Research on fault diagnosis for reciprocating compressor valve using information entropy and SVM method
| Autor: | Houxi Cui, Rongyu Kang, Xinyang Lan, Laibin Zhang |
|---|---|
| Rok vydání: | 2009 |
| Předmět: |
Engineering
Reciprocating compressor business.industry General Chemical Engineering Energy Engineering and Power Technology Pattern recognition Small sample Management Science and Operations Research Machine learning computer.software_genre Fault (power engineering) Industrial and Manufacturing Engineering Support vector machine ComputingMethodologies_PATTERNRECOGNITION Control and Systems Engineering Sample problem Pattern recognition (psychology) Artificial intelligence Safety Risk Reliability and Quality business computer Gas compressor Food Science |
| Zdroj: | Journal of Loss Prevention in the Process Industries. 22:864-867 |
| ISSN: | 0950-4230 |
| DOI: | 10.1016/j.jlp.2009.08.012 |
| Popis: | A method of compressor valve fault diagnosis using information entropy and SVM is proposed in this paper. The main obstacle in the fault diagnosis focuses on the low non-linear pattern recognition performance and small sample number. Therefore, the information entropy, which is flexible and tolerant to the non-linearity problem, is applied to analyze the characteristic of the signals. SVM is employed in the fault classification because of its superiority in dealing with smaller sample problem. The information entropy features and the optimization test of the SVM model are detailed analyzed. The experiment shows the good performance of the information entropy SVM method in compressor valve fault diagnosis. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect