Date of Award
2-10-2023
Document Type
Master Thesis
Degree Name
Master of Engineering (Research)
Department
IMaR Research Centre
First Advisor
Dr Daniel Riordan
Second Advisor
Krishna Panduru
Third Advisor
Abdullah Caliskan
Abstract
With the advent of Industry 4.0, Manufacturing operations have become more connected and automated. Modern industrial monitoring technology can monitor the performance of both the sensor utilised and the material under monitoring. This has led to a trend where legacy machines without this technology are being removed from production and/or not included in new reconfigurations of production lines, well before their productive lifespan had ended. A key driver for this is the difficulty in retrofitting Predictive Maintenance (PdM) systems to legacy machines, in part due to the machine-to-machine variation present in older machines. Accounting for this variation in PdM systems requires extensive amounts of data-capturing, fine-tuning and customisation for each machine before deployment in production. This thesis proposes the development of a data capture system and processing framework designed for an efficient Machine Learning PdM algorithm to be retrofitted to legacy machines. Given a limited number of data samples collected from a machine to be maintained. The aim is to predict a failure or/and maintenance time by making use of the difference between the characteristics of the variation of the healthy and unhealthy data collected from the machine. The goal is to measure the healthiness of the machine by using a Siamese network, trained with a public data set, and fine-tuned with data samples obtained from machines with similar characteristics. Although the use of different training and testing data sets coming from completely different sources is involved, reasonable results are obtained due to the proposed Transfer Learning technique.
Recommended Citation
O Brien, Conor, "The Development of a One-Shot Learning Machine Learning for Industrial Use with Predictive Maintenance" (2023). Theses [online].
Available at: https://sword.cit.ie/allthe/823
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.
Access Level
info:eu-repo/semantics/openAccess
Project Identifier
info:eu-repo/grantAgreement/SFI/SFI Research Centres Programme (Programme) Phase 1 (Sub-programme)/16/RC/3918/IE/Confirm Centre for Smart Manufacturing/CONFIRM
