Join us at Sandvik Coromant Tools in Gimo, Sweden, for an exciting master's-level thesis project. This work is an extension of previous research, focused on creating a groundbreaking proof of concept for automating the validation of simulated 3D models against their master counterparts.

Background and scope of the project

Our mission at Sandvik Coromant involves the production of cutting-edge tools utilizing multi-axis machine tools and parametrically generated NC-programs. Since the 1980s, we have strived for automation, harnessing methods to generate 3D models and NC-programs. Our CAD/CAM system, Siemens NX, plays a central role, with access to its API. Presently, our in-house programming platform, built on Java, is the backbone of this process.

The crafted 3D model serves as the gold standard, representing our desired production outcome. Yet, real-world production may slightly deviate due to manufacturing constraints and specific requirements. To bridge this gap, we utilize Vericut software to generate simulated 3D models of our manufactured products. Human assessment has traditionally been the method for evaluating differences between these models. Our ambition is to replace this with a machine learning-based verification process.

Work description

Your mission is to create and train a machine learning algorithm that can determine the acceptability of differences between 3D models. This assignment is to deliver a "proof of concept" tailored to one specific tool family. Furthermore, you'll explore the implementation of this concept within our automatic batch environment.

To succeed in this mission, several key activities lie ahead:

• Project planning.
• In-depth comprehension and understanding of:
  • NC preparation process (CAD/CAM).
  • Simulation process.
  • Automation process.
• Proposal of a robust methodology.
• Design of an experimental setup.
• Execution of the proof of concept.

Your profile

This assignment is ideally suited for Master of Science candidates in IT, Data Science, or Technical Physics. A strong programming foundation, particularly in machine learning, is indispensable to craft an effective solution. We encourage collaboration, and this project is designed for two students to work together.

At Sandvik Coromant, we firmly believe that a diverse blend of experiences, perspectives, and backgrounds enriches our work environment, strengthens our business, and benefits our customers. We look forward to welcoming students who contribute to this diversity.

Time and place

The thesis project is planned commencing in January 2024 or as soon as feasible. The project will be based in Gimo, Sweden.

Contact information and application

For more information about the thesis program, please contact:

Jonas Järvström, jonas.jarvstrom@sandvik.com, +46 730 28 54 23

Please send your application to the contacts above as soon as possible, and no later than 2023-11-19, as selections are made on an ongoing basis.