Do homes that are more energy efficient consume less energy? · Standard Assessment Procedure...

Do homes that are more energy efficient consume less energy?

EPRG Presentation

Scott Kelly

18th October 2010

OutlineMotivation

Context

Review of literature (modeling residential energy)

Data-sources and variables

Simple bivariate regression (the first case)

Overview of Structural Equation Modelling (SEM)

Methodology and application of SEM

Results

Policy implications

Motivation (PhD)Building an energy and CO2 simulation model of the UK residential sector.

Historical data, forward looking, holistic, bottom-up

Technology adoption by households (choice theory)

How this transforms to CO2 savings – how much?

Development of a deterministic engineering thermodynamic model

Assumption that more efficient homes use less energy.

Explain what the dominant factors are that drive residential energy consumption.

OutlineMotivation

Context






Results

Policy implications

Context

GHG emissions by source in the UK 2008

Total = 627 MtCO 2eq

(MtCO2eq)

GHG emissions by end-use in the UK 2008

Source: DECC (http://decc.gov.uk/en/content/cms/sta tistics/climate_change/data/data.aspx)

73% of UK GHG attributable to household consumption dem and.

(Druckman & Jackson 2009)

Context

Average energy consumption by end-use

Source: UK Domestic Energy Fact File (2006)

of total ENERGY

consumption is used forHEATING

40%

OutlineMotivation

Context






Results

Policy implications

Review of literature

DeCARB (BREDEM)

BREHOMES

UKDCM

CDEM

OutlineMotivation

Context






Results

Policy implications

Data from 1996 EHCS and FES‣ English House Condition Survey (EHCS) - 12,131 cases

‣ Fuel and Energy Survey (FES) - 2,531 cases

‣ EHCS contains information on physical properties.

‣ FES contains energy consumption characteristics + metered energy data!!!

‣ Economic status of occupants, demographics etc.

Explanatory variables identified from dataset‣ Number of HHLD occupants (cont.)

‣ HHLD income (cont.)

‣ Floor area (cont.)

‣ SAP (Standard Assessment Procedure) (cont.)

‣ Temperature difference (External - Internal) (cont.)

‣ Energy pattern (0-5 ) (categorical)

‣ Dwelling energy expenditure (and consumption) (cont.)

DATA

‣ Age of head of HHLD (cat.)

‣ Heating Degree Days (cont.)

‣ Urban dummy

‣ Owner dummy

‣ Economic status dummy

Standard Assessment Procedure

‣ SAP is the governments standard assessment procedure

for rating the energy performance of buildings.

‣ The adopted methodology in L1A and L1B (existing)

‣Measured on scale from 0 - 120

‣Multiple evolutions of procedure - 1996, 1998, 2001,

2005, 2008

Factors used to calculate SAP‣ Materials used for construction

‣ Thermal insulation of building fabric

‣ Ventilation characteristics of the dwelling and equipment

‣ Efficiency and control of the heating system

‣ Solar gain through windows and openings

‣ The type of fuel used to provide heating

‣ Any renewable energy technologies installed.

SAP

Descriptive StatisticsTable 1: Descriptive statistics for model variables

1. Std.Error and Std.Deviation calculations were c alculated from the re-calibrated effective sample s ize of (n c = 1025).

1.570.053.9561Age of head of household (categorical)

0.450.020.571.000.00Economic Status (dummy)

0.460.020.691.000.00Owner of house (dummy)

0.390.010.811.000.00Urban Recode (dummy)

2006.25208923671749Degree Days (categorical)

1.240.043.155.01.0Energy Pattern (categorical)

5.020.1612.127.1-20.0Temperature Difference (°C)

4.750.156.9039.1-9.2Outside temperature (°C)

2.790.0919.036.90.3Living room temperature (°C)

2848.87642333274.2Annual Energy Expenditure (£)

15.60.4944.41090SAP Rating

10,07231515,317103,8252,340HHLD Income (£)

31.20.9781.3252.420.0Floor Area (m2)

1.350.042.5110.001.00Number in household

Std. Deviation1Std. Error1MeanMaximumMinimum

OutlineMotivation

Context






Results

Policy implications

Bivariate regression

Table 2: bivariate regression results

OutlineMotivation

Context






Results

Policy implications

Structural Equation Modelling (SEM)

Also known as:

� Causal modelling, path analysis, simultaneous equation modelling, LISREL

(linear structural relations), Covariance Structure Analysis.

SEM is an umbrella for three processes

1. Path analysis (uses measured variables)

2. Confirmatory Factor Analysis (latent variables)

3. Structural Regression Models (combination of 1 & 2)

The SEM ProcessSEM is an extension of the GLM

SEM techniques require LARGE datasets

“Normal” methods:

‣ Start with data and then use statistics to explain data (mean, standard deviation, covariance)

SEM methods use confirmatory technique

‣ Build model on previous knowledge and theory

‣ Test theory with underlying data

Many statistical methods are shown to be special cases of SEM

‣ Correlation, T-Test, ANOVA

‣ Two way ANOVA, Regression, MANOVA, CFA etc.

Theory

Results & Fit Indices

Interpretation

Model Construction

Data Preparation

Model Estimation

SEM Vocabulary

Latent Variable (Y,X)‣A variable in the model that is not directly measur ed. i.e. intelligence, democracy

Exogenous Variable (X)‣Variable that is not caused by another variable

‣Usually causes one or more variables in the model

Endogenous Variable (Y)‣Variable that is caused by one ore more variables i n the model.

‣May cause another endogenous variable.

Disturbance (error, residual) (ε)‣Unspecified causes of endogenous variables.

Symbols used in SEM

Manifest variable (indicator variable)

Unobserved or latent variables

Presumed direct causal effect

Presumed non-recursive (reciprocal) causal effect

(A causes B but B also causes A)

Covariance / Correlation between pair of exogenous

variables

ε Measured error in observed variable

Typical SEM layout

YY

XX11

XX22

XX33

DD

STANDARD REGRESSION MODEL

OutlineMotivation

Context






Results

Policy implications

Data preparation

Outliers -> Type I and and Type II errors.

‣ Univariate outliers

‣Multivariate outliers - Cook’s distance & centred leverage

‣ HHLD Income, Floor Area, Energy Expenditure, truncated to 5 std. from mean.

Missing Data

‣ Problematic in SEM if not handled correctly (Lee, 2005)

‣ Less than 5% missingness.

‣MNAR, MAR & MCAR. (Rubin, 1976)

‣ Listwise deletion, pairwise deletion, mean substitution, regression based

imputation, pattern matching & expectation maximisation.

‣ Tested effect of missingness in data -> EM Method.

Data preparation

Grossing weights -> minority groups over represented (social)

Large groups under represented (owners)

Original data is bias towards under-represented groups

Assumption of normality

Normality transformations showed that original data was

robust to small deviations from normality.

(((( ))))2

1

2

1

1,025

n

ii

c n

ii

wn

w

====

====

= == == == =∑∑∑∑

∑∑∑∑. ( )

c

Ss e x

n==== (Dorofeev, 2006)

Model specification

1 1 1 1 1

2 2 4 2 2 21

3 3 6 3 8 32

4 4 12 4 4

5 5 5 9 10 11 5 3 7 5

0 0 0 0 0 0

0 0 0 0 0

0 0 0 0 0

0 0 0 0 0 0

0

y A y w

y A w y wx

y A w y wx

y A w y

y A w w w w y w w

εεεεε

= + + += + + += + + += + + +

Importing data into AMOS 18.0

Pearson Correlation

Energy

Expenditure

HHLD

Income

Floor

Area

Number of

Occupants

Temperature

Difference SAP

Energy

Pattern

Degree

Days

Energy Expenditure 1 0.375** 0.420** 0.452** 0.085** 0.031 0.188** 0.012

HHLD Income 0.375** 1 0.436** 0.475** 0.104** 0.110** 0.100** 0.013

Floor Area 0.420** 0.436** 1 0.352** 0.021 0.106** 0.109** -0.004

Number of Occupants 0.452** 0.475** 0.352** 1 0.053 0.104** .131** 0.004

Temperature Difference 0.085** 0.104** 0.021 0.053 1 0.034 .093** 0.123**

SAP 0.031 0.110** 0.106** .104** 0.034 1 .084** 0.012

Energy Pattern 0.188** 0.100** 0.109** 0.131** 0.093** 0.084** 1 0.04

Degree Days 0.012 0.013 -0.004 0.004 0.123** 0.012 0.04 1

**. Correlation i s s igni ficant at the 0.01 level (2-ta i led).

*. Correlation i s s igni ficant at the 0.05 level (2-ta i led).

Table 3: Correlation matrix for model variables

OutlineMotivation

Context






Results

Policy implications

ResultsTable 4: standardised direct effects

Model Results

Annual Energy Expenditure

0.31**

e4

Temperature Difference

0.19**

0.15**

0.27**

0.33**

0.09** 0.12**

0.05

0.09**

0.23**-0.22**

0.29

-0.050.23

0.02

0.11

e5

e1

0.27**

0.38**

0.19**

0.31**Number of Occupants

Floor Area

0.15**

0.38**

0.13**

e3e2

0.12**Household Income

Energy Pattern

0.23**-0.22**

SAP

0.13**

Explaining SAP

High propensity to

consume energySAP Rating

Low propensity to

consume energy

Measured energy

consumption

Floor Area

HHLD income

Number of occupants

Energy pattern

High energy

Low energy

Actual energy

consumption

Indirect & total effects

Table 5: standardised indirect effects

Table 6: standardised total effects

BootstrappingTable 7: Bootstrapping results

Model fit statisticsModel fit statistics in SEM are still widely debated

In SEM, the null-hypothesis (H0) is that the model is correct.

The alternative (Ha) is that it is not.

Therefore (and p-value) measures probability that model

fits perfectly to the population.

If P<0.05 we can’t reject null-hypothesis that the model is

correct and therefore have evidence the model may explain

reality.

2χ

Table 8: Model fit statistics

Results

£27.70Bedroom heated weekEnergy pattern

-£185.0030 -> 80 SAPSAP

Energy pattern

Temperature

Floor area

Number of occupants

HHLD income

Variable

£27.70Living room heated week

£2.50Each 1°C increase

£23.44Each extra 10m2

£88.32each extra person

£67.80Increase £10,000

Annual HHLD Energy

Expenditure

Effect

Table 9 : Total real effects on energy expenditure (1996)

Average income = £15,317 Average SAP rating = 44.4

Average occupancy = 2.51 Average annual energy expenditure = £642(1996) £1167(2009)

OutlineMotivation

Context






Results

Policy implications

Policy implicationsHomes with a propensity to consume more energy are shown to have relatively higher SAP rates.

� The scope for further savings from these homes may be limited.

� Homes with relatively high SAP ratings are subject to the law of diminishing returns.

Homes with a propensity to consume less energy, have lower SAP rates and therefore have greater potential to benefit from energy efficiency measures.

� These homes already consume relatively less energy.

� These homes are also more likely to be affected by the rebound effect.

This suggests an Energy Efficiency Barrier that must first be overcome.

This calls for more comprehensive and larger energy efficiency measures.

Different strategies for different energy consumers. Dual policy approach.

Summary

Serious lack of high quality, high resolution data for modellingresidential energy demand at the disaggregated level.

Limited number of disaggregated (bottom-up) statistical models of residential energy consumption.

Using SEM it is possible to isolate both direct and indirect effects to explain residential energy consumption.

At face value homes with high SAP rates do not automatically lead to low energy consumption.

If other factors are held constant (ceteris paribus) then increased energy efficiency does lead to lower energy consumption.

ThanksScott Kelly

[email protected]

References� Vernon, D T., 2007. Editorial Board/Publication Information. Personality and Individual Differences, 42(5), CO2. Available at:

http://www.sciencedirect.com/science/article/B6V9F-4N0YTYJ-1/2/633cc4f1810315c0c44b9aff88933071 [Accessed August 16, 2010].

� Aigner, D.J., Sorooshian, C. & Kerwin, P., 1984. Conditional Demand Analysis for Estimating Residential End-Use Load Profiles. The Energy Journal. Available at: http://econpapers.repec.org/article/aenjournl/1984v05-03-a06.htm [Accessed July 12, 2010].

� Aydinalp, M., Ugursal, V. & Fung, A., 2003. Modelling of residential energy consumption at the national level. International Journal of Energy Research, 27(4), 441-453. Available at: http://dx.doi.org/10.1002/er.887 [Accessed December 31, 2009].

� Aydinalp-Koksal, M. & Ugursal, V.I., 2008. Comparison of neural network, conditional demand analysis, and engineering approaches for modeling end-use energy consumption in the residential sector. Applied Energy, 85(4), 271-296. Available at: http://www.sciencedirect.com/science/article/B6V1T-4PPWMKT-1/2/00e584c85a9ea29aa33a1ffc7de0fd33 [Accessed May 13, 2010].

� Azadeh, A., Saberi, M. & Seraj, O., 2010. An integrated fuzzy regression algorithm for energy consumption estimation with non-stationary data: A case study of Iran. Energy, 35(6), 2351-2366. Available at: http://www.sciencedirect.com/science/article/B6V2S-4YRXKC2-1/2/6941e7f45306604323da8868deb7d7df [Accessed May 13, 2010].

� Baker, K.J. & Rylatt, R.M., 2008. Improving the prediction of UK domestic energy-demand using annual consumption-data. Applied Energy, 85(6), 475-482. Available at: http://www.sciencedirect.com/science/article/B6V1T-4R7D01S-1/2/84bb8c06ef691adfe81f5690b9d314a5 [Accessed July 14, 2010].

� Baker, P., Blundell, R. & Micklewright, J., 1989. Modelling Household Energy Expenditures Using Micro-Data. The Economic Journal, 99(397), 720-738. Available at: http://www.jstor.org/stable/2233767 [Accessed July 14, 2010].

� BERR, 2008. Renewable Energy Strategy, Available at: http://www.berr.gov.uk/energy/sources/renewables/strategy/page43356.html.

� Blunch, N.J., 2008. Introduction to Structural Equation Modelling Using SPSS and AMOS, London: SAGE.

� Böhringer, C. & Rutherford, T.F., 2009. Integrated assessment of energy policies: Decomposing top-down and bottom-up. Journal of Economic Dynamics and Control, 33(9), 1648-1661. Available at: http://www.sciencedirect.com/science/article/B6V85-4VWHVX8-2/2/6bb86cc14dee1381c0b012812016e89d[Accessed July 12, 2010].

� BRE, 2005. The Governments Standard Assessment Procedure for Energy Rating of Dwellings,

� Centre for alternative technologies, 2007. Zero Carbon Britain, Available at: http://www.cat.org.uk/news/news_release.tmpl?command=search&db=news.db&eqSKUdatarq=37990&home=0 .

� Communities and Local Government, 2006. Building a Greener Future: Towards Zero Carbon Development - Consultation - Planning, building and the environment - Communities and Local Government, Available at: http://www.communities.gov.uk/archived/publications/planningandbuilding/buildinggreener [Accessed April 1, 2009].

� Communities and Local Government, 2010. English Housing Survey (EHS) - Housing - Communities and Local Government. Available at: http://www.communities.gov.uk/housing/housingresearch/housingsurveys/englishhousingsurvey/ [Accessed August 23, 2010].

� DECC, 2009. Heat and Energy Saving Strategy Consultation, Available at: http://hes.decc.gov.uk/ [Accessed March 26, 2009].

� DETR, 1996. English House Condition Survey: User Guide

� Dorofeev, S., 2006. Statistics for Real-Life Sample Surveys: Non-Simple-Random Samples and Weighted Data, Cambridge: Cambridge University Press.

� ECI, 2005. 40% house, Environmental Change Institute. Available at: http://www.eci.ox.ac.uk/research/energy/downloads/40house/40house.pdf [Accessed May 17, 2009].

� Elliott, A., 2007. Statistical analysis quick reference guidebook : with SPSS examples, Thousand Oaks Calif.: Sage Publications.

References� Energy Efficiency Partnership, 2008. An assesment of the size of the UK household energy efficiency market, Element Energy Ltd. Available at:

http://www.eeph.org.uk/uploads/documents/partnership/Assessment%20of%20the%20UK%20household%20energy%20efficiency%20market_171108.pdf [Accessed May 6, 2009].

� Fiebig, D.G., Bartels, R. & Aigner, D.J., 1991. A random coefficient approach to the estimation of residential end-use load profiles. Journal of Econometrics, 50(3), 297-327. Available at: http://www.sciencedirect.com/science/article/B6VC0-45828KT-15/2/468b7905ac68f93a2c0bb72622daa437 [Accessed July 12, 2010].

� Great Britain., 2007. Building a greener future : policy statement., London: Dept. for Communities and Local Government.

� Hitchcock, G., 1993. An integrated framework for energy use and behaviour in the domestic sector. Energy and Buildings, 20(2), 151-157. Available at: http://www.sciencedirect.com/science/article/B6V2V-47XF6WC-49/2/bdd2bf07c9bd94b31c8af1b85a31d453 [Accessed May 13, 2010].

� Hoogwijk, M., van Vuuren, D.P. & Scrieciu, S., 2008. Secotoral emission mitigation potentials: Comparing bottom up and top down approaches, Available at: http://www.env.go.jp/press/file_view.php?serial=12478&hou_id=10316 [Accessed May 15, 2009].

� Jebaraj, S. & Iniyan, S., 2006. A review of energy models. Renewable and Sustainable Energy Reviews, 10(4), 281-311. Available at: http://www.sciencedirect.com/science/article/B6VMY-4DS7NYR-1/2/9eaf57b390ee97d59e4b83255470dd8b [Accessed May 13, 2010].

� Johnston, D., 2003. A physically Based Energy and Carbon dioxide emission model of the UK Housing Stock. Available at: http://www.leedsmet.ac.uk/as/cebe/assets/djthesis.pdf [Accessed May 15, 2009].

� Kaplan, D., 2009. Structural Equation Modeling: Foundations and Extensions 2nd ed., London: SAGE.

� Kavgic, M. et al., 2010. A review of bottom-up building stock models for energy consumption in the residential sector. Building and Environment, 45(7), 1683-1697. Available at: http://www.sciencedirect.com/science/article/B6V23-4Y889R7-1/2/58548662baa14e28ab75d32ad0b8ee22 [Accessed July 9, 2010].

� Keith, T., 2006. Multiple Regression and Beyond, Boston, Mass: Pearson Education.

� Kline, R.B., 2005. Principles and Practice of Structural Equation Modeling 2nd ed., New York, NY: Guilford Press.

� Larsen, B.M. & Nesbakken, R., 2003. How to quantify household electricity end-use consumption. The Energy Journal. Available at: http://econpapers.repec.org/scripts/a/abstract.plex?h=repec:ssb:dispap:346 [Accessed July 12, 2010].

� Lee, S., 2007. Structural Equation Modeling: A Bayesian Approach, Chichester: Wiley.

� Levine, M. & Urge-Vorsatz, D., 2008. Chapter 6 - Resdiential and Commercial Buildings. In IPCC AR4 WGIII. Available at: http://www.ipcc.ch/pdf/assessment-report/ar4/wg3/ar4-wg3-chapter6.pdf [Accessed April 20, 2009].

� McFarland, J.R., Reilly, J.M. & Herzog, H.J., 2004. Representing energy technologies in top-down economic models using bottom-up information. Energy Economics, 26(4), 685-707. Available at: http://www.sciencedirect.com/science/article/B6V7G-4CPM1NF-1/2/78915c04f7b0f9800c127ebd28e7ff5e [Accessed May 17, 2009].

� McKinsey, 2009. Pathways to a Low-Carbon Economy, Available at: file:///D:/Scotts%20Stuff/PhD%20Material/Litterature/Reports/MCkinsey/PathwayToLowCarbonEconomy_FullReportA.pdf [Accessed May 5, 2009]

� Mckinsey, 2008. Sustainable Urban Infrastructure Study - London Edition a view to 2025, London. Available at: http://w1.siemens.com/entry/cc/features/sustainablecities/all/pdf/SustainableUrbanInfrastructure-StudyLondon.pdf.

References� Parti, M. & Parti, C., 1980. The Total and Appliance-Specific Conditional Demand for Electricity in the Household Sector. The Bell Journal of Economics,

11(1), 309-321. Available at: http://www.jstor.org/stable/3003415.

� Perron, D. & Lafrance, G., 1994. Evolution of Residential Electricity Demand by End-Use in Quebec 1979-1989: A Conditional Demand Analysis. Energy Studies Review, 6(2). Available at: http://digitalcommons.mcmaster.ca/esr/vol6/iss2/4 [Accessed July 12, 2010].

� Rubin, D.B., 1976. Inference and Missing Data. Biometrika, 63(3), 581-592. Available at: http://www.jstor.org/stable/2335739 [Accessed July 21, 2010].

� Rylatt, R., Gadsden, S. & Lomas, K., 2003. Methods of predicting urban domestic energy demand with reduced datasets: a review and a new GIS-based approach. Building Services Engineering Research & Technology, 24(2), 93-102. Available at: http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=9925511&site=ehost-live&scope=site [Accessed March 30, 2010].

� Schumacker, R.E. & Marcoulides, G. eds., 1996. Full information estimation in the presence of incomplete data. In Advanced structural equation modeling: issues and techniques. L. Erlbaum Associates.

� Shorrock, L., 2003. A detailed analysis of the historical role of energy efficiency in reducing carbon emissions from the UK housing stock. ECEE. Available at: http://www.bre.co.uk/filelibrary/rpts/eng_fact_file/Shorrock.pdf [Accessed May 21, 2009].

� Strachan, N. & Kannan, R., 2008. Hybrid modelling of long-term carbon reduction scenarios for the UK. Energy Economics, 30(6), 2947-2963. Available at: http://www.sciencedirect.com/science/article/B6V7G-4SDPX7W-1/2/b2ca447237df5e8683b6a726320f820f [Accessed July 14, 2009].

� Summerfield, A.J., Lowe, R.J. & Oreszczyn, T., 2010. Two models for benchmarking UK - domestic delivered energy. Building Research & Information, 38(1), 12. Available at: http://www.informaworld.com/10.1080/09613210903399025 [Accessed June 9, 2010].

� Swan, L.G. & Ugursal, V.I., 2008. Modeling of end-use energy consumption in the residential sector: A review of modeling techniques. Renewable and Sustainable Energy Reviews, In Press, Corrected Proof. Available at: http://www.sciencedirect.com/science/article/B6VMY-4VBC3X3-2/2/338f45a5b335f8a8ee44523322033fd1 [Accessed April 7, 2009].

� Tabachnick, B.G., 2007. Using Multivariate Statistics 5th ed., Boston [Mass.]: Pearson/A&B.

� Tso, G.K. & Yau, K.K., 2007. Predicting electricity energy consumption: A comparison of regression analysis, decision tree and neural networks. Energy, 32(9), 1761-1768. Available at: http://www.sciencedirect.com/science/article/B6V2S-4MR1P3S-1/2/541d5b6c12ec4c9bb28cf2da82fc53b9 [Accessed June 21, 2010].

� Tuladhar, S.D. et al., 2009. A top-down bottom-up modeling approach to climate change policy analysis. Energy Economics, 31(Supplement 2), S223-S234. Available at: http://www.sciencedirect.com/science/article/B6V7G-4WVT0KJ-1/2/e678cbcc8418613e6e47a0ab5604af5e [Accessed July 12, 2010].

� UK Government, 2008. Climate Change Act 2008, Available at: http://www.opsi.gov.uk/acts/acts2008/ukpga_20080027_en_1 [Accessed May 12, 2009].

� World Business Council for Sustainabe Development, 2009. Energy Efficiency in Buildings - Transforming the market, Available at: http://62.50.73.69/transformingthemarket.pdf [Accessed April 27, 2009].

� WWF, 2007. WWF Building Stock Report, Available at: http://assets.wwf.org.uk/downloads/how_low_report.pdf [Accessed April 14, 2009].

Key Facts

Average HHLD energy demand is 22 MWh / year

Every 1 degree increase in heating season temp. leads

to a reduction of 1MWh / year

Energy price elasticity is measured at -0.2 this means a

50% increase in energy prices leads to 10% reduction

in energy demand.

(A. J. Summefield et al, 2010)

Non-recursivityStationarity assumption

‣ Requires the causal structure of the model not to c hange substantially over time.

‣ e.g. large houses will consume more energy.

Equilibrium assumption

‣ Any changes underlying the feedback relationship ha ve already manifested and come to equilibrium.

‣ e.g. high income HHLD’s effect on energy.

Do homes that are more energy efficient consume less energy? · Standard Assessment Procedure...

Documents

Transcript of Do homes that are more energy efficient consume less energy? · Standard Assessment Procedure...