Abstract:In order to quickly and accurately check the harmonic problems of electric vehicle charging,a neural network is proposed based on triangle function of current real-time harmonic analysis method,which uses the weighted harmonic component of current estimates that first uses current weight,the negative gradient descent method which is used to update weights,calculate the trigonometric function of the optimal weights of the neural network.Using the optimal weight can have fast and accurate estimation system harmonic component and related parameters,good robustness,and can quickly realize the analysis of the current every harmonic parameter.Finally,the method is simulated and analyzed by MATLAB,and the fast Fourier transform (FFT) technique is compared.The results show that the current harmonic data can be detected quickly and effectively.