RESUMEN
Reciprocating compressors and centrifugal pumps are rotating machines used in industry, where fault detection is crucial for avoiding unnecessary and costly downtime. A novel method for fault classification in reciprocating compressors and multi-stage centrifugal pumps is proposed. In the feature extraction stage, raw vibration signals are processed using multi-fractal detrended fluctuation analysis (MFDFA) to extract features indicative of different types of faults. Such MFDFA features enable the training of machine learning models for classifying faults. Several classical machine learning models and a deep learning model corresponding to the convolutional neural network (CNN) are compared with respect to their classification accuracy. The cross-validation results show that all models are highly accurate for classifying the 13 types of faults in the centrifugal pump, the 17 valve faults, and the 13 multi-faults in the reciprocating compressor. The random forest subspace discriminant (RFSD) and the CNN model achieved the best results using MFDFA features calculated with quadratic approximations. The proposed method is a promising approach for fault classification in reciprocating compressors and multi-stage centrifugal pumps.
RESUMEN
The optimum pressure ratio for the stages of a multistage compression process is calculated with a well known formula that assigns an equal ratio for all stages, based on the hypotheses that all isentropic efficiencies are also equal. Although the derivation of this formula for two stages is relatively easy to find, it is more difficult to find for any number of stages, and the examples that are found in the literature employ complex mathematical methods. The case when the stages have different isentropic efficiencies is only treated numerically. Here, a step by step derivation of the general formula and of the formula for different stage efficiencies are carried out using Lagrange multipliers. A main objective has been to maintain the engineering considerations explicitly, so that the hypotheses and reasoning are clear throughout, and will enable the readers to generalise or adapt the methodology to specific problems. As the actual design of multistage compression processes frequently meet engineering restrictions, a practical example has been developed where the previous formulae have been applied to the design of a multistage compression plant with reciprocating compressors. Special attention has been put into engineering considerations.
RESUMEN
As an essential mechanical device in many industrial applications, reciprocating compressors have a high demand for operating efficiency and availability. Because the temperature of each part of a reciprocating compressor depends considerably on operating conditions, faults in any parts will cause the variation of the temperature distribution, which provides the possibility to distinguish the fault type of reciprocating compressors by differentiating the distribution using infrared thermal imaging. In this paper, three types of common fault are laboratory experimented in an uncontrolled temperature environment. The temperature distribution signals of a reciprocating compressor are captured by a non-contact infrared camera remotely in the form of heat maps during the experimental process. Based on the temperature distribution under baseline condition, temperature fields of six main components were selected via Hue-Saturation-Value (HSV) image as diagnostic features. During the experiment, the average grayscale values of each component were calculated to form 6-dimension vectors to represent the variation of the temperature distribution. A computational efficient multiclass support vector machine (SVM) model is then used for classifying the differences of the distributions, and the classification results demonstrate that the average temperatures of six main components aided by SVM is a promising technique to diagnose the faults of reciprocating compressors under various operating conditions with a classification accuracy of more than 99%.