Skip to Main Content
According to the characteristics of harmonics in power system, formation method of training samples is researched. Based on the principle of harmonic measuring equipment, which is the type of parallel filtering, a new measuring method of harmonics based on wavelet neural network is proposed. The wavelet neural network is optimized by improving scale factor and displacement factor of wavelet basis function, adaptive learning rate to increase convergence rate and measuring accuracy of network. The measuring method proposed has been simulated by program, simulation result shows the feasibility and validity.