Investigation Of Cavitation Diagnosis In A Centrifugal Pump Using Vibration Method

Centrifugal pumps are widely used in a range of applications,which has the advantages of convenient maintenance,simple structure and wide application range.Cavitation is a challenging flow abnormality,it not only limits the development trend of pump,but also affect the effective operation and stability of the centrifugal pump.In order to ensure reliable operation of pumps under the cavitation state,an understanding of the onset and mechanism of the phenomenon resulting in unsteady cavitation is necessary.If we can detect the stage of incipient cavitation in a centrifugal pump accurately,then some effective measures can be taken to reduce the harm of cavitation and to extend the service life of the centrifugal pump.At first,the test system of cavitation in centrifugal pump is introduced,including the test device and data acquisition system,and describes the experiment scheme and test steps.Data characteristics of the centrifugal pump were analyzed,and two kinds of numerical statistics methods(the root mean square value(RMS)and the power spectral density(PSD)of the cavitation induced vibration signal)are used to analysis and research on the cavitation induced vibration characteristics.Next,the random and discrete of the vibration signal lead to the traditional method of Fourier transform is not suitable for the cavitation induced vibration signal.The wavelet analysis has the advantages of self-adaptability and multiresolution,so it is more suitable for the processing of cavitation induced vibration signal.The wavelet analysis was used to detect the states of internal cavitation in this study,and the analytic procedure consist of the wavelet decomposition for the cavitation induced vibration signal and the reconstruction frequency spectrum of wavelet coefficients of each order.Finally,this study focuses on cavitation instability at normal flow rate,at which point the unsteady cavitation occurs as the net positive suction head available falls below 5.61 m.An ameliorative algorithm–united algorithm for cavitation vibration analysis is proposed based on short time Fourier transform(STFT)and Wigner–Ville distribution(WVD).The STFT–WVD method is then tested using vibration data measured from a centrifugal pump.As indicated by the relationship between vibration and suction performance,the inception and development of cavitation can be effectively detected by the distribution and intensity of the united algorithm at the testing points.The intermediate frequency components around 3.5 kHz fluctuate initially with the development of cavitation.The time–frequency characteristic is found to be conducive to the monitoring of the cavitation performance of centrifugal pumps.