Scribd
Upload
46 views1 page

Master Thesis Proposal Unsupervied Learning Website

This document describes a master's thesis opportunity focusing on using deep learning techniques for anomaly detection in industrial processes. The thesis will investigate architectures like autoencoders and generative adversarial networks for anomaly detection and develop a concept applying these methods. As a use case, the thesis will implement typical fault scenarios on process simulators like the Tennessee Eastman process to validate the deep learning approach. The goal is to develop an unsupervised deep learning solution for anomaly detection to help predict failures and suggest maintenance in industrial systems.

Uploaded by

Ines B.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
46 views1 page

Master Thesis Proposal Unsupervied Learning Website

This document describes a master's thesis opportunity focusing on using deep learning techniques for anomaly detection in industrial processes. The thesis will investigate architectures like autoencoders and generative adversarial networks for anomaly detection and develop a concept applying these methods. As a use case, the thesis will implement typical fault scenarios on process simulators like the Tennessee Eastman process to validate the deep learning approach. The goal is to develop an unsupervised deep learning solution for anomaly detection to help predict failures and suggest maintenance in industrial systems.

Uploaded by

Ines B.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

Master Thesis/Bachelor Thesis

Deep learning based anomaly


detection
MASTER THESIS AT THE AUTOMATION TECHNOLOGY LAB

Prof. Dr.-Ing.
Andreas Schwung Motivation
Department of Modern production systems have evolved rapidly to meet the demands for higher
Electrical Energy
productivity and production quality. The recent research agenda of “Industry 4.0” has
Technology
provided ever more traction for sophisticated and intelligent systems. AI methods,
specifically Deep Learning is one of the hottest and most researched topics for
optimisation and diagnosis of industrial processes. Deep learning algorithms are based
on the concept of how a human mind works. It learns to perform tasks using only data
from the system without the need of expert knowledge about the underlying process.
This task of anomaly detection helps the machine operator or the manufacturer for
early prediction of failures in the system and suggestion for maintenance action to be
Contact performed. Due to the rare occurence of an anomaly in an actual industrial process
and therefore, not many examples for the machine learning algorithm to learn from,
Gavneet Singh Chadha unsupervised learning techniques fit aptly to the anomaly detection research problem
Lübecker Ring 2 being. Therefore, the objective of this master is to implement an unsupervised deep
59494 Soest learning approach for anomaly detection in industrial processes.
Tel.: 02921 378 3491
E-Mail:
chadha.gavneetsingh@fh-
Task
swf.de
During the offered master thesis, the capabilities of existing deep learning
or
structures like Autoencoders and Generative Adversarial Networks (GANs) as well
as its applicability for anomaly detection shall be investigated. Then a concept and
system structure based on these architectures shall be derived and implemented.
Prof. Dr.-Ing.
Andreas Schwung
As an application example the well known Tennessee Eastman process and the
Lübecker Ring 2
59494 Soest Bulk Good Laboratory Plant in the department of automation technology shall be
applied.Typical scenarios for fault shall be defined using the process and
Tel.: 02921 378 3419 implemented. Finally, validation and test shall underline the applicability of the
E-Mail:
schwung.andreas@fh-
approach.
swf.de

Requirements
Eagerness to learn about innovative and modern technologies. Good knowledge
about machine learning, in general, will be an advantage. We are looking for master
students of the South Westphalia University of applied sciences who want to
complete their master thesis or bachelor thesis.

You might also like