We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Non-destructive online condition monitoring and trend prediction of lubricating oil in a steam turbine.
- Authors
Xiaojian Xu; Xinping Yan; Chenxing Sheng; Chengqing Yuan; Baobao Song
- Abstract
This paper proposes a non-destructive online condition monitoring system for the lubricating oil of a steam turbine that does not damage the machine or disturb its normal operation. The monitoring system can collect various oil parameters through the use of different types of sensor and display the variation trends of each parameter in real time. The data collected by the online condition monitoring system can be shared with distributed analysis centres via the Internet and then further deeply analysed. The relationship between the working load of a steam turbine and the number of solid particles over 4 pm in the lubricating oil is expressed by a piecewise function, which is then applied in the fault diagnosis of the steam turbine. The variation trends of moisture are extracted and predicted using a qualitative trend extraction method and a sliding window. The Akaike information criterion (AIC) is used to maintain the balance between the complexity and accuracy of the prediction model and to determine the most suitable length of the sliding window. Experiments conducted on the No 3 steam turbine in a power plant indicate that the condition monitoring system can successfully detect faults in the steam turbine and that the future trend of moisture can be correctly predicted by the prediction model.
- Subjects
LUBRICATING oils; STEAM-turbines; NONDESTRUCTIVE testing; TURBINES; AUTOMOTIVE chemicals
- Publication
Insight: Non-Destructive Testing & Condition Monitoring, 2016, Vol 58, Issue 12, p649
- ISSN
1354-2575
- Publication type
Article
- DOI
10.1784/insi.2016.58.12.649