Digital twin head

How digital twins improve the efficiency of industry


The technology behind digital twins may lead to fundamental changes within industry. Investments can become more profitable, resources can be used more efficiently and costly production downtime can be prevented. “This is an incredibly exciting aspect of the digital revolution,” says Jens Sandström, Head of Digital Business at AFRY.

A digital twin is a real time connected, digital simulation model of an actual product, process, social infrastructure or system of the same. The twin’s properties enable it to describe the actual physical original or process under different conditions, and to present the results to the user in a value-based way. 

“For example, it makes it possible for companies within the manufacturing industry to better predict when a particular component or machine will need to be serviced or replaced,” explains Jens Sandström.

In this way, costly and unforeseen production downtime can be avoided, and repairs and downtime can instead be planned in plenty of time, allowing for a more even flow.

However, the potential applications go much further. A digital twin can also make investments more profitable, streamline the use of resources and optimise the urban environment in the smart cities of the future. 

“This technology offers a wealth of opportunities,” continues Jens Sandström. “We’re already creating a great deal of added business value by helping our clients to improve their efficiency and cut costs.”

Building a digital twin requires cutting-edge expertise within a wide range of different fields: in-depth engineering knowledge within the industry’s applications, simulating proposed objects, mechanics, chemistry, mathematics, statistics, IT/connectivity and graphic design of user interfaces, as well as an understanding of how an application is designed and produced.

Few players currently have all necessary expertise under one roof. This makes AFRY unique.

“As a leading engineering and design company, AFRY has all the expertise needed in order to create digital twins, and can therefore offer both individual components and a turnkey concept: Digital Twin by AFRY,” says Martin Kwasniewski, Business Unit Manager West, CAE, AFRY. “We build digital twins that can interact with both people and digital and physical assets.”

Just like a person, Digital Twin by AFRY has different stages of development at which it has different levels of competence for predicting what will happen in the future as a result of what it has experienced, different actions it intends to carry out and available or predicted conditions.

“The digital twin can also learn from its mistakes and build up a library of experience that it can re-use to deliver better value for its users in future,” adds Martin Kwasniewski.

The concept is adapted to suit the client’s needs and wishes.

“Digital twins come with different levels of intelligence, from basic connected machines to the most advanced twin with built-in AI that runs an entire factory – fully automated, and without the need for operators,” concludes Jens Sandström.  

Digital twins

Among other things, a digital twin can:

  • Make investments more profitable
  • Save money by scheduling maintenance downtime correctly
  • Produce at a higher quality, with lower resource consumption and lower risk

AFRY’s model includes five different levels of intelligence for digital twins:

Level 0: Unconnected digital simulation model of a real product, process, machine, infrastructure in a society or system of the same, here called the object. Used by the user as an aid to, under assumed conditions, design the object or simulate how the object responds to different scenarios.

Level 1 (Smart): A digital simulation model, which is connected to the real object in real time. The twin’s properties enable it to describe the actual physical original or process over time, under different conditions, and to present the results to the user in a value-based way.

Level 2 (Smart): At this level, the digital twin can be used to diagnose the object, such as to diagnose how the item is operating right now.

Level 3 (Smart): The digital twin can forecast and calculate when parts of the object will i.e. have a reduced function or getting a downtime. Based on this outcome, an optimal use of the object can be achieved which minimizes the costs and maximizes the return of any necessary investments.

Level 4 (Intelligent): With built-in AI, the twin proposes actions itself. The digital twin continually learns from its own and others' experiences and can propose actions, for example how to control the object for best performance

Level 5 (Intelligent): The digital twin takes its own decisions and handles the object itself. At this level, it replaces operators and employees.

Martin Kwasniewski

Market Area Manager, Technical Analysis

Hans Bjarnehed

Sales Manager, Business Area Specialized Technical Services