BBG relies on artificial intelligence in production automation

As part of a research project, we are working on developing production automation based on artificial intelligence (AI). The project "EKI - Engineering for AI-based Automation in Production Environments" is funded by the "dtec.bw - Center for Digitization and Technology Research of the German Armed Forces". Other partners besides BBG are Helmut Schmidt University/University of the Federal Armed Forces Hamburg and Weidmüller Interface GmbH & Co KG in Detmold as software specialist.

The project is intended to enable self-learning adaptation of production systems to changing requirements and environmental conditions, for example to manufacture new variants of a product.

The results are verified at BBG in Mindelheim

In our production facility in Mindelheim, a complete plant for encapsulation glass with polyurethane (PUR) is being set up to demonstrate and test the research results. This includes various automation modules for preparing the PUR encapsulation process, such as priming and flashing off the components. Added to this is the PUR foaming station, consisting of a metering machine, mold carrier system and encapsulation tool with automated release agent feed. The feeding of inserts and the removal of the finished components are also automated. At the end of the process chain are further automation modules equipped with AI algorithms for post-processing and quality control.

In setting up the smart plant, we can draw on our experience from other projects. For example, we have already presented smart tools for use in Industrie 4.0 applications and in the smart factory environment in 2020.

AI - we put it to the test

automated manufacturing cell with robot for release agent application

Together with the other project partners, we have defined five concrete use cases in which AI is to be applied. It is important that the research results are implemented in everyday industrial life. To this end, they will be subjected to practical testing with the help of our fully automated production plant.

The five AI practice tests are:

  1. Self-learning processes for primer application as part of production preparation
  2. The automated trimming of the sprue and the parting surface during rework
  3. The implementation of a cloud-based recipe management
  4. The autonomous detection of the need for preventive maintenance
  5. The optimization of resource consumption via an energy management system

The various tasks will be scientifically accompanied in individual dissertations at HSU. The project is initially scheduled to run until August 31, 2024.

Industrial manufacturers are under pressure - will AI bring solutions?

The background to the research project is the increasing pressure on manufacturers. The demand for new products with a wide range of variants is contributing to this, as are demands for increased productivity, resource conservation and cost reductions.

So far, mainly individual aspects such as plant modularization, intuitive software, the improvement of mechatronic components and parameter optimization have been investigated. What is currently missing is a framework that brings the different approaches together.

Together with the project partners, we are working on combining the individual building blocks into overall solutions that can be used in as many industrial fields of application as possible. For this reason, the creation of AI/ML algorithms is also being used to test their usability in as many machine environments as possible. Thus, a new engineering approach with open interfaces is being developed in the project, which enables the integration of ML and AI software components into different manufacturing processes. A core idea is assistance functions to support plant manufacturers in the rapid selection, adaptation and integration of AI-based automation systems.


Go back