Robust Identification of Electrical Drive Models – Master Thesis

About the role

We are looking for a master‑thesis student to investigate robust identification of compositional electrical drive models. The work combines physics‑based modelling, Python development and experimental design to improve our system‑identification pipeline.

Key responsibilities

• Study existing physics‑based models of electric drives and the associated Python component library.
• Review state‑of‑the‑art literature on robust system identification and experiment design.
• Develop and implement robustness diagnostics addressing parameter identifiability, sensitivity and convergence.
• Analyse excitation content of datasets, evaluate criteria and relate them to achievable identification quality.
• Validate the proposed approach on a practical benchmark use case and document the results.
• Extend the methodology to handle partially observable effects such as temperature.

Required profile

• Enrolled in a Master programme in Cybernetics, Engineering, Mathematics, Computer Science or a comparable field.
• Basic knowledge of dynamic systems, differential equations and the physics of electrical machines.
• Proficiency in Python programming; optional experience with MATLAB/Simulink, JAX or PyTorch.
• Ability to work independently, systematic and analytically on complex problems.
• Very good command of English.

Required skills

• Python
• MATLAB/Simulink
• JAX
• PyTorch