Cavitation Detection and Condition Monitoring in Centrifugal Pumps

Cavitation in centrifugal pumps arises when local pressure falls below the vapour pressure of the fluid, forming and subsequently collapsing vapour bubbles within the impeller and volute. This phenomenon undermines hydraulic performance, induces noise and vibration, accelerates material erosion and may precipitate mechanical failure. Condition monitoring seeks to detect onset and progression of cavitation through non-intrusive measurements of vibration, acoustic emissions, fluid-borne noise and electrical signals. Early recognition allows operators to adjust operating parameters—such as flow rate, speed and net positive suction head—to mitigate damage and extend service life. Recent advances encompass sophisticated signal processing techniques, from time-frequency decomposition to bispectral analysis, as well as data-driven algorithms including convolutional neural networks and transfer learning. Integration of sensor arrays at strategic locations combined with online diagnostics has enabled real-time monitoring and predictive maintenance in a wide range of industrial applications, from water treatment facilities to petrochemical plants and power generation. Continued research converges on enhancing sensitivity under noisy, variable operating conditions, reducing false alarms and broadening applicability to pumps of differing designs and scales.

Research from Nature Portfolio

No recent Nature Portfolio content available.

Cavitation Detection and Condition Monitoring in Centrifugal Pumps publication trend

The graph below shows the total number of articles in cavitation detection and condition monitoring in centrifugal pumps across all publications each year (not limited to Nature Index journals).

Technical terms

Cavitation: Formation and collapse of vapour bubbles in regions where local fluid pressure falls below vapour pressure.

Net Positive Suction Head (NPSH): Difference between inlet pressure and vapour pressure, expressed as head, critical for preventing cavitation.

Fluid-borne noise: Acoustic emissions transmitted through the fluid medium, used to infer flow anomalies such as cavitation.

Wavelet packet decomposition (WPD): Signal processing method that splits a signal into frequency sub-bands for detailed time-frequency analysis.

Vibration bispectrum: Higher-order spectral technique capturing phase relationships between frequency components, enhancing detection of non-linear interactions.

Transfer learning: Application of a neural network pretrained on one task to a related task, reducing training data requirements and improving generalisation.

References

1. Detection of Inception Cavitation in Centrifugal Pump by Fluid‐Borne Noise Diagnostic. Shock and Vibration (2019).
2. Cavitation Detection in Centrifugal Pump Based on Interior Flow‐Borne Noise Using WPD‐PCA‐RBF. Shock and Vibration (2019).
3. Cavitation Analysis in Centrifugal Pumps Based on Vibration Bispectrum and Transfer Learning. Shock and Vibration (2021).