Stability Analysis And Bearing Capacity Prediction Of K6 Aluminum Alloy Reticulated Shell With Random Defects

As a typical spatial structure,reticulated shell has been widely used in long-span structures due to its unique structural characteristics.In recent years,aluminum alloy reticulated shell is proposed on the basis of steel reticulated shell.This article studies the nonlinear buckling behavior of rectangular tube section aluminum alloy shell through to the thousands of cases of K6 type examples,the influence of random initial defects to the shell is discussed,and the effect of random defects on the bearing capacity of reticulated shells is also quantified.At the end of this paper,an artificial neural network model for bearing capacity prediction of reticulated shells with defects is proposed.The main research contents and conclusions are as follows:(1)The linear and nonlinear buckling analysis of aluminum alloy reticulated shells with rectangular tube sections with different parameters is carried out.The results show that the linear buckling instability mode can be classified as the top unstability,buckling and overall buckling at the bottom,the nonlinear buckling instability mode can be divided into the bottom of the instability type,main ribs instability type and diagonal area instability type.(2)Node installation deviation has an obvious effect on nonlinear stability bearing capacity of rectangular tube section aluminum alloy reticulated shell,and the vertical deviation of node installation is more adverse,while the horizontal deviation node installation for structure is impact.The theory of configuration vulnerability is used to explain the finding rules.(3)The coordinate transformation formula of initial bending bar is developed,the latticed shell with initial bending of rectangular tube cross section aluminum bar is built,and the results show that the influence of bearing capacity of structure with bar bending is far less than the effects of node installation deviation.The joint action of the two control the stability bearing capacity of the structure by the installation deviation at the time point.(4)Considering the defects of the random sample size are too much,the decreasing quantity of imperfection shell is hard to described,this article is based on the statistical law of scatter data,useing double parameter Weibull distribution to describe the degree of influence of the defects.Based on the probability theory,the corresponding calculation method of reduction coefficient is established.(5)For rectangular tube cross section aluminum alloy shell bearing capacity formula is difficult to derived and described,this paper puts forward artificial neural network can be used for fitting,and then uses the trained neural network to the known parameters of geodesic bearing capacity prediction.The results of train show that the network has higher prediction accuracy through a lot of training and learning from the sample data.This network can be used as a quick method to calculate the bearing capacity of reticulated shells.(6)In order to extend the research methods in this paper,a parametric reticulated shell modeling platform named Shell Builder and two interactive defect introduction plug-ins Node Imperfection and Bending Bar have been developed.These three plug-ins make the studied problems diversified and provide technical support for further exploration of multiple types of problems.