Off-Design Operation and Cavitation Detection in Centrifugal Pumps Using Vibration and Motor Stator Current Analyses

Area of Science:

• Mechanical Engineering
• Industrial Monitoring

Background:

• Centrifugal pumps are vital industrial components requiring accurate operational diagnosis for reliability.
• Lack of flowmeters and pressure sensors hinders determination of optimal efficiency point (BEP) in many applications.

Purpose of the Study:

• To investigate the detection of off-design operation and cavitation in centrifugal pumps using accelerometers and current sensors.
• To evaluate the effectiveness of vibration and stator current signals for pump operation diagnosis.

Main Methods:

• Collected simultaneous vibration (three-axis accelerometer) and stator current (Hall-effect sensors) data.
• Employed signal processing techniques: wavelet thresholding, variational mode decomposition (VMD), Park vector modulus transformation, and marginal spectrum for feature extraction.
• Evaluated seven machine learning classification algorithms for identifying off-design conditions and cavitation.

Main Results:

• Both vibration and current signals proved effective for operation diagnosis.
• Combining both signal types yielded the most reliable diagnostic results.
• Machine learning algorithms successfully identified off-design operation and cavitation.

Conclusions:

• Accelerometers and current sensors are viable tools for diagnosing centrifugal pump operational states.
• Integrating vibration and current data enhances the accuracy and reliability of pump diagnostics.
• This approach offers a practical solution for monitoring pumps lacking traditional sensors.