Estimation of EPC of French Housing Stock

8/18/2019 Estimation of EPC of French Housing Stock

1/17

1

Estimation of the Energy Performance Certificate of a housing stock

characterised via qualitative variables through a typology-based approach

model: a fuel poverty evaluation tool.Pietro FLORIO1*, Olivier TEISSIER2

1LESO, EPFL, Station 18, CH-1015 Lausanne, Switzerland

2DESH, CSTB, 14 Boulevard Isaac Newton, 77420 Champs-sur-Marne, France

*corresponding author: tel : +41 21 69 33363, e-mail : [email protected]

Abstract

The European Union stresses the accent on the need of energy consumption and expenditure data related

to housing stock (1). Most of housing and living conditions databases of Member States investigate the

housing status of a concerned region through a survey to be submitted to the sampled households. Theassessment of energy performance then requires a simplified energy performance certification method,

based on qualitative variables.

In this paper the French Enquête Nationale Logements (abbr. ENL) is considered. A conversion algorithm is

elaborated to refer each of the ENL housing units to a reference building and a reference HVAC system of

the European Typology Approach for Building Stock Energy Assessment database (TABULA - EPISCOPE) for

France. The ENL housing stock is better specified in its technical and energetic features through a

typological data crossing. As a result, an energy label and an energy performance index expressed in

[kWh,ep/(m2 year)] is issued for every single ENL row.

The calculation outcomes are assessed through a sensitivity analysis and compared to other national

statistics; finally the energy labels distribution is discussed. Many purposes of results exploitations are

cited, concerning in particular the fuel poverty evaluation and the energy expenditure per household

estimation.

Introduction and purposes

The elaboration of an efficient energy saving policy requires a more and more accurate estimation of the

energy consumptions and expenditure related to building stocks. The European Union, with the publication

of the EPBD Recast (1) stresses the accent on the need of an assessment procedure taking into account the

energetic cost of buildings. The aim is to set “minimum energy performance requirements for buildings orbuilding units with a view to achieving cost-optimal levels”.

The present calculation process produces an Energy Performance Certificate (EPC), instructed by the

international standard EN 15217 and transposed into the national legislations. The energy performance

indicators are issued as a result of a national evaluation method, with reference to a well-established

European set of standards concerning the building envelope and equipment. This method underpins a

complex range of technical variables to be collected by a professional or a qualified person. The amount of

data required is complex to collect at national stocks scale. Usually this burden exceeds the data collection

budget and purposes of a national statistic survey.

Most of housing and living conditions databases investigate the housing status of a concerned region

through a survey to be submitted to the sampled households. The structure of this survey includes then a

PRE-PRINT version © 2014 Elsevier B.V. All Rights ReservedEnergy and Buildings 89 (2015) 39–48 DOI: http://dx.doi.org/10.1016/j.enbuild.2014.12.024

http://dx.doi.org/10.1016/j.enbuild.2014.12.024http://dx.doi.org/10.1016/j.enbuild.2014.12.024http://dx.doi.org/10.1016/j.enbuild.2014.12.024http://dx.doi.org/10.1016/j.enbuild.2014.12.024


2/17

2

set of easy-to-answer questions, accessible and understandable to anyone, that does not need a

professional support to be filled. The outcome is a qualitative characterisation of the building stock that

makes it difficult to quantify the energy performance and compare it.

The present paper investigates an approach consisting in a simplified calculation procedure, which only

focuses on the most characterising features of the building and associate them to default energy needs, butthis would lack in accuracy. Otherwise an energy performance assessment could be committed but a

smaller sample is then imposed. This is the case of the United Kingdom: the English Housing Survey consists

of a first part including 13300 interviews per year and a second part involving a physical inspection of a

housing subsample (6200 households) (2). Also France has a national housing enquiry (Enquête Nationale

Logements) concerning almost 37000 households and a more detailed energy performance survey (Phébus)

based on 2500 rows (3).

The article 11 paragraph 7 of the EPBD Recast asserts that “certification for single-family houses may be

based on the assessment of another representative building of similar design and size with a similar actual

energy performance quality”. Within this spirit, the aim of this paper is to experiment a data crossingbetween a national large-sampled housing survey (in this case the French Enquête Nationale Logements,

abbr. ENL) and a residential typology database, the European Typology Approach for Building Stock Energy

Assessment (4) for France. A specific algorithm is intended to associate a reference building from TABULA

to each housing unit of the ENL in order to merge the gross qualitative characterisation provided by the

household interview with a consolidated set of technical and geometrical features. In this way the building

envelope and the HVAC system are defined in detail and in accordance with the European Standards.

Energy needs and requirements are then estimated and an energy performance label is issued for each row

of the whole ENL sample. The results are then compared to other national data distributions concerning

energy use. The adoption of a European validated calculation tool allows the use of this dataset for

international comparisons and analysis too. The conclusions of this paper present a possible further

exploitation of this methodology and the related results.

Presentation of the data sources: ENL and TABULA

This section presents the structure of the chosen data sources to clarify the purposes of this paper. A

detailed focus on the concerned variables is provided in the next section.

Enquête Nationale Logements, ENL 2006

The French Enquête nationale logements (ENL) purpose is to describe housing conditions of households and

their spending in housing at a national scale. It is one of the most important surveys of the national

statistical institute (INSEE), by its age (since 1955), its frequency (every 4/5 years) and the size of its sample(roughly 40 000 households). The survey gives lots of details both on housing units (size, number of rooms,

age, comfort, equipment, surroundings and urban position, rent, loan, charges, energy spending, etc.) and

on households’ characteristics (number and ages of people, education, professional activity and distance

from work, incomes and social benefits, etc.). In this condition, the thermal quality of the dwelling is not a

central subject, but a lot of information can be used to make estimations (see below).

Typology Approach for Building Stock Energy Assessment, TABULA 2011

Firstly the IEE project DATAMINE (2006-2008) collected 19000 Energy Performance Certificates across the

European partners and transferred them to a common database by use of commonly defined data fields.

For different age and size groups “average buildings” were defined which are representative for therespective sample subsets. During the following IEE project TABULA (2009-2012) residential building


3/17

3

typologies have been developed for 13 European countries. Each national typology consists of a

classification scheme grouping buildings according to their size, age and further parameters and a set of

exemplary buildings representing the building types. A selection of HVAC (heating, ventilation and air-

conditioning) systems is also provided to be associated together with the building types. Each “reference

building” is then characterised with typical energy consumption values and an estimation of the possible

energy savings through the implementation of different refurbishment measures.

The method is focused on the energy use for space heating and domestic hot water of residential buildings.

Cooling, air conditioning, lighting, electric appliances are until now not considered in the concept but can of

course be supplemented later. The results of this process have been published by the project partners in

national “Building Typology Brochures”, written in their respective languages and enclosed with statistical

data for buildings and supply systems.

Each building type in TABULA is identified through a code which resumes its main features, in particular:

the national and regional relevance (e.g.: according to climate zones), the housing size (single family house,

terraced house, multifamily house, apartment blocks), the building age (in 10 classes, from 1800 to 2000),the reference type (Real example building, real average building, theoretical statistical model) and the

increasing refurbishment package code (001 to 003). The reference HVAC system as well is defined by a

national code and by three components related to the concerned heating, DHW and ventilation system.

More details can be found in the TABULA final report (5): an outline of the code composition is presented in

Figure 1.

Figure 1: code structure of the TABULA reference building (left) and the TABULA reference HVAC system (right), resuming theirmain features.

Variables employed in the calculation process

The ENL 2006 survey provides a set of variables that can be exploited for an energetic characterisation of

the concerned buildings. As can be remarked, most of them state a broad qualitative description of the

building, with banded values. Three variables (SOURCE, SOURCEECS and MODE) were designed on purpose

to resume the output of several other variables. Even if these data are not detailed enough to attempt an

energy performance evaluation, they can be used to select a reference building from the TABULA database.

FRnational code

[FR; IT; ...]

H1climatic zone

[N; H1; H2; H3]

THhousing type

[SFH; TH; MFH; AB]

07building age

[01-10]

Genbuilding size

[SUH; MUH;Gen]

Re.Exreference type

[Re.Ex; Re.Av;Sy.Av]

001 refurbishmentcode [001-003]

FR national code[FR; IT; ...]

heatingsystem type

DHW systemtype

ventilationsystem type

Genbuilding size

[SUH; MUH;Gen]


4/17

4

In fact, the features listed in Table 1 are comparable with the ones outlined in Figure 1. A conversion

algorithm described in the following sections of this paper links each row of ENL to a reference building and

a reference HVAC system of TABULA. This is done by the use of the TABULA code structure, according to

the variables listed in Table 1.

Table 1: list of ENL variables providing a building characterisation, used to refer to TABULA database.

VARIABLE TYPE VALUES DESCRIPTION

GENERALS

IAAT Qual 1-10 classes of building construction achievement

HTL Qual 1-8 housing unit type (collective building, single family house…)

IMI Qual 1-3 individual housing type (detached, terraced, …)

INDCOLL Qual 1-2 independent room location (collective or individual building)

INE Nr 0-99 number of building stories

INL Nr 2-9338 number of building housing units

DEP Nr 01-97 local authority department number

GEOMETRY

HAUT Qual 1-3 classes of ceiling height in the housing unitHSH1 Nr 1-999 floor surface of the housing unit

ENVELOPE

GFACRBIS Qual 1-7 type of main façade coating materials

GTOIT2 Qual 1-5 roof thermal insulation

GVIT1 Qual 1-3 double-glazing windows

HEATINGKCLN qual 1-11 heating mode (district heating, collective heating, individual heating…)

KMOD1EN qual 1-4 renewable heating source

SOURCE built source1-4 heating source (gas/electricity/oil and others)

KEMEA_1 qual 1-2 presence of standing radiators / coils

KEMEA_2 qual 1-2 presence of mobile radiators / coils

KEMEA_4 qual 1-2 heating by floor / ceiling

KEMEA_5 qual 1-2 heating by radiative panels

MODE built 1-2 individual / collective heating

KCC1_4 qual 1-2 use of a coal as heating source

DHW

KAO_3 qual 1-2 presence of independent DHW devices

KCEC_N nr 1-10 number of DHW energy sources

SOURCEECS built 1-5 main DHW energy source

KCEC_8 qual 1-2 use of solar energy as DHW energy source

VENTILATIONGVMC qual 1-3 presence of a mechanical ventilation

ENL to TABULA algorithm description

First step: associating each housing unit of ENL with a reference building in TABULA

With reference to Figure 1, the TABULA code is assumed as a standard to identify every row of ENL. The

aim of the described algorithm is then to assign a TABULA code to each of the 36.955 ENL housing units and

pair them to a specific reference building equipped with different reference HVAC systems. This process

can be easily automated and it returned 1,1% of uncertainties at this stage.

As a first step, the number of the department which hosts the housing unit (variable DEP in ENL) is related

to a climatic zone according to a ministerial table (6). Three climatic zones are settled in France: H1, H2, and

H3 (colder to warmer).

The second point focuses on the building construction achievement date. It identifies the technologic

typology of the building, as well as the usually employed materials. This information is contained in variable


5/17

5

IAAT of ENL, grouped by classes. The ENL housing units, gathered in ten classes from “before 1871” to

“1999 and beyond”, are re-organised in ten classes as well from “1800-1900” to “1990-2000”, according to

the TABULA classification.

After that, the building size typology is defined by a complex set of conditions. First, it considers the

housing type mentioned in variable HTL of ENL and consequently it considers the sub-specificationprovided by variable IMI in case of individual housing and by variable INDCOLL in case of independent

rooms. So the individual or collective setting is better specified (detached or terraced houses, studio-

apartments or family flats, …). Then, the number of stories (INE) and of the housing units (INL) crossed with

the constructional period (IAAT) determine the final type assignment in TABULA code (“SFH” for single

family house, “MFH” for multifamily house, “TH” for terraced house, “AB” for apartment blocks). In this

way the building typology changes according to the achievement date and to the size: for example an

apartment block dating 1970s counts a larger number of housing units than a 1870s one.

Figure 2: French reference buildings table, according to the constructional period (rows) and to the building size (columns). (4).


6/17

6

At this point, each ENL row is represented by a reference building, characterised itself by a typical geometry

and a set of materials related to the constructional date. This base case might have been refurbished with

an insulation envelope or by a windows set replacement to save thermal energy. Variable GVIT1 of ENL

marks the presence of double glazing windows in most of the housing rooms, while GTOIT2 identifies a

qualitative roof-insulation degree of the housing unit (“recent insulation, within the last 10 years”; “aged

but adequate insulation”; “aged and poor insulation”; “no insulation”). The base case, marked as “001” by

the TABULA refurbishment code, is upgraded to “002” if double glazing windows are installed in the

apartment, and raised to “003” if both the double glazing windows installation and a satisfying roof-

insulation degree are attained (recent or aged but adequate insulation).

Second step: associating each HVAC system of ENL with a reference system in TABULA

Modelling an HVAC system composed by a heating, a Domestic Hot Water (DHW) and a ventilation

component allows an estimation of the primary energy requirements of the housing unit according to an

Energy Performance Certification method. Even if this represents just theoretical household energy

consumption and does not take into account the specific consumption of electrical appliances, an EPC

rating can be easily employed to compare the energy performance within different housing groups, at a

national or regional level. The most difficult task of the algorithm construction is the choice of an HVAC

system, from the list proposed by TABULA (Table 2), according to the few information provided by ENL.

Firstly, the heating system is selected through the analysis of different variables. The most important one

concerns the heating energy source: the dummy variables from KCC1_1 to KCC1_7 of ENL about each fuel

use were resumed in a unique variable named SOURCE. Three groups of heating energy sources were

identified: gas, electricity, oil or other fuels. After that, the type of heating equipment is considered: this

can be a district heating exchanger, a single or multi family unit boiler, an independent heater or a heat

pump. Variable KCLN of ENL provides the employed heating production mode; KMOD1EN identifies the

presence of a renewable heating system as a heat pump possibly coupled with solar captors. Finally, the

type of heating emission expressed in variables from KEMEA1 to KEMEA6 of ENL is taken into account. In

fact radiators, fan coils and radiative panels require different hot water loop temperatures.

At this point most of necessary information to identify a heating system is assumed. The installation date

can better specify the global efficiency of heat production (for example in presence of a condensing boiler,

strongly developed in the last decade). Unfortunately, ENL does not collect any data concerning the last

heating system set-up. However, for the purposes of this research, a modelling assumption can be made:

the heating system is intended to be installed when the building construction was achieved, and never

replaced. This statement is confirmed by a strong correlation between the building constructional date and

the heating consumptions, as showed by the last CEREN report (7).

A similar process is settled to select the DHW system from the TABULA proposition list. As a first step, the

system structure of each ENL housing unit is observed: if the DHW and the heating production are

combined, a unique boiler is normally installed. In that case the chosen DHW system is the same as the

heating one. Otherwise, the independent DHW appliances are selected as before, according to the fuel use

marked in variable SOURCEECS (which reassembles variables from KCEC_0 to KCEC_9 of ENL). Gas, electric

or oil independent water heaters are likely outputs, possibly coupled with solar captors. The installation

date is also considered with the assumption mentioned above.

Finally, a standard air change rate between 0,4 and 0,8 [vol/h] is assumed according to the building typeand the expected infiltrations (8). The installation of mechanical ventilation is identified by variable GVMC


7/17

7

of ENL, which specifies the occurrence of air handling devices in all rooms or just in humid rooms

(bathrooms and kitchens). If the presence of mechanical ventilation is marked, a standard ventilation

system that needs 3 [kWh,el/(m2 year)] is considered according to the TABULA assumptions (8).

Table 2: Heating and DHW systems. Energy expenditure by energy source and by system typology. (4).

ENERGY EXPENDITURE gas electricity oil other

(Energy demand / Heatproduction), [-]

singleunit

multiunit

singleunit

multiunit

singleunit

multiunit

single unitmultiunit

renewable

wood coaldistrictheatingHEATING geothermal air/water

independent coils1,06

stove3,76 3,76

heat pump0,29 0,45

boiler

period 1 2,12 2,23 1,45 2,12^ 2,62

period 2 1,65 1,87

period 3 2,12

period 4 1,83

condensingboiler 1,41 1,41

lowtemperatureloop 1,46 1,46

heat exchanger1,53

DHWheat pump0,86

water heater

period 1 2,07 2,121,59;

1,44* 1,57 3,27 3,44*

period 2 1,84 1,99 1,48 1,41

period 3 3,16 1,87

period 4 1,98 1,37condensingheater 1,51* 1,89

lowtemperature

loop 1,57* 1,96*heat exchanger

2,39

AUXILIARY SYSTEM

solar captors

* =implementedsystem * * * *

wood stove

^ =implementedsystem ^

Energy performance certification method

Usually an Energy Performance Certificate (EPC) issues an “energy label” from A to G, according to sevendifferent primary energy requirement classes, expressed in [kWh,ep/(m2 year)] bands. The aim of this


8/17

8

research is to rate each ENL housing unit from A to G following its primary energy needs for heating, DHW

and ventilation purposes.

The first step is to calculate the thermal energy requirements for space heating, including ventilation, of the

housing unit. This can be done by knowing the heating end-use energy needs (the energy to be delivered by

the heating system in the housing unit in order to reach a certain set-point temperature). Such a quantity isinfluenced by some technical, geometrical and climate-related variables.

In the previous section both a technical and a climate profile were defined for each of the ENL rows.

Construction materials, age, size typology of the building on one side and climate zone on the other side

were chosen and assembled in the TABULA reference models list to be referred in ENL. As a result, TABULA

already provides a heating end-use energy need value. Unfortunately, this takes into account the geometry

of the TABULA reference building. Therefore, an adjustment factor is needed to represent the ENL

geometry instead. This factor is intended as the ratio between the geometrical indicators of the ENL

housing unit and the TABULA reference building. In particular, such a geometrical indicator is the

“compactness ratio” between the outdoor-exposed surface and the heated volume of the building (alsoknown as “S/V ratio”). The S/V ratio can be easily calculated for each of the TABULA reference buildings,

while the absence of information concerning the outdoor-exposed surface in ENL forces the adoption of

another assumption. In place of the outdoor-exposed surface, the living area of the housing unit is retained

and corrected by the ratio between the outdoor-exposed surface and the living area of the TABULA

selected building for that ENL row. As a result, the simplified equation appears like this:

, = , ∙ ℎℎ �kWh, th

m2 ∙ year

assuming “Qh” as the heating end-use energy need and “h” as the housing unit height. In this way, if an ENL

housing unit is more compact than its TABULA reference (greater height with an equal volume): the heating

need “Qh” will be higher because the heated volume is bigger, but it will be adjusted by a lower

“compactness ratio”. This resulting thermal energy amount per surface unit is multiplied by an energy

expenditure coefficient that depends on the selected heating system (see Table 2).

Concerning the DHW energy requirements, 10 [kWh/(m2 year)] are considered as the standard need for a

single-family house (15 for a multi-family house to take into account a higher number of occupants per

square meter) (8); these needs are multiplied, as before, by an energy expenditure coefficient related to

the chosen DHW heating system. Finally, the selection of a particular ventilation system already provides an

electrical energy need per square meter (as stated before).

In order to merge all these energy amounts to a unique primary energy output, the heating, the DHW and

the ventilation needs must be multiplied each by a fuel conversion factor, according to the employed

energy source (2,58 for electricity and 1,00 for every other source in France) (9 p. 893). The resulting

primary energy requirements are classed according to a list of energy amount ranges, marked with a letter

(from “A” to “G”) (10).

Results

In order to be successfully extended to the whole French population, the results are weighted through the

QEX statistical factor of ENL before being aggregated. The first step of the algorithm produces the building

typology distribution presented in Table 3.


9/17

9

Table 3: building typology distribution of the modelled ENL (weighted sample), according to the constructional date and thebuilding size.

construction year class period SFH TH MFH AB sum

1 < 1915 6,00% 6,18% 4,24% 0,76% 17%

2 1915…1948 3,88% 5,01% 3,64% 0,81% 13%

3 1949…1967 3,83% 3,74% 6,28% 3,42% 17%

4 1968…1974 3,52% 2,33% 4,54% 2,91% 13%5 1975…1981 4,74% 2,48% 2,30% 2,35% 12%

6 1982…1989 4,44% 2,67% 1,99% 1,00% 10%

7 1990…1999 3,18% 2,03% 2,03% 1,29% 9%

8 2000…2005 3,70% 2,16% 1,65% 0,91% 8%

sum 33% 27% 27% 13% 100%

The second step of the calculation process concerning the HVAC system can be compared either to the

heating energy table of the CEREN statistics (7) or to the Energy Performance Certificate National database

“Observatoire DPE ” (11). As can be seen the modelled values are very similar to the CEREN ones for 2006.

The Energy Performance Certificate National database is not a statistical sample, but it just collects the

input data from the EPC redaction in the whole France – corresponding to transactions (rent or sale) –, so

the results must be weighted and treated carefully.Table 4: energy source usage quotas of the modelled ENL (weighted sample), of the CEREN statistics (7), and of the EnergyPerformance Certificate National database (11).

Energy source Modelled ENL CEREN, 2006 CEREN, 2013 Observatoire

DPE,12/12/2013

Gas 38,1% 40,2% 44,1% 35,0%

Electricity 29,4% 29,2% 31,5% 42,0%

Oil+Others 32,5% 30,5% 24,5% 23,0%

The energy labels output of the calculation process is shown in Figure 3 (1,1% is missing due to modelling

uncertainties). These data can be compared to other French national references (12), (11) to validate the

calculation model. The results are compliant with the ANAH report (12 p. 10) and close to the “Obsevatoire

DPE ” ones (see Table 5). In fact ANAH used a fairly similar method: it combined data from ENL 2002 and a

collection of 600 energy performance audits. Observatoire DPE is quite different, as explained above; it is

rather normal that some differences occur.

Figure 3: calculation results: DPE labels of the modelled ENL French housing stock (weighted sample; 1,1% of modelling errors).

2'589'238

3'574'064

6'156'735

8'232'976

4'295'402

741'588

373'726

9.9%

13.6%

23.5%

31.4%

16.4%

2.8%

1.4%

- 2'000'000 4'000'000 6'000'000 8'000'000 10'000'000

>450

>3312311519151


10/17

10

Table 5: energy labels comparison of the modelled ENL (weighted sample), of the ANAH report (12), and of the EnergyPerformance Certificate National database (11).

Energy label Modelled ENL ANAH, 2008 Observatoire DPE,12/12/2013

Modelling uncertainties 1,1% 0% 0,0%

G >450 9,9% 15%* 5,3%

F >3312311519151


11/17

11

Variables with classed values are assessed with increasing or decreasing class scenarios, with unchanged

values in highest or lowest class; quantitative continuous variables are assessed with 10% value increase or

decrease.

Results of the sensitivity analysis are shown in Figure 4. Boxplot figures reveal a sprawled distribution with

few values above 400 kWh/m2 year of primary energy requirements. This can be avoided by refining thestatistical data validation process, but it doesn’t affect the results and the conclusions since all non-

performing dwellings are together in “G” class. Most of the scenarios have the same interquartile range,

staying between the “D” and “E” label.

Building age has a moderate impact on the final consumption with index variations between 5% and 10%.

Compactness variations influence the S/V ratio, increasing the energy need by rising the outdoor exposed

surface. Among HVAC variations, heating system efficiency is the most impacting as expected, with 5% to

10% shift.

Nevertheless, the decisive variable is the refurbishment class, with a 20% to 30% impact on median. Theidentification of the refurbishment class, as described above, can not rely on dedicated variables in ENL.

Therefore, the accuracy of this operation is not guaranteed, but the comparison with other databases (see

Table 5) presents compliant results, showing a resilient algorythm adaptation. Further refining tools will be

provided by Phébus, but such a high sensitivity of the model to refurbishment variables suggests a growing

need of on-site surveys, able to track the fast evolution of performance of existing buildings in line with

more and more binding legislative requirements.


12/17

12

Figure 4: results of the sensitivity analysis. The upper part shows the boxplots of primary energy requirements according to thedifferent variation scenarios (DPE labels are marked on the right axis). The lower part shows the shift (%) of distribution indexes

such as 1st

quartile, minimum value, median, maximum value, 3rd

quartile, average value compared to the base case.

800

1300

1800

2300

B A S E C A S E

O L D E R B U I L D I N G

M O R E R E C E N T

B U I L D I N G

H I G H E

R P E R F O R M A N C E

L O W E

R P E R F O R M A N C E

1 0 % H I G H E R

C O M P A C T N E S S

1 0 % L O W E R


1 0 %

H I G H E R H . S . E F F .

1 0 %

L O W E R H . S . E F F .

1 0 % H

I G H E R D H W E

F F .

1 0 % L O W E R D H W E

F F .

A L L M . V E N T . O N

A L L M . V E N T . O F F

Q1 MIN

MED MAX

Q3 AV

0

50

100

150

200250

300

350

400

450

500

B A S E C A S E

O L D E R B U I L D I N G

M O R E R E C E N T

B U I L D I N G

H I G H E R P E R F O R M A N C E

L O W E R P E R F O R M A N C E

1 0 % H I G H E R


1 0 % L O W E R


1 0 % H I G H E R H . S . E F F .

1 0

% L O W E R H . S . E F F .

1 0 %

H I G H E R D H W E

F F .

1 0 %

L O W E R D H W E

F F .

A L L M . V E N T . O N

A L L M . V E N T . O F F

E P ( k W h / m 2 y

e a r )

ABC

D

E

F

G

BASECASE

OLDER

BUILDING

MORE

RECENTBUILDING

HIGHER

PERFORMANCE

LOWER

PERFORMANCE

10%HIG

HERCOMPACTNESS

10%LOW

ERCOMPACTNESS

10%HIG

HER H.S.EFF.

10%LOW

ER H.S.EFF.

10%HIG

HERDHWEFF.

10%LOW

ER DHWEFF.

ALL

M.VENT.ON

ALL

M.VENT.OFF

Q1 0.00% 8.27% -5.97% -20.85% 25.28% -7.29% 7.22% -7.29% 7.22% -2.31% 2.47% 2.32% -2.67%

MIN 0.00% 4.05% 0.00% 0.90% 3.01% -0.28% 0.28% -0.28% 0.28% -9.72% 9.72% 27.61% -14.19%

MED 0.00% 7.99% -5.72% -20.00% 30.41% -7.73% 7.70% -7.73% 7.70% -2.18% 2.06% 1.70% -1.84%

MAX 0.00% 0.00% 0.58% -30.83% 0.00% -9.55% 9.55% -9.55% 9.55% -0.45% 0.45% 0.58% 0.00%

Q3 0.00% 6.36% -7.79% -21.41% 18.08% -8.45% 8.44% -8.45% 8.44% -1.38% 1.51% 0.76% -1.70%

AV 0.00% 7.43% -4.48% -22.44% 19.79% -8.22% 8.22% -8.22% 8.22% -1.64% 1.64% 1.66% -1.43%

-40.00%

-30.00%

-20.00%

-10.00%

0.00%

10.00%

20.00%

30.00%

40.00%

s h i f t ( % )


13/17

13

Results exploitation and conclusion

One of the advantages of the ENL survey is the combination of information concerning both the housing

quality and the financial situation of households, referring to a large sample (37000 households vs. 2500 in

Phébus). The model elaborated in this paper can be used for many purposes to drive public policies:

•

this is a tool to elaborate and analyse several scenarios of “energetic transition” at differentgeographic scales. Used at national or regional level, it is a very detailed diagnostic of energy

consumption in the building sector. This is a first step to identify the main potentials for energy

savings and elaborate a strategy of refurbishment. Upgrading progressively the refurbishment code

(001 to 002 and 003) in certain types of buildings enables to derive scenarios of transition towards

a low emission and consumption sector in 2050. In comparison with existing models (ENERTER for

instance) (13), this method also gives economic information on the households. It means the

refurbishments choice can be analysed for each household, considering levels of investments,

savings and impacts on the household budget. Moreover, economic incentives (national or regional

subvention, tax credit, etc.) can be directly taken into account, in addition to regulatory measures;

• this is a powerful model to evaluate public policies impacts on households’ incomes, in particular

those related to energy. As ENL contains very precise information on incomes and energy

expenditures, this model can help to analyse redistributive effects on households. In particular, the

effects of incentives to energy efficiency (carbon tax and energy taxes as subventions and tax

credit) can be evaluated in a very precise way. Their impact on the refurbishment process and the

energy efficiency improvement of the building stock can be roughly estimated. One specific

application would be the elaboration of an incentive conditioned by the building energy label.

Coupling detailed analysis (for a lot of households’ types) on economic, housing and energetic

features is barely possible.

An analysis of “fuel poverty” demonstrates the exploiting opportunities offered by the model. Between 3

and 5 million households in France are considered as “fuel poor”, depending on the applied indicator; the

phenomenon ensues from low income, poor thermal quality of housing and high price of energy. The

evaluation of the fuel poverty condition needs precise information both on revenue and energy spending.

As it was mentioned in the introduction, the Phébus enquiry has been made available in 2014, and a first

exploitation has been made by the national statistical institute (INSEE) (14); the detailed results will provide

a crossed estimation of the energy performance of buildings and the living conditions and incomes of

families. An historical comparison of these data is then strongly needed in order to obtain fuel poverty

trends and to formulate suitable coping policies; moreover, up to now, the 2006 data are the most updated

ones and still represent a reliable reference.

The use of the energy labels distribution elaborated in this paper coupled with a fuel poverty indicators list

allows an assessment of the phenomenon and its correlation with the energy performance of housing. In

Figure 5 the estimated number of fuel poor households in France is displayed, according to different kinds

of indicators:

• the 10% threshold, as the number of units above the 10% ratio between the energy expenditure

and the household’s income;

• the Hills “LIHC” indicator (for low income high cost of energy), as the number of units beneath the

60% of the median income, at the same time affected by a higher-than-median energyexpenditure;


14/17

14

• a declarative indicator like the “feeling of cold”, as the number of people declaring having felt cold

at least 24h during the previous winter (because of different reasons, including poor quality of

housing and budget restriction).

Concerning the 10% threshold and the feeling of cold, both the entire population and the lowest 3 deciles

of income are considered. In this way the fuel poor households can be better identified. Furtherinformation concerning the fuel poverty indicators and the fuel poor profiles for France can be found in the

2014 ONPE annual report (15).

Figure 5: estimated number of fuel poor housing units in the modelled ENL according to different indicators.

These results are particularly important as they refer to a well described population, characterised by a

national enquiry both on the technical and on the social profile. The existence of housing refurbishment

works is tracked as well as the households living conditions (education, employment, urban mobility, social

aids, relationship with the territory, etc…). As a consequence, this approach allows a more completeawareness in the fuel poor identification process compared with other methodologies (e.g. “Précariter”

software by Energies Demain based on the census of the population 2008, which can just roughly estimate

the incomes and the energy expenditure) (13).

In conclusion, this is a simple but reliable method to estimate the energy performance at a large scale,

suitable for public policies analysis and evaluation. It can be set-up on a spreadsheet, the logical sequence

is pragmatic and operational. In spite of this, it must be refined to deal with building stock managing

purposes and more detailed technical issues.

Further considerations could be made by implementing the model with an estimation of the energetictheoretical expenditure per household. Some standard data concerning the specific electricity use of

appliances and equipment (16) would be added to the HVAC consumptions to better approximate the total

- 500'000 1'000'000 1'500'000

>450

>3312311519151


15/17

15

household energy needs. After that, a set of standard tariffs (17) would be applied to the energy

consumptions in order to model a theoretical expenditure, to be compared with the real expenditure

charged to the household. The determinants of the gaps should be analysed in order to reveal other

correlations (18). For example a lower real expenditure would indicate the presence of energy restrictions,

which is a form of fuel poverty particularly difficult to identify.

Acknowledgements

This paper is produced through the help of the entire Economics and Social Sciences Department (DESH) of

the Building Scientific and Technical Centre (CSTB). Academic support has been given by the Solar Energy

and Building Physics Laboratory of the Swiss Federal Institute of Technology (LESO-PB at EPFL).


16/17

16

References1. EP, European Parliament. DIRECTIVE 2010/31/EU OF THE EUROPEAN PARLIAMENT AND OF THE

COUNCIL on the energy performance of buildings (recast). Official Journal of the European Union. 18 06

2010.

2. DCLG, Department of Communities and Local Government. English Housing Survey. [Online] 2013.

[Cited: 09 10 2013.] https://www.gov.uk/government/collections/english-housing-survey.

3. MEDDE, Ministère de l'Ecologie, du Développement Durable et de l'Energie. Enquête Performance de

l’Habitat, Équipements, Besoins et Usages de l’énergie (Phébus). [Online] 2013b. [Cited: 09 10 2013.]

http://www.statistiques.developpement-durable.gouv.fr/sources-methodes/enquete-

nomenclature/1541/0/enquete-performance-lhabitat-equipements-besoins-usages.html.

4. TABULA, Typology Approach for Building Stock Energy Assessment. Welcome page. [Online] 2013b.

[Cited: 09 10 2013.] http://www.building-typology.eu/.

5. —. Main Results - Final project report. 2012.

6. MEDDE, Ministère de l'Ecologie, du Développement Durable et de l'Energie. La répartition des

départements par zone climatique. site web du Ministère de l'Ecologie, du Développement Durable et de

l'Energie. [Online] 11 10 2013c. [Cited: 19 12 2013.] http://www.developpement-durable.gouv.fr/La-

repartition-des-departements.html.

7. CEREN, Centre d'etudes et de recherches economiques sur l'energie. Données statistiques du CEREN.

CEREN website. [Online] 06 2013. [Cited: 10 12 2013.] http://www.ceren.fr/stats/stat1.aspx.

8. TABULA, Typology Approach for Building Stock Energy Assessment. Common Calculation Method.[Online] 01 2013a. [Cited: 12 12 2013.]

http://episcope.eu/fileadmin/tabula/public/docs/report/TABULA_CommonCalculationMethod.pdf.

9. MEDDE, Ministère de l'Ecologie, du Développement Durable et de l'Energie and METL, Ministère de

l'Egalité des Territoires et du Logement. Arrêté du 30 avril 2013 portant approbation de la méthode de

calcul Th-BCE 2012 [...]. Réglementation Thermique 2012. [Online] 2013. [Cited: 12 12 2013.]

http://www.rt-batiment.fr/batiments-neufs/reglementation-thermique-2012/textes-de-references.html.

10. RF, République Française. Arrêté du 15 septembre 2006 relatif au diagnostic de performance

énergétique pour les bâtiments existants proposés à la vente en France métropolitaine. Journal Officiel de

la République Française. 28 09 2006, 25.

11. ADEME, Agence de l'Environnement et de la Maitrise de l'Energie. Répartition des DPEs. Espace

Diagnostic de Performance Energétique. [Online] 08 2013. [Cited: 12 12 2013.] http://www.observatoire-

dpe.fr/index.php/graphique/.

12. ANAH, Agence Nationale de l'Habitat. Modélisation des performances énergétiques du parc des

logements - Etat énergétique du parc en 2008. [ed.] Julien Marchal. 01 2008.

13. Energies Demain. ENERTER@ modélisation énergétique territorale. PRECARITER@ is a declination of

ENERTER@ developed for ErDF (EDF Group) to tackle with fuel poverty and help to identify the mostsensitive areas. [Online] 06 2013. [Cited: 12 12 2013.] http://www.energies-

demain.com/spip.php?article13.


17/17

17

14. INSEE, Institut National de la Statistique et des Etudes Economiques. Le parc des logements en France

métropolitaine, en 2012 : plus de la moitié des résidences principales ontune étiquette énergie D ou E.

Chiffres et Statistiques. [Online] 07 2014. http://www.statistiques.developpement-

durable.gouv.fr/publications/p/2099/1041/parc-logements-france-metropolitaine-2012-plus-moitie.html.

15. ONPE, Observatoire Nationale de la Précarité Énergétique. Premier rapport de l'ONPE. Définitions,indicateurs, premiers résultats et recommandations. [Online] 09 2014.

http://onpe.org/sites/default/files/pdf/documents/rapports_onpe/onpe1errapportsynthese.pdf.

16. ODYSSEE. Consumptions per dwelling for electric appliances. Energy Efficiency Indicators in Europe.

[Online] 2011. [Cited: 12 12 2013.] http://www.odyssee-indicators.org/online-indicators/.

17. MEDDE, Ministère de l'Ecologie, du Développement Durable et de l'Energie. Base de données PEGASE

(Pétrole, Electricité, Gaz et Autres Statistiques de l'Energie). MEDDE website. [Online] 10 12 2013a. [Cited:

12 12 2013.] http://www.statistiques.developpement-durable.gouv.fr/donnees-ligne/r/pegase.html.

18. Allibe, Benoit. Modélisation des consommations d'énergie du secteur résidentiel français : amélioration

du réalisme comportemental et scénarios volontaristes. [ed.] Jean-Charles Hourcade. EHESS. Paris : Thèse

de doctorat en Économie de l'environnement, 2012.

19. TABULA, Typology Approach for Building Stock Energy Assessment. Welcome page. [Online] 2013.

[Cited: 09 10 2013.] http://www.building-typology.eu/.

20. INSEE, Institut National de la Statistique et des Etudes Economiques. Résultats du recensement de la

population 1999. INSEE website. [Online] 2000. [Cited: 12 12 2013.] http://www.recensement-

1999.insee.fr/default.asp?asp_action=produit&c_typeprod=TAB&c_prod=LOG1&c_theme=LOG&c_codgeo

=2&c_nivgeo=F.

Estimation of EPC of French Housing Stock

Documents

Transcript of Estimation of EPC of French Housing Stock