Experimental investigation on centrifugal compressor blade crack classification using the squared envelope spectrum

Area of Science:

• Mechanical Engineering
• Rotating Machinery Diagnostics
• Condition Monitoring

Background:

• Centrifugal compressors are vital in modern industry, with impellers being critical components.
• Impeller blade cracks pose a significant risk, necessitating effective monitoring and classification methods.
• Existing methods struggle to detect subtle crack-induced signal changes.

Purpose of the Study:

• To develop and validate a method for classifying impeller blade cracks in centrifugal compressors.
• To utilize pressure pulsation (PP) and casing vibration signals for crack detection.
• To extract weak signal features indicative of blade cracking.

Main Methods:

• Employing pressure pulsation (PP) sensors near the crack and casing vibration sensors.
• Developing a signal processing technique based on the squared envelope spectrum (SES).
• Conducting experimental investigations to verify the classification method's effectiveness.

Main Results:

• The squared envelope spectrum (SES) method successfully extracts weak signal characteristics from PP and vibration data.
• Experimental results demonstrate the effectiveness of the proposed method for blade crack classification.
• The approach provides a reliable means for identifying and categorizing impeller blade cracks.

Conclusions:

• The squared envelope spectrum (SES) based method is an effective tool for blade crack classification in centrifugal compressors.
• This technique enhances the condition monitoring capabilities for critical rotating machinery.
• Accurate classification of blade cracks improves operational safety and reduces maintenance costs.