TUM Project Opportunities
← Back to projects

idp-092 · Informatics IDP Hub

Foundation Models for Process Monitoring in Vibration-Assisted Drilling

Automatically extracted from the linked project document. Check the original PDF before relying on a detail.

Document language: English · Pages: 1 · Extracted: 25 Sept 2026, 23:18

Offer 1

Foundation Models for Process Monitoring in Vibration-Assisted Drilling

English summary

The project investigates foundation models for process monitoring in vibration-assisted drilling and compares them with classical machine-learning models. Students will review open-source models, evaluate data formats for process data, benchmark models, and investigate fine-tuning strategies and their effects on accuracy and robustness.

Overview

Project details
Foundation Models for Process Monitoring in Vibration-Assisted Drilling
Source excerpt · Page 1
  • “# **Foundation Models for Process Monitoring in Vibration-Assisted Drilling**” Page 1 ↗
Organization
Not stated in the PDF
Project formats
Not stated in the PDF
Degree levels
Not stated in the PDF
Project goal
Systematically investigate the suitability of foundation models for process monitoring in VAD and compare them with small, classical machine-learning models.
Source excerpt · Page 1
  • “The goal of this work is to systematically investigate the suitability of such foundation models for process monitoring in VAD and compare them with small, classical machinelearning models.” Page 1 ↗

Topics and work

Subjects
  • Foundation models for time-series and tabular data
    Source excerpt · Page 1
    • “Large, pre-trained models for time-series and tabular data could also be used for tool-wear estimation with only a small amount of process-specific data.” Page 1 ↗
  • Process monitoring in vibration-assisted drilling (VAD)
    Source excerpt · Page 1
    • “The goal of this work is to systematically investigate the suitability of such foundation models for process monitoring in VAD and compare them with small, classical machinelearning models.” Page 1 ↗
  • Tool-wear estimation
    Source excerpt · Page 1
    • “Large, pre-trained models for time-series and tabular data could also be used for tool-wear estimation with only a small amount of process-specific data.” Page 1 ↗
Application areas
  • Aerospace industry
    Source excerpt · Page 1
    • “In the aerospace industry, holes in multi-layer material stacks are generally produced using industrial robots.” Page 1 ↗
  • Vibration-assisted drilling of multi-layer material stacks
    Source excerpt · Page 1
    • “During vibration-assisted drilling (VAD), multi-channel sensor signals and quality data are recorded.” Page 1 ↗
Methods and tools
  • Open-source models
    Source excerpt · Page 1
    • “Literature review and overview of suitable open-source models” Page 1 ↗
  • Multi-channel sensor signals and quality data
    Source excerpt · Page 1
    • “During vibration-assisted drilling (VAD), multi-channel sensor signals and quality data are recorded.” Page 1 ↗
  • Classical machine-learning models
    Source excerpt · Page 1
    • “Benchmarking against classical machine-learning models” Page 1 ↗
  • Fine-tuning strategies
    Source excerpt · Page 1
    • “Investigation of fine-tuning strategies and their influence on accuracy and robustness” Page 1 ↗
Activities
  • Review literature and survey suitable open-source models
    Source excerpt · Page 1
    • “Literature review and overview of suitable open-source models” Page 1 ↗
  • Evaluate data formats and applicability to process data
    Source excerpt · Page 1
    • “Evaluation of data formats and applicability to process data” Page 1 ↗
  • Benchmark against classical machine-learning models
    Source excerpt · Page 1
    • “Benchmarking against classical machine-learning models” Page 1 ↗
  • Investigate fine-tuning strategies and their influence on accuracy and robustness
    Source excerpt · Page 1
    • “Investigation of fine-tuning strategies and their influence on accuracy and robustness” Page 1 ↗
Kinds of work
  • Literature research
    Source excerpt · Page 1
    • “Literature review and overview of suitable open-source models” Page 1 ↗
  • Data analysis and ML
    Source excerpt · Page 1
    • “Benchmarking against classical machine-learning models” Page 1 ↗
Expected outputs
Not stated in the PDF

Requirements

Required skills
  • Programming skills
    Source excerpt · Page 1
    • “Programming skills, ideally in Python and PyTorch” Page 1 ↗
  • Interest in machine learning and process monitoring
    Source excerpt · Page 1
    • “Interest in machine learning and process monitoring” Page 1 ↗
  • Initiative
    Source excerpt · Page 1
    • “Initiative, a structured approach to work, reliability, and commitment” Page 1 ↗
  • A structured approach to work
    Source excerpt · Page 1
    • “Initiative, a structured approach to work, reliability, and commitment” Page 1 ↗
  • Reliability
    Source excerpt · Page 1
    • “Initiative, a structured approach to work, reliability, and commitment” Page 1 ↗
  • Commitment
    Source excerpt · Page 1
    • “Initiative, a structured approach to work, reliability, and commitment” Page 1 ↗
Recommended skills
  • Python and PyTorch
    Source excerpt · Page 1
    • “Programming skills, ideally in Python and PyTorch” Page 1 ↗
  • Basic knowledge of time-series data
    Source excerpt · Page 1
    • “Basic knowledge of time-series data is advantageous, but not required” Page 1 ↗
Eligible study fields
Not stated in the PDF
Programming in the project
Not stated in the PDF
Programming prerequisite
Required
Source excerpt · Page 1
  • “Programming skills, ideally in Python and PyTorch” Page 1 ↗

Practical details

Team size
Not stated in the PDF
Duration
Not stated in the PDF
Start
Not stated in the PDF
Application deadline
Not stated in the PDF
Work location
Not stated in the PDF
Location mode
Not stated in the PDF
Working language
Not stated in the PDF

Learning and support

Learning opportunities
Not stated in the PDF
Support offered
  • Supervision by Prof. Dr.-Ing. Michael F. Zäh
    Source excerpt · Page 1
    • “**Supervising Professor:** Prof. Dr.-Ing. Michael F. Zäh” Page 1 ↗

Application and contacts

Contacts
How to apply
Not stated in the PDF
Further information
Not stated in the PDF