Reduction of Material Noise in Ultrasonic Nondestructive Evaluation Using Synthetic Frequency Diversity Algorithms

Lars Ericsson

Licentiate Thesis, Report UPTEC 94108R, October 1994.

Abstract:

Ultrasonic nondestructive evaluation of coarse materials, such as stainless steel, is often complicated by the backscattering from the material microstructure. This material noise can, however, be substantially reduced by making use of either spatial diversity or frequency diversity methods.

The thesis concerns signal processing techniques for implementing the latter approach using a single received ultrasonic response signal. The investigated class of Synthetic Frequency Diversity Algorithms process the signal in the time-frequency domain, using a filter bank followed by a statistical processor. The Split Spectrum Processing (SSP) technique, belonging to this category of algorithms, is elaborated and novel filter banks and statistical processors are developed and compared to the original SSP implementation.

The algorithms are evaluated using real ultrasonic signals, originating from test blocks with structures similar to those found in nuclear reactor vessels.

Related publications:

[Eric95] L.Ericsson, T. Stepinski,
Ultrasonic Pulse Detection with Split Spectrum Processing and Consecutive Polarity Coincidence, Proc. of the 13th Int. Conf. on NDE in the Nuclear and Pressure Vessel Industries, May 1995, pp. 201-208