Algorithmic treatment of structure deviations in gamma ray tomography data sets

The measurement of void fraction distribution in electrically heated fuel element bundles suffers from the problem of structure shifts due to thermal expansion and magnetic pinching of the constructive elements during heating. To cope with such displacement problems for subchannel void fraction determination we developed special image processing algorithms. The core feature of these algorithms is a precise tracking of the rods in the images by means of a pattern recognition algorithm. In this way rod displacements can be determined automatically. This allows us to automatically generate a sub-channel integration mask that is subsequently used to determine the void fraction values in the unaffected sub-channel areas.