Condition Monitoring with Time Series Data Based on Probabilistic Model
- Resource Type
- Conference
- Authors
- Soh, Jaehyun; Kim, DaeEun
- Source
- 2021 24th International Conference on Electrical Machines and Systems (ICEMS) Electrical Machines and Systems (ICEMS), 2021 24th International Conference on. :2630-2634 Oct, 2021
- Subject
- Fields, Waves and Electromagnetics
Power, Energy and Industry Applications
Transportation
Analytical models
Time series analysis
Hidden Markov models
Focusing
Maintenance engineering
Probabilistic logic
Data models
Gaussian mixture model (GMM)
Data selection
Condition monitoring
Condition-based maintenance (CBM)
Prognostics and health management (PHM)
- Language
- ISSN
- 2642-5513
As many systems become automated, system maintenance is becoming more critical. It is important always to monitor the system condition to maintain the system more efficiently and stably. In this paper, we propose a probability-based algorithm that analyzes time-series data of a complex system. We evaluate various system conditions with high accuracy by analyzing critical data among time-series data with GMM-based probability.