Optimization Of2D RVM For CFRP By Grading And Reconstruction Of3D Void Morphology

Ultrasonic attenuation method is commonly used to predict porosity in Carbon Fiber Reinforced Plastics (CFRP). Models were built by scholars at home and abroad to analyze ultrasonic scattering attenuation due to voids. The key of modeling is weather the void morphology and distribution can be described reasonably. Most models are based on the theory of isotropic medium, in which the void morphology and distribution were assumed regular. Results from theory prediction and experiment fit well only at special frequency. Based on random medium theory and statistical method, Random Void Model (RVM) was proposed to describe CFRP with voids by our research group. Voids are regarded as random disturbance in small scale on the matrix medium with average characteristics and are described by stationary random process and other parameters (correlation length, roughness factor, disturbance standard deviation). Compared with former models, the results of void morphology and predicted ultrasonic attenuation based on RVM are closer to the reality. But the model has some disadvantages at present:①For samples in which void size dispersions are large, the model can’t take both large voids and small voids into consideration. The information of large voids will be lost.②The model is planar, while the real material is three-dimensional. It also causes the deviation between theoretical results and experimental results. This paper mainly includes the optimization of2D RVM and reconstruction of3D void morphology.For a sample with the porosity P=4.08%, in which the range of void length L is5-217um and the dispersion of void length is large, the voids were graded:voids with L<50μm belonged to grade1and voids with L>50μm belonged to grade2. Model for each grade was built, and then all the models were superposed together to achieve the RVM for this sample. Both large and small voids were reflected. The distribution of void size was more similar to the fact. The better description of void morphology with the use of grading enhanced the applicability of RVM.Five regions were selected with the help of ultrasonic C-scan system from unidirectional CFRP laminate. The range of attenuation coefficient is2.49-4.42dB·mm-1and the porosity is from0.88%to3.04%. Based on102serial sections for each region achieved using metallographic method,3D void morphologies were reconstructed. Along the fiber the void size is large like root, and the voids mainly distribute on the interfaces of liber layers. The morphology is complex. Some voids have branches. In region2and10, the Voids distribute almost everywhere. In the3D RVM, the void morphologies are similar to the reconstructed results. The void morphology can affect the scattering attenuation seriously. The reconstruction of voids can help us grasp the morphology characteristics of the real voids clearly. It’s of great significant for guiding modeling. It provides the modeling basis for making clear the ultrasonic scarring mechanism of voids and analog computation.