Reliability design of structures
Date of Issue2016
School of Mechanical and Aerospace Engineering
It is known that uncertainties or variabilities exhibit in structures. Parameters such as loading, material properties, geometric size, and material strength can be viewed as uncertainties and should be considered in structure design and analysis. Uncertainties are dealt with by applying conservative values of the uncertain quantities in the framework of deterministic design. Deterministically designed structures are however, often over-designed with unnecessary cost and weight due to the heavy reliance of the conservative values made. Such approach offers a safety measure that is only relative and is insufficient to establish the reliability of a structure. This report presents reliability-based analysis to appropriately account for the uncertainties through probabilistic and statistical considerations to evaluate the performance of a structure. Uncertainties are modelled as random variables with mean values, standard deviations, probability density and distribution functions. By defining a failure mode, termed as limit state function, reliability analysis takes these uncertainties into consideration and estimates the failure probability. Three methods discussed are: First-order reliability (FORM), Second-order reliability (SORM), and Monte carlo simulation (MCS). Each method uses the statistical data in a different manner involving various assumptions and limitations. A comparison between deterministic and reliability approaches, particularly MCS, is demonstrated through an example of a cantilever beam structure with computational software MATLAB to show that the deterministic approach using a safety factor is insufficient to provide a full picture of the reliability of a structure. A 13 truss bar case study is conducted with CAD software SolidWorks and its finite element analysis (FEA) function for analysing structures that require non-trivial computation effort and time. The procedures on the simulation for a static study is studied and results showed that the SolidWorks simulation is able to analyse the stress of the structure in a greater precision. Recommendations such as the respond surface method (RSM) and advanced non-linear dynamic FEA simulation software ANSYS can be carried out in the future.
Final Year Project (FYP)
Nanyang Technological University