AI-guided optimization of manufacturing protocols for AHSS coils

ABSTRACT

Manufacturing protocols or processing parameters in a coiling mill affect multiple desired properties of advanced high-strength steel (AHSS) coils. These properties include yield strength (YS), ultimate tensile strength (UTS), elongation percent, hardness, etc. In this work, attempts were made to maximize YS, UTS, and elongation percent for AHSS coils while determining the operating parameters that can be helpful in achieving those properties. Additionally, operating parameters were also determined for a few specific grades of AHSS steel with respect to desired properties of interest. Actual plant data from a coiling mill was analyzed through a set of statistical and artificial intelligence (AI) based algorithms. Predictive models were developed through k-Nearest Neighbor (k-NN) algorithm. Optimization of multiple properties was performed through a non-dominated sorting genetic algorithm (NSGA2). The concept of parallel coordinate chart (PCC) was used for visualization as well as identifying operational parameter that can be helpful in achieving a desired property. The research methods and findings presented in this article are of industrial significance and can be applied to other manufacturing processes.

KEYWORDS:

• Steels
• optimization
• automotive
• manufacturing
• simulations
• ahss
• ai
• ys
• uts
• elongation
• processes

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/10426914.2022.2105871