基于部分满充放电数据的锂离子电池梯次利用SOH估计OA

Accurate and efficient state of health(SOH)estimation of retired lithium-ion batteries in echelon utilization is crucial for maximizing their value throughout the entire lifecycle.Existing studies predominantly focus on improving estimation accuracy and reducing time and energy consumption,while overlooking equipment,experimental,and implementation costs,thus lacking engineering considerations.This research is conducted around feature selection,model development,and charging-discharging strategy optimization.Firstly,four effective peak-related health features are extracted from the incremental capacity curve during partial discharge after a full charge.Secondly,echelon utilization SOH estimation is achieved based on a deep feedforward neural network.The proposed method achieves root mean square error below 3.85%and mean absolute error below 3.65%in tests across multiple temperatures and battery types,with a maximum error of 2.85%for LiFePO4 batteries,demonstrating high accuracy,robustness,and adaptability.Furthermore,combining the proposed method,a SOH estimation charge-discharge scheme considering storage and reuse is developed.While ensuring estimation accuracy,it fully aligns with the practical application scenarios in echelon utilization,significantly reducing time,energy,equipment,experimental,and implementation expenses.The minimum energy consumption is approximately 40%of the total battery energy,providing novel insights and implementation pathways for SOH estimation in echelon utilization.