SASBE 360: Climate Change–Induced Overheating Risk in Irish Dwellings: A Machine Learning–Based Asse

Climate Change–Induced Overheating Risk in Irish Dwellings: A Machine Learning–Based Assessment

Author: Omprakash Ramalingam Rethnam (1,3), Ciara Ahern (2,3), Mark Mulville (1)
(1) Built Environment Research and Innovation Centre, Technological University Dublin, Ireland;
(2) MaREI Research Ireland Centre for Marine Energy and Climate, Technological University Dublin, Ireland; 3Irish Building Stock Observatory, Dublin, Ireland

Abstract:
Most overheating research has focused on extreme heat events in tropical climates, while dwellings in temperate regions have traditionally been optimised to reduce heating energy demand. However, climate change is expected to increase summertime temperatures in cooler climates, introducing significant overheating risks to housing stocks that are largely unprepared for such conditions. This study investigates climate change–induced summertime overheating risk in Irish dwellings using a mid-floor apartment typology. A large parametric dataset was generated using dynamic building energy simulations under short-term (2021–2050), mid-term (2041–2070), and long-term (2071–2100) future climate scenarios for Dublin, consistent with Met Éireann projection conventions. An ensemble-based machine learning surrogate modelling approach was employed to quantify overheating risk and identify dominant contributory building features.
The results reveal a clear migration of dwellings across overheating risk categories over time. Under short-term future conditions, most dwellings exhibit low to moderate overheating risk, whereas under mid-term conditions, a substantial 10% of configurations transition into higher-risk categories. Feature importance analysis indicates that window solar gain is the most influential factor under long-term future conditions, accounting for approximately 35% of the relative importance, followed by thermal mass at around 25%. These findings suggest that even highly energy-efficient dwellings today may become increasingly vulnerable to overheating under future climate scenarios, highlighting the need for climate-resilient housing design strategies in temperate regions.

Видео SASBE 360: Climate Change–Induced Overheating Risk in Irish Dwellings: A Machine Learning–Based Asse канала Digital Roads of the Future