Abstract

The proliferation of Low-Carbon Technologies (LCTs), such as electric vehicles, heat pumps, and solar panels, poses a challenge for distribution system operators (DSOs). LCTs are installed within the household premises, i.e., behind the meter, and do not usually require a formal registration process. Since DSOs are generally unaware of their presence, automated detection techniques that can identify LCTs from electricity smart meters (SMs) measurements are becoming essential for network planning. Recent advances in Machine Learning (ML) have shown the potential of supervised learning techniques in energy demand forecasting, load profiling and segmentation, and smart metering. ML models can identify and classify usage patterns, providing immediate data insights and real-time feedback. This paper investigates the application of supervised learning approaches in LCT detection and develops two classifiers that can reliably detect their presence in SM profiles. The performance of both classifiers is demonstrated using a publicly available dataset SM measurements collected in Belgium and compared with other state-of-the-art techniques. Additionally, the generalisation ability of each classifier is examined on an unseen dataset from Ireland.