中文摘要：

为了准确地分析具有初始几何缺陷的单层网壳结构整体稳定性,运用随机缺陷模态法、一致缺陷模态法和N阶特征缺陷模态法,对4个不同矢跨比的K8型单层网壳进行了近1300例弹塑性荷载-位移全过程分析,探讨不同分析方法的合理性和可行性。研究表明：随机缺陷模态法能较为科学地评估初始几何缺陷对结构稳定性的影响,但计算量较大;对稳定承载力系数样本进行统计特性分析时,空间样本数量n不应小于100;运用随机缺陷模态法确定网壳结构最终的稳定承载力系数时,建议采用＂3σ＂原则;采用最低阶屈曲模态模拟初始几何缺陷分布,求得的稳定承载力并非最不利,其保证率得不到有效保证;N阶特征缺陷模态法能够通过较少的计算量,得到满足＂3σ＂原则要求的稳定承载力,并能较为合理、安全地评估网壳结构的稳定性能;在运用N阶特征缺陷模态法时,建议N=20。

英文摘要：

To accurately analyze the integral stability of single layer latticed shells with initial geometric imperfections, nearly 1300 elasto-plastic load-displacement analyses of K8 single layer latticed shell were carried out to investigate the accuracy and feasibility of different analysis methods. Four different rise-to-span ratios of K8 single layer latticed shell were considered, and the analysis methods employed herein were the random imperfection mode method, the consistent mode imperfection method and the N-order eigenvalue imperfection mode method. The results show that the random imperfection mode method can evaluate the influence of initial geometric imperfections on structural stability more reasonably, but the calculation cost is quite large. To obtain the statistical characteristic of the stability bearing capacity coefficient samples, the number of space samples should be no less than 100. ＇3σ＇ principle is suggested to determine the stability bearing capacity coefficient by using the random imperfection mode method. Utilizing the first-order buckling mode to simulate the distribution of initial geometric imperfection may fail to obtain the most unfavorable stability bearing capacity. The N-order eigenvalue imperfection mode method can generate a reasonable stability bearing capacity which meets the requirement of ＇3σ＇ principle with less calculation, and moreover, it can evaluate the stability performance of single layer latticed shell reasonably. The N is suggested to be 20 when using the N-order eigenvalue imperfection mode method.