Research Trends in Predictive Maintenance: A text mining and topic modeling based literature analysis
- Type:Bachelor-/Master's Thesis
- Date:ab sofort
- Supervisor:
-
Background:
Predictive Maintenance (PdM) is a real-world industrial AI application, offering great potential for use. Additional to AI techniques, it demands a holistic embedding into industrial service systems. Especially in manufacturing systems, there are various best practices from researchers and practitioners. Through digitalization, the establishment of connected products allows an expansion of smart maintenance services escorting physical products. Vehicle applications are one example of a fast-growing aspect of smart maintenance services containing huge value potentials for customers and the automotive industry.
The literature on Predictive Maintenance is extensive. However, it is still hard to judge what state-of-the-art of PdM is at the moment. It is often difficult to tell, if presented PdM solutions have been brought into practice.
This Master thesis shall bring more clarity, insight and structure into the vast PdM literature.Research Goal:
Based on a prior Master thesis applying only topic analysis on PdM literature, this research work shall dig deeper into the contents of the selected articles. Analysis should go down to the word level. We have already developed a descriptive framework for PdM which could be taken as a basis for defining a number of keyword vocabularies along which the literature articles are classified. The goal is to have a characterization of the articles according to industries, technical components, physical measurements, AI/ML methods, and progress of implementation.
Your Profile:
- Interest in industrial AI research and applications
- Experience with Python, ideally text mining techniques
- Interest in working alongside with researchers and towards scientific publication
- High motivation to work on interesting real-world problems
- Self-organized and goal-oriented working mode, including the motivation to bring own ideas
We offer you a challenging research topic, close supervision, and the opportunity to develop practical and theoretical skills. If you are interested, please send an email to Yannic Wolf (yannic.wolf@tum.de) with cc: hansjoerg.fromm@kit.edu along with a transcript of records and CV.