GEROTOR PUMP CAVITATION MONITORING AND FAULT DIAGNOSIS USING VIBRATION ANALYSIS THROUGH THE EMPLOYMENT OF AUTO-REGRESSIVE-MOVING-AVERAGE TECHNIQUE

Gerotor pumps, as well known, are widely used in lubrication circuits of internal combus- tion engines for their simplicity, high efficiency and low costs. In this paper an experimental characterization of a Gerotor pump is shown. The re- search is a collaboration between the Hydraulic Power Research Group (HPRG) of the Uni- versity of Naples “Federico II”and the Istituto Motori of CNR. The aim of this research is to investigate the possibility of using simple data to acquire, as vibrational data, on the pump under investigation to detect possible cavitation problem by implementing a proper mathematical procedure to this aim. As been demonstrated that these pumps are particularly subject to cavitation. The gerotor pumps are much subject to cavitate than expected, especially under par- ticular operating conditions of the engine. Cavitation is also correlate to vehicle dynamics particularly with the recent tendency to reduce the mass of oil in the sump. Therefore, the research to avoid cavitation is crucial nowadays. Consequently this study has been firstly performed with an experimentation on the pump by monitoring the delivered oil flow-rate and the adsorbed torque on the pump shaft. An accurate analysis of the pressure oscilla- tions in cavitation conditions, has been, also, conducted. Then, an accelerometer sensor has been properly located to study the cavitation with a fault diagnosis system based on vibration detection. The experimental tests have been performed on a test bench of the Hydraulic Labora- tory of the University of Naples “Federico II”in Italy. The bench allows testing the pump working by varying the shaft speed, the oil temperature, the suction and the delivery pres- sure. The main measured parameters are shaft torque, oil flow rate, mean suction and de- livery pressure and the instantaneous suction and delivery pressure. As expected, tests revealed the high influence of the suction pressure on the deliv- ered oil flow rate, while no significant influence has been noted on the adsorbed torque. Furthermore, the pressure oscillations in the pump delivery are highly influenced by the suction pressure. A non intrusive accelerometer has been installed during the experimentation on the oil pump. Since the vibration due to cavitation is the main concern of this study, the ac- celerometer has been mounted at the suction port in the radial direction. More precisely, the paper presents a fault diagnosis system based on vibration detec- tion. Firstly, a Fast Fourier Transform of the vibration signal has been computed. The in- vestigation has been made with and without the presence of cavitation varying the pump rotation speeds. Limitations due to the detecting of the on line cavitation problems by monitoring the FFT vibration spectra have been overcame by implementing an alternative method based on stochastic approach. This diagnosis method of accelerometer time series analysis based on an Auto Regres- sive and Moving Average (ARMA) method has been used to determine the pump failure. The diagnosis results have demonstrated the ability of the proposed mathematical tech- nique in the identification of cavitation phenomena proving that the proposed approach is a useful methodology to detect the presence of fault. The approach can predict, with good accuracy, pump failure in real time operation. In addition, a threshold vibration level in decibel scale is also fixed.