Event Detection Non-Intrusive Load Monitoring Based on FCNN

Non-intrusive load monitoring is a technique used to monitor the consumption of loads in an electrical system without the need for any physical intervention or modification of circuits or devices. Traditional methods often involve time-series analysis, whereas this paper proposes an event-based approach to load analysis, introducing an algorithm based on fusion convolutional neural network (FCNN) for event detection and appliance classification. The experiments in this paper primarily use the public dataset LIT-SYN. In the experimental results, for switch detection, the method achieved nearly $\mathbf{9 9. 9 \%}$ accuracy for single appliances, and over $\mathbf{9 0 \%}$ detection accuracy under mixed loads.