The grant will be used for advancing the manufacturing of composite materials. Associate professor Vedrana Andersen Dahl and Professor Jens Wenzel Andreasen will develop a system for the characterization of material properties in real time – that is, while the material is being manufactured. Their characterization will provide immediate feedback to optimize the manufacturing process on-the-fly.
To develop so fast a system, they need a new (faster) data acquisition approach. Instead of rotating the sample to take projections from many angles, they will take just a few projections but simultaneously. They also need a new data representation. Their representation will be pre-trained for a specific type of material. Thanks to pre-training, they will be able to extract information from the sparse data.
In this project, they have decided to work with composite materials which are used in sustainable energy (wind turbine blades), transportation (automotive and structural aerospace material), biomechanical prosthetics, and many other areas. However, the methodology developed in the project may later be adapted to other materials.
“I have worked a lot with the off-line characterization of composite materials. For this, we analyzed small material samples, and it usually took a long time. With this project, I hope to provide a similar characterization in-line, while the material is produced. So we need a much, much faster analysis. But if we succeed, we can impact material production more than with our earlier work.” says Vedrana Andersen Dahl, Associate Professor, DTU Compute.
And Jens Wenzel Andreasen adds:
“Our vision is to create a new paradigm for experimental data acquisition, by combining our knowledge of the desired end result with machine learning techniques. Similar to the famous advice by one of the Atomic Force Microscopy pioneers, Paul Hansma, we want to do ‘every experiment as poorly as possible’. In our case, we will use prior knowledge and machine learning to guide our experiments to provide the data we need and nothing more, thus enabling the fastest possible in-line measurements,” says Jens Wenzel Andreasen, Professor, DTU Energy.
The project received DKK 2.989.609 from VILLUM FONDEN.
Read more about the project: https://people.compute.dtu.dk/vand/renner/
