Research on Battery Capacity Estimation Based on GRNN Model
In order to estimate the remaining capacity of 18650 batteries more accurately under the condition of high efficiency and small data statistics,based on the radial basis neural network model of GRNN nonlinear regression theory,the kernel function parameters of the model were improved by using grid search,so that the remaining capacity of 18650 batteries could still be accurately estimated under the condition of less data.The battery was charged and discharged in cycles,the relevant parameters in the cycle were extracted,and the data of Yijie battery were used for training,and then the capacity of 4 of them was estimated to obtain the results.The results show that the ohmic internal resistance of the battery is negatively correlated with the capacity of the battery,and the discharge time of iso voltage drop is positively correlated with the capacity.The improved GRNN model is obtained,and the kernel function parameter is 5,and the new model is more accurate in the case of a small amount of data.
small amount of databattery capacity estimationGRNN modelkernel function parametersgrid search
万方数据
维普
