Research Of AE Signal Processing And Analysis Technique

1.introduction Acoustic Emission(AE) is a kind of dynamic and nondestructive testing technique, namely test nondestructively while the inside structure defect of the component or the material or the potential defect change in sports. The characteristic of AE testing technique is that the object being tested can capably take part in the testing course, it utilizes the object internal defect to emit sound waves under the function of external force, then we can receive the position and situation of the defect by inference form the sound wave. According to the characteristics of the sound waves emitted and the external conditions which cause the AE signals, not only can find out about defect present state but also can find out about forming history of the defect, and the expansion trend that this defect will take place under the actual service condition of future, which will realize the supervision of the component or the material and measuring task. In AE testing technique, the AE source is generally very faint. It also has the characteristic of variety, suddenness and uncertainty. It is through passive testing that the AE testing technique can dynamically analyze the rupture behavior of materials and dynamically test and assess the structure integrality, so the tested signals include many components and interference. In many situations, the outside interferences and noises can be far greater than the signal which people really want, with the influence of many kinds of factors such as the propagation route from AE source to transducer, characteristic of the transducer, AE testing instrument, etc, the AE signal finally received is a compound of complicated multi-mode waves and noises. From the complicated signal, how to obtain the description of the AE characteristics people really want is the key but difficult question. 2.Research Content The aim of this paper is to use signal processing methods to separate signals and noise. In this article, based on a great deal of resources from home and foreign country, the AE basic technique is briefly generalized, and the characters of the AE signal is analyzed and discussed. Introduced the application of nondestructive testing in projects. And pointed out that along with the development of national economy, industrial production and progress of science and technology, people pay close attention to the production safety question further, so the AE testing technique as a kind of new-type dynamically nondestructive testing technique occupy a important position in nondestructive testing. And then introduced its development and application at home and abroad, put forward the content and main work of the thesis. On the basis of analyzing the AE technique's characteristics, introduced several kinds of signal processing methods at present, and the application and development of each method. Pointed out the limitation of each kind of analytical methods and difficulties faced, and analyzed its development prospect. Then introduced the basic theory of the signal processing using wavelet method. Analysed the deficiencies of traditional Fourier transformation, and pointed out that wavelet transform is the most effective method in AE signal processing for its nicer time-frequency part analyse performance. Explained the basic theory of wavelet transform and introduced the idea of multi-resolution analysis. On the basis of analyzing the AE signal's characteristics sufficiently, studied the problems of signal recovery, separation of signal and noise, character parameters'obtaining of AE signals, pattern-recognition, acoustic emission source location, etc. The main work can be described as follow: (1) Aim at the complicated transmitting process, an idea of signal recovery was proposed. That is using standard lead break signal as a pulse signal, which responding function can be regarded as the system's transfer function. The emulation testing showed that the actual signal did can be recovered. Much detailinformation about AE source could be recovered, and the disturbances among transfer process could be greatly reduced and the recognition capability greatly improved. (2) To various kinds of interference composition of the signals, wavelet transform was used to assist the noise removing. First the feasibility of wavelet using in the AE signal processing was studied. Then deeply discussed the choice of wavelet bases and maximal decompose level. The appropriate wavelet bases, which can effectively separate the signal and the noise, and the decompose level which can effectively save the characteristics of the signal were selected. (3) As to the obtainment of the AE signal character parameters, a combined method of parameters analyse and waveform analyse was used. Except the traditional method of character parameters obtaining, the coefficient of each frequency band energy to all energy was used, to improve the precision of the pattern-recognition. (4) Because the relationship of AE signal parameters and the type of the defects are always influenced by many complicated factors like the nature of the material and defect, the environment, equipments, etc, the fuzzy pattern-recognition method was used in this thesis as the pattern-recognition method of AE sources to suit the relationship of AE parameters and defect types. (5) To the question of the AE location accuracy, on the basis of analyzing the defect of traditional location methods, pointed out that because sound wave of different frequency range would decay in various degree in the course of propagating, the time and order that the sound wave reached the sensors would be influenced, then the precision of AE source location would be influenced too. By comparing the several methods of obtaining time delay between different sensors, an effective AE source location method is proposed, which is the wavelet packet analysis combined with the cross-correlation function, which can reduce the error effectively and improve the precision of AE source location.