The Research And Implementation On Life Prediction Of VRLA Battery For Power Grid

Automation, intelligent and unmanned duty is the main trend in the development of power grid. Because of the substation automation, intelligent degree is higher and higher and the expansion of unattended unmanned, substation DC power supply to assume the role of more and more important. VRLA batteries are the core in substation DC power system, its performance related t o security and stability of the entire substation quality operation. With VRLA batteries substation DC system is used widely, a series of problems caused by life beginning to emerge, such as battery failure in advance and causing the whole backup DC power failure at the substation can not give power to protection devices and circuit breaker tripping mechanism, causing the loss of pressure the entire station accident. So the research on service life of VRLA battery is increasingly significant.This article starting of the VRLA works, llustrates the importance and necessity of VRLA battery life studies; secondly, it introduces the main technical parameters of VRLA batteries, focusing on the detection method of VRLA batteries; then discusses the factors that affect the life of VRLA details, including the failure mechanisms of VRLA battery, float voltage, ambient temperature, corrosion of the positive plate, dehydration and the thermal runaway, and analyzes the life degradation processes of VRLA batteries, obtained VRLA battery life degradation is the actual capacity of degradation, out in the characterization is the terminal voltage decreased and the internal resistance increased. Through the VRLA battery life prediction problem analysis and the comparison of several kinds of prediction methods, this paper use the neural network forecasting method to predict the life of VRLA batteries. Then, on the basis of comparing the BP neural network and wavelet neural network VRLA batteries life prediction methods, establish a life prediction model for VRLA batteries based on wavelet neural network and through MATLAB modeling and simulation to verify the validity and accuracy of the model. Finally, this paper designs VRLA batteries life prediction system for power grid.