Improved Pufferfish-Optimized KNN-SVM Technique for State-of-Health Estimation of Electric Vehicles Lithium-ion Battery

State of Health (SOH) is one of the important parameters for Battery Management System. It is the ratio of present maximum capacity of the battery to the maximum capacity of a new battery. Most commonly battery used in Electric Vehicles is the Lithium-ion Battery (LIB). The estimation of SOH for LIBs is important for better functioning of Electric Vehicles. To estimate the SOH, this paper proposed a novel improved pufferfish optimized K-nearest neighbor and support vector machine (KNN-SVM) techniques. The hyperparameters of KNN-SVM are optimized using an improved pufferfish optimization algorithm (IPOA). This proposed method will analyze the relationship between aging factors and capacity degradation. The SoH estimation framework has been confirmed with the NASA battery dataset. The batteries studied in this research are identified as B05, B06, and B07, corresponding to battery #5, battery #6, and battery #7, respectively, and they are all lithium iron phosphate batteries. The result demonstrate that the SoH could be reliably estimated from a single dataset using the proposed hybrid KNN-SVM module. The proposed method suggest a smaller error value for batteries of B06 and B07 but this shows a successful SoH estimate for battery B05 that obtains an MAE of 0.03057, an MSE of 0.00120 and an RMSE of 0.989115.