Work

Continual learning for building energy forecasting via hybrid elastic weight consolidation and synaptic intelligence

Year: 2026

Type: article

Abstract: Accurate building energy forecasting is crucial for smart grid management and carbon reduction, but conventional static models suffer from performance degradation due to non-stationary environments and concept drift. To address this catastrophic forgetting challenge, this study proposes a dynamic energy forecasting framework based on a hybrid continual learning approach. The model combines an asymmetric encoder-decoder convolutional gated recurrent unit backbone with an attention mechanism to extract complex spatiotemporal features. Crucially, a dual-regularization mechanism that merges Elastic Weight Consolidation (EWC) and Synaptic Intelligence (SI) is introduced to balance network stability and plasticity without requiring storage of historical data. Empirical evaluations across four real-world public building datasets demonstrate that the framework consistently outperforms static baselines and vanilla regularization implementations. The proposed approach achieves high forecasting accuracy, with an average R² of 0.94 and a mean absolute percentage error of 5% under the optimal monthly update configuration, while also yielding a 78% reduction in maximum forecasting errors during severe environmental transitions. This research establishes a robust replay-free regularization methodology for continuous, adaptive energy management in long-term dynamic operations.

Source: Energy and AI

DOI: 10.1016/j.egyai.2026.100882

Authors: Siqi Han, Zhiyuan Gao, Xin Xin, Jianxin Hao, Xuewei Zhang

Institutions: City University of Hong Kong, PLA Army Engineering University, Chinese People's Liberation Army

Language: English

Topic: Building Energy and Comfort Optimization

Subfield: Building and Construction

Field: Engineering

Domain: Physical Sciences

SDGs: Climate action, Affordable and clean energy

Open Access status: gold

Estimated APC paid: $3,630