Actual Problems in Machine Building 2024 Vol.11 N1-2

Actual Problems in Machine Building. Vol. 11. N 1-2. 2024 Innovative Technologies in Mechanical Engineering ____________________________________________________________________ 38 APPROACH TO AUTOMATION DESIGN OF TECHNOLOGICAL PROCESSES FOR PROCESSING PARTS BASED ON GRAPHS Yanyushkin A.S.1, D.Sc. (Engineering), Associate Professor, e-mail: yanyushkinas@mail.ru Yanyushkin A.R.1, Post-graduate Student, e-mail: andreyyanyushkin@gmail.com Mikhalev O.N.2, Ph.D. (Engineering), e-mail: mih_tm@mail.ru 1I.N. Ulianov Chuvash State University, 15 Moskovsky Prospekt, Cheboksary, 428015, Russian Federation 2Moscow, Russian Federation Abstract Many problems are solved when designing technological processes. Some of such problems are difficult to formalize, for example, designing a part processing route, assigning optimal cutting modes, etc. These problems are solved using a creative approach, which is characteristic of humans and is difficult to automate. However, the rapid development of information technology demonstrates high capabilities in automating such tasks. It is possible to use graph neural networks and generative technologies to build a part processing route. In this case, the part is represented as a graph, which is directly fed to the network and with the help of which the network can get a complete picture of the part. Then it becomes possible to find similar parts, its individual elements and technical processes, on the basis of which recommendations for optimal parts processing are formed. And also synthesize the information found, which can help build a processing route for a new part. The introduction of this approach into various CAD systems can help developers solve many other technical design problems. Keywords automation of technological process design, CAD/CAM/CAPP-systems, part graph, graph neural network, 3D model recognition.

RkJQdWJsaXNoZXIy MTk0ODM1