Structured Non-probabilistic Robust Reliability Study

After more than half century development , reliability had becomed an important quality measure for product or structure. Now classical reliability theory is developed perfectly and has been applied in engineering extensively. But based on the usual engineering test data, tail distribution of random variable could not be available precisely. Then, it's very difficult for traditional probalistic model to analyise the reliability for the complex atructure.For the product with very low probability, simplification and computation of mechanical model could not be obtained accurately. It is necessary to search a new method of structural reliability quantitative analysis in order to complete the reliability theory.In the traditional reliability model ,the probability density function of random variable is important.So,in chapter 2, the methods to get probability density function of engineering random variable are discussed. Three methods, which include approximate method, Monte-Carlo simulation method and multiple parameter method, are compared by some examples.In chapter 3, some non-probabilistic reliability models are discussed. In engineering applications, the uncertainties of the variables are discribed by intervals engineering. First a simple non-probabilistic safe index is presented after considering the influence of geometrical non-probalistic uncertainties. Then non-probabilistic reliability method based on interval quanlity of variable is researched and fuzzy assessments of non-probabilistic reliability methods are accomplished.Robust reliability measures the safety of product or structure by the resistance of them to uncertainties. In chapter 4, robust reliability theory is studied initially. An index is proposed and a solution is provided. For multiple factors uncertainties problems, Signal-to-Noise ratio is put forward. By test, optimization of input uncertainties can be found to make structural robust.