Modeling Ni-MH battery based on immune evolutionary network

In order to overcome the defect of conventional neural networks, computational algorithm is used to train RBF network to model the Ni-MH battery. First, RBF network centre is identified by the artificial immune data clustering method. A new immune algorithm, adaptive parallel immune evolutionary strategy, PIES is used to train RBF network and RBF neural network training steps are designed. Finally, under the state of constant current discharging and FUDS discharging, validity of the battery model is verified within an error of 0.3V.