Model rane procjene kvarova rotacijske opreme primjenom dubokog strojnog učenja

Kolar, Davor

prikaz prve stranice dokumenta Model rane procjene kvarova rotacijske opreme primjenom dubokog strojnog učenja

Download
PDF 3.5 MB

doctoral thesis

Model rane procjene kvarova rotacijske opreme primjenom dubokog strojnog učenja

Zagreb: University of Zagreb, Faculty of Mechanical Engineering and Naval Architecture, 2019. urn:nbn:hr:235:773286

Kolar, Davor

University of Zagreb
Faculty of Mechanical Engineering and Naval Architecture

Cite this document

APA 6th Edition

Kolar, D. (2019). Model rane procjene kvarova rotacijske opreme primjenom dubokog strojnog učenja (Doctoral thesis). Retrieved from https://urn.nsk.hr/urn:nbn:hr:235:773286

MLA 8th Edition

Kolar, Davor. "Model rane procjene kvarova rotacijske opreme primjenom dubokog strojnog učenja." Doctoral thesis, University of Zagreb, Faculty of Mechanical Engineering and Naval Architecture, 2019. https://urn.nsk.hr/urn:nbn:hr:235:773286

Chicago 17th Edition

Harvard

Kolar, D. (2019). 'Model rane procjene kvarova rotacijske opreme primjenom dubokog strojnog učenja', Doctoral thesis, University of Zagreb, Faculty of Mechanical Engineering and Naval Architecture, accessed 27 March 2023, https://urn.nsk.hr/urn:nbn:hr:235:773286

Vancouver

Kolar D. Model rane procjene kvarova rotacijske opreme primjenom dubokog strojnog učenja [Doctoral thesis]. Zagreb: University of Zagreb, Faculty of Mechanical Engineering and Naval Architecture; 2019 [cited 2023 March 27] Available at: https://urn.nsk.hr/urn:nbn:hr:235:773286

IEEE

D. Kolar, "Model rane procjene kvarova rotacijske opreme primjenom dubokog strojnog učenja", Doctoral thesis, University of Zagreb, Faculty of Mechanical Engineering and Naval Architecture, Zagreb, 2019. Available at: https://urn.nsk.hr/urn:nbn:hr:235:773286

Cite this item: https://urn.nsk.hr/urn:nbn:hr:235:773286

Please login to the repository to save this object to your list.

Metadata

Title	Model rane procjene kvarova rotacijske opreme primjenom dubokog strojnog učenja
Title (english)	Deep learning-based early fault diagnosis model for rotary machinery
Author	Davor Kolar
Mentor	Dragutin Lisjak (mentor)
Committee member	Danko Brezak (predsjednik povjerenstva)
Committee member	Dragutin Lisjak (član povjerenstva)
Committee member	Danijel Pavković (član povjerenstva)
Committee member	Tomislav Staroveški (član povjerenstva)
Committee member	Tomislav Šarić (član povjerenstva)
Granter	University of Zagreb Faculty of Mechanical Engineering and Naval Architecture Zagreb
Defense date and country	2019-11-15, Croatia
Scientific / art field, discipline and subdiscipline	TECHNICAL SCIENCES Mechanical Engineering
Universal decimal classification (UDC)	658 - Business management, administration. Commercial organization
Abstract	Trend razvoja industrije prema konceptu 4.0 u području održavanja promovira ideju donošenja odluka o aktivnostima održavanja temeljenih na analitici velike količine podataka. Tradicionalni algoritmi strojnog učenja omogućavaju opisivanje i otkrivanje složenih odnosa unutar velike količine podataka, no pritom zahtijevaju ekspertno znanje iz domene problema tijekom obrade i analize ulaznih signala s ciljem određivanja relevantnih značajki na temelju kojih bi se s dovoljnom točnošću opisala zakonitost izlaznog signala. U okviru ove disertacije, a imajući u vidu trenutno stanje i trendove u području održavanja u industriji, istražena je mogućnost računalnog generiranja značajki ulaznih vibracijskih signala u cilju rane procjene nastanka i klasifikacije kvarova rotacijske opreme uslijed različitih tipova vibracijskih opterećenja. Pripremljen je i opisan eksperimentalni postav te kreiran i proveden eksperiment na temelju plana. Simulirano je osam različitih stanja opreme te su prikupljeni podaci za dvije različite brzine vrtnje. Modeliranje je provedeno primjenom algoritama dubokog strojnog učenja s ciljem točne klasifikacije tipa kvara, a na temelju automatski generiranih značajki vibracijskog signala. Tijekom procesa modeliranja, optimizirani su hiperparametri modela primjenom Bayesove optimizacije. Konačni rezultati dokazuju mogućnost primjene dubokih struktura neuronskih mreža za automatsko generiranje značajki ulaznog signala i klasifikaciju stanja opreme.
Abstract (english)	Recent trends focusing on Industry 4.0 concept arise a data-driven fault diagnosis as key topic in condition-based maintenance. Traditional data-driven techniques of fault diagnosis require signal processing for feature extraction, as they are unable to work with raw signal data, consequently leading to need for expert knowledge and human work. In the context of this dissertation, considering the current state and trends in the field of maintenance in the industry, the possibility of computer-generated vibration signal input features was investigated in order to assess early on the occurrence and classification of rotational equipment failures due to different types of vibration loads. An experimental set-up was prepared and described, and a plan-based experiment was created and conducted. Eight different equipment conditions were simulated and data were collected for two different rotational speeds. The modeling was performed using deep machine learning algorithms with the aim of accurately classifying the fault type based on the automatically generated vibration signal features. During the modeling process, the hyperparameters of the model were optimized using Bayesian optimization. The final results prove the possibility of applying deep neural network structures to automatically generate input signal features and classify equipment states.
Keywords
Keywords (english)
Language	croatian
URN:NBN	urn:nbn:hr:235:773286
Study programme	Title: Mechanical Engineering and Naval Architecture Study programme type: university Study level: postgraduate Academic / professional title: doktor/doktorica znanosti, područje tehničkih znanosti (doktor/doktorica znanosti, područje tehničkih znanosti)
Type of resource	Text
File origin	Born digital
Access conditions	Open access
Terms of use
Repository	Repository of Faculty of Mechanical Engineering and Naval Architecture University of Zagreb
Created on	2020-01-28 09:31:12