Research on static VAR generator based on the neural network

Along with the social requirements of power quality improvement, power quality problems are becoming the higher demands of social, reactive power compensation increasingly become the research focus. Based on neural network, non-linear mapping characteristics, puts forward a suitable for direct current control of static VAR generator detecting reactive and harmonic current method, established the neural network predictive model, and gives corresponding algorithm. Simulation shows that the method has a real-time and accurate, adaptive tracking of power system load change characteristics.