Turning factory simulations into better decisions at Scania

From factory layouts and energy efficiency to ergonomics and decision support, VF-KDO has explored how simulation and optimisation can help manufacturers make better-informed choices. Scania has been one of the industrial partners involved in the research.

“Long before we started VF-KDO, we had already observed that industrial companies are highly interested in extracting knowledge, from their data to support better decision-making.”

For Amos Ng, Professor of Automation Engineering at the University of Skövde and project leader of VF-KDO, the key is to use virtual factory models to generate knowledge that cannot easily be obtained from real production systems alone.

“However, many overlook that more valuable insight can often only be obtained by running optimization on virtual factory models at different levels rather than relying solely on data from real factories.”

Within VF-KDO, researchers have therefore investigated how simulation-based optimisation can support decisions when there is more than one possible answer.

When there is more than one right answer

“In VF-KDO, we collaborated with manufacturing companies to explore how simulation-based optimization and knowledge discovery can support decision making,” says Sunith Bandaru, Professor of Industrial Engineering at the University of Skövde.

The challenge is that industrial problems rarely have a single objective – or a single optimal solution.

“Industrial problems often involve several optimal solutions, making it difficult to choose one.”

Sunith Bandaru, Professor of Industrial Engineering at the University of Skövde.

Instead of simply identifying an optimum, the researchers developed techniques for extracting knowledge from the optimisation process itself.

“To address this, we developed techniques that extract useful knowledge from the optimization process, thus helping stakeholders make more informed decisions.”

That knowledge can also have a value beyond the immediate problem.

“The obtained knowledge also makes it easier to solve similar problems in the future.”

Simulating the human side of production

One area where this approach has been applied is ergonomics. Within VF-KDO, researchers began using simulations of human movement to investigate and improve workstation designs.

Aitor Iriondo Pascual, Post-Doctor in Product Design Engineering at the University of Skövde.

“When I started in VF-KDO, it was the first time that we started doing simulations of humans, and with that we started improving the workstation designs,” says Aitor Iriondo Pascual, Post-Doctor in Product Design Engineering at the University of Skövde.

“And with that we started with simulation-based multi-objective optimization of ergonomics.”

From research to the factory floor

Scania has supported several industrial PhD students within the research profile, covering a range of challenges connected to production.

Fredrik Ore, Automation Competence Leader at Scania Group.

“At Scania, we have supported several industrial PhD students within VF-KDO, focusing on areas such as factory layout optimization, energy efficiency, and simulation-based decision support,” says Fredrik Ore, Automation Competence Leader at Scania Group.

The research has also been connected directly to Scania’s own development and validation work.

“In our Smart Factory Lab, VF-KDO provides validation frameworks for multi-object optimization that we can apply directly in our operations.”

That means different production objectives can be considered together rather than separately.

“This includes balancing productivity with musculoskeletal risks in assembly and optimizing energy consumption alongside cycle times in the machining lines.”

The work also points towards the next generation of decision support. As VF-KDO enters its final phase, the research is increasingly connected to artificial intelligence and new ways of managing and using knowledge.

“Beyond further advancing human-centric well-being and learning in our continued research and collaborations, the rapid rise of AI, particularly generative AI and its disruptive impact across industries, has promised to fundamentally rethink how decision support systems are developed,” says Amos Ng.

Amos Ng, Professor of Automation Engineering at the University of Skövde and project leader of VF-KDO.

“As our response, during the final phase of VF-KDO over the past two years, we have focused on designing and implementing a large language model power knowledge management system that enables more traceable, transparent and explainable decision support.”