Robust, Co-design Exploration of Multilevel Product, Material, and Manufacturing Process Systems

• Published in: Integrating materials and manufacturing innovation Vol. 13; no. 1; pp. 14 - 35
• Main Authors: Baby, Mathew, Rama Sushil, Rashmi, Ramu, Palaniappan, Allen, Janet K., Mistree, Farrokh, Nellippallil, Anand Balu
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.03.2024; Springer Nature B.V
• Subjects: 7th World Congress on Integrated Computational Materials Engineering; Characterization and Evaluation of Materials; Chemistry and Materials Science; Co-design; Collaboration; Composite materials; Decision making; Decision support systems; Designers; Hot rolling; Machine learning; Manufacturing; Manufacturing industry; Materials Science; Mechanical properties; Metallic Materials; Microstructure; Multilevel; Nanotechnology; Neural networks; Optimization; Robustness (mathematics); Self organizing maps; Simulation; Steel making; Structural Materials; Surfaces and Interfaces; Systems design; Thematic Section: 7th World Congress on Integrated Computational Materials Engineering; Thin Films; Uncertainty; Co-design; interpretable Self-Organizing Map (iSOM); coupled-compromise Decision Support Problem (c-cDSP) construct; Robust design
• ISSN: 2193-9764; 2193-9772
• DOI: 10.1007/s40192-023-00324-4

Achieving targeted product performance requires the integrated exploration of design spaces across multiple levels of decision-making in systems comprising products, materials, and manufacturing processes—product-material-manufacturing process (PMMP) systems. This demands the capability to co-design PMMP systems, that is, share ranged sets of design solutions among distributed product, material, and manufacturing process designers. PMMP systems are subject to uncertainties in processing, microstructure, and models employed. Facilitating co-design requires support for simultaneously exploring high-dimensional design spaces across multiple levels under uncertainty. In this paper, we present the Co-Design Exploration of Multilevel PMMP systems under Uncertainty (CoDE-MU) framework to facilitate the simultaneous exploration of high-dimensional design spaces across multiple levels under uncertainty. The CoDE-MU framework is a machine learning-enhanced, robust co-design exploration framework that integrates robust, coupled compromise Decision Support Problem (rc-cDSP) construct with interpretable Self-Organizing Maps (iSOM). The framework supports multidisciplinary designers to (i) understand the multilevel interactions, (ii) identify the process mechanisms that affect material and product responses, and (iii) provide decision support for problems involving many goals with different behaviors across multiple levels and uncertainty. We use an industry-inspired hot rod rolling (HRR) steel manufacturing process chain problem to showcase the CoDE-MU framework’s efficacy in facilitating the simultaneous exploration of the product, material, and manufacturing process design spaces across multiple levels under uncertainty. The framework is generic and facilitates the co-design of multilevel PMMP systems characterized by hierarchical product-material-manufacturing process relations and many goals with different behaviors that must be realized simultaneously at individual levels.