Study on the construction and application of Cloudization Space Fault Tree

• Published in: Cluster computing Vol. 22; no. Suppl 3; pp. 5613 - 5633
• Main Authors: Li, Sha-Sha, Cui, Tie-Jun, Liu, Jian
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.05.2019; Springer Nature B.V
• Subjects: Characteristic functions; Computer Communication Networks; Computer Science; Continuity (mathematics); Decision making; Design optimization; Electric components; Fault trees; Kernel functions; Monte Carlo simulation; Operating Systems; Processor Architectures; Randomness; Reliability aspects; System reliability; Uncertainty; Cloud model; Space fault tree; Safety system engineering; System reliability
• ISSN: 1386-7857; 1573-7543
• DOI: 10.1007/s10586-017-1398-y

The Space Fault Tree (SFT) is a theoretical and technical framework proposed by the author in 2012. The SFT measures the reliability with fault probability, and analyzes the relationship between system reliability and influencing factors. The characteristic function (CF) of SFT is constructed by the relationship between the component fault probability and the influencing factors, and is the basis of the SFT. The system fault data is different from the general monitoring data, which has large discreteness, randomness and fuzziness, that is, uncertainty. The existing characteristic function is a continuous function with finite discontinuities. The function can be considered as a kernel function of the fault data, but it is difficult to express the uncertainty of fault data. To this end, using cloud model (CM) to transform the characteristic function, so that it has the ability to express the uncertainty of data, called the cloudization characteristic function (CCF). The cloudization SFT (CLSFT) is constructed by using the CCF, and enables the relevant theory and methods of SFT to express the data uncertainty, thus the expression of the fault data characteristics and rules are more accurate. Firstly, the construction process and rationality analysis of CLSFT are given. Secondly, the concepts of SFT are reconstructed by cloud model. Use these definitions and methods to analyze the fault data of a simple electrical component. The relationship is studied between the component fault probability and the using time and using temperature. The results reflect the discreteness, randomness and fuzziness of the fault data to a large extent. In summary, the paper provides the reference for analyzing and controlling the uncertainty of the fault data and reliability in practical application.