State-of-Charge Estimation Based on Immune Evolutionary Networks

Based on Clonal Selection Theory, an adaptive Parallel Immune Evolutionary Strategy (PIES) is presented. Compared with conventional evolutionary strategy algorithm (CESA) and immune monoclonal strategy algorithm (IMSA), experimental results show that PIES is of high efficiency and can effectively prevent premature convergence. A three-layer feed-forward neural network is presented to predict state of charge (SOC) of Ni-MH batteries. Initially, partial least square regression (PLSR) is used to select input variables. Then, five variables, battery terminal voltage, voltage derivative, voltage second derivative, discharge current and battery temperature, are selected as the inputs of NN. In order to overcome the weakness of BP algorithm, PIES is adopted to train weights. Finally, under the state of dynamic power cycle, the predicted SOC and the actual SOC are compared to verify the proposed neural network with acceptable accuracy (5%).