A physics-enhanced online joint estimation method for SOH and SOC of lithium-ion batteries in eVTOL aircraft applications

The state of health (SOH) and state of charge (SOC) of lithium-ion batteries are critical indicators for safe operation and maintenance. However, the high C-rate discharge conditions encountered in electric vertical take-off and landing (eVTOL) aircraft applications present significant challenges for accurate and stable long-term state estimation. To address these issues, this paper proposes a physics-enhanced online joint estimation framework for SOH and SOC. A novel autoencoder-Mamba network (AMN) with unsupervised feature extraction and long-sequence temporal modeling capabilities is developed to address the dynamic high C-rate conditions. The integration of physics-informed health features (PIHFs) with the feature space derived from an autoencoder aims to enhance the robustness and accuracy of online SOH estimation. Furthermore, one-step SOH prediction values and PIHFs are employed as inputs and combined with the unscented Kalman filter (UKF) to reduce long-term cumulative errors in online SOC estimation. The performance and effectiveness of the proposed methods are validated using a publicly accessible eVTOL battery aging dataset. The experimental results indicate that the proposed methods achieve a root mean square error of less than 0.5 % for SOH estimation and 0.78 % for SOC estimation across all experimental groups. In comparison to other algorithms, our methods exhibit significant long-term accuracy and stability, indicating their potential for online implementation.