Lithium-Ion Battery RUL Indirect Prediction Based on GAAA-ELM

Since direct prediction of remaining useful life (RUL) for lithium-ion battery is difficult, the indirect RUL prediction method based on Genetic Algorithm Ant Algorithm-Extreme Learning Machine (GAAA-ELM) model is proposed in this paper. Equal-voltage drop discharge time is selected as indirect life characteristic parameter of lithium-ion battery. The first-order partial correlation coefficient method was used to verify the correlation between the indirect life parameter and direct life parameter. Build the relational model of equal-voltage drop discharge time and actual capacity and the prediction model of the equal-voltage drop discharge time are built based on GAAA-ELM. GAAA is applied to optimize the weight and threshold of ELM. Lithium-ion battery RUL is predicted by the proposed method based on NASA lithium-ion battery data sets. By comparing with BP prediction method, ELM prediction method and GA-ELM prediction method, it is shown that the proposed method can effectively predict lithium-ion battery RUL and get better prediction accuracy.