Search academic texts

Information × Registration Number 0222U003451, 0119U001323 , R & D reports Title Creating a new generation of methods for fractodiagnostic of materials and structures based on the use of neural networks. popup.stage_title Head Maruschak Pavlo O., Доктор технічних наук Registration Date 24-03-2022 Organization Ternopil National Technical University named after Ivan Puluj popup.description2  Methodical instructions on the use of neural networks for fractodiagnostic analysis of mechanisms of destruction of materials and technical objects of oil and gas and aerospace complex, identification, recognition and calculation of parameters of separation wells, their further statistical analysis are offered. The topographic description of fracture surfaces developed and refined is usually based on the use of convolutional semantic neural network segmentation. A number of modern deep learning algorithms for semantic segmentation have been developed, which contain two main parts: encoder and decoder, where encoder is represented by the architecture of some of the most efficient classifiers of convolutional neural networks, such as ResNet. Methodological bases of formation of a new scientific direction on the basis of fractodiagnosis of responsible constructions on the basis of approaches of artificial intelligence are created. The methodology of digital recognition and parameterization of fractures of materials and structures has been created and successfully tested in practice, which provides quantitative assessment of morphological structures of fracture materials of samples, as well as to compare physical and mechanical conditions of their formation. A new method of fractodiagnosis was studied on a sample of 1.0 million electron microscopic images of viscous separation pits. Approbation of the method on the control sample of wells indicates its recognition accuracy at the level of 98.0%. Product Description popup.authors Andriy Babiy V. Baran Denys Ya. Valiashek Volodymyr Bogdanovych Galan Yuriy Ya. Kobelnyk Oksana Kozbur Ihor R Kozbur Halyna V Lytvynenko Yaroslav V. Lyashuk Oleh Leontiyovych Senchyshyn Viktor S. Sorochak Andrii P. Третяков Oleksandr O. Khvostivska Liliya Tsitsyura Olena I. Yasniy Oleh P. popup.nrat_date 2022-03-24 Close