Research On Reliability Based On Uncertainty And Method Of System State Assessment

As hi-tech products with rich functions are heavily used in daily life and industrial production, the robustness and security of products are attracting growing attention. Since the reliability and performance of a product are measured by its stability and safety, methods to comprehensively evaluate the reliability and performances of products are intensively researched and practiced. With the developments of studies and applications, limits are gradually discovered in methods to assess the reliability and system states of products and these limits are mainly restricted in the following aspects: 1) in researches on the reliability of random uncertainty, an improved grasp of system information extends types of uncertainty from being a single random uncertainty to an integration of the fuzzy uncertainty, grey uncertainty and some other uncertainties, rendering the traditional reliability methods inefficient in the analysis of failures triggered by product uncertainty. Innovations are required to more effectively tackle problems with product reliability; 2) In the comprehensive evaluation of the system state, the comprehensive evaluation of state system based on AHP method and its derivative methods are the most practiced methods. Yet despite the many advantages and widespread usages of these methods, room for improvement still exists in the conversions of evaluation results and the efficiency of the evaluation.Based on the above discussions, the uncertainty of reliability in product performance degradation, the fuzzy system reliability in discrete time state, the system reliability of the unascertained uncertainty, the rapid methods of comprehensive evaluation of system status, the qualitative transformation method of system state evaluation are researched in this paper. The contents of the paper are as follows:1)Given the many uncertainties in the performance-degradation failure of products,three theories,fuzzy theory,blind number theory and cloud theory,are introduced in the research of product reliability that involves uncertain information. Research work is conducted in the following three aspects.(1) The threshold of performance degradation is traditionally considered a fixed value,yet in actual fact the degradation fuzzily falls between a certain range. Therefore a fuzzy threshold of degradation is proposed and a fuzzy threshold based reliability analysis method is achieved(2)Apart from the random uncertainty, several other uncertainties co-exist in the process of product degradation. so the blind number theory is referred to and a blind theory based practice is invented, which proves to be a satisfying solution to the coexistence of uncertainties.(3)Considering the randomness and fuzziness in the reliability research, the cloud theory is incorporated. combined with the Monte Carlo method and the minimum membership threshold, a numerical simulation method,CloudMonte Carlo approach is established based on data of performance degradation and the problem, the concurrence of the random and fuzzy uncertainty in numerical simulation is properly resolved.2) Integration operation is inevitable in the reliability computation of fuzzy systems and a great part of the computation of the reliability distribution functions could hardly arrives at their primitive functions. in view of this inconvenience, a discrete method is put forward to compute the reliability of fuzzy systems. this method works through a discretization of the operation time of products and switches the common practice in the numeration of fuzzy reliability to a discrete practice, which,as illustrated in the final example,claims its calculation precision when with enough discrete points.3) An incomplete understanding of information which arises from artificial reasons makes the traditional random theory useless in solving problems of such type. Through the introduction of unascertained theory in the calculation of system reliability, the influence of load on the system component failure is analyzed based on generalized stress-strength interference model and structural model of system reliability(series and parallel structures). According the method, the calculation method of system reliability is obtained based on failure independence and failure correlation in unascertained theory of system component and the calculation of system reliability is realized about the dependent and related failure of components of the system in considering the unascertained uncertainty.4) Due to a lack of theoretical support in the quantificationof the evaluation result of system state, a qualitative state conversion method of comprehensive evaluation result of system state is proposed based on the fuzzy inverse transform. First, the quantitative results of system state is calculated by system state comprehensive evaluation which is based on fuzzy AHP method, then, the quantitative results of system state comprehensive evaluation are converted into the qualitative results via the fuzzy inverse transformation, solving the problems of rough processing and no theoretical support in the traditional method of qualitative transformation.5) Considering the inconvenience in the evaluation of system state comprehensive ranging from the requirement of several people’s participation to complex processing to a lack time efficiency, an evaluation method of system state is proposed. First, the AHP method is used to obtain the relative weight of underlying element about the top level evaluation index, then, the membership degree of underlying element on the state level can obtained by the method of grey correlation, last, the final state grade of system can achieved though the calculated value for the compound operation; it realizes the participation system state level rapid evaluation in a minority staff and reaches the goal of rapid evaluation.