Napovedovanje odpovedi stroja za proizvodnjo tesnil z metodami strojnega učenja

MARKEŽIČ, ALJAŽ (author), Bosnić, Zoran (mentor)

Cilj diplomske naloge je izdelava sistema za napovedovanje odpovedi indu- strijskih strojev v sodelujočem podjetju. Reševanja problema smo se lotili z uporabo pristopov in metod strojnega učenja. Najprej smo se lotili preo- blikovanja in filtriranja podatkov, da jih lahko uporabimo kot učne podatke za klasifikacijske modele, nato smo izvedli učenje klasifikatorjev in ovredno- tili uspešnost modelov. V prvem pristopu smo uporabili samo nabor zadnjih časovnih podatkov in modele brez sposobnosti pomnjenja; v drugem pristopu smo uporabili zgodovinsko obogateno množico podatkov; v tretjem pristopu pa smo uporabili nezgodovinske podatke z modeli, ki imajo sposobnost po- mnjenja (LSTM in GRU). Na podlagi rezultatov smo ugotovili, da je tretji pristop najbolj uspešen.

• Language: Slovenian
• Keywords: strojno učenje, klasifikacija, časovne vrste, napovedovanje, napoved odpovedi.
• Work type: Bachelor thesis/paper
• Organization: FRI - Faculty of Computer and Information Science
• Year: 2018
• PID: 20.500.12556/RUL-102764
• Publication date in RUL: 07.09.2018

MARKEŽIČ, ALJAŽ, 2018, Napovedovanje odpovedi stroja za proizvodnjo tesnil z metodami strojnega učenja [online]. Bachelor’s thesis. [Accessed 24 March 2025]. Retrieved from: https://repozitorij.uni-lj.si/IzpisGradiva.php?lang=eng&id=102764

Secondary language

• Language: English
• Title: Predicting failure of rubber seal production machine using machine learning methods

Abstract

The goal of the thesis is implementation of a predictive system for detecting failures in industrial machines. We tackle the problem by using different machine learning approaches and methods. Initially, we transformed the received data into a representation for supervised learning. In the next step we trained the classifiers and evaluated their performance. We applied three different approaches, as follows. In the first approach we trained memoryless models without using historical data; in the second approach we extended the data with additional historical attributes; in the third approach we trained memory-retaining models (LSTM and GRU) with a non-historic dataset. On the basis of our experimental results we discovered that the third approach gives as the best results.

• Keywords: machine learning, classification, time sequence, forecasting, failure prediction.