Fraunhofer Institute for Industrial Mathematics ITWM
Fraunhofer Institute for Industrial Mathematics ITWM

Development and Transfer of AI Methods

KEEN Innovation Platform Accelerates Use of AI in Process Industry

BMWi Project KEEN »Artificial Intelligence Incubator Laboratories in the Process Industry«

Artificial intelligence recognizes patterns of complex processes and helps to integrate similarities, statistical anomalies and simulations and to derive decision recommendations. Our researchers support the BMWi project KEEN with expertise in the development of AI methods and their application in industry. The project was among the winners of the Artificial Intelligence (AI) innovation competition of the Federal Ministry of Economics and Energy (BMWi). KEEN was awarded in September 2019 and started in April 2020.

 

The KEEN innovation platform is being developed in the project. It networks 20 industrial and scientific institutions with the aim of accelerating the use of AI technologies and AI methods in the process industry and comprehensively assessing their potential. The process industry, which includes the chemical and pharmaceutical industries, is the third largest industrial sector in Germany.

Expertise of the Fraunhofer ITWM

In concrete terms, this means for our team to develop models for substances and substance mixtures. "Because only when I know how substances behave can I design a process, e.g. for separation," explains PD Dr. Michael Bortz, head of the department "Optimization - Technical Processes" at the Fraunhofer ITWM. The group is also working on models for the processes: "Process simulations are often very time-consuming because they are very computationally intensive. We are setting up AI models that are trained from simulation data, and which can then calculate much faster and even be real-time capable," says Bortz.

The Innovation Platform KEEN

The KEEN consortium is researching the implementation of AI procedures in three major thematic areas:

  • The modelling of processes, product properties and plants
  • Engineering (especially the support of complex planning processes and safety engineering)
  • The realization of self-optimizing plants

KEEN consists of seven subprojects:

  • Technology roadmap
  • Common data basis
  • Substance data, surrogate models
  • Process data
  • Smart engineering
  • Self-optimising system

Artificial Intelligence in the Process Industry

In order for the process industry to remain competitive in international markets, production must become more forward-looking. Ever shorter product life cycles on the one hand and the high demand for sustainability and the responsible use of resources on the other hand form a field of tension that traditional plant planning and process control can no longer cope with. Artificial intelligence has the potential to meet these specific challenges. This is because it can record, understand and analyze large amounts of data and thus better forecast complex processes.

Project Partners

KEEN is coordinated by the TU Dresden and connects 20 industrial and scientific institutions.                         

  • ABB
  • AirLiquide 
  • Bayer
  • Capital-Gain Consultants
  • Covestro    
  • DDBST
  • DECHEMA
  • Evonik
  • Fraunhofer ITWM
  • GVT
  • INOSIM
  • KROHNE
  • LeiKon
  • Merck
  • MetaTwin
  • NAMUR
  • SAMSON                                                                                       
  • ScaDS
  • TU Berlin
  • TU Dortmund
  • TU Dresden
  • TU Kaiserslautern
  • XVisual
German Federal Ministry for Economic Affairs and Energy (BMWK – Bundesministerium für Wirtschaft und Klimaschutz)
© BMWK
German Federal Ministry for Economic Affairs and Energy (BMWK – Bundesministerium für Wirtschaft und Klimaschutz)

Related Links: