Every new car contains roughly 250 to 300 sheet metal parts, and an AI platform now aims to design them so they are lighter, stronger, and cheaper to manufacture. The transport sector generates about a quarter of global CO₂ emissions, and lightweighting vehicle body parts can cut fuel consumption by up to 30 percent. But current sheet forming processes—essential for making those parts—are plagued by manufacturing defects and a disjointed design workflow. Engineers cannot easily predict how a complex 3D shape will behave during forming, so they often discover problems too late. This project will build the world’s first AI-driven Design for Forming (AI-DfF) platform, using advanced Graph Neural Networks to simulate intricate physical systems. The goal is to integrate manufacturability and performance metrics early in the design process, before costly mistakes are made. If successful, the platform could accelerate the creation of novel vehicle parts by tenfold, while boosting their performance and eco-friendliness. That would strengthen UK manufacturing competitiveness in an era where AI is reshaping industrial design. The work is applied, not fundamental science—it directly targets a bottleneck in automotive production, with clear routes to commercial impact.
View original technical description
The ultimate goal of this project is to pioneer the world's first AI-driven Design for Forming (AI-DfF) platform, facilitating versatile applications in design development of high-performance manufacturable vehicle components. The transport sector accounts for about a quarter of global CO2 emissions. Lightweighting vehicles, especially their body parts, is crucial to increase fuel efficiency, which can reduce fuel consumption by up to 30%. Sheet forming processes, contributing to approximately 250-300 parts in every car, are essential in this regard because of their cost-effectiveness and superior stiffness-to-weight ratios. However, challenges such as manufacturing defects and a disjointed design process complicate achieving desired outcomes. An AI-driven solution is proposed to integrate manufacturability and performance metrics early in the design process. AI applications are emerging in the forming sector to enhance efficiency and product quality. Despite initial successes in AI-based modelling of forming processes, current models struggle with complex real-world scenarios, particularly intricate 3D geometries in automotive designs. Advanced Graph Neural Networks (GNNs) offer hope for simulating intricate physical systems, hinting at the potential for broader AI application in the forming sector. The project intends to meld knowledge from various domains, such as materials, mechanics, and AI, to tackle challenges in the forming space. Key objectives include creating an efficient database for forming, developing methods for complex geometric representation in AI-compatible formats, developing GNN-based surrogate models for AI-driven simulations, and establishing a platform that synergises data and AI models for optimal system design and performance. The envisioned AI platform aims to make design processes vastly more efficient, potentially accelerating the creation of novel vehicle parts by tenfold and significantly boosting their performance, manufacturability, and eco-friendliness. This innovation is crucial for the UK's manufacturing competitiveness in the rapidly advancing age of AI.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know