMS Chapter 19.3 Part A, 1st edition, 10.5. The reference wind speed shall also be measured using a pitot tube. Each measurement of wind speed, Vi , for the velocity profile shall be accompanied by a simultaneous measurement of reference wind speed, Vref .

C.3.3 DATA ANALYSIS AND RECORD-KEEPING

The wind speed, Vi , shall be divided by the reference wind speed, Vref , to obtain a normalized wind speed for each location of the velocity profile. The data shall then be summarized in a table, as illustrated by Table C.2.


Table C.2—Velocity Profile Survey Data

Location Vi Vref (Vi /Vref)

1 Nominal Wind Speed ____ mph

2 Survey Measurement Station No. _________

3 Reference Measurement Station No._______

.

Sample standard deviation = Sample standard deviation = ______ mph

.

.

36

All locations (V1 through V36)

minimum

maximum

mean

V1 through V16 (i.e., excluding the boundary)

minimum

maximum

mean

Display the normalized wind speeds, (Vi /Vref ), on a grid of the cross section as shown in Figure C.2.

Figure C.2—Grid Display of Normalized Wind Speeds

Vi V– 2

i 1=

36

35

-------------------------------

0.5


The range of measured values for the reference wind speed, Vref , shall be within ±5 % of their mean, . The normalized wind speeds, (Vi /Vref ), for all locations shall not vary by more than ±40 % from their average value, and the normalized wind speeds for nonboundary locations (i.e., V1 through V16 ) shall not range by more than ±20 %.

C.4 Location of Wind Speed Sensors

The instruments for measuring wind speed shall be positioned in the wind tunnel as specified in API MPMS Chapter 19.3 Part A, 1st edition, 10.5.1. This includes a requirement for each sensor to be located in a position such that it is

measuring a value that is within ±5 % of the geometric average wind speed, . An acceptable cross-sectional position shall be located by measuring the wind speed, Vj , at a trial location, j, while simultaneously measuring a reference wind speed, Vj (ref) . The requirement for the measured wind speed to be within ±5 % of the average is then evaluated as follows:

If this requirement, as well as the other requirements of API MPMS Chapter 19.3 Part A, 1st edition, 10.5.1, are met, then the trial location is acceptable. Document this determination by reporting the measured wind speed, Vj, and the corresponding reference wind speed, Vj (ref), for the location selected for each wind speed sensor.

Vref

V

0.95Vj

Vj ref ---------------

Vref

V---------

1.05


ANNEX D—TESTING OF STANDARD DEVICES

D.1 General

This annex illustrates appropriate procedures for testing standard devices, as required by section 6.4. Procedures are specified below for the testing of deck-fitting standard devices in accordance with the test methods of API MPMS Chapter 19.3, Parts A and E. The standard devices to be tested are selected, and formulas and tables are provided to facilitate the comparison of test results with the reference values.

D.2 Selection of Standard Devices

D.2.1 Deck Fitting Standard Devices

The standard devices for the testing of deck fittings shall be as given in D.2.1.1 and D.2.1.2.

D.2.1.1 Deck Fitting Standard Device Number 1

Standard device number 1 shall be a 3 in diameter adjustable deck leg for the pontoon area of an external floating roof. The deck leg shall be ungasketed and without a sock, and the test assembly shall be constructed in accordance with fitting number 27 as shown in Figure number B-9 of API Publication 2517D, Documentation File for API Publication 2517, 1st Edition, March 1993. The freeboard of the leg sleeve above the surface of the test liquid shall be 18 in.

D.2.1.2 Deck Fitting Standard Device Number 2

Standard device number 2 shall be a slotted guidepole with no well gasket, no float, no pole wiper, and no pole sleeve. The slotted guidepole test assembly shall be constructed in accordance with fitting number 21 as shown in Figure numbers B-1 and B-2 of API Publication 2517D, 1st Edition, March 1993.

D.2.2 Rim Seal Standard Device

The standard device for internal floating roof rim seals tested in accordance with the weight loss test method of API MPMS Chapter 19.3 Part C shall be a liquid-mounted single resilient-filled seal with a 5 ½ inch rim space, constructed as described in the API report Loss Factor Measurements of Internal Floating Roof Rim Seals, February 1, 1995. An alternative method of testing floating roof rim seals is presented in API MPMS Chapter 19.3 Part B that utilizes an Air Concentration Test (ACT) facility that had been owned by API but which no longer exists. A standard device has not been selected for the ACT method.

D.2.3 Deck Seam Standard Device

The standard device for deck seams shall be an ungasketed bolted seam for a non-contact deck as described for Test Pan Number 2 in paragraph 4.1.2 of the API report Development of the Testing Protocol for the Measurement of Deck Seam Loss Factors Using the Fugitive Emission Test Method, September 14, 1999.

D.3 Testing Standard Devices to Calibrate the Test Facility

D.3.1 General

Standard devices shall be tested as part of the preparation of a testing laboratory for the testing of deck fittings, in order to compare the test results of the test facility to the reference values. This testing shall be performed in accordance with the procedure given in D.3.2 to D.3.4.


D.3.2 Perform The Evaporative Loss Testing

D.3.2.1 Deck Fitting Testing

Test each deck fitting standard device at each of the following wind speeds: 0 mph, 4.3 mph, 8.5 mph, and 11.9 mph. Standard device number 1, the deck leg, shall be tested at least three times at each wind speed. Standard device number 2, the slotted guidepole, shall be tested at least three times at 0 mph, and three times at each of three orientations at each of the nonzero wind speeds. The orientations are defined by the position of the slots with respect to the direction of the wind. Tests shall be conducted with the slots oriented at 0° (i.e., a line through the center of the slots oriented parallel to the wind direction), 45°, and 90° (i.e., a line through the center of the slots oriented perpendicular to the wind direction). Testing of standard devices shall be conducted in full compliance with the testing protocol of API MPMS Chapter 19.3 Part A, except as specified in D.3.3. Deck fittings to be used only on internal floating-roof tanks may be tested only at a wind speed of 0 mph.

D.3.2.2 Rim Seal Testing

Test the rim seal standard device at 0 mph at each of the following gap conditions: 0 in2/ft dia, 1.2 in2/ft dia, 2.8 in2/ft dia. The standard device shall be tested at least three times at each gap condition. Testing of the rim seal standard device shall be conducted in full compliance with the testing protocol of API MPMS Chapter 19.3 Part C, except as specified in D.3.3.

D.3.2.3 Deck Seam Testing

Test the deck seam standard device at least three times at a wind speed of 0 mph. Testing of the deck seam standard device shall be conducted in full compliance with the testing protocol of API MPMS Chapter 19.3 Part D, except as specified in D.3.3.

D.3.3 Evaluate the Variability of the Test Results

Estimate the percent error (percent confidence interval) of the test results as specified in B.3.7. The standard devices are not required to be disassembled and reassembled between tests, but shall be moved to a different measuring station after each test (i.e., consecutive tests of a given standard device shall not be performed at the same measuring station). The percent error (percent confidence interval) for each level of wind speed at which the standard device is tested shall be within the levels specified for 10 mph in Table B.2. For any test series displaying greater variability than the specified allowable, three more tests of the same standard device at the same level of wind speed shall be tested. The data from all six tests shall then be used in the comparison to reference values.

D.3.4 Comparison to Reference Values

The results of the standard device testing shall be compared to reference values by performing a statistical test of the null hypothesis that there is no difference between the means of the two samples (i.e., the testing laboratory’s test results and the reference data) at a 0.10 level of significance. The reference data and required formulas are presented in Tables D.1 and D.2.

D.4 Periodic Testing of Standard Devices

D.4.1 General

Section 6.4 requires that at least every twentieth protocol test conducted at a test facility be of a standard device.

D.4.2 Sequence for Testing Standard Devices

The testing laboratory shall perform tests at 8.5 mph for standard devices for use with external floating roofs and at the nominal zero miles per hour wind speed for standard devices for use only with internal or covered floating roofs.


For deck fittings, the first periodic test shall be of standard device number 1, the deck leg, followed by standard device number 2, the slotted guidepole, with the slots oriented first at 0°, then at 45°, and finally at 90°. The selection shall then return to standard device number 1, again at 8.5 mph, and the sequence shall be repeated.

D.4.3 Evaluation for Trends

The testing laboratory shall maintain records of all results of standard device testing. Following the third periodic test of a standard device, the testing laboratory shall maintain records of trends in the test results. Trends shall be investigated by performing a linear regression of the data after each additional periodic test of that device, and comparing the slope of the regression curve to zero.

Table D.1A—Deck Fitting Standard Device Number 1

Deck Leg—EFR Pontoon Area (18 in sleeve freeboard) No Gasket or Sock Wind Speed = 0 mph

Reference Data

Test Prog. Ftg. No. Test No. Loss Rate

L (lb/h)

Test Liquid Temp., (T,

°F)

P (psia)

Pa

(psia)

P* (dimensionle

ss)

Mv

(lb/lb-mole) Loss Factor

Kf (lb-mole/yr)

1993 5 Z25 0.0007506 67.75 2.07 14.4 0.039 86.2 1.97

1984 27 15B 0.000675 75.3 2.49 14.4 0.047 86.2 1.45

= 1.71

=

0.136

nref = 2

Calibration Data


L (lb/h)


°F)

P (psia)

Pa

(psia)

P* (dimensionle

ss)

Mv


Kf (lb-mole/yr)

initial SD-1 A1 KfA1

“ SD-1 A2 KfA2

“ SD-1 A3 KfA3

.

.

.

.

.

.

.

.

.

.

.

.

“ SD-1 An KfAn

where Loss Rate, L = the slope of the correlated weight loss curve, from the test data (lb/h), (T, °F) = the average temperature of the test liquid, from the test data (°F), P = the true vapor pressure of the test liquid (psia).

where (T, °R) is the temperature of the test liquid (°R); (T, °R) = (T, °F) + 459.67, Ap is a constant (dimensionless). Ap = 13.824 for n-hexane, Bp is a constant (°R). Bp = 6907.2 (°R) for n-hexane. Pa = the average atmospheric pressure, from the test data (psia), Pa for the reference data is taken as 14.4 psia, based on an elevation of 630 ft. for Plainfield, IL, P* is a vapor pressure function (dimensionless).

Mv = the molecular weight of the test liquid vapor, assumed to be 86.2 lb/lb-mole for technical grade n-hexane (lb/lb-mole). Loss Factor, Kf = the normalized loss rate on a yearly basis (lb-mole/yr).

Xref

Sref2

P Ap

Bp

T R -----------------–exp=

PP Pa

1 1 P Pa – 0.5+

2-------------------------------------------------------=

KfL (24 hr/day) (365.25 days/yr)

P Mv ---------------------------------------------------------------------------------=


Statistical Calculations:

= the mean of the loss factors for the calibration tests, KfAi .

ncal = the number of calibration tests, ncal ≥ 3,

= the mean squared error of the calibration tests.

= the pooled variance.

t = the test statistic.

Degrees of freedom for the student’s t-distribution: (nref + ncal – 2).

The calibration test is acceptable if .

Table D.1B—Deck Fitting Standard Device Number 1

Deck Leg—EFR Pontoon Area (18 in sleeve freeboard) No Gasket or Sock Wind Speed = 4.3 mph

Reference Data


L (lb/h)


°F)

P (psia)

Pa

(psia)

P* (dimensionle

ss)

Mv


Kf (lb-mole/yr)

1993 5 61 0.001331 71.11 2.25 14.4 0.042 86.2 3.19

1993 5 61R 0.002059 77.1 2.60 14.4 0.050 86.2 4.21

1993 5 61RR 0.001419 78.88 2.71 14.4 0.052 86.2 2.77

1984 27 15A 0.001296 77.1 2.60 14.4 0.050 86.2 2.65

= 3.20

=

0.503

nref = 4

Calibration Data


L (lb/h) Test Liquid

Temp., (T, °F) P

(psia) Pa

(psia)

P* (dimensionle

ss)

Mv


Kf (lb-mole/yr)

initial SD-1 B1 KfB1

“ SD-1 B2 KfB2

Xcal

Xcal

KfAi

i 1=

ncal

ncal

-------------------=

Scal2

Scal2

KfAi Xcal– 2

i 1=

ncal

ncal 1–

------------------------------------------=

Sp2

Sp2 nref 1– Sref

2ncal 1– Scal

2+

nref ncal 2–+----------------------------------------------------------------------=

tXref Xcal–

Sp1

nref--------

1ncal---------+

----------------------------------=

t t0.95 nref nca l 2–+

Xref

Sref2


“ SD-1 B3 KfB3

.

.

.

.

.

.

.

.

.

.

.

.

“ SD-1 Bn KfBn

Variables defined as for Table D.1A.


= the mean of the loss factors for the calibration tests, KfBi . ncal = the number of calibration tests, ncal ≥ 3,





The calibration test is acceptable if: .

Table D.1C—Deck Fitting Standard Device Number 1


Reference Data


L (lb/h)


°F)

P (psia)

Pa

(psia)

P* (dimensionle

ss)

Mv


Kf (lb-mole/yr)

1993 5 65 0.001632 76.32 2.55 14.4 0.049 86.2 3.40

1993 5 65R 0.002677 80.6 2.83 14.4 0.055 86.2 4.99

= 4.20

=

1.252

nref = 2

Calibration Data


L (lb/h)


°F)

P (psia)

Pa

(psia)

P* (dimensionle

ss)

Mv


Kf (lb-mole/yr)

initial SD-1 C1 KfC1

“ SD-1 C2 KfC2

“ SD-1 C3 KfC3

.

.

.

.

.

.

.

.

.

.

.

.

Xcal

Scal2

Scal2

KfBi Xcal– 2

i 1=

ncal

ncal 1–

------------------------------------------=

Sp2

Sp2 nref 1– Sref

2ncal 1– Scal

2+

nref ncal 2–+----------------------------------------------------------------------=

tXref Xcal–

Sp1

nref--------

1ncal---------+

----------------------------------=


Xref

Sref2


“ SD-1 Cn KfCn



= the mean of the loss factors for the calibration tests, KfCi .

ncal = the number of calibration tests. ncal ≥ 3,






Table D.1D—Deck Fitting Standard Device Number 1


Reference Data


L (lb/h)


°F)

P (psia)

Pa

(psia)

P* (dimensionle

ss)

Mv


Kf (lb-mole/yr)

1993 5 69 0.001821 72.02 2.3 14.4 0.043 86.2 4.26

1993 5 69R 0.002897 76.82 2.58 14.4 0.049 86.2 5.96

1993 5 69RR 0.003474 78.12 2.67 14.4 0.051 86.2 6.91

1984 27 11 0.001721 66.0 1.98 14.4 0.037 86.2 4.73

1984 27 11R 0.001371 68.1 2.09 14.4 0.039 86.2 3.56

1984 27 11RR 0.001917 70.9 2.24 14.4 0.042 86.2 4.62

= 5.01

=

1.484

nref = 6

Calibration Data


L (lb/h)


°F)

P (psia)

Pa

(psia)

P* (dimensionle

ss)

Mv


Kf (lb-mole/yr)

initial SD-1 D1 KfD1

“ SD-1 D2 KfD2

Xcal

Xcal

KfCi

i 1=

ncal

ncal

-------------------=

Scal2

Scal2

KfCi Xcal– 2

i 1=

ncal

ncal 1–

------------------------------------------=

Sp2

Sp2 nref 1– Sref

2ncal 1– Scal

2+

nref ncal 2–+----------------------------------------------------------------------=

tXref Xcal–

Sp1

nref--------

1ncal---------+

----------------------------------=


Xref

Sref2


“ SD-1 D3 KfD3

.

.

.

.

.

.

.

.

.

.

.

.

“ SD-1 Dn KfDn



= the mean of the loss factors for the calibration tests, KfDi .





Degrees of freedom for the student’s t-distribution: (nref + ncal – 2)

The calibration test is acceptable if:

Xcal

Xcal

KfDi

i 1=

ncal

ncal

-------------------=

Scal2

Scal2

KfDi Xcal– 2

i 1=

ncal

ncal 1–

------------------------------------------=

Sp2

Sp2 nref 1– Sref

2ncal 1– Scal

2+

nref ncal 2–+----------------------------------------------------------------------=

tXref Xcal–

Sp1

nref--------

1ncal---------+

----------------------------------=



Table D.2A—Deck Fitting Standard Device Number 2

Slotted Guidepole—No Well Gasket, Float, Wiper, or Sleeve Wind Speed = 0 mph

Reference Data


L (lb/h)


°F)

P (psia)

Pa

(psia)

P* (dimensionle

ss)

Mv


Kf (lb-mole/yr)

1993 1 Z1 0.01445 61.79 1.78 14.4 0.0330 86.2 44.5

Calibration Data


L (lb/h)


°F)

P (psia)

Pa

(psia)

P* (dimensionle

ss)

Mv


Kf (lb-mole/yr)

initial SD-2 A1 KfA1

“ SD-2 A2 KfA2

“ SD-2 A3 KfA3

.

.

.

.

.

.

.

.

.

.

.

.

“ SD-2 An KfAn

where Loss Rate, L = the slope of the correlated weight loss curve, from the test data (lb/h). (T, °F) = the average temperature of the test liquid, from the test data (°F). P = the true vapor pressure of the test liquid (psia).

where (T, °R) is the temperature of the test liquid (°R); (T, °R) = (T, °F) + 459.67 Ap is a constant (dimensionless). Ap = 13.824 for n-hexane. Bp is a constant (°R). Bp = 6907.2 (°R) for n-hexane. Pa = the average atmospheric pressure, from the test data (psia). Pa for the reference data is taken as 14.4 psia, based on an elevation of 630 ft. for Plainfield, IL. P* = a vapor pressure function (dimensionless).



= the arithmetic average (mean) of the loss factors for the calibration tests, KfAi .



P Ap

Bp

T R -----------------–exp=

PP Pa

1 1 P Pa – 0.5+

2-------------------------------------------------------=


P Mv ---------------------------------------------------------------------------------=

Xcal

Xcal

KfAi

i 1=

ncal

ncal

-------------------=

Scal2


= the unbiased estimate of variance. t = the test statistic.

where


Table D.2B—Deck Fitting Standard Device Number 2

Slotted Guidepole—No Well Gasket, Float, Wiper, or Sleeve Wind Speed = 4.3 mph

Reference Data

Test Prog. Ftg. No. Test No. Loss Factor

Kf (lb-mole/yr)

1993 1 all .......................................................................average of all orientations = 2388

1984 21 all .......................................................................average of all orientations = 2296

= 2342

=

4284

nref = 2

Calibration Data

Orientation (in degrees)

Ftg. No. Test No. Loss Rate

L (lb/h)


°F)

P (psia)

Pa

(psia)

P* (dimensionle

ss)

Mv


Kf (lb-mole/yr)

0 SD-2 B1 KfB1

45 SD-2 B2 KfB2

90 SD-2 B3 KfB3

Average of all orientations, test series I ..................................................................................................................

0 SD-2 B4 KfB4

45 SD-2 B5 KfB5

90 SD-2 B6 KfB6

Average of all orientations, test series II ................................................................................................................

0 SD-2 B7 KfB7

45 SD-2 B8 KfB8

90 SD-2 B9 KfB9

Average of all orientations, test series III ...............................................................................................................

Scal2

KfAi Xcal– 2

i 1=

n

ncal 1–

------------------------------------------=

Scal2

ncal

tXcal 0–

variance------------------------

Xcal 0–

Scal2

ncal--------------------------

Xcal 0–

Scal ncal -------------------------------= = =

0 44.5 lb-mole/yr.=

t t0.95 n 1–

Xref

Sref2

KfBI

KfB1 KfB2 KfB3+ + 3

--------------------------------------------------=

KfBII

KfB4 KfB5 KfB6+ + 3

--------------------------------------------------=

KfBII

KfB7 KfB8 KfB9+ + 3

--------------------------------------------------=




= the mean of the loss factors for the calibration test series, KfBj .

ncal = the number of calibration test series, ncal = 3,

= the mean squared error of the calibration test series.





Table D.2C—Deck Fitting Standard Device Number 2


Reference Data


Kf (lb-mole/yr)

1993 1 initial ....................................................................... average of all orientations = 6563

1993 1 repeat ....................................................................... average of all orientations = 6024

= 6293

= 144962

nref = 2

Xcal

Xcal

KfBj

j I=

III

ncal

------------------=

Scal2

Scal2

KfBj Xcal– 2

j I=

III

ncal 1–

-----------------------------------------=

Sp2

Sp2 nref 1– Sref

2ncal 1– Scal

2+

nref ncal 2–+----------------------------------------------------------------------=

tXref Xcal–

Sp1

nref--------

1ncal---------+

----------------------------------=


Xref

Sref2


Calibration Data



L (lb/h)


°F)

P (psia)

Pa

(psia)

P* (dimensionle

ss)

Mv


Kf (lb-mole/yr)

0 SD-2 C1 KfC1

45 SD-2 C2 KfC2

90 SD-2 C3 KfC3

Average of all orientations, test series I .................................................................................................................

0 SD-2 C4 KfC4

45 SD-2 C5 KfC5

90 SD-2 C6 KfC6

Average of all orientations, test series II .........................................................................................................

0 SD-2 C7 KfC7

45 SD-2 C8 KfC8

90 SD-2 C9 KfC9

Average of all orientations, test series III ........................................................................................................



= the mean of the loss factors for the calibration test series, KfCj .







KfCI

KfC1 KfC2 KfC3+ + 3

--------------------------------------------------=

KfCII

KfC4 KfC5 KfC6+ + 3

--------------------------------------------------=

KfCIII

KfC7 KfC8 KfC9+ + 3

--------------------------------------------------=

Xcal

Xcal

KfCj

j I=

III

ncal

------------------=

Scal2

Scal2

KfCj Xcal– 2

j I=

III

ncal 1–

-----------------------------------------=

Sp2

Sp2 nref 1– Sref

2ncal 1– Scal

2+

nref ncal 2–+----------------------------------------------------------------------=

tXref Xcal–

Sp1

nref--------

1ncal---------+

----------------------------------=



Table D.2D—Deck Fitting Standard Device Number 2


Reference Data


Kf (lb-mole/yr)

1993 1 all ....................................................................... average of all orientations = 7307

1984 21 all ....................................................................... average of all orientations = 8396

= 7852

= 593612

nref = 2

Calibration Data



L (lb/h)


°F)

P (psia)

Pa

(psia)

P* (dimensionle

ss)

Mv


Kf (lb-mole/yr)

0 SD-2 D1 KfD1

45 SD-2 D2 KfD2

90 SD-2 D3 KfD3

Average of all orientations, test series I ...........................................................................................................

0 SD-2 D4 KfD4

45 SD-2 D5 KfD5

90 SD-2 D6 KfD6

Average of all orientations, test series II .........................................................................................................

0 SD-2 D7 KfD7

45 SD-2 D8 KfD8

90 SD-2 D9 KfD9

Average of all orientations, test series III .......................................................................................................



= the mean of the loss factors for the calibration test series, KfDj .



Xref

Sref2

KfDI

KfD1 KfD2 KfD3+ + 3

---------------------------------------------------=

KfDII

KfD4 KfD5 KfD6+ + 3

---------------------------------------------------=

KfDIII

KfD7 KfD8 KfD9+ + 3

---------------------------------------------------=

Xcal

Xcal

KfDj

j I=

III

ncal

-------------------=

Scal2






Scal2

KfDj Xcal– 2

j I=

III

ncal 1–

------------------------------------------=

Sp2

Sp2 nref 1– Sref

2ncal 1– Scal

2+

nref ncal 2–+----------------------------------------------------------------------=

tXref Xcal–

Sp1

nref--------

1ncal---------+

----------------------------------=



Table D.3A—IFR Rim Seal Standard Device

Single Resilient-Filled Rim Seal—No Secondary Wind Speed = 0 mph

Reference Data

Test Prog. Test. No. Rim Seal Gap Area

(in2/ft-diam)

Seal Length LS (ft)

Loss Rate L (lb/h)

Test Liquid Temp., (T, °F)

P (psia)

Pa

(psia) Mv


Kr (lb-mol/ft-yr)

1994 18 0 11.48 0.0048164 84.47 5.848 14.4 85.97 1.0654

Calibration Data


(in2/ft-diam)

Seal Length LS (ft)

Loss Rate L (lb/h)


P (psia)

Pa

(psia) Mv


Kr (lb-mole/ft-yr)

initial A1 KrA1

“ A2 KrA2

“ A3 KrA3

.

.

.

.

.

.

.

.

.

“ An KrAn


where (T, °R) is the temperature of the test liquid (°R); (T, °R) = (T, °F) + 459.67 Ap is a constant (dimensionless). Ap = 13.78 for the reference isohexane. Bp is a constant (°R). Bp = 6513 (°R) for the reference isohexane. Pa = the average atmospheric pressure, from the test data (psia). Pa for the reference data is taken as 14.4 psia, based on an elevation of 630 ft. for Plainfield, IL. P* = a vapor pressure function (dimensionless).

Mv = the molecular weight of the test liquid vapor, calculated to be 85.97 lb/lb-mole for the reference isohexane. Loss Factor, Kr = the normalized loss rate on a yearly basis (lb-mole/ft-yr).

𝐾𝐿 24 hr/day 365.25 day/yr 𝜋

𝐿 𝑃∗ 𝑀


= the arithmetic average (mean) of the loss factors for the calibration tests, KrAi .

𝑋∑ 𝐾

𝑛



𝑆∑ 𝐾 𝑋

𝑛 1

P Ap

Bp

T R -----------------–exp=

PP Pa

1 1 P Pa – 0.5+

2-------------------------------------------------------=

Xcal

Scal2



where μ0 = 1.0654 lb-mol/ft-yr


Table D.3B—IFR Rim Seal Standard Device


Reference Data


(in2/ft-diam)

Seal Length LS (ft)

Loss Rate L (lb/h)


P (psia)

Pa

(psia) Mv


Kr (lb-mol/ft-yr)

1994 C4 1.2 11.48 0.0077423 82.88 5.353 14.4 85.97 1.9165

Calibration Data


(in2/ft-diam)

Seal Length LS (ft)

Loss Rate L (lb/h)


P (psia)

Pa

(psia) Mv


Kr (lb-mole/ft-yr)

initial B1 KrB1

“ B2 KrB2

“ B3 KrB3

.

.

.

.

.

.

.

.

.

“ Bn KrBn





𝐿 𝑃∗ 𝑀


= the arithmetic average (mean) of the loss factors for the calibration tests, KrBi .

Scal2

ncal

tXcal 0–

variance------------------------

Xcal 0–

Scal2

ncal--------------------------

Xcal 0–

Scal ncal -------------------------------= = =

t t0.95 n 1–

P Ap

Bp

T R -----------------–exp=

PP Pa

1 1 P Pa – 0.5+

2-------------------------------------------------------=

Xcal


𝑋∑ 𝐾

𝑛



𝑆∑ 𝐾 𝑋

𝑛 1




Table D.3C—IFR Rim Seal Standard Device


Reference Data


(in2/ft-diam)

Seal Length LS (ft)

Loss Rate L (lb/h)


P (psia)

Pa

(psia) Mv


Kr (lb-mol/ft-yr)

1994 19 2.8 11.48 0.0107510 84.04 5.665 14.4 85.97 2.4772

Calibration Data


(in2/ft-diam)

Seal Length LS (ft)

Loss Rate L (lb/h)


P (psia)

Pa

(psia) Mv


Kr (lb-mole/ft-yr)

initial C1 KrC1

“ C2 KrC2

“ C3 KrC3

.

.

.

.

.

.

.

.

.

“ Cn KrCn



Scal2

Scal2

ncal

tXcal 0–

variance------------------------

Xcal 0–

Scal2

ncal--------------------------

Xcal 0–

Scal ncal -------------------------------= = =

t t0.95 n 1–

P Ap

Bp

T R -----------------–exp=




𝐿 𝑃∗ 𝑀


= the arithmetic average (mean) of the loss factors for the calibration tests, KrCi .

𝑋∑ 𝐾

𝑛



𝑆∑ 𝐾 𝑋

𝑛 1




PP Pa

1 1 P Pa – 0.5+

2-------------------------------------------------------=

Xcal

Scal2

Scal2

ncal

tXcal 0–

variance------------------------

Xcal 0–

Scal2

ncal--------------------------

Xcal 0–

Scal ncal -------------------------------= = =

t t0.95 n 1–


Table D.4—Deck Seam Standard Device

Bolted Seam—Non-Contact Deck, No Gasket

Wind Speed = 0 mph

Reference Data

Test Prog. Test Pan

No. Test No.

a (lb-mol/ft-yr)

b (lb-mol/ft-yr

in.wc)


P (psia)

Pa

(psia) P*

(dimensionless) Mv

(lb/lb-mole)Loss Factor

Kd (lb-mol/ft-yr)

1999 2 7 0.027 0.950 63.18 2.19 14.4 0.041 86.2 0.0746

1999 2 8 0.005 0.263 66.33 2.37 14.4 0.045 86.2 0.0177

= 0.0462

=

0.0016

nref = 2

Calibration Data

Test Prog. Test Pan

No. Test No.

a (lb-mol/ft-yr)

b (lb-mol/ft-yr

in.wc)


P (psia)

Pa

(psia) P*

(dimensionless) Mv

(lb/lb-mole)Loss Factor

Kd (lb-mol/ft-yr)

initial SD-4 1 Kd1

“ SD-4 2 Kd2

“ SD-4 3 Kd3

.

.

.

.

.

.

.

.

.

.

.

.

“ SD-4 n Kdn

where a = the y-intercept of the loss factor versus pressure difference correlation equation, Kd = a + b Pd , from the test data (lb-mol/ft-yr), b = the slope of the loss factor versus pressure difference correlation equation, from the test data (lb-mol/ft-yr in.wc), (T, °F) = the average temperature of the test liquid, from the test data (°F), P = the true vapor pressure of the test liquid (psia).

where (T, °R) is the temperature of the test liquid (°R); (T, °R) = (T, °F) + 459.67, Ap is a constant (dimensionless). Ap = 13.824 for n-hexane, Bp is a constant (°R). Bp = 6907.2 (°R) for n-hexane. Pa = the average atmospheric pressure, from the test data (psia), Pa for the reference data is taken as 14.4 psia, based on an elevation of 630 ft. for Plainfield, IL, P* is a vapor pressure function (dimensionless).

Mv = the molecular weight of the test liquid vapor, assumed to be 86.2 lb/lb-mole for technical grade n-hexane (lb/lb-mole). Loss Factor, Kd = the normalized loss rate on a yearly basis (lb-mol/ft-yr). Kd = a + b Pd Pd is the pressure difference between the pressure inside the test pan vapor space and the pressure inside the test enclosure (in.

wc), specified as: Pd = 0.050 in.wc

Xref

Sref2

P Ap

Bp

T R -----------------–exp=

PP Pa

1 1 P Pa – 0.5+

2-------------------------------------------------------=



= the mean of the loss factors for the calibration tests, Kdi .

𝑋∑ 𝐾

𝑛



𝑆𝑐𝑎𝑙2

∑ 𝐾𝑑𝑖 𝑋𝑐𝑎𝑙2𝑛𝑐𝑎𝑙

𝑖 1

𝑛𝑐𝑎𝑙 1




The calibration test is acceptable if .

Xcal

Scal2

Sp2

Sp2 nref 1– Sref

2ncal 1– Scal

2+

nref ncal 2–+----------------------------------------------------------------------=

tXref Xcal–

Sp1

nref--------

1ncal---------+

----------------------------------=



ANNEX E—RANGE OF VARIATION IN DEVICE DESCRIPTION

E.1 General

This annex provides procedures for evaluating variations in the description of devices proposed for protocol testing. Device descriptions shall be evaluated for two purposes. The first is to determine whether the device is substantially similar to any other devices for which a loss factor has previously been developed. The second purpose is to identify requirements for the construction of the test device. In each case, the evaluation shall assess the configuration or arrangement of parts, the tolerance on dimensions, and the substitutions of material that may be associated with the device.

E.2 Substantially Similar Devices

E.2.1 General

Sections 6 and 8 require the results of all valid protocol tests of a device to be included in a report of loss factor determination, and the device shall be identified in sufficient detail to distinguish it from similar devices. The testing laboratory is to record the existence of any other protocol tests that may be deemed to be of the same device.

The substantive distinguishing features shall be included in the physical description that is reported with the loss factor for the device. The determination of whether claimed differences are substantive will include the considerations discussed below.

E.2.2 Arrangement of Parts

A given device is typically comprised of certain components assembled in a particular arrangement. In order for a difference in the configuration or arrangement of the components to constitute a substantially different device, it would typically involve a change in one of the following:

a. The total length of seams or area of openings.

b. The presence or absence of gaskets.

c. The area of liquid surface tributary to openings or seams.

d. Whether openings or seams are in contact with the liquid surface.

Greater care in the fabrication or assembly of the device than that specified in the description of the device or controlled by the written installation procedure shall not constitute a substantive difference.

E.2.3 Dimensional Tolerances

A given device may consist of the same arrangement of parts as another, but differ in its dimensions. Substantive dimensional differences would typically involve one of the following:

a. The allowable size of gaps in seams or between parts.

b. The height of openings or seams above the liquid surface.

c. The diameter of sleeves or wells.

Differences in the thickness of material or in the size or spacing of fasteners shall not constitute a substantive difference unless shown to cause a change in the fit-up of the device.


E.2.4 Substitutions of Material

Substitutions of material would typically only constitute a substantive difference if the change involved a difference in the permeability of the component to hydrocarbon vapors.

E.3 Construction of the Test Device

Section 6.2 requires identification of the range of variation in the device description that is to be included in the report of the loss factor. This range includes acceptable tolerances on dimensions, alternate arrangements of the parts, and substitutions of material. The options expected to result in the greater rate of evaporative loss would typically be selected.


ANNEX F—STEADY STATE CONDITIONS

F.1 General

The test methods of API MPMS Chapter 19.3, Parts A through E, specify that the test data to be used in calculating the evaporative loss rate shall be measured during a period of a steady rate of weight loss. This requirement is in recognition of the potential for initial loss rates to be unstable for high loss-rate devices, in that rapid cooling of the test liquid surface during initial evaporation may cause a decay in the rate of evaporation. This annex provides a method of evaluating whether observed loss rates are steady for the weight loss test methods of API MPMS Chapter 19.3, Parts A and E. The term steady, in this context, is understood to mean a nearly-linear rate of weight loss over time.

F.2 Determination of Uncertainty in the Loss Rate

Section B.3.5 summarizes the determination of the loss rate, L, by a linear regression of the corrected weight loss, Wci , on time, tmi , resulting in a correlated, or fitted, weight loss, Wai .

Wai = L tmi

where:

Wai = correlated, or fitted, value of corrected weight loss at time tmi , also known as .

The coefficient of determination, r2, may be used as an indication of the proportion of variation in the data that is explained by the variable time, tmi .

F.3 Test for Stable Conditions

Stable conditions are determined to have been achieved when the loss rate is substantially linear. The stability of the loss rate may be estimated for a given period of time by calculating r2 on the basis of the readings obtained during that period. Stable conditions are adequately demonstrated when r2 determined from six consecutive readings is greater than 0.95, or some other level approved by API. The readings obtained during and subsequent to the period in which stability is demonstrated shall then constitute the test data for that test.

Wci


ANNEX G—DE MINIMIS LOSS FACTOR

G.1 General

The test methods of the API MPMS Chapter 19.3, Parts A through E, are not valid for measuring loss rates lower than the specified tolerance of the instruments or, for weight loss methods, the observed drift of the load cells or scales. Such a lower bound on the validity of the test method may be termed a minimum detect level of the test apparatus. The testing laboratory shall determine and record the minimum detect level for each test apparatus used at its facility.

G.2 Loss Rates below the Minimum Detect Level

When the measured loss rate for a device is below the minimum detect level for the test apparatus, document that the loss rate for the device is below the minimum detect level.

The device may be retested using a test apparatus capable of measuring lower rates of loss, if such equipment is available and certified for use. A lower minimum detect level may be achieved by using more sensitive instruments or equipment, instruments characterized by lower drift rates, or some other change in the test apparatus. If the results of the retest, using a more sensitive test apparatus, are again determined to be below the minimum detect level, then document the lower de minimis value. If the loss rates measured in the retest are above the minimum detect level of the more sensitive test apparatus, then these test results shall be used to develop a device-specific loss factor.

All of the test results for the same device shall be reported, as discussed in E.2. In the event that the measured loss rates are below the minimum detect level for one or more test apparatus, however, the report shall designate which test apparatus the loss factor is to be based on. Only the results of tests using that test apparatus will be included in the data base for calculating the loss factor.


ANNEX H—UNCERTAINTY IN THE LOSS FACTOR

H.1 General

This annex provides the procedure to be used in estimating the 95 % uncertainty parameters of the loss factor, as required by section 8. The uncertainty shall be calculated in relation to the reference value for the generic version of the device in question, by testing the null hypothesis that there is no difference between the reference value and the average of the test data, at a 0.05 level of significance. Reference values at a given wind speed shall be the average of the API test values used in the development of the generic loss factors, when available, rather than the predicted values from the loss factor equation.

The test for difference shall be applied at the nominal 0 mph wind speed level for every device, and additionally at the nominal 10 mph wind speed level for devices to be used with external floating roofs.

H.2 Calculation Procedure

This procedure is illustrated for deck fittings in Table H.1. The nominal 0 mph wind speed includes all test results at wind speeds less than 0.5 mph, and the nominal 10 mph wind speed to be used in the evaluation of deck fittings is 8.5 mph. The reference value to be used in Table H.1 is selected from Table H.2 for the appropriate deck-fitting description and wind speed. The results of all tests at the specified wind speed for the device in question are then filled in on Table H.1, and the statistical calculations performed. If the absolute value of the test statistic, t, is less than or equal to the student’s t-distribution, t0.975, n–1, then there is no difference at that wind speed between the reference value and the average value for the tested device, at a 0.05 level of significance.

Table H.1—Comparison to the Generic Loss Factor

Deck Fitting Description ________________________ Wind Speed = ____ mph

Reference Value

Loss Factor

Kf (lb-mole/yr)

Reference value from Table H.2 for this deck fitting description at this wind speed =

Test Data


L (lb/h)


°F)

P (psia)

Pa

(psia)

P* (dimensionle

ss)

Mv


Kf (lb-mole/yr)

1 Kf1

2 Kf2

3 Kf3

. . .

. . .

n Kfn


where (T, °R) is the temperature of the test liquid (°R); (T, °R) = (T, °F) + 459.67, Ap is a constant (dimensionless). Ap = 13.824 for n-hexane, Bp is a constant (°R). Bp = 6907.2 (°R) for n-hexane. Pa = the average atmospheric pressure, from the test data (psia). P* is a vapor pressure function (dimensionless).

P Ap

Bp

T R -----------------–exp=




= the arithmetic average (mean) of the loss factors, Kfi .

= the mean squared error.


where = the reference value. There is no difference at this wind speed between the reference value and the average value for the tested device,

at a 0.05 level of significance, if: .

Table H.2—Deck Fitting Reference Values

Deck Fitting Description Fitting No.

0 mph 8.5 mph

Test No. Ref. Value (lb-mole/yr)

Test No. Ref. Value (lb-mole/yr)

Access hatches

unbolted, ungasketed cover 12 Z16 35.62 44R 97.14

unbolted, gasketed cover 16 Z20 30.74 56R 84.60

bolted, gasketed cover – a 1.60 a 1.60

Fixed-roof support columns

pipe column; ungasketed sliding cover – a 31.00 n/a

pipe column; gasketed sliding cover – a 35.00 n/a

pipe column; flexible-fabric sleeve seal – a 10.00 n/a

built-up column; ungasketed sliding cover – a 47.00 n/a

built-up column; gasketed sliding cover – a 33.00 n/a

PP Pa

1 1 P Pa – 0.5+

2-------------------------------------------------------=


P Mv ---------------------------------------------------------------------------------=

X

X

Kfi

i 1=

n

n

----------------=

S2

S2

Kfi X– 2

i 1=

n

n 1–

--------------------------------=

S2

n

tX 0–

variance------------------------

X 0–

S2

n-----------------

X 0–

S n ------------------= = =

0

t t0.975 n 1–


Gauge floats (automatic gauge)

unbolted, ungasketed cover 11 Z15 14.09 43R 63.27

unbolted, gasketed cover 15 Z19 4.16 55R 41.38

bolted, gasketed cover – a 2.80 a 2.80

Gauge hatch / sample ports

weighted lid, ungasketed 9 Z13 2.29 41R 2.44

weighted lid, gasketed 13 Z17 0.46 53R 0.74

slit fabric seal (10% open area) – a 12.00 n/a

Vacuum breakers

weighted lid, ungasketed 10 Z14 7.61 42R 48.46

weighted lid, gasketed 14 Z18 6.04 54R 15.35

Deck drains

3-inch diameter, open 21 Z21 1.43 77 9.38

3-inch diameter (10% open area) 22 Z22 1.80 79 2.74

1-inch diameter, open – a 1.20 n/a

Deck legs—IFR type – a 7.90 n/a

Deck legs—EFR type, double deck & center area of pontoon deck

ungasketed, no sock 6 Z29R 0.81 88 1.20

gasketed, no sock – a 0.50 a 0.70

with sock, no gasket – a 0.50 a 0.70

Deck legs—EFR type, pontoon area of pontoon deck

ungasketed, no sock 5 Z25 1.97 65R & 65RR 4.20

gasketed, no sock 7 Z30 1.32 89 1.52

with sock, no gasket 8 Z31 1.16 90 1.73

Rim vents

weighted pallet, ungasketed – a 0.70 a 16.00

weighted pallet, gasketed – a 0.70 a 1.60

Vertical ladders

ungasketed sliding cover – a 76.00 n/a

gasketed sliding cover – a 56.00 n/a

Unslotted guidepole

ungasketed sliding cover; no wiper or sleeve 18 Z7 30.48 18 4153.62

gasketed sliding cover; no wiper or sleeve 27 Z11 24.32 31 1064.68

ungasketed sliding cover with sleeve; no wiper 28 Z10 24.36 32 157.68

gasketed sliding cover with sleeve; no wiper 19 Z6 8.44 19 80.36

gasketed sliding cover with wiper; no sleeve 17 Z5 13.39 17 39.13

Slotted guidepole

ungasketed sliding cover; no float, wiper, or sleeve 1 Z1 44.51 5b 6297.51

gasketed sliding cover; no float, wiper, or sleeve 25 Z12 39.77 29b 4699.93

ungasketed sliding cover; float—no wiper or sleeve 3 Z3 35.11 7b 4095.49

gasketed sliding cover; float—no wiper or sleeve 26 Z9 25.04 26b 2117.61

gasketed sliding cover; sleeve—no float or wiper 2 Z2 16.00 6b 1971.76

gasketed sliding cover; float & wiper—no sleeve 4 & 23 Z4, Z23 20.58 8b, 78b 490.75

gasketed sliding cover; float, sleeve, & wiper C01, 24, 29 CBI 4Z, Z24, Z32 10.79 CBI 3b 80b, 86b 76.74

aTest data not available—reference values taken from the API MPMS Chapter 19.2, 1st Edition. bTest no. represents a series of tests at multiple orientations.


ANNEX I—METRIC UNITS

I.1 General

To convert the inch pound units (U.S. customary units) employed in the text to equivalent SI units of the International System of Units, the guidelines of the API MPMS Chapter 15 shall be followed. The pertinent conversion factors are summarized in Table I.1 below.

The unit of length is either the kilometer, km, or the meter, m. The unit of mass is the kilogram, kg. The unit of time is either the hour, h, or the year, yr. The unit of temperature is the degree Celsius, °C, or the kelvin, K. The unit of electromotive force is the volt, v.

I.2 Velocity

The unit of velocity is the kilometer per hour, km/h.

I.3 Pressure

The unit of pressure is the kilopascal, kPa.

I.4 Loss Factors

The text employs the pound-mole per year, designated lb-mole/yr, as the unit of loss rate. The loss rate is determined as the product of the dimensionless coefficients, P* and Kc , times a loss factor (Kf or Kr ) as described in 3.2.2. The loss factor (Kf or Kr ), while not a loss rate, is also expressed as lb-mole/yr. The equivalent SI unit for the loss rate and the loss factor is the kilogram-mole per year, designated kmol/yr. As with inch pound units, the loss factor in kmol/yr have to be multiplied by the dimensionless coefficients P* and Kc to obtain a loss rate in kmol/yr.

Table I.1—Metric Conversion Table

Quantity To convert from: To: Multiply:

Length foot meter (m) (Length, ft) × 3.048 E – 01

Mile kilometer (km) (Length, mi) × 1.609 344 E + 00

Mass pound kilogram (kg) (Mass, lb) × 4.535 924 E – 01

Temperature degrees Fahrenheit degrees Celsius (°C) [(Temperature, °F)-32] × 5/9

degrees Rankine kelvin (K) (Temperature, °R) × 5/9

Velocity mile per hour kilometer per hour (km/h) (Velocity, mph) × 1.609 344 E + 00

Pressure lbf / in2 absolute kilopascal (kPa) (Pressure, psia) × 6.894 757 E + 00

Loss Factor pound-mole / year kilogram-mole / year (kmol/yr) (Loss Factor, lb-mole/yr) × 4.535 924 E – 01


BIBLIOGRAPHY

API

Manual of Petroleum Measurement Standards (MPMS), Chapter 19.1, Evaporative Loss from Fixed-Roof Tanks

Publ 2517 Evaporative Loss from External Floating-Roof Tanks

Publ 2517A Addendum to API Publication 2517—Evaporative Loss from External Floating-Roof Tanks

Chicago Bridge & Iron Technical Services Company, Loss Factor Measurements of Internal Floating Roof Rim Seals, Interim Report Number 2, Prepared for the American Petroleum Institute Committee on Evaporation Loss Measurement, February 1, 1995.

Chicago Bridge & Iron Company, Development of the Testing Protocol for the Measurement of Deck Seam Loss Factors Using the Fugitive Emission Test Method, Final Report, Prepared for the American Petroleum Institute Committee on Evaporation Loss Estimation, September 14, 1999.

Manual of Petroleum Measurement Standards Chapter 19.3 ...ballots.api.org/copm/cele/ballots/docs/API...

Documents

Transcript of Manual of Petroleum Measurement Standards Chapter 19.3 ...ballots.api.org/copm/cele/ballots/docs/API...