Karriere am WZL


Diplomarbeit oder Masterarbeit oder Bachelorarbeit

am Forschungsbereich Technologie der Fertigungsverfahren, Abteilung Zerspantechnologie, Gruppe Modellierung und Bewertung von Zerspanprozessen

A hybrid approach using machine learning to model the broaching for fir tree slot in turbine discs

Broaching is state of the art for manufacturing fir tree slot in turbine discs in aerospace industry. The broaching processes can be divided into roughing, semi-finishing, and finishing. The functionality of the manufactured slot is mainly determined by the finishing process with regard to the groove geometry and the surface integrity. The broaching tool and process design is nowadays mainly based on the experiences. In order to reduce the development time, a hybrid approach using machine learning to model the broaching process has been proposed at WZL within a research project funded by German Research Foundation (DFG).

Development of an approach based on hybrid data from experiments and simulations using different Machine Learning algorithms to model the broaching process

1. 2D chip formation simulations with Finite Element Method in microscopic level with ABAQUS
2. Experimental determination of the cutting forces in orthogonal cutting on a broaching machine
3. Extension of an Artificial Neutral Network (ANN) with TensorFlow
4. Analyzing the hybrid data with different machine learning algorithms
5. Implementation of the trained ML-program in macroscopic level to model the broaching for fir-tree in turbine discs
We offer:
• Intensive support and training in the subject areas (cutting, FEM, TensorFlow, machine learning)
• Free and independent work with a good working atmosphere in a highly motivated team
• Contact to a large number of companies at national and international level

• High motivation and commitment
• Studies in mechanical engineering, materials science, CES or computer science
• Interest in interdisciplinary works
• Basic knowledge of FEM and a program language
• German or English
Zeitaufwand: 200,00 Arbeitsstunden

Bingxiao Peng, M.Sc. RWTH
Cluster Produktionstechnik 3A 238
Tel.: +49 241 80-28181
Fax: +49 241 80-22293
Mail: B.Peng@wzl.rwth-aachen.de