Methods of diagnostic of pipe mechanical damage using functional analysis, neural networks and method of finite elements

Abstract

The main goal of the study is to analyze methods and diagnose mechanical damage to the pipeline using functional analysis, neural networks and the finite element method. In the work, mathematical formulations of the corresponding geometrical inverse problems of the theory of shells on reconstruction of defects of lateral surface are formulated according to measurement data obtained from sensors located in a given section of the shell. The statement was given and a method for solving inverse geometric problems for a shell of Tymoshenko type was developed. The authors have offered methods for solving inverse geometric problems of identifying volumetric and crack-like defects in extended underground structures and pipelines based on the analysis of responses to unsteady elastic-wave perturbations using the mathematical apparatus of wavelet signal transformation, the finite element modeling method and intelligent software system based on neural network.

Keywords

• ballistics
• inverse problems
• regularization method
• variation principle
• euler equation