A new method based on RBFNN in SOC estimation of HEV battery

In order to mend the prediction of battery´s SOC in hybrid electric vehicles, by analyzing varying rule of electromotive force, residual capacity and equivalent inner resistance of battery during discharge, it has been put forward that neural network model should take time based characteristics into consideration. So the model could reflect the dynamic characteristics of battery more exactly. The battery model based on Radial Basis Function Neural Network is set up and trained with the curve altering with time. The pertinent experiments show that improved ability of the established model to estimate SOC of battery has been achieved.