Nonintrusive load monitoring based on particle swarm optimization

This paper improves a nonintrusive load monitoring (NILM) algorithm based on Particle Swarm Optimization (PSO). The initialization part of PSO is optimized, and the algorithm consumption time is reduced when the accuracy is guaranteed. At the same time, the distortion power is introduced into the fitness function, and two new parameters are added to constrain the optimal solution. In order to verify the algorithm, load disaggregation tests were performed on eight commonly used household appliances. The experimental result results show that the average accuracy rate is 95.6% and the highest accuracy rate is 98.8% without a large amount of training data.