AI for Sustainable and Smarter Metal Casting
Rohollah (Amir) Ghasemi is the researcher aiming to improve quality assurance in high-pressure die casting (HPDC) through artificial intelligence and machine learning. His research project, AI-CAST, seeks to strengthen the competitiveness of Swedish industry while contributing to a more sustainable society.
High-pressure die casting is widely used in sectors such as automotive and manufacturing, particularly for producing lightweight high-performance components in aluminum and magnesium alloys. But the process is highly complex. Small porosities—tiny gas-filled or shrinkage cavities inside the material—can weaken components and lead to high rejection rates.
”Traditionally, manufacturers have relied on trial and error to mitigate these issues—a method that is slow, expensive, and often ineffective. The AI-CAST project instead aims to create an AI-driven, real-time quality assurance system that monitors and optimizes process parameters as the casting takes place” says Amir Ghasemi, Senior Lecturer in Mechanical Engineering at the University of Skövde.
From academy to industry, and back again
After completing a postdoctoral position at Jönköping University, Amir joined Husqvarna Group in 2019, working as a materials specialist. There, he gained first-hand experience of the real challenges facing the industry. This became a turning point:
“I realized that traditional approaches are often too time-consuming and costly. I saw an opportunity to apply AI to develop robust and intelligent solutions to truly complex production problems,” says Amir.
Collaboration Across the Value Chain
The project brings together the University of Skövde, Husqvarna Group, NovaCast Systems, and UB Verktyg. Each partner contributes unique expertise: the University leads the research with its knowledge of AI and materials science; Husqvarna provides practical casting expertise, machinery, and real-world data; UB Verktyg contributes know-how in mold design and tooling; and NovaCast Systems supplies advanced simulation software to predict outcomes before physical trials are performed.
This integrated approach ensures that the project is both scientifically innovative and firmly grounded in industrial needs.
Research That Makes a Difference
AI-CAST combines data-driven analytics and machine learning techniques with the goal of identifying a “robust process window” that ensures consistently high-quality production, with fewer defects and lower energy use. The target is to reduce porosity levels by at least 20 percent.
For industry, this means lower production costs, less material waste, and higher product quality—factors that improve competitiveness in the global market. For academia, the project generates new knowledge about how AI can be applied to complex physical processes, which is also integrated into teaching and thesis projects at the University of Skövde.
“Our vision is to establish a new foundation for how complex alloys are manufactured. In the long run, we want to contribute to a more sustainable industry, where resources are used more intelligently and Swedish manufacturing becomes more competitive and innovative,” Amir explains.
Looking Ahead
For Amir, the combination of academic and industrial collaboration is key to success.
“Technical expertise alone is not enough. You also need mutual understanding and strong communication if research is to be transformed into practical solutions,” he concludes.
