Joint Estimation Method of State of Health and Remaining Useful Life for Lithium-ion Batteries Based on Hybrid Networks

To efficiently and accurately predict the state of health (SOH) and remaining useful life (RUL) of lithium-ion batteries, we propose a hybrid network-based joint estimation method of lithium-ion batteries SOH and RUL. First, we develop a framework for the indirect health factor (HF) extraction of lithium batteries and form a convolutional neural network (CNN)-recurrent gated unit (GRU) battery SOH estimation model using a CNN and GRU with HF as the input and capacity as the output. Second, we build a CNN-GRU battery RUL prediction model using the SOH estimation results and the true SOH values to predict the RUL. Experimental results show that the maximum root mean square error of SOH estimation is 2.31%, and the RUL prediction error is 5.29%. Therefore, the method can comprehensively assess the SOH and RUL of lithium batteries.