About this Abstract |
Meeting |
MS&T23: Materials Science & Technology
|
Symposium
|
Materials Processing and Fundamental Understanding Based on Machine Learning and Data Informatics
|
Presentation Title |
Machine Learning-based Prediction of the Mechanical Properties of Microalloyed Steel Subjected to Thermomechanical Controlled Processing |
Author(s) |
Sushant Kumar Sinha, Denzel Guye, Xiaoping Ma , Kashif Rehman, Stephen Yue, Narges Armanfard |
On-Site Speaker (Planned) |
Sushant Kumar Sinha |
Abstract Scope |
Microalloyed steels constitute about 12% of global steel, dominating oil and gas extraction, construction, and transportation industries. The as-rolled mechanical properties of microalloyed steels result from their chemical composition and thermomechanical processing history. The ability to predict mechanical properties from process parameters and composition has offline and online applications. Offline prediction can facilitate alloy design with lean chemistry and enhanced mechanical properties. Online prediction can enable dynamic control of the rolling mill and reduce the need for mechanical testing. This study proposes an artificial neural network methodology for predicting the lower yield strength (LYS) and ultimate tensile strength (UTS) of microalloyed steels. The feature engineering approach transforms raw data into features typically used in physical metallurgy. SHAP values explain the effect of thermomechanical controlled processing which can be explained by physical metallurgy. |