Transient Event Detection Method Based on IPSO-CUSUM

The performance of event detection is crucial for Non-Intrusive Load Monitoring (NILM). In order to enhance the effectiveness of non-intrusive load monitoring methods, this study proposes a sliding window-based bilateral CUSUM changepoint detection model. Additionally, an improved Particle Swarm Optimization (IPSO) algorithm is introduced to optimize the threshold of CUSUM. The experiment utilizes voltage and current data from eight typical appliances, which are then input into the IPSO-CUSUM event detection model. The results demonstrate that this method enables event detection for nonintrusive load monitoring, leading to significantly improved accuracy and the ability to accurately reflect the operating state of the load.