A Comparative Study of Random Forest and LSTM Models for Battery Remaining Useful Life  Prediction

Monias Tapiwanashe Munhamoh; Peter Bukelani Musiiwa

doi:10.71097/IJSAT.v16.i3.7244

A Comparative Study of Random Forest and LSTM Models for Battery Remaining Useful Life Prediction

Author(s)	Mr. Monias Tapiwanashe Munhamoh, Mr. Peter Bukelani Musiiwa
Country	Zimbabwe
Abstract	In critical areas requiring reliable power such as electric vehicles, renewable energy storage systems, aerospace and aviation, medical devices and substation DC systems Lithium-ion batteries find many applications. These critical and sophisticated systems require reliable operation and higher safety considerations thereby preventing unwarranted failures, which may be catastrophic. The uninterrupted power requirement is very critical for the safety operation of these systems. Real time monitoring of the critical battery parameters such as capacity, voltage, current and temperature becomes very important. This is crucial for predictive maintenance resulting in planning for the necessary routine maintenance and replacement at the end of life of a system. Failures are thus detected before total system collapse. Routine inspections and checking of critical parameters done in most cases requires a lot of human intervention and the regular maintenance does not suit the unexpected failures which in most cases occur suddenly. On the other hand, Machine Learning models offer predictive maintenance techniques according to the model built from the model features. Machine Learning based techniques such as Decision Tree Regression, Random Forest, Support Vector Regression, Gaussian Process Regression and Long Term Short Memory are used to predict the Remaining useful life (RUL) of Lithium-ion batteries. This paper looks at two machine-learning models used to predict the remaining battery useful life. The Random Forest (RF), representing ensemble methods class of machine learning and the Long Term Short Memory (LSTM) representing the deep learning/sequence models class are discussed. The selected models chosen on the basis that they are a good representative of the respective class of Machine – Learning models. The methodology used in this study include downloading and loading in MATLAB the publicly available online NASA data set. Preparation of the data for modelling is done through exploratory Data Analysis in MATLAB. The model features such as battery capacity, voltage, current and temperature are considered in this study. These parameters chosen on the basis of their great influence in the determination of battery remaining useful life. The two Machine –learning models are implemented in MATLAB. The performance parameters Root Mean Square Error (RMSE) and the Statistical Correlation Coefficient R^2 are obtained to find the Model performance in predicting RUL. The simulated results in this paper proved that Random Forest is a better model than the LSTM when used with NASA data set for RUL prediction. The LSTM is more complex and slower to train, although the accuracy of the model increases with continuous training. This conclusion is based on the comparison of simulation results of RMSE and R^2obtained. It is noted that with continuous training the performance parameters of the LSTM model do improve greatly. This may imply that the model can be a better RUL predictor. The study provides simulation techniques in the use of Machine-Learning Models in predictive maintenance. This is necessary to avoid unwarranted system failures, minimising maintenance costs and reducing plant/system downtime. Similar simulations and analysis can be done for all systems where predictive maintenance is required. Such system which require very high reliability and system security. The availability of accurate measurable data (affecting system deterioration) is crucial for such simulations if the results are to be generalised. More studies are therefore required for other such systems.
Keywords	Regression, Random forest, Long-term short memory
Field	Engineering
Published In	Volume 16, Issue 3, July-September 2025
Published On	2025-07-25
DOI	https://doi.org/10.71097/IJSAT.v16.i3.7244
Short DOI	https://doi.org/g9vdcx

View / Download PDF File

doi

CrossRef DOI is assigned to each research paper published in our journal.

IJSAT DOI prefix is
10.71097/IJSAT

Downloads

Research Paper Format Copyright Permission Form and Undertaking Form Cover Page Vol 16 Isu 3 Cover Page Vol 16 Isu 2 Cover Page Vol 16 Isu 1

All research papers published on this website are licensed under Creative Commons Attribution-ShareAlike 4.0 International License, and all rights belong to their respective authors/researchers.

CC-BY-SA

About IJSAT Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us	Message on WhatsApp	+91-9687-182-185	editor@ijsat.org

International Journal on Science and Technology

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

A Comparative Study of Random Forest and LSTM Models for Battery Remaining Useful Life Prediction

Share this