Numerical Modeling Of Laser Ultrasound In Layered Systems Using Finite Element Method And Determination The Materials Parameters Using The Neural Networks

In this paper, by using the finite element method (FEM), a theoretical model of laser-generated ultrasound thermoelastically in coating/substrate systems is studied and established. And using the same method, the generation of laser ultrasonic wave and its propagation, waveform characteristics are also studied. As well as, with artificial neural network inverse determination material parameter method, theoretical analysis and numerical simulation is studied.In accordance with classic thermal conduction equation, this paper analyzed the changing of material's optical and thermal physical properties under laser radiation, a result has been obtained: transient temperature field inside the material is being distributed in a very small localized region, where great temperature gradient exists quite unevenly; a method of linear elastic finite element numerical simulation laser thermal stress in a quasi-static process is suggested: we conclude that, as time passes, axial direction stress wave will propagate towards the inner part of the material and radial direction stress wave will propagate in radial directions; further, we conclude that ultrasonic wave generated in the material is the mutual result of surface thermal source and body thermal source, namely the mutual result of surface thermal dipole and normal force.Since the temperature field induces the stress and displacement fields and the effect of the stress and displacement fields on the temperature field is negligibly small. the sequential field-coupling is used in the present modeling. An optimized finite element is developed to simulate the laser thermoelastic generation and propagation of surface acoustic waves in coating/substrate systems based on a well understanding of the influences of element size and integration time step on the solution stability. Besides the surface skimming longitudinal waves, a normal dispersive surface acoustic wave is observed in slow coating and fast substrate combination, while an anomalous dispersive surface acoustic wave is observed in fast coating and slow substrate combination, which confirm the validity of this model. Furthermore, a pseudo surface acoustic wave is observed in fast coating and slow substrate combination by decreasing the size of the laser spot, which allows generating higherfrequency than the cutoff frequency of this system.Based on the time frequency analysis method of Pseudo Wigner-Ville Distribution, we analyzed transient displacement response in a coating/substrate system. Using smoothed PWV distribution, transient laser ultrasonic wave's time frequency dynamic characteristics is provided. On the time frequency plane, we have given wave's mode, location, shape and other spreading characteristics. And according to group delay, group velocity can be estimated from single waveform.An inverse method based on artificial neural network (ANN) is presented to determine the elastic properties of films from laser-generated surface waves. The surface displacement responses are used as the inputs for the ANN model; the outputs of the ANN are the material parameters of the film. Levenberg Marquardt algorithm is used as numerical optimization to speed up the training process for the ANN model. The materials parameters are not recovered from the dispersion curves but rather directly from the transient surface displacement. It is found that this procedure is very efficient for determining the materials parameters of layered systems.The returns in this paper may provide theoretical base for laser ultrasonic theory and numerical simulation research, as well as lossless inspection technology and widening its application scope.