Implementation of predictive maintenance for industrial machinery using neural networks.
Školitel
Ing. Michal Hodoň, PhD.
Študijný program
Aplikovaná informatika
Študijný odbor
Informatika
Detailnejší opis problému
The problem with using neural networks for predictive maintenance in industrial machinery is about predicting when machines will fail and planning maintenance to prevent downtime and save costs. Industrial machines are essential for many industries but can wear out and break, causing production problems. Traditional ways of maintaining machines—like doing it on a set schedule or fixing things only after they break—often aren't very effective. They can lead to unnecessary maintenance or costly breakdowns. Neural networks help by using data from sensors on the machines to spot signs of upcoming failures. These sensors track things like vibrations, heat, and pressure. The data helps train the neural networks to see patterns that indicate a machine might soon break down. The big challenge is to accurately use this data to predict problems and figure out the best time for maintenance. Using neural networks for predictive maintenance allows companies to fix things before they break, not just after. This approach makes machinery more reliable, uses maintenance resources better, and involves putting new technology and training into place. The result is a more efficient maintenance process that cuts down on unexpected machine stoppages and makes machines last longer.
Predpokladaný vedecký prínos (algoritmy, metodológie, ...)
- Predictive maintenance extends machinery life, improving performance and longevity. - Predictive maintenance ensures maintenance is only performed when necessary, optimizing the use of resources and cutting costs. - Predictive maintenance prevents breakdowns by foreseeing failures, enhancing machine reliability
Odporučány postup vedeckého bádania
Literature Review: Review of literature on neural networks and predictive maintenance for industrial equipment. Data Collection: Gather and analyze data from the defined industrial machinery, including temperature, vibrations, and other operational parameters important for predicting failures. Model Development: Develop and train neural networks based on collected data to identify patterns and predict potential failures.
Druh výskumu
aplikovaný výskum a experimentálny vývoj
Výskumná úloha, ktorej súčasťou bude riešená téma
APVV-24-0429 - Experimentálne skúmanie a počítačové modelovanie prúdenia vzduchu pri požiari v cestnom tuneli - podaný projekt APVV-24-0653 - Inovatívne riešenie pre dočasné elektrické a vodovodné prípojky v stavebníctve - podaný projekt UNIZA MVP - Nízkoprikonový senzor pre aplikácie inteligentnej dopravy
Doterajšie výsledky riešenia na školiacom pracovisku, vrátane odkazov na dostupné zdroje
Solar energy harvesting for the mobile robotic platform [electronic] / Michal Hodoň ... [et al.]. In: Innovations for community services [print, electronic] : proceedings. - 1. vyd. - Cham: Springer Nature, 2022. - ISBN 978-3-031-06668-9 (online). - s. 17-27 [print, online]. Zaradené v: SCOPUS, Spôsob prístupu: https://link.springer.com/chapter/10.1007/978-3-031-06668-9_4 Compressed sensing - a way to spare energy in WSN for UAV [electronic] / Ondrej Karpiš ... [et al.]. In: 17th IFAC Conference on Programmable Devices and Embedded Systems PDES 2022 — Sarajevo, Bosnia and Herzegovina, 17-19 May 2022. - 1. vyd. - Amsterdam: Elsevier, 2022. - s. 17-176. Zaradené v: SCOPUS ; Web of Science Core Collection Spôsob prístupu: https://www.sciencedirect.com/science/article/pii/S2405896322003433?via%3Dihub Impact of external phenomena in compressed sensing methods for wireless sensor networks [print] / Michal Kochláň, Michal Hodoň. In: Proceedings of the 2017 Federated Conference on Computer Science and Information Systems [electronic]. - 1. vyd. - Varšava: Polskie Towarzystwo Informatyczne, 2017. - ISBN 978-83-946253-7-5. - s. 857-863 [print]. Zaradené v: Web of Science Core Collection ; SCOPUS Spôsob prístupu: https://ieeexplore.ieee.org/abstract/document/8104651/ Fordal, Jon Martin, et al. "Application of sensor data based predictive maintenance and artificial neural networks to enable Industry 4.0." Advances in Manufacturing 11.2 (2023): 248-263. Zonta, Tiago, et al. "A predictive maintenance model for optimizing production schedule using deep neural networks." Journal of Manufacturing Systems 62 (2022): 450-462. Krenek, Jiri, et al. "Application of artificial neural networks in condition based predictive maintenance." Recent developments in intelligent information and database systems (2016): 75-86.
V prípade otázok sa obráťte na vedúceho témy:
hodon6@uvp.uniza.sk