When the product drives production, industry can become more flexible

Anders Nilsson’s research shows how automated production systems can become both more flexible and more accessible, for example for customized products. Photo: University West.

Today’s manufacturing industry is facing a clear challenge. Many customers demand tailored solutions, increasing the need for variation in production, while manufacturing processes must still remain efficient and profitable. Existing automation solutions are often designed for stable mass production, not for rapid change.

“Almost no one who manufactures directly for customers can fully automate, simply because the level of variation is so high. Reprogramming systems every time becomes too expensive and too complex,” says Anders Nilsson, who recently earned his PhD at University West.

Before becoming a lecturer and doctoral student in Production Technology at University West in 2009, Nilsson worked for many years as a project engineer programming industrial control systems. At University West, he has collaborated closely with industry in automation and manufacturing projects. In his thesis, he investigates how automated systems can become more adaptable without requiring advanced specialist expertise every time something changes.

Starting from the product – not the machine

A central premise of the research is that today’s automation systems are strongly machine focused. Each machine is programmed in detail to carry out specific tasks. When the product changes, the code often needs to be rewritten – a time-consuming and costly process. Instead, Anders has developed a framework that shifts the focus to the product and the manufacturing process.

“Today, much of the ‘thinking’ is embedded in the machines. You program exactly how each step should be carried out. We move that to the product instead, so that it ‘knows’ what needs to be done to it – for example where a hole should be drilled or how something should be assembled.”

This approach makes it possible to use product information, such as data from CAD models, directly in production. In this way, production can be driven by the needs of the product rather than by the programming of individual machines.

Multi‑agent systems as support

To determine how tasks should be executed, Anders uses what is known as a multi‑agent system. This means that both products and resources – such as machines, robots, and tools – are represented by digital agents that can communicate with one another.

“If several machines are capable of performing the same operation, the system decides which one should do the job. Quite simply, it selects the one that is available or best suited at that moment.”

This makes it possible to connect and use resources as needed. The result is a solution that is significantly easier to adapt when conditions change, without having to reprogram the entire system from scratch.

Manageable by in‑house personnel

An important component of the thesis is what is referred to as a human‑centric Plug & Produce system. The idea is to make better use of the knowledge that already exists within the company. In traditional systems, changes often require external experts. Here, the ambition is for operators and process planners to be able to configure the system themselves.

“The people working in production already have knowledge of both products and processes. It makes sense that the systems should be designed around that knowledge, rather than around advanced programming.”

When product‑specific information can be removed from machine control systems and instead managed in configurable structures, the same equipment can be used for multiple products without reprogramming. It also becomes possible to connect new resources to production as needed. The system can then automatically begin using them – for example, to increase capacity or relieve other parts of the production flow.

Tested in realistic environments

The framework has been tested in several so‑called testbeds, including timber wall production and material kitting systems. In both cases, the applications involve a high degree of product variation, placing strong demands on flexibility. This has made it possible to evaluate how the system responds to change, such as when new resources are added or production conditions shift.

“It’s in these situations that you really see whether the solution holds up. When every order is slightly different, the system has to constantly adapt. We’ve been able to modify the setup and immediately see the impact on production. That provides a completely different understanding than purely theoretical work.”

Enabling flexible automation systems

Some work remains before the solution can be widely adopted in industry. More types of resources need to be integrated, and the systems must be scaled up to larger production environments. However, the results clearly point toward a different way of organizing automated manufacturing. The research shows that it is possible to build automation systems that are both flexible and accessible – by starting from people, products, and processes rather than from the limitations of machines.

Flexibility for customized products

The conclusions may be particularly important for companies working with customized products and small production series, where automation is often difficult to justify today.

“Above all, I would like this to be used by smaller companies, so they can become more competitive in the market. But it’s useful for any company with a large and diverse product portfolio.”

At University West, Anders now continues his work as a lecturer and research engineer, while also developing research on human–machine interaction.

“It’s mainly about how we can control processes instead of letting them control us,” Anders concludes.

Read Anders Nilsson’s doctoral thesis:
"A human‑centric framework for next‑generation Plug & Produce manufacturing"

Contact: Anders Nilsson, PhD in Production Technology, University West

At University West, we conduct research in collaboration with society and industry to help create a better future.