Adapting to change: A comparison of continual and transfer learning for modeling building thermal dynamics under concept drifts

• Investigation of how to incorporate new data into building thermal dynamics models. • Impact of seasonal, retrofit, and occupancy changes on model accuracy studied. • Comparison of Transfer Learning, Continual Learning, and models trained from scratch. • Propose Seasonal Memory Learning (SML) that uses data from past year, same season. • SML outperforms initial fine-tuning (IFT) by 48.3% when concept drifts present. Transfer Learning (TL) is currently the most effective approach for modeling building thermal dynamics when only limited data are available. TL uses a pretrained model that is fine-tuned to a specific target building. However, it remains unclear how to proceed after initial fine-tuning, as more operational measurement data are collected over time. This challenge becomes even more complex when the dynamics of the building change, for example, after a retrofit or a change in occupancy. In Machine Learning literature, Continual Learning (CL) methods are used to update models of changing systems. TL approaches can also address this challenge by reusing the pretrained model at each update step and fine-tuning it with new measurement data. A comprehensive study on how to incorporate new measurement data over time to improve prediction accuracy and address the challenges of concept drifts (changes in dynamics) for building thermal dynamics is still missing. Therefore, this study compares several CL and TL strategies, as well as a model trained from scratch, for thermal dynamics modeling during building operation. The methods are evaluated using 5–7 years of simulated data representative of single-family houses in Central Europe, including scenarios with concept drifts from retrofits and changes in occupancy. We propose Seasonal Memory Learning (SML), a CL strategy that provides greater accuracy improvements than existing CL and TL methods, while maintaining low computational effort. SML outperformed the benchmark of initial fine-tuning by 42.5% without concept drifts and 48.3% with concept drifts.