Photo: Mikal Schlosser

Digital twins to take industry into the future

Information technology Data analysis Mathematical modelling Physics Wind energy Materials Production and management
Tests and experiments can provide new knowledge and the ability to predict the behaviour of materials and durability with ever greater precision. This is achieved by means of digital copies of processes and products.

As a vital part of Industry 4.0, so-called ‘digital twins’ play a crucial role in driving change and improvement in industrial manufacturing processes and products.

Digital twins or digital copies are virtual prototypes that can be used throughout the production chain—from design to end product. This is achieved by combining huge volumes of data from sensors and 3D scans with advanced mathematical models that describe the complex physics that take place in the manufacturing processes.

The technology is used in the major industries and is increasingly spreading to all industries.

DTU is involved in research and is collaborating on several fronts. Most recently, a collaboration between DTU Wind Energy, manufacturers, and subcontractors from the wind energy sector has received a grant of EUR 11 million (DKK 83 million) to develop and demonstrate techniques for the production of digital twins of wind turbine blades.

One of the leading experts in the field is Professor Jesper Hattel, Head of Section at DTU Mechanical Engineering, where for many years he has conducted research on numerical modelling of manufacturing processes.

According to Jesper Hattel, DTU’s strong position can be attributed to the fact that advanced mathematical modelling is embedded in the University’s DNA and used in a wide range of disciplines to simulate the consequences of changes in materials and process conditions.

"A digital twin of the entire production process provides knowledge of the entire production chain and thus better predictions so that production can be adapted and made more efficient and profitable. At the same time, the use of digital twins results in better and more sustainable products."
Professor Jesper Hattel, DTU Mechanical Engineering

Ancient principle
The basic principle in digital twins—namely combining theoretical knowledge and practice in the best possible way to predict the behaviour of a structure or process—is as old as engineering itself and has been continually refined down through the centuries. That being said, digitization and big data have significantly moved the fence posts for manufacturing companies, says Jesper Hattel. 

“Previously, simulations have mostly been used in the design phase to predict, for example, how blade materials respond to the stresses exerted on it. However, a digital twin of the entire production process provides knowledge of the entire production chain and thus better predictions so that production can be adapted and made more efficient and profitable. At the same time, the use of digital twins results in better and more sustainable products,” he says.

Simulation of the entire production chain ensures enhanced design calculations for a product such as a wind turbine, providing information about materials, maintenance, and service life—knowledge which is extremely useful early in the design phase. The designs will bemore realistic and you can reduce the error rates significantly with less wastage, says Jesper Hattel.

Remember the real world 
“In the design phase, a great deal is based on assumptions about the material properties in the finished product, and there will be uncertainties in production that are impossible to predict. You therefore need a safety margin, and more often than not the trend is to over-dimension the structure. It may be in the foundation and in the tower of a wind turbine. A digital twin provides knowledge, reducing redundancy and thus waste—as well as undesirable effects,” he says.

But it is not just in the production chain that digital twins will lead to change—also in education, training, and instruction we will see unlimited potential for testing and experimenting prior to production. Future operators can try their skills and test their ideas without the risk of costly mistakes.

But before we get too far ahead of ourselves, Jesper Hattel urges a little caution.

“The challenge of simulations is that you can be seduced by how easy everything looks and the seeming accuracy they provide. Naturally, we must use all necessary aids, but you need to know the reality of what you’re designing, and a computer model can never replace an advanced laboratory.

The biggest mistakes are often made when you forget the fundamental physical contexts.”