直流充电桩计量误差预测技术的应用研究OACHSSCD

In order to solve the problem of large online measurement errors and insufficient measurement accuracy of DC charging piles,this paper proposes an improved grey wolf algorithm and recovery mechanism integrated DC charging pile measurement error prediction technology.It firstly uses the random differential mutation algorithm to improve the search range,introduce non-linear mutation probability,increase the sensitivity of local search in the later stage,and solve the problem of falling into local optima;proposes improved GWO optimization model,which uses the minimum relative error as the background value to further reduce analysis errors.Finally,through comparative verification,it analyzes the accuracy of the new measurement error prediction technology for DC charging piles.The results show that the new measurement error prediction technology can reduce the mean square error by 70.62％and the average absolute value error by 28.4％,effectively improving the measurement accuracy of DC charging piles.