混合动力系统直流电网绝缘电阻测量误差补偿

摘要

In order to solve the problem of DC power grid insulation detection in the hybrid system,the unbalanced branch method was adopted for the measurement of insulation resistance.Hardware circuit was built to simulate the DC bus of the hybrid power system and the relevant experimental data acquisition was completed. According to the experimental results of error analy-sis,the least squares algorithm and BP neural network algorithm were used to compensate errors.The simulation result shows that both of two compensation effect is good.And the compensation effect based on neural network algorithm is better.%为了解决混合动力系统直流电网的绝缘检测问题,采用不平衡支路法进行绝缘电阻的测量.搭建硬件电路来模拟混合动力系统的直流母线并完成相关实验数据的采集.根据实验结果进行误差分析,采用最小二乘法以及BP神经网络算法对测量误差进行补偿.仿真结果表明,这两种方法的补偿效果良好,且神经网络算法的补偿效果更优.