Building retrofitting towards net zero energy under climate change

Abstract The challenge in Net-Zero Energy Building (NZEB) retrofitting is to identify the most effective measures to address energy performance issues. This paper presents a machine learning model for optimizing retrofit measures to achieve NZEB under the influence of climate change. Specifically, the non-dominated sorting genetic algorithm (NSGA-III) minimizes energy consumption and the predicted percentage of dissatisfaction (PPD) while achieving a NZE balance, thereby obtaining the Pareto front. The Order of Preference by Similarity to Ideal Solution (TOPSIS) ranking technique is then applied to the Pareto front to obtain an optimal solution. Various passive energy retrofit measures are investigated, and renewable retrofit measures are employed to cover the required energy. This process is repeated for different time frames to consider the impact of climate change on selecting retrofit measures. The results clearly indicate that, for retrofitting a residential NZEB, higher insulation values are needed for future scenarios compared to the present scenario due to the effects of climate change. Using the future scenarios defined by the Shared Socioeconomic Pathways framework, a higher level of envelope insulation and renewable retrofit measures are required to achieve NZEB in the Sustainable Future scenario compared to the Fossil Fuel-dependent Future scenario, with increases of 35% and 50%, respectively.