Research On Several Key Techniques Of Pulsed Eddy Current Nondestructive Testing

Pulsed eddy current (PEC) is a new nondestructive testing (NDT) technique developed in recent years. In view of more flaw information can be get from detection of complex structure by using wideband exciting, it has become a hot research area in aviation NDT. On the basis of analyzing the state of the art of the research on PEC technique both inland and overseas, several key techniques about the dimension evaluation of crack, imaging detection of corrosion and classification of different simultaneous occurrence flaws are studied in this paper. The brief structure of the research and the novel approaches are as follows:An improved pulsed eddy current NDT method is proposed. The exciting field is transformed into a uniform eddy current field by using new structure of probe, thus the improved PEC method has the self-differential feature. The pick-up sensor measures the effective magnetic field, which consists of the field mainly generated by the flaw. As a result, it is not demanded to subtract a reference signal from the PEC transient response signal and to implement signal synchronization in hardware design, which makes the process in signal analysis very simple and testing sensitivity very high. This method can apply to the detection of both crack and corrosion.The main purpose for evaluation of cracks is to get the information about length and depth. A new method for evaluation of the length of crack is proposed, which utilizes the corresponding relation between maximum and minimum values of PEC peaks in scanning wave and the two ends of crack to get the length information. Compared with conventional PEC methods, it is dispensable to establish a perplexing mathematic model while the high precision of detection is obtained.The zero-crossing time of PEC time domain signal is extracted to test the depth of cracks. However, there is a different variation in zero-crossing time for the surface and subsurface cracks, as a result of which the classification of different kinds of cracks must be carried out beforehand. Based on the principle of PEC, surface cracks mainly affect the high frequency components and subsurface crack mainly affect the low frequency components, so the PEC signals can be analyzed in frequency domain, which can eliminate the influence of noises during measurement. A method of frequency spectrum separating point (FSSP) is proposed for flaw classification, which enhances the accuracy in classification of defects. The agreement between theory and experiment shows that the present method is valid and applicable.With the purpose of overcoming problems that signal changes variably and eigenvalue is difficult to pick up in the testing of corrosion, a new bevel array probe is proposed, experimental results show that the new probe effectively improves the characteristics of the induction signal and greatly increases the measurement precision. The method of FSSP is used to classify corrosion, a feature is extracted for the detection of corrosion, and an inversion mathematic formula to inversely evaluate the depth and volume is derived.Array pulsed eddy current (APEC) is used to detect corrosion. In order to solve the problem of low identification accuracy in conventional NDT that only extractes features in the time domain, on the basis of PEC that has a broadband frequency, features are extracted both in time and frequency domain. A method based on principal component analysic (PCA) is proposed to enhance the accuracy in identification of corrosion fringe, the agreement between theory and experiment shows that the present method is correct.It is also needed to detect the depth and volume after the corrosion fringe was ascertained, multi-sensor data fusion is carried to improve the precision of detection. An improved method for consensus multi-sensor data fusion is proposed. The symmetry of confidence distance measure matrix is achieved by defining a new confidence distance measure, and weighted average consensus fusion algorithm is used to accomplish the data fusion process. The imaging results of corrosion are presented at last.In the course of real-life detection of fuselage, the paint thickness variations or small variations in fastener head protrusion may lead to the lift-off. Corrosions as well as lift-off brings about similar influence on PEC signals and makes signal interpretation very difficult. Time-Frequency analysis (TFA) method is applied to the analysis of PEC signals with the aim of discriminating the occurrence of corrosion from the simultaneous variations of lift-off, smoothed pseudo Wigner-Ville distribution (SPWVD) is used to get the signal spectrum map. The method of extracting the lowest energy area as the classification eigenvalue is proposed, and a good classification result is acquired.