Improved carbide volume fraction estimation in as-cast HCCI alloys using machine learning techniques

Abstract

An improved approach is presented for the estimation of carbide volume fraction (CVF) in as-cast High Chromium Cast Iron (HCCI) alloys using Machine Learning (ML) techniques. The limitations of existing formulae for CVF estimation in HCCI alloys, which relied on a limited number of alloy compositions, are addressed. A comprehensive dataset comprising 320 distinct alloy compositions from 60 different sources was compiled. ML models trained on this dataset revealed the significant influence of carbon (C), chromium (Cr), and molybdenum (Mo) on CVF determination. By leveraging ML algorithms, a predictive model was developed that offers enhanced accuracy in estimating CVF across a wider range of compositions. This ML-based approach provides researchers with a valuable tool for determining CVF in as-cast HCCI alloys, minimizing the need for resourceintensive and time-consuming experimental procedures. The results obtained demonstrate improved CVF estimation accuracy and broader applicability, thus facilitating more efficient and reliable CVF determination in HCCI alloys.