MACH4.0 – Machine learning and data science in Metal Cutting Manufacturing

The tutorial MACH4.0 – Machine Learning and Data Science in Metal Cutting Manufacturing is aimed at Master's students of mechanical engineering and industrial engineering who want to experience data science in production technology in a practical way. The tutorial uses AI to predict the component quality of milled workpieces. Participants will gain a comprehensive insight into data recording during production on machine tools as well as the subsequent data processing and training of machine learning models with Python.

Organizational

Event-type Tutorium (Master)
Cycle Summer semester
Time frame 4 Credit Points part-time
Language German – Slides: English
Begin Tuesday, 30 March 2027
Dates Total period:
Tue, 30 March to Fri, 08 May 2026 (part-time)

Lecture:
Tue, 30 March to Fri, 02 April 2027 and
Tue, 06 April to Thu, 08 April 2027
08:30 a.m. – 11:30 a.m. each day
Exercises:
Tue, 30 March to Fri, 02 April 2027 and
Tue, 06 April to Thu, 08 April 2027
1:00 p.m. – 4:00 p.m. each day
Data Challenge:
Thu, 08 April to Fri, 07 May 2027 (free group work)
Examination period:
Thu, 13 May / Tue, 18 May / Wed, 19 May 2027 (final presentation)
Lecturer Willi Wünschel M.Sc.
Lucas Gräff M.Sc.
Room L1|17 108
Notes A weekly consultation hour will take place during the project work/data challenge.
Microsoft Teams will be used for this.

Registration

Registration period 01.01.2027 until 26.03.2027
Registration process Participation in the tutorial is only possible if you have registered in time via e-mail with Lucas Gräff M.Sc.or Willi Wünschel M.Sc.

You should only register for the course and the examination in TUCaN once you have received a confirmation email from the department. If you only register via TUCaN and are not registered at the department, your registration will be considered invalid and will be removed from the TUCaN registration. If there are more registrations than the maximum number of participants, a selection will be made with regard to the order of registration.
Participant numbers Maximum 20

Exam

Examination notes The examination for the tutorial Machine learning and data science in Metal Cutting Manufacturing consists of two practical examinations, which are completed during the tutorial. This is followed by a personal reflection and a final presentation of the project work carried out.

Examination performance:

30 % – Exercises on the topics of data processing and machine learning
(individual assessment; period: 30.03. – 08.04.2027)
60 % – Execution of a project work/ data challenge in small groups
(evaluation of group performance; period: 08.04. – 07.05.2027)
10 % – Personal reflection on the project work
(individual assessment; period: 07.05. – 12.05.2027)
opt. Grade bonus - Above-average presentation of results
(individual assessment; period: 13.05. – 19.05.2027)
Announcement of notes The grades will be announced via TUCaN (opens in new tab).

Content

Aims The aim of the MACH4.0 tutorial is to apply modern data science methods in a practical way to analyse production data on machine tools as part of a project. To this end, the requirements for the automation and networking of machine tools in the context of Industry 4.0 are taught and linked with a profound understanding of processes and specific domain knowledge from machining.

The detection of quality-determining process deviations serves as a practical use case.

  • Understanding of data acquisition, processing and storage: Students gain knowledge about the acquisition of production data on machine tools in the context of Industry 4.0.
  • Analysing and visually processing time series data: Using the Python programming language and version management with Git, time series data from production is analysed and prepared.
  • Use of data science and machine learning methods: Methods from data science and machine learning are used to predict the quality of milled workpieces based on machine data.
  • Application of domain knowledge: The knowledge acquired in the field of machining is used to interpret production data and gain valuable information about the production process.
  • Development and validation of solution approaches: Solution approaches for technical problems are developed in a team.
  • Visual preparation and presentation: The results of the project work are clearly prepared and then presented.
Contents
  • Basics of CNC machine tools
  • Basics of NC programming
  • Basics of industrial network technology and data connection of machine tools
  • Process data acquisition and data fusion (alternatively: contextualisation of sensor data with information from the machine control system)
  • Quality control in machining (alternatively: surface quality, component dimensioning and tolerances)
  • Generation of label data
  • Explanation of the data science software stack (Pandas, Numpy, Scikit-Learn, Tensorflow)
  • Use of the version management software Git
  • Construction and application of a machine learning pipeline in Python
    • Data processing & explorative data analysis
    • Feature engineering & feature selection based on domain knowledge
    • Model selection & training
    • Model fine-tuning
  • Deep learning and advanced analytics in manufacturing engineering
  • Critical evaluation and practical validation of the results in the test field for manufacturing technologies (TEC-Lab)
Previous knowledge / training Basic knowledge of Python is required, as fundamental concepts cannot be taught within the scope of the tutorial. Further experience in machining processes, data acquisition on machine tools, data processing, and machine learning is advantageous but not mandatory.
Skript, Tutorials All documents are made available via a GitLab repository. This includes lecture slides, Python scripts and further documents.
Recommended textbook Fundamentals of the control loop of machine tools, data processing and acquisition in machine tool control systems and the importance of machine data:
Brecher, C.; Weck, M. (2021)
Werkzeugmaschinen Fertigungssysteme 3 – Mechatronische Systeme, Steuerungstechnik und Automatisierung

Introduction to machine learning with the Python packages Scikit-Learn, Keras and TensorFlow incl. practical examples
Géron, A. (2023)
Praxiseinstieg Machine Learning mit Scikit-Learn, Keras und TensorFlow – Konzepte, Tools und Techniken für intelligente Systeme
Projects TensorMill | Intelligente Vernetzung zur autonomen Fräsbearbeitung von Strukturbauteilen
Pay-per-Stress | Belastungsorientiertes Bezahlmodell für den Einsatz bei Werkzeugmaschinen
Publications Quality prediction using machine data and machine learning
Fertig, A. (2023):
Datenbasierte, prozessparallele Qualitätsprognose für spanend hergestellte Werkstücke mittels maschinellen Lernens

Bauerdick, C.; Fertig, A.; Wöhrle, S.; Abele, E. (2018)
Untersuchung der Möglichkeiten einer Online-Qualitätsüberwachung (Quality Monitoring) basierend auf Maschinendaten

Processing and contextualisation of machine data
Fertig, A.; Kohn, O.; Brockhaus, B.; Weigold, M.: (2021)
Consistent Contextualisation of Process and Quality Information for Machining Processes

Contact

  Name Working area(s) Contact
Willi Wünschel M.Sc.
Research Lead TEC | Manufacturing Technology
+49 6151 8229-639
L1|01 47
Lucas Gräff M.Sc.
TEC | Manufacturing Technology
+49 6151 8229-621
L1|01 24