Reconstruction of Defect Configuration in Plane Construction Using Numerical Modeling and Neural Networks at Positional Scanning

Abstract

AbstractThe results of modeling a plane construction with defect are presented. The aim of the work is to identify the parameters of the defect based on the use of finite element modeling (FEM) and artificial neural network (ANN). The properties and configuration of the defect differ from the layer in which it is located. The calculation of construction oscillations is carried out during its impulse loading at a certain point of the construction surface. The shape of the defect is considered as identifiable parameter of the defect. Neural networks are used as a tool for analysis. Surface waves excited by impulse loading are considered as input parameters. The results of the restoration of defect parameters obtained on the base of a combination of FEM and ANN are presented.KeywordsLayer constructionFEMImpact loadSurface wavesDefectANN