Compressive sensing of ultrasonic array data with full matrix capture in nozzle welds inspection

Area of Science:

• Nondestructive Testing
• Ultrasonic Testing
• Signal Processing

Background:

• Phased array ultrasonic technique (PAUT) with full matrix capture (FMC) is crucial for high-accuracy imaging and defect characterization in welded structures.
• Large data volumes in FMC hinder real-time monitoring of nozzle welds.
• Compressive sensing (CS) offers a potential solution for data reduction in PAUT-FMC.

Purpose of the Study:

• To propose and evaluate a compressive sensing (CS) based data compression method for PAUT-FMC in nozzle weld defect monitoring.
• To investigate sparse representations and reconstruction algorithms for FMC data.
• To assess the effectiveness of CS in improving defect detection efficiency.

Main Methods:

• Simulations and experiments were conducted on nozzle welds using PAUT with FMC.
• Full matrix capture (FMC) data were compressed using compressive sensing (CS).
• Reconstruction performance was compared between Orthogonal Matching Pursuit (OMP) and Basis Pursuit (BP) algorithms. Empirical Mode Decomposition (EMD) was used to construct a sensing matrix.

Main Results:

• A suitable sparse representation for nozzle weld FMC data was identified.
• Accurate image restoration was achieved with significantly reduced measured values.
• Flaw identification was successfully maintained despite data compression.
• Experimental results showed effective defect detection, though not perfectly matching simulations.

Conclusions:

• Compressive sensing (CS) effectively reduces data acquisition, storage, and transmission for PAUT-FMC in nozzle weld monitoring.
• The proposed CS method enables accurate image reconstruction and reliable flaw identification.
• CS significantly improves the efficiency of phased array defect detection in practical applications.