Research On Improvement Of Space Fault Tree

System reliability theory is an important basic theory in the saftey science.It plays an important role in various industries,especially for complex systems such as industrial,transportation,medical,military,et al.But with the development of information science and data science,the system reliability analysis is facing some new problems.It includes the reliability analysis under the influence of multiple factors,relationship analysis between the reliability and the influence factors in the fault data,and the stability analysis of the system reliability.These problems require the collaborative research and resolution with the saftey system engineering theory and the data science and information science.The dissertation develops and improves the existing space fault tree theory,and combines cloud model and factor space theory to solve the above problems.The space fault tree can analyze the system reliability under the influence of multiple factors.The cloud model can represent the discretization,randomness and fuzziness of the fault data.Factor space theory can analyze the logical relationship between fault and influence factors.Meanwhile,cloud model and factor space can also deal with fault data.The cloud model and factor space are introduced into the space fault tree,so that the modified space fault tree has the ability of the analysis,reasoning and presentation for the reliability and influence factors considering the multiple factors and fault data.The dissertation is divided into four parts.(1)Introduce the cloud model to modify the space fault tree.Taking the fault probability to measure the reliability,the cloud space fault tree inherits the ability of the SFT to analyze the influence of multi-factor on reliability,and also inherits the data uncertainty ability of cloud model.So that the cloud space fault tree is suitable for the analysis and processing of the actual fault data.(2)Establish the method of causality analysis between the reliability and influence factor in space fault tree.Construct the random variable decomposition of reliability data.Research the causal concept extraction and the background relation analysis method,and dimension reduction method for influence factors and fault data compression method.(3)Under the space fault tree theory,we put forward the concepts of factor effect path and factor effect history.Attempt to use the stability theory of motion system,and describe the stability of reliability system.(4)Under the space fault tree theory,some methods of system reliability analysis are proposed.We propose a similarity calculation method of cloud model based on envelope;Propose a fuzzy comprehensive evaluation method to consider the influence of different factor values on system reliability;The reliability evaluation method is modified under the influence of work environment factors.The dissertation is a theoretical research.Based on the space fault tree,the ability of fault data processing and logical analysis is improved,and the theory of reliability change description and stability is developed.The dissertation protrudes the processing procedures with some simple examples.The research provides useful theoretical methods for the development and application of information science and data science in the safety system engineering.It also provides a new basic theory for the reliability research in the safety system engineering.The dissertation has 34 figures,17 tables and 197 references.