A High Performance Ultrasonic System for Flaw Detection

In ultrasonic Nondestructive Evaluation (NDE), high frequency acoustic waves are used to test the integrity of materials. The detection of flaw echoes in the presence of high microstructure scattering noise is a challenging problem that requires advanced signal processing methods such as statistical analysis and pattern recognition algorithms. In this study, we designed and implemented a reconfigurable, high performance and low cost ultrasonic NDE platform based on Xilinx ZYNQ SoC. The system can generate high voltage pulses for exciting the ultrasonic transducers, receive the low voltage ultrasonic backscattered echoes, process the acquired data, and transmit and store the processed data to a host computer. in this study, we used machine learning algorithms as an alternate to conventional target echo recognition methods. In particular, Multilayer Perceptron Neural Network (MLPNN) is designed for ultrasonic flaw echo detection. The input to MLPNN is segments of backscattered signals, and regions within the Split Spectrum Processing (SSP) 2D distribution. The experimental results show that MLPNN can detect the flaw echo in the backscattered signal with very high precision.