Modal Analysis Of Centrifugal Compressor And Research On Fault Diagnosis Method

Centrifugal compressor is the important equipment in the industry. The workingenvironment is very complex, failure often happened. So operation safely and steadily is animportant security factor to the whole industrial production. It is very necessary andmeaningful to find an effective method of centrifugal compressor fault diagnosis. In thispaper, self-organizing feature map neural network was used to classify the single fault andparallel fault based on the vibration characteristics of the centrifugal compressor.The concept, classification and application of the modal analysis were discussed. Inthis paper, a large centrifugal compressor impeller3D model was made with Pro-E. Themodal analysis was made with the finite element method. The top10order naturalfrequencies and vibration modes were got. With the modal parameters, the dynamiccharacteristics of the impeller were known. It can avoid resonance and be sure the stableoperation.Several common failure types of centrifugal compressor were introduced. They arerotor unbalance, rotor misalignment, rotor bending, oil film oscillation, surge, and rotatingstall. The failure mechanism, corresponding frequency characteristics and solutions ofsome common fault types were introduced specially. And list the spectrum characteristicsof common faults.Then the structure and algorithm of the self-organizing feature map (SOM) neuralnetwork were illustrated briefly. The SOM neural network has its own advantages what isbecause of the characteristics of no supervision, self learning and lateral association ability.In this paper, the self-organizing feature map neural network was used to identifycentrifugal compressor faults.