Auto Regressive Exogenous modeling of pseudonoise thermography for the nondestructive evaluation of aerospace materials

An Auto-Regressive Exogenous model is proposed for the feature extraction in Random Photothermal active Thermography. The prosed approach is applied to the nondestructive evaluation of a Carbon Fiber Reinforced Plastic composite sample in which delamination was simulated by embedding Teflon tape pieces between the different plies of the structure. It is shown that the application of the Auto Regressive Exogenous model allows the impulse response of the sample to be estimated, thus simulating the result of a standard Pulsed Thermography experiment but by using a pseudo-noise modulated low-power heating source. Such technique opens the possibility to realize handier setup and to achieve higher SNR values than Pulsed-Thermography. For these reasons the proposed approach can find use in the nondestructive evaluation of aerospace composite components.