Jonathan Chambers PhD Candidate UCL Energy Institute Centre of Energy Epidemiology E nergy in homes climate and health UK homes account for about 25 of energy use and CO2 emissions Key sector to address to meet climate targets ID: 537637
Download Presentation The PPT/PDF document "Thermal characterisation of UK dwellings..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Thermal characterisation of UK dwellings at scale using smart meter data
Jonathan ChambersPhD CandidateUCL Energy Institute, Centre of Energy EpidemiologySlide2
Energy in homes: climate and health
UK homes account for about 25% of energy use and CO2 emissionsKey sector to address to meet climate targets:
energy efficiency is the ‘first fuel’ of the transition to a low carbon economyNot only about climate5million people living in fuel poverty in UK9000 excess winter deaths attributed to cold homes in 2014Slide3
Measuring thermal performance
Track:Building stock state and evolutionI
mpact of renovationsRegulatory compliance CertificationPerformance based contractingMost common approach: EPC/SAPSignificant divergence between predicted and measured demand: ‘Credibility Gap’
Expensive to scale, slow,
intrusiveSlide4
How can thermal characterisation of UK dwellings be performed rapidly and non intrusively at scale?
Q1.1 What parameters are needed for the thermal characterization?Q1.2 What data sources can be used to perform this assessment?
Q1.3 How can the parameters be inferred from data?Q2.1 How precise, accurate and reliable is the method?Q2.2 What is the relation between the characterization parameters determined by the new method and those determined by intrusive monitoring methods?Q2.3 To what extent do as-built thermal properties correspond to EPC labels/SAP assessments? What are the implications?Slide5
Data-driven grey-box characterisation
Availability of new consumption data enables scalable thermal characterisation of individual homesSmart Meters measure total gas and electricity consumption
hourlyPhysically based grey box modelResponse of energy demand to weatherDefine energy balance equations based on known building physics and fit to dataSlide6
Simplified Physical Demand Model (SPDM)
Describe building thermal processes using established approximations from literatureConduction
Radiative gainsWind-driven losses Baseload gainsLink functions Relate physical processes to energy demand
Make assumptions explicit
Additional data to refine resultsSlide7
DECONSTRUCT algorithm
DECONSTRUCT is an approach for estimating SPDM parametersPrincipal: select subsets from metered data which best represents each physical processes in SPDM
As opposed to estimating them all at onceConceptually less robustPotentially biases certain parametersSensitive to outliersMany degrees of freedom means you might find local instead of globally optimal solutionSlide8
Data pipeline
Smart
meter data from a variety of existing sets Gridded weather data from MetOffice and CFSRGeospatial data for weather and metadata lookup for locations
Extensive cleaning, error checking, format
conversionSlide9
Example site
– daily winter demand and low sun sample
ResultsSlide10
Bulk estimates
Sensible results for heat loss ratesGeneral shortage of base data to perform comparison againstInitial comparison using internal temperaturesImprovements to algorithm will reduce this
Next stepsExtend weather parametersidentify predictors of error
Histogram of difference between estimated and measured average internal temperatures for 25 sitesSlide11
Publications
[1] J. Chambers, T. Oreszczyn, and D. Shipworth, “Quantifying Uncertainty In Grey-box Building Models Arising From Smart,”
Build. Simul. Conf., vol. 0, no. 1, pp. 2947–2954, 2015.[2] J. Chambers, V. Gori, P. Biddulph, I. Hamilton, T. Oreszczyn, and C. Elwell, “How solid is our knowledge of solid walls? - Comparing energy savings through three different methods,” in CISBAT 2015 - International Conference Future Buildings & Districts Sustainability from Nano to Urban Scale, 2015.
[3] J. Chambers and A. Stone, “EPC toolkit.” RCUK-CEE, 2016.
https://github.com/RCUK-CEE/epctk
(software)
[4] J. D. Chambers and D.
Shipworth
, “Energy ergonomics as a framework for the analysis and understanding of smart device data : key definitions and experimental approach in dwelling heating demand,” in
BEHAVE 2016
, 2016.
(presentation, forthcoming
)Slide12
Conclusion and Future work
Aim to estimate the insulation levels of any home equipped with a smart meterUse grey box modelling to extract physical properties from noisy dataBuilt data pipeline to ingest diverse energy data sources and match with local weather
Promising initial resultsNext steps – measuring errors, cross validation, determining predictors of result qualitySlide13
Acknowledgements
Supervisors Prof Tadj Oreszczyn
Dr David ShipworthEDF Energy R&DFunding from EDF R&D industrial case studentship and the EPSRC