Optimized DELM Based on Multi-Strategy Improved Secretary Bird Optimization Algorithm for SOH Estimation of Li-Ion Battery

The SOH(state of health) of lithium-ion batteries is an important indicator for the reliable operation of lithium-ion batteries. In this paper, a prediction model based on multi-strategy improved secretary bird optimization algorithm and deep extreme learning machine(MISBOA-DEML) was proposed to solve the problem of low prediction accuracy of SOH of lithium-ion batteries. The MISBOA-DELM model was based on deep extreme learning machine(DELM), while the secretary bird optimization algorithm(SBOA) with multi-strategy fusion was employed to optimize the weight of DELM by searching the optimal solution in space. To verify the validity of the MISBOA-DELM model, experiments were conducted using CALCE data set of lithium-ion batteries and a company dataset. The results indicate that the proposed MISBOA-DELM can improve MAE on company battery dataset by 52% and 1.67% compared to DELM and SBOA-DELM. The effectiveness of the proposed method is verified, and it has better stability.