Domestic Energy Performance Certificates (EPCs) are used in the UK to provide energy efficiency ratings for use in policy and investment decisions on individual dwellings and at a stock level. There is evidence that the process of creating an EPC introduces measurement error such that repeat assessments of the same property give different ratings, compromising their reliability. This study presents a novel error analysis to estimate the size of this effect, using repeated EPC assessments of 1.6 million existing dwellings in England and Wales. A statistical model of how measurement error contributes to variation between repeated measurements is set out, and exploratory data analysis is used to decide how to apply this model to the available data. The results predict that the one standard deviation measurement error decreases with EPC rating, from around ± 8.0 EPC points on a rating of 35 to ±2.4 on a rating of 85. This predicted error is higher than the limit recommended in UK guidance except in very efficient buildings; it can also result in dwellings being rated in the wrong EPC band, for example it was estimated that 24% of band D homes are rated as band C.

Energy ratings and national housing energy models are useful for energy policy evaluation and development, but limited empirical validation of energy demand estimates is available across residential sub-sectors. This study used data from a sample of over 2.5 million gas-heated dwellings in England from the National Energy Efficiency Data-Framework (NEED) to compare with estimates of 2012 gas consumption from the Cambridge Housing Model (CHM), a national energy stock model. The analysis quantified differences by dwelling type, size, and age band. It also compared variations in gas consumption from NEED dwellings with that expected from Energy Performance Certificate (EPC) bands. The findings show that the CHM overestimates average gas consumption from NEED for all dwelling types built before 1930, most notably for large detached dwellings. For other dwellings built since 1930, the model estimates were in relatively close agreement with NEED data. Furthermore, a simple comparison between estimated gas consumption and NEED data suggests savings from upgrading dwellings to at least EPC band C would be substantially lower than expected. Findings raise question regarding assumptions used in models and EPC ratings, including occupancy and space heating patterns, and have implications for development of energy models and policy regarding energy efficiency programmes.

The sizing of district energy systems involves a trade-off between reliability and continuity of service, and avoidance of capital and running costs associated with oversizing. Finding the most appropriate sizing requires a thorough understanding of energy demand. However, empirical data necessary to support such an understanding is scarce, and district energy systems are typically oversized. This study uses smart meter data from the largest field trial to analyse residential energy consumption in the UK. It presents graphically the seasonal and daily variations in energy consumption patterns, the weather dependence of energy loads, and peak hourly demand during particularly cold weather conditions. It also explores the diversity effect in residential energy consumption and computes the after diversity maximum demand at different levels of aggregations. Results show that, peak hourly gas consumption was nearly seven times higher than electricity consumption during the cold spells, while diversity reduced gas and electricity maximum demand per dwelling up to 33% and 47%, respectively. This empirical quantitative analysis of energy demand and diversity can support improved design and operation of district energy, and in particular, enable reduced capital and running costs, and an improved understanding of economies of scale for district heating networks.