Application of Data Mining and Machine Learning Methods to Industrial Heat Treatment Processes for Hardness Prediction

Abstract

This work presents a data mining framework applied to industrial heattreatment (bainitization and case hardening) aiming to optimize processes and reduce costs. The framework analyses factors such as material, production line, and quality assessment for preprocessing, feature extraction, and drift corrections. Machine learning is employed to devise robust prediction strategies for hardness. Its implementation in an industry pilot demonstrates the economic benefits of the framework.