Application of Model-Based Estimation to Time-Delay Estimation of Ultrasonic Testing Signals

Time-Delay-Estimation (TDE) has been a topic of interest in many applications in the past few decades. The emphasis of this work is on the application of model-based estimation (MBE) for TDE of ultrasonic signals used in ultrasonic thickness gaging. Ultrasonic thickness gaging is based on precise measurement of the time difference between successive echoes which reflect back from the back wall of the test piece. The received echoes are modelled by Gaussian pulses and the desired system response is estimated using Gauss-Newton and Space Alternating Generalized Expectation Maximization (SAGE) algorithms. In addition to the model-based estimation approach, five other TDE techniques including peak-to-peak measurement, cross-correlation, cross-correlation with interpolation, phase-slope, and cross-correlation with Wiener filtering are also considered and compared with the SAGE. The main advantage of the SAGE algorithm, in addition to its higher accuracy, is its ability to deconvolve the overlapping echoes.