Hilbert spectrum analysis of unsteady characteristics in centrifugal pump operation under cavitation status

Cavitation has negative influence on nuclear pump operation. In order to detect incipient cavitation effectively, experimental investigation was conducted for acquisition of current and vibration signals during cavitation process. In this research, a centrifugal pump was modeled for nuclear pump research. The data was analyzed by HHT method. The results show that Torque oscillation resulted from unsteady flow during cavitation process could result in energy variation. RMS of IMF in current signal is sensitive to cavitation occurrence. It could be regarded as the indicator of incipient cavitation. RMS variation of IMF in radial vibration signals experiences monotone increasing tendency when cavitation gets severe. Such variation regulation could be selected as the indicator of cavitation stage recognition. Hilbert-Huang transform is suitable for transient and non-stationary signal process. Time-frequency characteristics could be extracted from results of HHT process to reveal pump operation condition. The contents of current work could provide valuable references for further research on centrifugal pump operation detection. (C) 2018 Elsevier Ltd. All rights reserved.