Bachelor/Master Thesis: Model Reduction by Neural Networks
The development of machine tools requires detailed knowledge of the product to be designed. Already during the design phase it is therefore necessary to simulate the system to be developed, to analyze the behavior and to incorporate the findings into the further development. The finite element method is a standard procedure in mechanical engineering to predict the static and dynamic behavior of machines already in the development stage. Unfortunately, the exact mapping of the dynamic machine behavior in particular requires a very long computing time. Therefore, on the way to the digital twin of fill processing machines, model reduction techniques are needed to be able to use FEM models in mechatronic co-simulations. A mechatronic co-simulation should be able to represent both the drive systems in the control engineering sense and the machine dynamics, and thus reproduce the entire dynamic behavior. An extremely promising and innovative approach of model reduction are neural networks, which are trained with the results of the FEM -full model and are then able to predict the machine behavior in a fraction of the time. The goal of this work is to create a finite element model of a machine tool in Siemens NX or Ansys Workbench that can be used to calculate mechanical transfer frequency responses, and to create and train a neural network that can predict the simulated transfer functions.
- Clarification and specification of the task
- Literature study and familiarization with the task
- System analysis of existing simulations and the machine tool
- Model building, simulation and training of the neural network - focus on transfer frequency response
- Comparison with other model reduction methods.
- Validation and verification (measurement data are already available)
- Ongoing studies at a UAS/university
- Motivation & reliability
- Assumption of printing costs in case of very good or good success
- Support from a supervisor from the relevant specialist department
- Get to know Fill as a potential employer and contribute your own ideas and knowledge
- Great opportunity to supplement your theoretical knowledge with practical experience
- Very good working atmosphere in a multi-award-winning family company
Timeframe: from fall 2023
We look forward to receiving your application!
Application as Bachelor/Master Thesis: Model Reduction by Neural Networks.
Please fill in all fields, the mandatory fields are marked with *!