A harmonic source location method based on simulated annealing genetic algorithm and WRELM

A new method based on simulated annealing genetic algorithm (SA+GA) and weighted regularization extreme learning machine (WRELM) is proposed to solve the problem of harmonic source location which has the characteristics of sparse distribution, few monitoring devices and high difficulty. Firstly, the optimal configuration model of the system measurement nodes is established and solved by SA+GA in order to achieve the system global observability. Then, based on the measured data obtained from monitoring nodes, a model of harmonic source location based on WRELM is constructed to identify the location of harmonic source. Subsequently, the parameters of WRELM location estimator are optimized. The training time and positioning effect of different positioning estimators are compared. Finally, the simulation results show that the harmonic source location method of WRELM combined with IGPM has the advantages of high accuracy and high speed.