A data-driven framework for lithium-ion battery RUL using LSTM and XGBoost with feature selection via Binary Firefly Algorithm

Algorithms have been developed to directly forecast the Remaining Useful Life (RUL) of Lithium-ion batteries, as opposed to the indirect methods that operate through the estimation of the State of Health (SOH). This approach facilitates a more detailed exploration of the connection between battery health indicators and RUL, potentially enhancing the precision of predictions compared to methods that solely model the RUL. In this study, the Long Short-Term Memory (LSTM) model is utilized to create the one-dimensional time-series forecasting model for individual battery health indicators. Following this, a machine learning model based on eXtreme Gradient Boosting (XGBoost) is constructed for SOH estimation by integrating the LSTM forecast results with actual observational data. Consequently, the computation of RUL is accomplished by applying the SOH model to the time-series forecasting outcomes of the health indicators (HIs).To improve prediction performance and minimize computational costs associated with the RUL forecasting, a feature subset that yields the highest prediction accuracy is identified through the Binary Firefly Algorithm, a heuristic optimization technique. The efficacy of the proposed methodology is validated using the Lithium-ion battery dataset from the NASA Prognostics Center of Excellence, and the qualitative ramifications of these findings are discussed.