News

Digital twin technology gains ground in food manufacturing

Food and beverage companies are increasingly turning to digital twins to optimise and automate manufacturing processes, but opportunities stretch far beyond this, say experts.

Digital twin technology uses a virtual replica of a physical asset, system, or process that updates in real-time. It is gaining ground in the food and beverage industry, particularly on the manufacturing side as companies seek to automate and optimise operations.

At a glance:

• There has been significant uptake of digital twin technology in recent years.
• Food and beverage companies are primarily using digital twins to optimise manufacturing.
• Digital twins can also provide traceability and visibility throughout supply chain.

FMCG major Unilever recently announced plans to scale the use of AI-enabled digital twins across its manufacturing network through an extended partnership with tech strategy specialist Accenture.

Unilever plans to add more than 40 digital twins to its global operations over the next 18 months, adding to existing models in Poland, India, Vietnam and the US. In Poland, for example, the company uses digital twins to improve the consistency of its mayonnaise production and cut down on waste.

Unilever said scaled integration of digital twin technology with AI-enabled insights and agentic capabilities will help more of its manufacturing teams identify issues sooner, simulate scenarios faster and make smarter decisions across the entire production cycle. As the system learns and advances, the rollout will also help factories improve quality, boost efficiency and respond more quickly to consumer demand.

How are digital twins used in food manufacturing?

Speaking to Ingredients Network, Peter Gorm Larsen, professor and deputy head of Software Engineering and Computing Systems at Denmark's Aarhus University, said it was an “impressive but doable” project, providing Unilever maintains full control over its manufacturing data.

“Digital twins can have difference purposes or services,” Larsen said, and there has been a “substantial uptake” of digital twin technology across many large companies and corporations in recent years.

For the food and beverage industry, he said many companies are using digital twins in manufacturing, such as Unilever, to assess pain points and optimise production, which “makes a lot of sense”.

But the opportunities provided by digital twin technologies don't stop there.

Lukas Esterle, associate professor of Electrical and Computer Engineering Software at the same university, said others are using the technology more broadly, across farming, processing, production and supply chain management.

“We have seen uptake of digital twins in the food industry; not only on the factory side of food production but also in the entire supply chain, from the field to shop,” Esterle said.

“While digital twins in the factory allow for improved production and lower wastage, digital twins for the entire supply chain allow for higher quality standards and improved insights for customers, enabling them to follow the food they consume,” he said.

Quality and traceability for consumers

For consumers, digital twin technology not only improves and optimises the products they are consuming, it also enhances the trustworthiness of these products, Esterle explained. “I would argue, enabling traceability for consumers provides them with unprecedented insights into modern food production, allowing them to have an understanding of what they are eating.”

Looking ahead, he said use of digital twin technology in food will likely shift from internal use to optimise operations and out to providing regulators and consumers with insights into production processes.

According to both experts, however, there are plenty of challenges ahead when it comes to scaling digital twin technology.

Integration, cyber-security, and data tracking

“In general, the biggest challenge I hear about from companies introducing digital twins in manufacturing is that their operational technology (OT) is mostly driven by programmable logic controllers (PLCs) and it is more difficult to move the learned models from the IT side with automated means,” said Larsen.

Cyber-security and uncertainties are two other broader challenges, he said.

However, if a company starts out with strong data tracking of its manufacturing processes, is able to bring in external experts to build the digital twins, and can lean into agentic AI to better understand issues when faults appear on production lines, he said integration should be far smoother.

Esterle agreed on the importance of data, particularly when developing digital twin technology for the multiple and varied processing steps within food production. “Having data already available goes a long way. Utilising it in the right way at each step of the process is key to unleash its full potential.”