Point-cloud acquisition in CFRP composites using ultrasonic location estimation with the phase-shift reference pulse

Abstract

Wrinkles in carbon fiber reinforced plastics (CFRP) degrade the mechanical performance. The numerical study on wrinkles requires internal structural point cloud for the meso-structure reconstruction. This paper explores the ultrasonic location estimation with the phase-shift reference pulse for the point-cloud acquisition. The point-cloud acquisition using pulse-echo signals is a time-delay estimation which is a multi-reflection and multi-delay problem. We theoretically simplify the multi-reflection problem as a single-reflection problem and propose a normalized-cross-correlation-peaks (NCCPeaks) method for the multi-delay problem. The proposed method can eliminate some amplitude decrease and balance the smearing and frequency-dependent effects. Simulation shows that the proposed method has the maximum detectable depth without missing inter-ply estimations and a stronger noise resistance than the instantaneous phase method. An experiment is implemented on the 32-ply CFRP sample with an induced wrinkle. The estimation results of the instantaneous phase and NCCPeaks methods are almost the same at the first thirteenth layers. After the thirteenth layer, the NCCPeaks method has a stabler estimation and lower false estimation index than the instantaneous phase method.