As part of the “National Energy Projections 2025”, SEAI's NEMF (National Energy Modelling Framework) team needed to update the archetypes of residential buildings in Ireland. The aim was to identify typical buildings to divide the 2 million existing dwellings into approximately one hundred categories, in order to carry out the long-term decarbonisation forecasting.
The main constraints of the project were that it had to be carried out in six weeks, using Python. In fact, this language enabled our client to automate the processing of a large quantity of information from the Building Energy Rating (BER) and the Central Statistics Office (CSO) Census databases. Data analysis was then used to consolidate and simplify existing databases, grouping them according to their current and future characteristics. The final objective was to have a set of building types described in a maximum of 100 archetypes, and whose modelled energy consumption corresponded to Ireland's national energy balance.
Enerdata was able to respond to the constraints of the project by setting up a reactive organisation for rapid implementation, combining three key skills: knowledge of the decarbonisation outlook for the residential building sector, data analysis and indicator production capabilities to identify the archetypes, and, finally, the data science skills, which enabled us to deliver an automatic Python script so that the client could easily update their database.
Enerdata, therefore, developed a script for collecting and processing raw data from the BER and CSO Census databases. We then proceeded to analyse approximately two million dwellings containing a wealth of information, such as dwelling type (e.g. terraced housing), building age, floor area, energy label, and primary and secondary heating systems. We then grouped the 2 million buildings into 82 statistically significant categories to explain both the existing housing stock and the challenges of a future stock with improved insulation and greater use of renewable energy. Finally, we implemented two calculation phases to correct the under- and over-consumption bias observed between the theoretical energy rating and actual consumption, and to calibrate the database to the national energy balance. To ensure that the project remained on schedule, we held weekly workshops with the SEAI team throughout the project to respond quickly to any issues encountered and to formalise decisions.
At the end of the project, Enerdata delivered a detailed report on the methodology used, including an analysis of the 82 residential building archetypes and a Python script containing all the calculations.