Research On Complex Structure Quality Detection Technology Based On Acoustic Resonance Spectroscopy

In engineering practice, the shell structure is widely used in various industries, the quality and reliability of which is related to the safety of production. Therefore, it has great significance for ensuring the reliability of the product and prolonging the service life of the product to test and analyses the subtle changes of quality state in the kind of structure. Conventional nondestructive detection technology can only detect significant structural damage, but for the subtle changes in structure state, it is difficult to identify. According to the modal test principle, any physical object has many vibration modes, which can show the change of the shape, the size, and physical properties of structure. So, as a structure damage identification method, the vibration modal testing has been widely used. However, many vibration modal testing technology, whose response frequency is usually less than 10 kHz, can be well used for the detection of structural damage, but is not enough to reflect the subtle change. Therefore, as to that the acoustic resonance spectroscopy detection method based on the modal test principle has higher response frequency, which is sensitive to the change of the structure state, so the detection signal of it can be used to identify the subtle changes in the structural state.In this paper, a cylindrical aluminum shell structure model is taken as the experimental object, and each 20 pieces of cylindrical object which is made of tungsten, copper, steel and aluminum are selected as four built-in components of the shell model, which represent four different types of structure state. Firstly, take the ultrasonic transducer as the signal excitation and pickup sensor, excite the testing model of static structure by the steady sinusoidal excitation method, and the acoustic resonance spectroscopy signals that reflect structure state which can be taken as the dynamic fingerprint feature that show the structure can be obtained, then analyses the correlation and the variation among acoustic resonance spectroscopy of different structure state. Secondly, after preprocessing the acoustic resonance spectroscopy data, we reduce the dimension of spectrum data and analyses it using principal components analysis. Because that the cumulative contribution rate of the top three order of principal component reaches 98.2%, by improving the elliptical control chart method of the multivariate control chart, conduct the outlier identification for the top three order principal component data, using the ellipsoid control chart method which based on 3 times standard deviation. The data point whose structure state is located outside the ellipsoid is identified as outliers. And then the detection for subtle changes of structure state of shell specimen which is built in different material is completed.Based on the further research and improvement of the acoustic resonance spectroscopy detection method in the research, the method is successfully applied to the experiments that analyze and identify the shell structure subtle changes. We initially build the prototype system for the detection and identification of shell structure state, and obtain effective results. The results of the research provide effective new method for the reliability analysis and product life extension of shell structure, which has important significance for preventing fault accidents.