A Fine-Grained Household Electricity Perception Method Based on Non-Intrusive Load Monitoring Technology

Non-Intrusive Load Monitoring (NILM) is capable to identify detailed electrical appliances by analyzing load data, thereby obtaining a list of appliance energy consumption and visualizing the carbon footprint of the devices. In this context, this paper proposes a fine-grained household electricity perception method based on NILM. To detect load events, the cumulative sum algorithm is employed. To enhance the accuracy of load event identification, the sparrow search algorithm (SSA) is introduced to improve the Self-Organizing Map (SOM) neural network by optimizing the initial weights. Furthermore, the depth-first search algorithm is utilized to match load events with the corresponding devices, which improves the speed and depth of the search path for perceiving device electricity consumption behavior. Finally, simulations are carried out, proving the effectiveness of the proposed fine-grained household electricity perception method.