Digital Signal Processing Technology Of Ultrasonic Echo Signal Model

Among the five kinds of NOT ways (UT, RT, MT, PT, ET), Ultrasonic Testing technology is being applied widely because of its advantages, such as good orientation, strong penetrating ability, higher energy and harmlessness.Digital signal of ultrasonic echo is a kind of signal depending on time, which includes deterministic signal and stochastic signal. A great deal of information is included in signals, such as organizational status of materials, the distribution, dimension and type of the defection, the morphology, mechanical property and stress state of the microscopic structure and so on. The information is reflected by dependent variables changing with time.Digital signal processing is a developing discipline. By use of computers or special processing devices, series are processed through digital pattern to get more contented signal formats.Traditional UT was conducted by hand. The operator can not analyse signals acquired correctly, because they judge waveforms shown on the detector of defects by their own experiences, which results that the positive connection between signals and objects can not be certified. To overcome the above-mentioned shortcomings, we construct math-models of ultrasonic signals in time domain and concentrate information on limited parameters to extract information easily, through digital signal processing technology. The model can be used not only to describe inherent changing laws and statistical properties, but also to forecast processes, monitor devices and diagnose faults. Models are more representative than time series or characteristic values.In the research, at first, 38CrMoAl samples with the same size have been heat-treated separately, these heat treatments are normalization, quench, quench & low temper, and quench & high temper separately; subsequently, digital signals of ultrasonic echo in different heat-treated states are acquired by ultrasonic testing installation; finally, the models of the exponential-periodicity function and the ARMA(p,q) time series are constructed by digital signal processing technology. The research results show that original signals in different heat-treated states can be simulated by corresponding models perfectly, and the effect of state discernment is striking.