A high precision detecting method for fundamental using BP neural network

High precision detecting on grid fundamental wave is the basis of energy measurement, power quality assessment and power system automation. This paper presents a high precision detecting method for frequency and amplitude of grid fundamental based on BP neural network. The results show that the intersection of symmetry two points which are between the both sides of signal’s zeros-crossing has monotone relation with frequency, but no linear relation. The intersection is independent with amplitude. Accordingly, we set the mapping of the intersection and the frequency with the BP neural network, and then propose the optimal selection method of symmetry points. These simulation results show that the precisions of frequency and amplitude are so high to 10-4 and 10-5, which are much higher than FFT and Hanning window algorithms. These influences of random noise and harmonic on this method are very small, so it has strong anti-interference.

Project Supported:

Clc Number: