Design of Large Space Structures with Imperfection Using ANN-Based Simulator

Abstract

Listen

Translate article

Based on the advanced computational plasticity and an artificial neural network (ANN) simulator, a new design strategy has been presented for large space structures with imperfections. Nonlinear system identification approach has also been greatly spread among the researchers and engineers in the past few years. The neural network simulators as a non-parametric system identification approach present a robust and efficient way to simulate the nonlinear behaviour of engineering systems. In the paper herein an artificial neural network (ANN) simulator, a general back error propagating perceptron, is use to simulate random imperfection for nonlinear dynamic analysis of large space structures. It is also desirable to search for a procedure for wind pressure calculation with accuracy and reliability. In this respect, attention is paid to the advanced computational fluid dynamics (CFD). The use of the advanced CFD analysis can help engineers to estimate the wind pressure for the design of large space structures with complex geometries. The characteristics of the new design method have been shown graphically using a full documented numerical example, which highlights the efficiency of the new simulation method. The purpose of this paper is to present a new design method, which takes into account the effects of imperfection on the resulting dynamic responses of large space structures under gravity, temperature and wind loadings.