Font Size: a A A

Research On Multi-stage Bayesian Evaluation Method Of Reliability Growth For Complex System

Posted on:2017-02-27Degree:MasterType:Thesis
Country:ChinaCandidate:L LiFull Text:PDF
GTID:2272330485988170Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
Quality development is the way for the prosperity of the nation and the strategy of powerful nation. Quality reflects the comprehensive strength of a country, is the embodiment of the core competitiveness of enterprises and industries and the degree of national civilization. Reliability is one of the inherent attributes of quality, our country’s products are severely restricted to participate in the international market competition due to low reliability. For large complex systems with high cost, large scale, long cycle of development and construction and high required reliability, if the reliability is not fully guaranteed in the course of its construction and operation, it will cause economic losses to the country and people that are difficult to estimate, therefore, improving the reliability of the products is of great significance to the development of our country at the present stage.In development and trial run stage of products or systems, in order to achieve the predetermined reliability index, the allocation of resources by the system arrangement of resources is needed according to the predetermined reliability value, which is called the reliability growth management. Reliability growth management can effectively shorten the development cycle, reduce development costs, and reduce the number of reliability tests, it is particularly significance for large complex systems with high cost, high reliability, large scale and long development cycle. The large complex system studied in this paper has the characteristics of small system level sample size in integration and trial run stage, and the integration and trial run stage process has many stages, the traditional reliability growth evaluation model is not applicable, at this point, we must reestablish the reliability growth model and combined with the engineering practice to evaluate the reliability. In this paper, on the basis of other existing research, two multi-stage Bayesian reliability growth evaluation methods for continuous and discrete complex systems are proposed to solve the problem of reliability growth evaluation of complex systems.Firstly, this paper introduces the background of the topic selection, the characteristics of large complex system, the purpose and significance of this paper, research status and so on, the 2nd chapter introduces the related concepts and indicators of reliability, AMSAA model and its research progress, Bayesian analysis method and its research progress. In the 3rd chapter, we use the NHPP process to build the reliability growth model for single subsystem to solve the modeling problem of the small sample size of the failure data for continuous complex system, then, the information transfer method based on the equivalent transformation growth factor is presented to solve the problem of the reliability growth information transfer between stages, finally, the hyper parameters of prior distribution of the model can be solved by maximum entropy method and the relationship of failure rate prediction of the later stages. In the 4th chapter, we use the discrete AMSAA model to build the reliability growth model for single subsystem to solve the modeling problem of the small sample size of the failure data for discrete complex system, then, the information transfer method based on the growth factor given by failure probability of binomial distribution is presented to solve the problem of the reliability growth information transfer between stages, finally, the hyper parameters of prior distribution of the model can be solved by maximum entropy method and the relationship of reliability prediction of the later stages. In the 5th chapter, summary and prospect is carried out.
Keywords/Search Tags:reliability growth, large complex system, Bayesian method, growth information, small sample size
PDF Full Text Request
Related items