Introducing the Default Effect: Reducing the Gap Between Theoretical Prediction and Actual Energy Consumed by dwellings through characterising data more representative of national dwelling stocks

Ciara Ahern, Technological University Dublin

Document Type Theses, Ph.D

Successfully submitted to the Technological University Dublin in fulfilment of the requirements for the reward of a Doctorate in Engineering.


Dwelling stock models that include the renovation status of the dwelling stock enable energy analyses of the stock. Using Ireland’s predominant single-family housing typology as a case study dwelling, the overarching objective of this research is to define a transparent generalisable methodology to create a stock model from a large empirical Energy Performance Certification (EPC) database, employing reference dwellings (RDs) defined using a ‘bottom-up’ approach. RDs are to be reported in compliance with Regulation No. (EU) 244/2012. The generalisable methodology defined allows for the development of stock models from EPC datasets.

Where obtaining the required data would be prohibitively costly, nationally applicable default-values are used in EPC assessments. To ensure that dwellings are not assigned wrongly-higher energy ratings, worst-case base-thermal-default U-values are used which, in the absence of empirical data, are determined by building type and prevailing building codes at time of construction date. A structural ‘default effect’ error in the EPC data was identified.

It was found that 58 % of walls and 67 % of roofs had significant retrofits of insulation, leading to; (i) less association between a dwelling’s age and its energy efficiency, and (ii) currently used default U-values being outmoded.

Outmoded default U-values; (i) decrease the credibility of both the EPC and its associated advisory report so inhibiting investment in energy efficiency, and (ii) lead to overestimation of potential residential energy savings, estimated at 22 % in post-thermal regulation dwellings and 70 % in pre-thermal regulation dwellings. To address this, generalisable methodologies have been developed to derive from an EPC dataset;

i) statistically relevant contemporaneous default U-values,

ii) a dwelling stock model derived from largely-default-free synthetically averaged RDs,

iii) the renovation status of a dwelling stock,

iv) a stochastically based energy-improvement payback calculation methodology, to be used where pessimistic default U-values are necessarily employed.

Use of the statistically-derived empirical data created will increase the validity and hence credibility of residential stock energy consumption models, the EPC, and its associated advisory report; enabling more valid quantification of the;

a) energy saving potential within Ireland’s predominant housing typology,

b) effect of default U-value use on the prebound effect, and

c) overall national building energy consumption.