Kisna Patel
Grade 12, National Senior Certificate (NSC) Candidate Worcester Gymnasium High School Worcester, South Africa
Download PDF
http://doi.org/10.37648/ijrst.v15i04.004
Predicting the operational lifespan of industrial components remains a critical challenge in manufacturing, where early failure leads to costly downtime and waste. This study addresses that problem by applying machine learning to estimate the lifespan of materials based on manufacturing process parameters and alloy composition. The dataset comprises 1,000 samples containing attributes such as component type, microstructure, cooling and heat-treatment parameters, alloy percentages, and defect counts. Seven regression models were developed and compared—Linear Regression, Polynomial Regression, Random Forest, XGBoost, Support Vector Regression, CatBoost, and LightGBM—using standardized numeric features and one-hot encoded categorical variables. Model performance was evaluated through mean squared error (MSE), mean absolute percentage error (MAPE), and the coefficient of determination (R²). Tree-based ensemble methods achieved superior results, with LightGBM delivering the best performance (MSE = 3757.36, MAPE = 4.13%, R² = 0.964). SHAP explainability analysis revealed that cooling rate, alloy composition, and defect counts were the most influential features. These findings demonstrate that gradient boosting ensembles, combined with explainability techniques, can provide accurate and interpretable predictions for material lifespan optimization in manufacturing environments.
Keywords: Machine Learning; Material Lifespan; Regression Models; SHAP; Manufacturing
Disclaimer: Indexing of published papers is subject to the evaluation and acceptance criteria of the respective indexing agencies. While we strive to maintain high academic and editorial standards, International Journal of Research in Science and Technology does not guarantee the indexing of any published paper. Acceptance and inclusion in indexing databases are determined by the quality, originality, and relevance of the paper, and are at the sole discretion of the indexing bodies.
