Classification of parts by artificial intelligence methods when choosing a mathematical model for solving plastic flow problems

To select an adequate mathematical model for a specifi c problem of continuous medium mechanics (CMM), it is often necessary to resort to classifi ers that allow dividing blanks or parts by certain characteristics. Usually, design and/or technological characteristics of parts are considered. Functional, parametric, design features of these parts, including geometric shape, can be selected as characteristics. The paper describes the possibility of constructing a classifi er of parts using artifi cial intelligence methods for further use of the classifi er when selecting a mathematical model for solving a number of CMM problems, including the problem of free fl ow in a thin layer or fl ow with the imposition of restrictions in one or more directions. For computer modeling, the method of machine learning of neural networks is used to select parts by geometric characteristics and combine these objects into the required category laid down in the program algorithm by the operator. Based on such “artifi cial selection”, a general classifi cation procedure is created. At the end of the computer modeling, the classifi er developed using artifi cial intelligence was verifi ed using ready-made 3D models and models created by a generative neural network algorithm. The proposed work examines the stages of creating a classifi er, its main parts and computer modeling methods, including using artifi cial intelligence methods.