Early-Stage State-of-Health Prediction of Lithium Batteries for Wireless Sensor Networks Using LSTM and a Single Exponential Degradation Model

Abstract

Highlights: What are the main findings? State-of-Health (SOH) battery prediction was tested under three different battery cycle consumption scenarios: 30%, 50%, and 65%. Training an LSTM model with only 50 records (equivalent to 30% of battery usage) enables accurate SOH prediction, achieving an MSE of (Formula presented.) and an RMSE of (Formula presented.). The best-performing model, trained with 110 records (65%), achieved an even lower MSE of (Formula presented.) and an RMSE of (Formula presented.). Two different training datasets were tested: one with raw sensor data (168 records) from NASA and another generated using a single exponential model (SEM) for curve fitting the battery degradation trend. Various LSTM architectures and hyperparameters were explored to optimize model performance. What is the implication of the main finding? Accurate SOH estimation with limited data: the ability to achieve high accuracy with only 50 records suggests that SOH prediction can be achieved early in a battery’s life, allowing proactive maintenance and failure prevention. Improved energy management: reliable SOH prediction contributes to better decision-making in smart grids, optimizing energy storage and distribution. Cost reduction and extended battery life: enhanced SOH estimation minimizes maintenance costs and prevents premature battery replacements. One of the most critical items from the reliability and the State-of-Health (SOH) point of view of wireless sensor networks is represented by lithium batteries. Predicting the SOH of batteries in sensor-equipped smart grids is crucial for optimizing energy management, preventing failures, and extending battery lifespan. Accurate SOH estimation enhances grid reliability, reduces maintenance costs, and facilitates the efficient integration of renewable energy sources. In this article, a solution for SOH prediction and the estimation of the Remaining Useful Life (RUL) of lithium batteries is presented. The approach was implemented and tested using two training datasets: the first consists of raw data provided by the Prognostics Center of Excellence at NASA, comprising 168 records, while the second is based on the curve fitting of the measured data using a single exponential degradation model. Long Short-Term Memory networks (LSTMs) were trained using data from three different scenarios, where battery cycle consumption reached 30%, 50%, and 65% correspondingly. Various architectures and hyperparameters were explored to optimize the models’ performance. The key finding is that training one of the models with only 50 records (equivalent to 30% of battery usage) enables accurate SOH prediction, achieving a Mean Squared Error (MSE) of (Formula presented.) and Root Mean Squared Error (RMSE) of (Formula presented.). The best model trained with 110 records achieved an MSE of (Formula presented.) and an RMSE of (Formula presented.).