Research On The Multi-source Information Fusion Techniques In The Process Of Reliability Assessment

Reliability information plays an important role in the assessment of product's reliability. In some circumstance, the correlative reliability information which issued form different sources can be ultilized in the process of reliability assessment when the reliability test data is scarce. The information fusion technique is an effective way to handle the multi-source information in the process of reliability assessment. It can take full advantage of the multi-source reliability information which be acquired at different time or from different place and get a more precise estimation of the product's reliability.The main task of this dissertation is to apply various qualitative and quantitative fusion methods to the fusion of different type reliability information issued from different sources. At the beginning of this dissertation, a general introduction to information fusion is given;it answers the question "what is information fusion" and covers briefly the theories, architectures, applications, history, and current status of information fusion technique. And then a framework of reliability assessment and some reliability information fusion models are put forward. Chapter 3 is an overview of some familiar fusion methods, and their applicability in the process of reliability assessment is also studied in this chapter. Chapter 4 is dedicated to the applications of Bayesian fusion method in reliability field. Prior reliability information from different sources may be represented in different forms, and then we show how to utilize different type of information in the process of reliability assessment. Also, the fusion of multi-source prior information in Bayesian method is studied in detail. Sometimes, the reliability information can not be described precisely, and then the fusion of imprecise reliability information is studied thoroughly in chapter 5. Also, multi-source reliability information fusion in the framework of possibility theory is studied, including the fusion of possibility distributions and reliability assessment based on the prior possibility distribution. Chapter 6 concentrates on the evidence theory in the field of reliability assessment. First, the evidence theory is improved and two new combination rules are proposed, it takes into account the weights of the sources in the fusion process. Then, the modelling and pooling of different type expert opinions are studied in the framework of evidence theory. Lastly, the fuzzy evidence theory is extended, and then it is applied to the qualitative reliability assessment successfully.Although the multi-source information fusion technique is still a developing theoretical project in the field of reliability engineering, but it will be certain to promote the development of the reliability assessment.