Reliability analysis of fabric structures

Abstract

This PhD thesis demonstrates a reliability analysis methodology to solve the safety issues existing in the current structural design of fabric structures: The safety coefficients proposed by different countries and academic institutes are not consistent, and the leading structural safeties are obscure and require a justification. A reliability tool specific for fabric structures is developed to estimate the structural safety and justify the safety coefficients for structural design based on the variations of the design variables like loads, material strength. The research work includes three main parts: the first one aims at a finite element formulation proposed for a highly accurate and efficient deterministic analysis. Four element types have been compared and discussed, and the linear strain triangle coupled with Dynamic Relaxation algorithm are shown to be most efficient with the satisfactory accuracy, The second part is focus on a probabilistic methodology to identify and analyze the material uncertainties based on the experiment data. The probabilistic models to qualify the variation in the fabric strength and Young’s modulus under uniaxial tension are demonstrated, and a practical algorithm to determine best data-fit distributions is also presented. The third part is the reliability formulation which consists of first order reliability method(FORM) and the finite element method based on the six node linear strain triangles. The analytical method with the principle of chain rule is applied in deriving the gradients of the limit state functions. The sensitivities of the structural safety corresponding to different uncertainties are compared and analyzed through numerical examples. Finally a case study based a realistic design example is undertaken, and safety factors for loads, materials, and other design elements are justified and discussed. This thesis demonstrates that the safety coefficients currently used in the fabric structure design may not be either economic or safe. Based on the uncertainty information of the design elements, the appropriate safety factors required for the structural safety standards can be evaluated using the reliability tool, and then an optimized design decision in consideration of safety and cost can be subsequently determined.