A research of digital signal phase difference algorithm for power system

The traditional DFT algorithm has a great spectrum leakage and picket fence effect with asynchronous sampling, which may lead to significant error about the phase and the frequency. In order to measure the accuracy phase difference of two sine signal, this paper provides an improved algorithm based on DFT. Firstly, the Hanning self-convolution window, which has a lower peak side-lobe level and a higher side-lobe rolloff rate, is studied in this paper to reduce the spectrum leakage. Then we introduce discrete spectrum correction technology to acquire a precise phase and frequency. Finally, frequency prediction approach based on discrete PID is proposed to update the sample frequency to gain upon the synchronous sampling, when the signal frequency changed. Simulation experiment shows that the improved DFT algorithm can reach a higher precision than DFT windowed by Hanning or Hamming, when signal frequency is changed linear or nonlinear.