Machine Learning-Driven Prediction of Microstructural Evolution and Mechanical Properties in Heat-Treated Steels Using Gradient Boosting

Author:

Saurabh Tiwari, Khushbu Dash, Seongjun Heo, Nokeun Park, and Nagireddy Gari Subba Reddy

Date Published:

01/15/2026

DOI:

https://doi.org/10.3390/cryst16010061

Journal/Conference:

Investigation of Microstructural and Properties of Steels and Alloys

Pages:

N/A

Summary:

In this study, gradient boosting models were created to predict microstructural phase fractions and mechanical properties in order to optimize heat treatment processes through machine learning.

How was DANTE Used?:

DANTE was one of the commercially available heat treatment simulation software, used to provide thermal history inputs, that was recommended to pair with their machine learning models for accurate results in industrial components.

Student/Trial License

Learn heat treatment. Explore simulation.

Start with the DANTE Student’s Guide to Heat Treating, a quick reference covering the fundamentals of heat treatment, steel microstructures, and common processes. Ready to go further? Get hands-on experience with DANTE through our student/trial license and explore real-world heat treatment simulations.

resources

Research that drives our decisions

Explore our peer-reviewed studies and technical papers developed over the past 50 years in the industry by our engineers.

Talk with our experts

Have a project or a question about heat treating? Connect with our engineers for to discuss how DANTE can fit into your workflow.

tools for heat treatment modeling

See how the DANTE suite enables engineers to simulate quenching, distortion, hardness, stress, and metallurgical phases

Proven industry results

Explore our case studies showing how DANTE FEA predicts distortion, hardness, phase changes, and residual stress.