Computational-Heuristic Algorithm for Tomographic Solution of Industrial Flaw Detection Problems

This paper presents the results of the development and study of the application of a computational heuristic algorithm for industrial tomography based on the use of a reference sample of an industrial product. Thus, based on a priori information about the reference sample of the product, it is possible to build a local tomographic reconstruction from limited initial data due to the selective sensitivity of the object under study to probing radiation. This approach is consistent with the principle of "strong methods", which has recently proven itself well in the field of artificial intelligence. The results obtained form the basis for the development of an appropriate information and computing technology, including algorithmic and software for computer processing and interpretation of measurement data. The paper also presents the designed architecture and software prototype with a graphical interface that implements local tomographic reconstruction based on the proposed algorithms. The presented experimental results testify to the efficiency and speed of the proposed algorithm in comparison with the classical algorithm of computed tomography. The proposed algorithm also makes it possible to detect defects of small thickness, when the classical tomography algorithm cannot cope. From an applied point of view, this study is of interest in the heuristic search for a defect in a product model and allows us to conclude that it is promising in the field of industrial flaw detection.