Ultrasonic Nondestructive Evaluation Of Steel Aging Microstructure Based On AR Model Parameters

Ultrasonic nondestructive testing (UNDT) is used extensively. It has been applied to many fields, such as avigation, petro-chemical engineering, vehicles, ship manufacture, nuclear, bridge architecture and so on. Ultrasonic nondestructive evaluation (UNDE) is the principal part to evaluate materials and a new developing direction of Nondestructive testing (NDT). It's well worth of studying and has extensive perspective.Under the premise of not injuring material capability, material nondestructive evaluation (NDE) is a subject evaluating completion of objects. By interaction between ultrasonic waves and elastic medium, it has carried much information of material, which are received in the form of ultrasonic echo signals. Therefore, analyzing of the signals becomes important.Signal is a series of number which is generated by D/A technique. As the carrier of information, it attracts information, materials and NDT researchers. Because it's random, it is the most important question for materials study and NDT researchers that we apply mathematics statistics theories and methods-Statistics signal processing techniques to analysis signal statistic methods and get effective picking-up and optimizing materials property information of material finally. Time series analysis is a method of studying stochastic data series by using mathematics statistics and stochastic process theories. It generally makes use of model methods to carry out, such as autoregressive (AR) model, autoregressive moving average (ARMA) model. The research technique route and tenet is to pick-up characteristic parameters by analyzing relevant ultrasonic echo signals and then found the relation between the structure configuration and evolvement process, till realize UNDE and UNDT of materials finally.The idea of this paper shows that based on the time series theory, the autoregressive model parameters which have been extracted and optimized are applied to the ultrasonic nondestructive evaluating of 20 steel aging micro structure at high temperature. The results show that it is feasible to evaluate 20 steel aging microstructure at high temperature using auto-regression model parameters, and the recognizing ratio is 84%. In addition, the recognizing ratio is 75% and 83.3% for Cr5Mo and 30Mn2SiV separately.