Study On Detection Method Of The Internal Defects In Carbon Fiber Reinforced Plastic

With the development of science and technology, performances of carbon fiber reinforced plastic composite material are improving rapidly. It is widely used in several areas. However, it is made of matrix and reinforcing material and has many uncertainty factors during the manufacturing operation, there are many kinds of defects in its internal structure during the process of manufacturing. The existence of these defects has a bad effect on performance evaluation, at the same time, the application of carbon fiber reinforced plastic composite material is restricted by these defects. Even worse, these defects lead to the entire structure out of use and cause great economic losses. It is important to study the defects of carbon fiber reinforced plastic composite material as well as the inspection and recognition of different defects.In recent years, ultrasonic phased array technology has widely used in industry due to its flexible, convenient and efficient features. In this paper, detection method of carbon fiber reinforced plastic composite material was studied. Firstly, through the beam computation module and defect response module in CIVA, the detection scheme was determined. In the simulation, according to the actual situation, set sample parameters, including the size and properties of the sample. Besides, the size and the center frequency of the probe and the type of the wedge were selected later.With the defect response, the feasibility of the experiment was proved.Finally, through the simulation, the experiments were conducted and obtained the A scan and B scan of three defects respectively. Based on the ultrasonic phased array system, three common defects(de-lamination, inclusion and de-bonding) in the carbon fiber reinforced plastic composite material is inspected. There are a few differences between B scan of three defects through the analysis of the experiment result but the result was not precise enough. Therefore, time domain signal and frequency domain signal of A scan waveform were analyzed. But there were no significant differences. An effective method for the above problems was put forward and verified.Three common defects in carbon fiber reinforced plastic composite material is identified by wavelet packet analysis and BP neural network method. Ultrasonic signal belongs to the non-stationary transient high frequency signal and the waveletpacket analysis is very suitable for these signals. With excellent nonlinear mapping ability, BP neural network technology was widely used in the field of pattern recognition area. The original A scan signals from these materials were analyzed by wavelet packet transform and the characteristic values of these samples were extracted. BP neural network was built and trained for identifying these defects. The recognition rate could reach 95.7%. The result shows that ultrasonic phased array technology can improve the inspection efficiency obviously and has the good imaging effect. BP neural network has good recognition rate of de-lamination, inclusion and de-bonding defects in carbon fiber reinforced plastic composite material.