A neural network model for Ni-Cd batteries

Ni-Cd batteries are nonlinear electrochemical devices with different characteristics at different charge currents. This paper proposes a neural network based method for simulation and modeling of Ni-Cd batteries that considers the impact of charge current. Simulations and measurements are performed for the RBF neural network and advantages and limitations of the model are presented. Inputs of the neural network are battery current ( Ibat ), no load voltage and time (t) while battery voltage ( Vbat ) is selected as the output. An experimental setup is used to validate the accuracy of the model at different charge rates. Computed and measured results show good agreements for a 7AH, size F, Ni-Cd battery, manufactured by SANYO. Theoretical and experimental results are compared and analyzed.