The Research Of Structural Hybrid Reliability Analysis Based On Probabilistic And Non-probabilistic

Uncertainties widely exist in practical engineering structures design and analysis,and how to deal with these uncertainties legitimately is the key to ensure structuralreliability analysis correctly. Probabilistic model can describe the uncertainty of theparameters, but it requires a large number of samples statistics, and a smallerparameter error may lead to a large deviation to the calculation results. So it limits theapplication of probabilistic methods in practical engineering. Non-probabilistic convexmodel use interval or ellipsoidal model to describe the uncertainty variable, and itrequires less sample information to create a more convenient uncertainty model. But inthe actual reliability analysis of structures, it is often related to the fuzziness of thefailure criteria, and this leads a great influence on the reliability analysis results of thestructures. Based on probability and non-probability model, and considering fuzzyfailure criteria, this paper aims at doing some researches and trials on the investigationof hybrid reliability. The major work of this dissertation are listed as follows：Firstly, based on the probability-interval hybrid uncertainty model, a structuralreliability analysis method considering fuzzy failure criterion is proposed. The mixedreliability model of the probability-interval structure considering fuzzy failurecriterion is equivalently transformed into a mixture model only with random andinterval variables through fuzzy stochastic theory. And an efficient calculation methodfor solving the equivalent maximum fuzzy failure probability is proposed. Finally,some numerical examples are provided to demonstrate the effectiveness of theproposed method.Secondly, based on the probability-convex hybrid uncertainty model, a structuralreliability analysis method considering fuzzy failure criterion is proposed. A hybridcomputing model is constructed based on the multi-dimensional ellipsoid convex setconsidering structural fuzzy failure criteria. And an efficient optimization iterationcalculation method is used to obtain the maximum fuzzy failure probability value ofthe mixed uncertainty model.Thirdly, based on the fuzzy random reliability model, a structural reliabilityanalysis method to calculate the fuzzy failure probability by response surface methodis proposed. The polynomial response surface is used to approximate the structuralperformance function, and the response surface function is iterated for several times.Finally, the fuzzy failure probability is calculated. At last, two numerical examples andan engineering application example are provided to demonstrate the effectiveness ofthe response surface method in fuzzy random reliability analysis...