Advanced

Lifetime prediction of sealing component using machine learning algorithms

Jansson, Olof (2016) MIO920
Production Management
Abstract
Tetra Pak is a world leader in the food packaging industry and has been so for a
very long time. In recent years however, they are experiencing increased competition
from low-cost suppliers selling their previously patented paper as a commodity.
This has forced Tetra Pak to focus more on selling complete systems and services.
One such potential service is condition monitoring coupled with predictive maintenance
of their packaging machines. In a packaging machine, there are electrical
components called inductors that are used for sealing packages.
In this thesis, a model for predicting the remaining useful life of an inductor
is built. Around 8 months of high resolution data is analysed and processed. The
primary tool for data... (More)
Tetra Pak is a world leader in the food packaging industry and has been so for a
very long time. In recent years however, they are experiencing increased competition
from low-cost suppliers selling their previously patented paper as a commodity.
This has forced Tetra Pak to focus more on selling complete systems and services.
One such potential service is condition monitoring coupled with predictive maintenance
of their packaging machines. In a packaging machine, there are electrical
components called inductors that are used for sealing packages.
In this thesis, a model for predicting the remaining useful life of an inductor
is built. Around 8 months of high resolution data is analysed and processed. The
primary tool for data processing is Matlab, and the predictive model is built using
Machine Learning algorithms in Microsoft’s analytics software Azure. In the data
there are clear and visible trends of the inductor degenerating, but the precision
of the predictive model is far too low to be useful in any real world-world scenario
- more data is probably needed. (Less)
Please use this url to cite or link to this publication:
author
Jansson, Olof
supervisor
organization
course
MIO920
year
type
M1 - University Diploma
subject
keywords
Analytics, Machine Learning, Microsoft Azure, Condition Monitoring, Predictive Maintenance
other publication id
16/5558
language
English
id
8895552
date added to LUP
2016-11-30 13:32:30
date last changed
2016-11-30 13:32:30
@misc{8895552,
  abstract     = {Tetra Pak is a world leader in the food packaging industry and has been so for a
very long time. In recent years however, they are experiencing increased competition
from low-cost suppliers selling their previously patented paper as a commodity.
This has forced Tetra Pak to focus more on selling complete systems and services.
One such potential service is condition monitoring coupled with predictive maintenance
of their packaging machines. In a packaging machine, there are electrical
components called inductors that are used for sealing packages.
In this thesis, a model for predicting the remaining useful life of an inductor
is built. Around 8 months of high resolution data is analysed and processed. The
primary tool for data processing is Matlab, and the predictive model is built using
Machine Learning algorithms in Microsoft’s analytics software Azure. In the data
there are clear and visible trends of the inductor degenerating, but the precision
of the predictive model is far too low to be useful in any real world-world scenario
- more data is probably needed.},
  author       = {Jansson, Olof},
  keyword      = {Analytics,Machine Learning,Microsoft Azure,Condition Monitoring,Predictive Maintenance},
  language     = {eng},
  note         = {Student Paper},
  title        = {Lifetime prediction of sealing component using machine learning algorithms},
  year         = {2016},
}