缺陷圆柱壳体外压屈曲的仿真分析

Simulation analysis of buckling of defective cylindrical shell

摘要:

基于Ansys workbench有限元仿真软件,对含初始缺陷的圆柱壳体进行了外压屈曲的仿真研究。首先进行了理想圆柱壳体的特征值屈曲分析,得出前30阶线性屈曲失稳模态作初始几何缺陷。然后基于单一模态缺陷法和组合模态缺陷法建立具有初始几何缺陷的模型,利用有限元仿真分析的非线性稳定算法和弧长法进行非线性屈曲分析。最后设计圆柱壳体静压屈曲实验,得出了临界屈曲载荷。将非线性屈曲分析结果与实验结果、承压结构工程分析方法计算结果作对比分析,结果表明:弧长法与非线性稳定算法相比,预测精度更高；承压结构工程分析方法求出的结果非常保守,临界屈曲载荷远低于实验结果；第1阶模态缺陷不是对圆柱壳体承压能力最不利的初始缺陷,最不利初始缺陷的模态阶数一般为低阶；模型的初始缺陷幅值大小对临界屈曲压力有较大的影响,但对最不利初始缺陷的模态阶数没有影响。2种初始缺陷构造方法都是可行的,但低阶多模态组合缺陷模型具有更精准的预测精度,误差约为3.55%。

Abstract:

Based on ANSYS, this paper conducts a simulation study on the external pressure buckling of a cylindrical shell with initial defects. First, the eigenvalue buckling analysis of the ideal cylindrical shell is carried out, and the first thirty order linear buckling modes are obtained as the initial geometric defects. Secondly, the single modal defect method and the combined modal defect method are used to establish the model with initial geometric defects. The nonlinear buckling analysis is performed by the non-linear stabilization algorithm and the arc-length method of the finite element method. Then, the critical buckling load is obtained by the static pressure buckling experiment of the cylindrical pressure shell. Finally, the results of nonlinear buckling analysis are compared with the experimental results and the calculation results of pressure-bearing structure engineering design methods. The results show that the arc-length method has higher prediction accuracy than the non-linear stability algorithm. The results obtained by the engineering design method of pressure-bearing structures are conservative, and the critical buckling load is much lower than the experimental results. The first-order modal defect is not the initial defect that has the greatest impact on the critical buckling load, the order of the most unfavorable initial defect is generally low. The amplitude of the initial defect in the model has a great influence on the critical buckling pressure, but it has little influence on the modal order of the most unfavorable initial defect. The above two initial defect construction methods are effective, but the low-order multi-modal combined defect model has higher prediction accuracy with an error of about 3.55%.