SUPPRESSING ULTRASONIC GRAIN NOISE USING NONCOHERENT DETECTION

L Ericsson , T Stepinski and M Gustafsson

First Int. Conf. on NDE in Relation to Structural Integrity for Nuclear and Pressurised Components, Amsterdam, The Netherlands, Oct. 20-22 1998

Abstract:

Ultrasonic nondestructive testing of metals with coarse material structure is difficult due to the scattering from the large grains. In pulse-echo inspection the backscattering from the grain boundaries appear in the received ultrasonic images as clutter, often referred to as grain noise. The effect becomes very distinct for ultrasonic waves with lengths in the same range as the grain size. Consequently, the most suitable frequency range for the ultrasonic transducer utilized for the inspection is determined by the material. Thus, in order to obtain adequate results during inspection a special transducer should be manufactured for each material composition. In practice, this approach is neither technically nor economically feasible; instead a suitable transducer must be selected from a limited set. The resulting ultrasonic images will in many cases be characterized by high grain noise level or low temporal resolution. An alternative approach is to utilize a transducer with a frequency range as wide as possible and then apply advanced digital signal processing in order to adapt the effective frequency range to the inspected material. In the paper, such a signal processing algorithm, based on the concept of noncoherent detection, is presented. The performance of the algorithm is illustrated using ultrasonic images acquired from specimens made of cast stainless steel and copper with large grains. The material structures are similar to those found in nuclear power plants and in canisters intended for long term storage of nuclear waste.

Source:

Postscript, 282 K ;