SOH Estimation Based on Multisource Feature Extraction and SSA-LSTM Algorithm.

Area of Science:

• Battery Technology
• Artificial Intelligence
• Data Science

Background:

• Accurate state of health (SOH) estimation is crucial for lithium-ion battery performance and safety.
• Complex operating conditions pose challenges for traditional SOH estimation methods.
• Insufficient feature extraction limits the precision of current battery health monitoring.

Purpose of the Study:

• To develop an advanced SOH estimation method for lithium-ion batteries.
• To improve feature extraction and selection for complex operational scenarios.
• To enhance the accuracy and reliability of battery health prognostics.

Main Methods:

• A multisource feature set was constructed by fusing empirical, statistical, and mechanistic features.
• Feature selection was performed using correlation and importance ranking to optimize feature quality and quantity.
• A Sparrow Search Algorithm-Long Short-term Memory Network (SSA-LSTM) was employed for SOH estimation.

Main Results:

• The proposed feature selection method successfully identified an optimal feature set.
• The SSA-LSTM algorithm demonstrated superior performance compared to other estimation methods.
• Maximum root-mean-square error (RMSE) and mean absolute percentage error (MAPE) were 0.73% and 0.53%, respectively.

Conclusions:

• The developed SSA-LSTM method provides accurate and reliable SOH estimation for lithium-ion batteries.
• The multisource feature extraction and selection strategy effectively addresses challenges in complex operating conditions.
• This approach contributes to the efficient, safe, and long-lasting operation of battery systems.