Detecting energy consumption anomalies with dynamic adaptive encoder-decoder deep learning networks

Abstract

Efficient management of building energy consumption is paramount for sustainability and cost-effectiveness, where anomalies in energy usage patterns can signify malfunctions, inefficiencies, or even potential hazards within the building systems. To address this problem, this study introduces an Asymmetric Hybrid Encoder-Decoder (AHED) anomaly detection architecture, designed to precisely forecast and identify point anomalies and collective anomalies within the domain of building energy usage. This architecture synthesizes both supervised and unsupervised learning approaches and utilizes an advanced decoder-encoder configuration for accurate prediction of energy consumption. Concurrently, the AHED framework applies sliding window techniques and cross-correlation analysis to convert multivariate temporal data into feature matrices, to detect anomalous patterns that manifest collectively within specified time intervals. The results demonstrate that the AHED model outperforms traditional anomaly detection techniques, achieving higher accuracy and improved generalization across diverse building environments, which affirms the efficacy and superiority of the asymmetric model in anomaly detection for building energy consumption. This study underscores the potential of dynamic adaptive deep learning networks in addressing the challenges of anomaly detection in building energy management, paving the way for more efficient and sustainable building operations.