Fault Classification in a Reciprocating Compressor and a Centrifugal Pump Using Non-Linear Entropy Features

Abstract

This paper describes a comparison of three types of feature sets. The feature sets were intended to classify 13 faults in a centrifugal pump (CP) and 17 valve faults in a reciprocating compressor (RC). The first set comprised 14 non-linear entropy-based features, the second comprised 15 information-based entropy features, and the third comprised 12 statistical features. The classification was performed using random forest (RF) models and support vector machines (SVM). The experimental work showed that the combination of information-based features with non-linear entropy-based features provides a statistically significant accuracy higher than the accuracy provided by the Statistical Features set. Results for classifying the 13 conditions in the CP using non-linear entropy features showed accuracies of up to 99.50%. The same feature set provided a classification accuracy of 97.50% for the classification of the 17 conditions in the RC.

Scale-Fractal Detrended Fluctuation Analysis for Fault Diagnosis of a Centrifugal Pump and a Reciprocating Compressor

“…Previous research performed in our research group has been reported concerning the proposal of several features useful for fault diagnosis in centrifugal pumps and reciprocating compressors. Research reported in [ 49 ] investigated the proposal of several feature sets useful for fault classification using classical machine learning models. Specifically, the research compared a statistical feature set composed of 12 features ( Statistical ), an information entropy feature set composed of 15 features ( InfoEntropy ), a non-linear entropy feature set composed of 14 features ( Entropy ), and the concatenation of all features previously mentioned, composed by 41 features Allfeat .…”

“…A scheme showing the sensor location concerning the internal components of the reciprocating compressor is shown in Figure 4 b. Concerning the accelerometer sensors, previous research using this test rig [ 49 ] has shown that sensor A1 , which is located close to discharge valve 1 of the first stage, provides excellent results concerning the classification of the valve faults. For this reason, we chose accelerometer A1 in this research.…”

Improved lightweight federated learning network for fault feature extraction of reciprocating machinery

“…Set the current model parameter of each participant as Wi , update the local model with the method of stochastic gradient descent, and get the updated parameter as W k i +1 . The stochastic gradient descent is shown in the formula (2). W k i +1 is sent to the coordinator, who uses the weighted average to get the aggregated parameter Wi+1 .…”

“…Examples of such machinery include reciprocating internal combustion engines, reciprocating pumps for high-pressure liquid transportation, and reciprocating compressors used in refining processes [1]. The safe and smooth operation of reciprocating machinery is vital for ensuring reliable industrial production [2]. Therefore, the study of reciprocating machine failures holds significant practical importance [3].…”

“…In recent years, some scholars have studied the feature extraction of vibration signals of diagnostic reciprocating machinery [1,[8][9][10]. Medina et al performed feature extraction and classification of 17 valve faults in reciprocating compressors [2]. Three different sets of features of the vibration signals were collected, which were 14 nonlinear entropybased features, 15 information entropy-based features and 12 statistical-based features.…”