State Evaluation of Centrifugal Compressor Unit Based on Parameter Distribution

Abstract

To evaluate the state of centrifugal compressor unit, this paper presents a method based on Logistic Regression Model and Gaussian Model. First, the improved failure mode and effect analysis (FMEA) is used to select important unit parameters on the basis of subsystem of centrifugal compressor unit. And weights of selected important parameters are determined by means of analytic hierarchy process (AHP). Thus, the state evaluation index is established. Then, the distribution models of the parameters in state evaluation index are determined according to its characteristics. Meanwhile, the corresponding parameter distribution model is constructed based on the alarm range of each parameter. Finally, the quantitative evaluation model of centrifugal compressor unit is established using the evaluation value of each parameter distribution model instead of the expert experience score. We applied the model to a motor-driven centrifugal compressor unit in a gas transmission station, and the evaluation result demonstrates the model can evaluate the state of compressor unit accurately.