Ultrasonic thickness measurement of the internal insulation layer of metal surfaces by using signal processing schemes

There are many two-layerd components where the outer layer is a metal of high acoustic impedance and the inner (insulating) layer is a non-metal with low acoustic impedance. In most cases, it is impossible to directly access the inner layer for inspection or thickness gaging. The key problem in thickness gaging of the inner layer from the outer surface is that the backwall echo from the insulation cannot be identified in the very many backwall echoes coming back from the metal layer. In this paper, we use the signal processing techniques of wavelet transform and EMD algorithm to overcome this problem. To demonstrate the repeatability of these methods, three two-layered samples called A, B and C are prepared. The backwall echo of the insulation layer cannot be identified in any of these samples without further processing. The results indicate that both methods can facilitate the measurement of the insulation thickness, however, due to higher sensivity of the wavelet method, it is more suitable than the EMD algorithm.