Hannover Messe 2022: Efficient and predictive planning of factory layouts using machine learning
What if companies could plan factory layouts at the touch of a button in such a way that material and energy flows as well as environmental factors are optimally taken into account? Researchers at Technische Universität Kaiserslautern (TUK) are developing an AI-supported tool to help production planners in the early stages of factory planning. The researchers will present their “Reinforcement Learning Algorithm” at the Hannover Messe from 30 May to 3 June at the Rhineland-Palatinate research stand (Hall 2, Stand B40). This algorithm is capable of designing and optimising factory layouts with regard to multiple factors.
The algorithm, which uses machine learning methods, is particularly beneficial if new production halls are planned or if production lines are frequently rearranged. Since layout planning of factories is extremely diverse, this AI tool can adapt to different scenarios, taking into account multiple factors such as production costs, available space, maximum flexibility and floor load-bearing capacity.
How is this possible? “The algorithm learns by simulating a wide variety of scenarios countless times and evaluating them on the basis of predefined factors,” explains Matthias Klar, scientific researcher at the Institute for Manufacturing Technology and Production Systems at TUK. “In this way, the software expands the knowledge of production planners. Especially in cases of new manufacturing processes for which no empirical values are yet available.” Another advantage of the algorithm is that it can also simulate dynamic conditions of the material flow. Matthias Klar gives an example: “Usually, companies position machines that are interlinked from a production point of view directly next to each other in the production line. However, the positioning can cause congestion in the material flow which disrupts the ongoing production process. These types of time and cost factors can be eliminated at the very beginning of the planning process with the help of the AI-supported software, which anticipates and plans ahead by using an integrated material flow simulation.”
In order to use the algorithm as an intelligent planning tool, neither immense computational capacities nor advanced IT knowledge are required. “Planning processes for production lines usually take several months,” says Matthias Klar, who has spoken to many companies as part of his doctoral studies. “The software needs a day to a maximum of a week to plan the best possible factory layout. Compared to the pure manual planning duration that is absolutely reasonable and does not require any additional investment in the company's own IT. It will also be very easy to use the tool. We are developing a demonstrator that guides users step by step through the planning process.” Therefore, the AI-supported tool also adapts to the circumstances of SMEs.
At the Hannover Messe, interested companies will be able to see an animation demonstrating how the algorithm sequentially places machines while achieving the specified parameters. This solution is especially interesting for small and medium-sized enterprises that manufacture products in individual and small series.
The project is funded by Deutsche Forschungsgemeinschaft (DFG) [German Research Foundation] as part of the International Research Training Group IRTG2057 “Physical Modeling for Virtual Manufacturing Systems and Processes”.
Questions can be directed to:
Matthias Klar
Institute for Manufacturing Technology and Production Systems
Phone: +49 631 205-5629
E-mail: matthias.klar(at)mv.uni-kl.de
+++
Klaus Dosch, Department of Technology and Innovation, is organizing the presentation of the researchers of the TU Kaiserslautern at the fair. He is the contact partner for companies and, among other things, establishes contacts to science.
Contact: Klaus Dosch, Email: dosch[at]rti.uni-kl.de, Phone: +49 631 205-3001
Wissenschaftlicher Ansprechpartner:
Matthias Klar
Institute for Manufacturing Technology and Production Systems
Phone: +49 631 205-5629
E-mail: matthias.klar(at)mv.uni-kl.de
Die semantisch ähnlichsten Pressemitteilungen im idw
