Algorithms for Pre-Processing of Eddy-Current Signals when Inspecting Heat-Exchanger Pipes of NPP Steam Generators

Abstract

Efficient algorithms for pre-processing of eddy-current signals that are typical when inspecting the condition of metal in the heat-exchanger pipes of steam generators of NPPs with VVÉR-440 and VVÉR-1000 reactor facilities are described. Flaws in these pipes can be reliably detected and classified, with their geometrical parameters evaluated, if one can tune out from multiple interfering factors and disturbances that affect the signal and emerge due to a number of structural irregularities in the test object (geometrical noise (pilger noise), spacer and antivibration grids, bends), specific detector-design features, and noises in the measuring tract of eddy-current apparatus. The tuning-out algorithms have been tested with experimental signals and incorporated in the PIRATE (Program for Identification & Recognition of defects through signal Analysis in Testing by Eddy currents) software intended for analyzing multifrequency signals.