Abstract
With the development of offshore wind power, three-level wind power converters have begun to be widely used in wind farms, but there has always been a problem of low-frequency fluctuations in bus voltage, which will seriously affect the power quality of wind farms. This article studies a genetic algorithm-back propagation (GA-BP) neural network optimization control method, injecting zero-sequence components on the basis of the sine pulsewidth modulation strategy of carrier stacking, and then optimizing the neural network weights and thresholds through genetic algorithm, and the bus capacitor voltage low-frequency fluctuations are tracked and compensated to improve the predictive ability of the control system. This combination of zero-sequence voltage injection and injection-compensated low-frequency fluctuation components solves the problem of bus voltage fluctuations. Through MATLAB/Simulink simulation analysis and 1140 V wind power converter experimental verification, the verification results show that the bus voltage fluctuation is reduced by 50%, and the grid current waveform is improved. The proposed scheme can well solve some of the influences caused by the fluctuation of the bus voltage in the wind power converter, ensure the operation reliability of the wind power system and improve the operation efficiency of the wind power system.
Funding Information
  • Provincial Natural Science Fund-Provincial and Municipal Joint Fund Project (2019jj60042)
  • Collaborative Innovation and Guidance Fund Project for Wind Power Equipment and Power Conversion (02019/03002)

This publication has 20 references indexed in Scilit: