Modelling and Evaluation of Li-ion Battery Performance Based on the Electric Vehicle Tiled Tests using Kalman Filter-GBDT Approach

A new market has opened up because to innovations in lithium-ion battery (LIB) technology, and that market includes the vehicle sector. Many LIBs are now praising the possibility of a transportation industry that produces less greenhouse gas emissions as a means to fight against climate change. In order to be put on the market, expensive electric vehicles that rely on batteries must fulfill stringent requirements for LIBs regarding safety, lifetime, energy density, and fast-charging. This approach outlines three steps: preprocessing the model text, feature extraction, and training. Battery big data analysis's distribution features are utilized during the pre-processing phase. By combining the cyclic and calendar losses of LIBs, the smart feature selection (SFS) method can accurately forecast these losses during feature selection. The model was trained using a Kalman Filter-GBDT. The average accuracy of the suggested method is 94.53%, which is superior than both the Kalman Filter and GBDT.