Condition monitoring and identification are effective ways to ensure the safe and reliable operation of nuclear power pumps. However, the condition monitoring of the cutting impeller is blank. In order to effectively monitor and identify the operating status of nuclear power pump impellers corresponding to different cutting amounts. The paper collects the measured stator current signals of nuclear power pumps with 6 cutting quantities under 13 operating conditions. Variational Mode Decomposition (VMD) and Empirical Mode Decomposition (EMD) methods are utilized to analyze the state characteristics of the collected signals. The effect of different blade cutting amount on the signal characteristics of nuclear electric pump is obtained. The research results indicate as follow: when a single method is used to identify large impeller flow, the diagnostic accuracy of EMD and VMD can reach more than 90%, while Total Harmonic Distortion (THD) is less than 70%, and even less than 20% in some areas. However, for different impeller diameters and different flow rates, the identification accuracy of EMD and VMD is relatively low, only 60%. Under special working conditions, it can even be lower, with only about 50% at low flow rates between 0.2Q0-0.3Q0. EMD-VMD can accurately identify impellers of different diameters and different flow rates, and the accuracy of fault identification can be improved to over 90%, even higher than 95% in the range of 0.7Q0-1.2Q0. At the same time, the minimum flow rates of 0.2Q0 can also achieve 80% accuracy, which can effectively achieve fault diagnosis. The research results can provide data support for monitoring the operating status of self-cutting centrifugal pumps, which is of great significance for safe and stable operation.
Read full abstract