RUL Prediction Method of Series Battery Based on Improved Limit Learning Machine

The conventional RUL prediction method for series type batteries mainly relies on particle filter algorithms, which have poor tracking effect on battery status, resulting in low prediction accuracy. Therefore, a RUL prediction method for series battery systems based on an improved extreme learning machine is proposed. Firstly, by analyzing the capacity decay process of the battery during charging and discharging, the decay parameters are extracted. Then, an improved extreme learning machine algorithm is used to establish a battery RUL prediction model, where the decay parameters are used as input information to obtain the predicted values, and the uncertainty of the predicted values is quantitatively analyzed. Finally, calculate the dispersion of battery capacity and track the battery status to achieve RUL prediction based on the distribution interval of the predicted results. The experimental results show that applying this method to RUL prediction of series batteries has high prediction accuracy.