This course combines the fundamentals of turbomachinery with modern machine learning methods for condition monitoring and anomaly detection. The first part of the course introduces the operating principles of radial turbomachines, with a particular focus on centrifugal compressors, radial turbines, compressor performance maps, and the physical mechanisms leading to compressor stall and surge. The second part introduces the fundamentals of artificial intelligence and machine learning, focusing on data-driven condition monitoring, outlier detection, and predictive methods for engineering systems.

A central element of the course is a hands-on laboratory and coding project. During two laboratory sessions, students operate a turbocharger test rig, intentionally drive the compressor into surge under controlled conditions, and record synchronized sensor data, including pressure, temperature, and acoustic measurements. Using the data they have collected themselves, students subsequently develop and evaluate Python-based machine learning models to detect and predict the onset of compressor surge. This project provides practical experience in the complete workflow, from data acquisition and preprocessing to model development and performance evaluation.

By combining theoretical foundations in turbomachinery with practical machine learning applications, the course equips students with the knowledge and programming skills required to develop data-driven monitoring solutions for turbomachinery systems.

Course Content

• Fundamentals of radial turbomachinery
• Radial turbines
• Centrifugal compressors
• Compressor performance characteristics and maps
• Compressor stall and surge
• Laboratory: Operation of a turbocharger test rig and surge experiments
• Introduction to artificial intelligence and machine learning
• Data preprocessing and feature engineering
• Outlier detection and anomaly detection methods
• Machine learning for turbomachinery condition monitoring
• Python-based coding project using self-recorded sensor data
• Development and evaluation of machine learning models for compressor surge prediction

The number of participants is limited to a maximum of 8 due to the physical constraints of the integrated lab sessions. To enroll in the lecture, please register using the following link:

https://plus.campus.kit.edu/signmeup/procedures/7122

Note: Places will be allocated on a first-come, first-served basis!

Registration is open from 30.09.2026 (09:00) to 28.10.2026 (23:59). Shortly after registration, applicants will receive notifications regarding their selection, considering the limited number of available spots.

• Number of paricipants is limited to a maximum of 8 (FCFS).
• Only master level students can be admitted to the course.
• Profound knowledge on thermodynamics and fluid mechanics is mandatory.
• Basic knowledge in python is strongly recommended.
• For MACH students, Machine and processes lecture is highly recommended before taking this course. For CIW and MECH students, attendance at the lecture is mandatory prior to registration!
• We expect students to be interested in applying theoretical knowledge and translate it into real world experiments.
• During registration, you have to submit a small motivation letter.
• Lecture is offered in English.

The lecture is part of the "Research Infrastructures in Research-Oriented Teaching (RIRO)" initative at KIT.