Heat treatment is widely used to modify the microstructure and mechanical properties of medium carbon steel. However, selecting suitable austenitizing temperature, soaking time, quenching medium, and tempering conditions can require extensive experimentation. This paper presents a proposed machine-learning framework for predicting and optimizing the mechanical properties of AISI 1045 medium carbon steel. The framework considers process variables and material composition as input features and hardness, tensile strength, yield strength, elongation, and impact energy as target responses. Random Forest, support vector regression, artificial neural networks, and gradient-boosting methods are identified as candidate models. The workflow comprises data acquisition, preprocessing, feature engineering, model training, validation, prediction, and multi-objective process optimization. The supplied manuscript includes illustrative comparisons showing close agreement between selected experimental and predicted values, and a feature-importance example in which austenitizing and tempering temperatures rank highly. However, the source does not document the dataset provenance, sample size, experimental protocol, train-test split, cross-validation results, or complete model metrics. Consequently, the numerical examples should be treated as illustrative rather than as independently verified experimental findings. The proposed framework offers a basis for more systematic heat-treatment decision support, but its industrial performance must be established through reproducible experiments and external validation.
Keywords: artificial intelligence, machine learning, medium carbon steel, heat treatment, mechanical properties