A Multi-Task Learning Method Integrating Multi-Information Fusion for Non-Intrusive Load Monitoring

The aim of non-intrusive load monitoring (NILM) is to determine the energy consumption of each electric load within a range by analyzing the total electricity meter data for a particular area. Since existing methods don’t consider the interactions between various loads, it is difficult to fully explore load features from the limited total meter data. To address this problem, NILM model is proposed based on multi-information fusion and multi-task learning (MIFMTL-NILM). Specifically, 1) Enhancing the features of samples by extracting the load features inherent in the total electricity meter data of various windows. 2) Enhancing the accuracy of load disaggregation and reducing the resource overhead of edge devices, multi-task learning is used to share network parameters across various electrical loads. Finally, the performance and effectiveness of the proposed model are validated using the public dataset UK-DALE and the constructed dataset in the laboratory.