A Multi-Index Fusion Adaptive Cavitation Feature Extraction for Hydraulic Turbine Cavitation Detection

Area of Science:

• Mechanical Engineering
• Fluid Dynamics
• Signal Processing

Background:

• Cavitation in hydraulic turbines causes mechanical damage, reducing efficiency and lifespan.
• Accurate cavitation detection is vital for operational reliability and performance.
• Extracting cavitation features is difficult due to noise and signal complexity.

Purpose of the Study:

• To propose a novel method for adaptive cavitation feature extraction and detection.
• To overcome challenges posed by noise and signal non-stationarity in hydroacoustic signals.
• To enhance the accuracy and reliability of cavitation monitoring in hydraulic turbines.

Main Methods:

• A multi-index fusion adaptive variational mode decomposition (VMD) algorithm was developed.
• The number of VMD decomposition layers was adaptively determined using cavitation indicators.
• Cavitation features were selected based on frequency characteristics for different cavitation degrees.

Main Results:

• The proposed method effectively retains cavitation information and improves feature extraction quality.
• Incipient and supercavitation were successfully detected.
• Experimental validation demonstrated high detection accuracy and a low false alarm rate.

Conclusions:

• The multi-index fusion adaptive method offers a robust solution for cavitation detection in hydraulic turbines.
• This approach enhances operational reliability and maintains energy conversion efficiency.
• The findings are significant for predictive maintenance and performance optimization of hydraulic machinery.