An Improved Cascaded Forward Back Propagation Neural Network for State of Energy Estimation of Lithium-ion Batteries

In order to improve the accuracy of lithium-ion battery state of energy (SOE) estimation, and to solve the traditional back propagation (BP) neural network is easy to fall into the local optimum, slow convergence speed and other problems. A cascade forward back propagation (CFBP) neural network based on Weighted Mean of Vectors Algorithm (INFO) optimization algorithm is proposed for SOE estimation of lithium-ion battery. Obtain the global optimal weights and thresholds of CFBP neural network by introducing INFO optimization algorithm. Improve the model generalization ability and convergence efficiency. The experimental results show that the SOE estimation error of the proposed INFO-CFBP method can be stably controlled within 5% under two different working conditions. Compared with the traditional CFBP neural network, the root-mean-square error is significantly reduced, and the ability of robustness is stronger.