Condition monitoring of a valve in a reciprocating compressor using machine learning approach
Publication Type:Journal Article
Source:International Journal of Applied Engineering Research, Research India Publications, Volume 10, Number 13, p.33078-33081 (2015)
Reciprocating compressors are used in industries to provide pressurized air, which in turn is used for a variety of production processes. Compressors are expected to be made available as and when required and any delay or downtime of the same will affect the production process. Machine Learning based fault diagnosis of a compressor-valve is proposed in this paper.In reciprocating compressors, valves contribute to a greater percentage of failure and a diagnostic method to detectthe cause of failure is required. Fault diagnosis followed by a remedial measure is widely welcome in industry to improve the productivity. Faulty conditions are classified using machine learning algorithms like LogisticRegression (LR), Support Vector Machine (SVM) and Random Forest Tree (RFT). Accuracy of classification of different valve conditions is improved by identifying the best statistical feature selection from Random Forest Tree.The results confirm that the proposed method can classify the valveconditions withgreater accuracy nearing 75% and reliability. © Research India Publications.
cited By 0
Cite this Research Publication
Related Research Publications
- Application of Support Vector Machine (SVM) and Proximal Support Vector Machine (PSVM) for fault classification of monoblock centrifugal pump
- Prediction of Tool Condition During Turning of aluminium/alumina/graphite Hybrid Metal Matrix Composites Using Machine Learning Approach
- Multicomponent fault diagnosis using statistical features and regularized least squares
- Automobile Gearbox Fault Diagnosis using Naive Bayes and Decision Tree Algorithm.
- Fault diagnosis of roller bearing using fuzzy classifier and histogram features with focus on automatic rule learning