»Event Shows High Degree of Networking at the Site«

Day of Process Engineering 2022: TU Kaiserslautern and Fraunhofer ITWM for Interdisciplinary Dialogue

In the field of process engineering, researchers from different chairs of the TU Kaiserslautern and the Fraunhofer ITWM cooperate intensively. On Friday, November 04, 2022, scientists from mathematics, computer science, and engineering met at Fraunhofer ITWM for the sixth »Process Engineering Day«. About 70 participants took the opportunity for an interdisciplinary dialog. The event was organized under the umbrella of the Center of Excellence Simulation and Software-based Innovation.

The two organizers, Prof. Dr. Michael Bortz from Fraunhofer ITWM and Prof. Dr.-Ing. Hans Hasse from TU Kaiserslautern, talk about the contents and goals of the event in an interview:

What is special about the event?

Michael Bortz: The event shows the high degree of networking at the Kaiserslautern site when it comes to digitization in process engineering: networking between the TU and Fraunhofer ITWM, but also networking between the disciplines of process engineering, computer science, and mathematics.

 

What does networking look like in practice?

Hans Hasse: The networking is close and takes place on many levels. The focus is on joint research projects, often together with partners from industry, but also on large basic research-oriented projects, such as the research group »Deep Anomaly Detection on Sparse Chemical Process Data«, which started on 01.10.2022 and is funded by the German Research Foundation (DFG) for four years – with the prospect of an extension for another four years. The exchange is close and multifaceted: meetings at the working level take place almost daily and strategic issues are also regularly discussed together. The Fraunhofer Performance Center »Simulation- and Software-based Innovation« provides an excellent framework for this.

 

Which current topics were discussed particularly intensively?

Michael Bortz: There is a lot of interest in the use of Machine Learning methods in the traditionally knowledge-based domain of process engineering. How can learning methods be adapted and used in such a way that real added value is created, for example, for raw material and energy efficiency?

 

Can you give an example of how Machine Learning can lead to greater raw material and energy efficiency? 

Hans Hasse: The energy transition and the changeover to a circular economy pose enormous challenges for industry. In plain language: a great many of the existing processes must be revised, and often completely new processes must be developed, in a short period of time. This is not at all feasible with classical methods alone. Machine Learning can help here in many ways: the spectrum ranges from the prediction of substance data from mixtures that have never been measured to the automatic synthesis of new process routes and their optimization. 

 

How and why has the thematic focus shifted since the last Process Engineering Day in 2019?

Michael Bortz: The last Process Engineering Day consisted of an exchange between the participating engineering chairs at TU Kasiserslauern and the departments at Fraunhofer ITWM. With the Process Engineering Day 2022, we are actively including the computer science and mathematics competences on Machine Learning (ML). In view of various approved cooperation projects that are just starting up between these actors, this year's Process Engineering Day can also be seen as a kick-off for these projects.

 

Can you say more about the content of these projects?

Hans Hasse: The research group mentioned above is concerned with detecting anomalies in chemical production processes. This is a tremendously important task, because we are also talking about the question of whether a plant is in safe operation or not. Of course, there are existing methods for this, but with Machine Learning, the extremely extensive and diverse data available on the condition of plants can be viewed and evaluated in the big picture for the first time. The results will be highly interesting not only for the chemical industry but also for many other industries. 

 

Where is the journey heading? When is the next Process Engineering Day planned and what questions will be of interest between now and then?

Michael Bortz: Next year, there will again be an international workshop on topics of digitalization in process engineering with MMiPE (Mathematical Methods in Process Engineering). The next Process Engineering Day will take place in 2024. It will then be exciting to see what new advances can be reported in the use of Machine Learning methods in process engineering.