Model-based signal processing for laser ultrasonic signal enhancement

The use of laser-based electronics in the testing of materials and structures offers various advantages over more traditional ultrasonic methods, but is often less sensitive when applied to real materials. Although high energy laser pulses can generate large ultrasonic displacements, nondestructive evaluation requires that the ablation regime be avoided, thus limiting the amount of optical energy which may be used. For this reason, signal processing of laser generated ultrasonic waveforms detected using laser interferometers may be required to extract the desired information from a nondestructive laser ultrasonic test. A model-based signal processing technique offers a way to enhance the signal-to-noise ratios significantly for ultrasonic waveforms obtained using laser-based systems with the generation of the ultrasound occurring in the nondestructive thermoelastic regime. Under ideal conditions, good signal-to-noise ratios can be achieved using laser-based ultrasonics. However, many materials which need to be tested have less than ideal surface finishes for optical detectors. The application of signal processing to laser-based ultrasonics may provide the necessary improvement in sensitivity. Aussel and Monchalin used cross-correlation methods to extract acoustic velocities and elastic constants from noisy measurements. Once the constraints are made available through experimentation of calculation, it is possible to enhance the noisy interferometer measurements even further by generating a predicted or reference response using a propagation model that captures the essence of the displace signal to be estimated. Using estimates of the required constants, a reasonable reference response can be generated that enables significant enhancement of the measured displacement. This work uses a model-reference approach to increase the signal-to-noise ratio in noisy laser-based ultrasonic waveforms