Forecasting of the Holes Shape Accuracy of Thin-Walled Body Parts Through the Application of Artificial Neural Networks

Abstract

Abstract. In this paper is proposed a method to predict deviations of accuracy of a basic shape of holes due to a thermal deformation (TD) at the design stage of the technological process. Proposed method of control of technological process consists of two stages: the first stage (auxiliary) is based on the finite element method (FEM), the second (main) -on the modeling of artificial neural network (Ann). In this paper is developed an algorithm of calculation of input and output parameters of the network using LS-DYNA. A structure of the Ann to predict and to adjust a trajectory of a movement of a tool at the preparation stage of the technological process for work pieces included in group process is invented in this article.