A Measurement Model Of Reliability Based On The Cloud Theory

Reliability measure gives a quantitative estimation of objects with some certain reliability method or model using processed data which is acquired by measuring generalized stress and generalized intensity of the object. Reliability measure is a basic problem in reliability engineering. A variety of uncertainties including randomness, fuzziness and incomplete credibility coexist in reality. Even though this situation is common, it is not fully researched. The measure results achieved by existing measurement methods are expressed by a specific value or range, which cannot fully describe the reliable performance trend of the products.There are some reliability measure methods for the condition that all kinds of uncertainties coexist. The research and comparison of these theories are listed. Then the interval valued Fuzzy sets based on Gaussian cloud is proposed, which can handle randomness, fuzziness and incomplete credibility at the same time. The method is based on conventional reliability, cloud theory and fuzzy theory. Three examples offering simulation diagrams for comparison are listed. What’s more, the reliability measure results are described by distribution functions.The modeling process is as follows: First, make the fuzzy stress or intensity variables’ incomplete credibility information cloudy by using the theory of cloud. And get the Gaussian cloud formed interval-valued Fuzzy sets which are described by the upper and lower membership functions at different credibility. Then, deal with upper and lower membership functions to make them have the mathematical characteristics of probability density function. In this way, the upper and lower limit values can be calculated with conventional stress-intensity interference model and a reliability interval can be achieved. At last, the reliability measure results curve which consists of a series of reliability intervals and their corresponding credibility can be obtained.The method not only can deal with a variety of uncertainties comprehensively, but also can keep more information in reliability measure results.