Identification and analysis of current signal characteristics of nuclear power pump units based on start-up process

Monitoring and identification are the effective ways to ensure the safe and reliable operation of nuclear power pumps. In order to monitor and identify the operating status of nuclear power pumps with different impellers effectively, this paper uses six types of impellers for analysis, and thirteen current signals are collected for each impeller in the start-up process of the experiment. Signal preprocessing is performed by resampling, combined with Fast Fourier Transformation (FFT), Variable Mode Decomposition (VMD), Empirical Mode Decomposition (EMD), and VMD-EMD joint method for feature analysis, and it explores the relationship between characteristic signals and operating conditions, revealing the variation patterns of current signals in the start-up process of different impellers under different operating conditions. The results show that the amplitude of the current signal increases with the increasing of the flow rates and the impeller diameter in the start-up process. The EMD method has an error of less than 20% under low flow rates, but the range of the error is 36%-47% in the efficient working area, with a maximum of nearly 50%. The VMD method has an accuracy of about 80% in an efficient working area, but has a significant error at low flow rates, reaching up to 50%. The VMD-EMD method can solve the problem of low accuracy of results, and the accuracy is guaranteed to be above 80%. The accuracy is above 90% within the efficient range of 0.8Q₀-1.2Q₀. This method can effectively achieve fault diagnosis.