Machine Learning-Enabled Composition Optimization and Industrial Production of Heat-Treatment-Free High-Strength Ductile Iron
Abstract Achieving high strength and adequate ductility in as-cast ductile iron without heat treatment remains a major challenge. In this work, a machine learning-assisted alloy design strategy integrating orthogonal experiments and ensemble learning was developed to establish composition-property relationships for composition optimization. Based on the optimized design and advanced casting technologies, including fully automatic cored-wire feeding spheroidization process, and triple inoculation, a heat-treatment-free high-strength ductile iron was successfully developed. The optimized alloy achieved a tensile strength of 974 MPa, a yield strength of 613 MPa, and an elongation of 6.2% in the as-cast state. Eliminating conventional heat treatment substantially reduces manufacturing time, energy consumption, and production cost while maintaining excellent mechanical performance. This work demonstrates the effectiveness of data-driven alloy design for developing high-performance ductile irons and provides a practical pathway toward sustainable and intelligent cast iron manufacturing.
Authors
- Depeng Shen (ORCID: https://orcid.org/0000-0002-6135-3884)
- Wenyu Liu (ORCID: https://orcid.org/0000-0002-4582-7488)
- Bingtao Tang (ORCID: https://orcid.org/0000-0001-5334-4403)
- Zhiwei Zhang
- Suxin Zhai
- Maozhen Zhang
- Jinguo Jiao
- Jian Ye
Institutions
- Qilu University of Technology (CN)
Publication Details
- Journal
- International Journal of Metalcasting
- Published
- 2026-11-14
- DOI
- https://doi.org/10.1007/s40962-026-02208-7
- Primary Topic
- Machine Learning in Materials Science
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- National Natural Science Foundation of China
- Shandong Academy of Sciences
- Qilu University of Technology