Advanced State of Health Prediction for Lithium-Ion Batteries Using Capacity Estimation and Feed-Forward Neural Networks: A Machine Learning Approach
Main Article Content
Abstract
Accurate SOH estimation is a critical goal in pursuing the safe use of lithium-ion batteries. This article uses a novel Feed-Forward Neural Network approach based on a capacity estimation method for SOH prediction. In addition, the algorithm utilized was created using Matlab® 2023 software and proposes a Feed-Forward Neural Network method to predict the battery ageing process. This article employed experimental data from the NASA PCoE Research Center to determine and compare the Actual State of Health (SoHs) and Predicted State of Health (SoHs) during battery charge and discharge cycles. The validity of the algorithm was determined by the effects of cell degradation by comparing the Machine Learning methods and, by simulating and comparing the results of the Training, Validation, and Test curves, algorithm was tested. Finally, the Mean Absolute Percentage Error (MAPE) and the Root Mean Squared Percentage Error (RMSPE) errors demonstrated that the simulations conducted in this paper correctly represent the state of degradation of the batteries and confirm the results and the validity of the Feed-Forward Neural Networks suggested.
Article Details
This work is licensed under a Creative Commons Attribution-NoDerivatives 4.0 International License.