Dissertation > Excellent graduate degree dissertation topics show

Reliability Modeling and Analyzing of Momentum Wheel Based on Bayesian Network

Author: LiHaiTao
Tutor: JinGuang
School: National University of Defense Science and Technology
Course: Military Equipment
Keywords: Reliability Small Sample Bayesian networks Momentum wheel
CLC: TB114.3
Type: Master's thesis
Year: 2007
Downloads: 165
Quote: 2
Read: Download Dissertation

Abstract

Momentum wheel is the key mechanical and electrical components of the long-life satellite attitude control system, with a kid like, long life, high reliability, failure data characteristics, very difficult to assess the reliability of the traditional large sample life data statistical inference method. How Small Sample Reliability Modeling and Analysis of momentum wheel failure condition is very concerned about the urgent need to address the problem of domestic aerospace units. On the other hand, although the momentum wheel field test samples is relatively small and the lack of failure data, but momentum wheel has accumulated in the process of development, production and use of a large number of prior information, including functional test, environmental test information in orbit telemetry data and engineering experience. Reliability characteristics for momentum wheel, using Bayesian network learning and reasoning techniques, fusion momentum wheel prior information to establish the reliability model momentum wheel and use the model analysis to assess the XX-3 momentum wheel and reliable level to address reliability modeling and analysis of the problem of the momentum wheel. The major contents of study are as follows: (1) by the momentum wheel failure analysis, access momentum wheel Bayesian network structure, based on a variety of ground test data using Bayesian network learning and momentum wheels, build momentum wheel discrete Bayesian network model. Using the model and Bayesian network inference techniques, assessment on the level of reliability of the momentum wheel ground conditions, proposed to affect the momentum wheel and sensitivity analysis and composite fault diagnosis method based on Bayesian network momentum wheel single fault The reliability factor. (2) in discrete Bayesian network model based on the momentum axle temperature, current performance data as well as data in orbit, the establishment of the mixed Bayesian network model based on the performance data of the momentum wheel consolidated under laboratory conditions. Between the use of the mixed Bayesian network model to assess the level of reliability of the conditions in orbit momentum wheel proposed fault diagnosis method of the momentum wheel in-orbit state, In addition, the article also performance sensitivity analysis of quantitative study of the performance and reliability of the momentum wheel relationship. (3) in mixed Bayesian network model on the basis of, time series model combined with the momentum wheel performance data to establish, build momentum wheel dynamic Bayesian network model, the model to solve the problem of the momentum wheel life and reliability expected. Meanwhile, this paper also proposed the concept of dynamic sensitivity analysis, dynamic sensitivity analysis of typical failure modes of the momentum wheel. At the end of the article summarizes the contents of the full text and further prospects.

Related Dissertations

CLC: > Industrial Technology > General industrial technology > Engineering and basic science > Engineering Mathematics > The application of probability theory, mathematical statistics > Reliability theory
© 2012 www.DissertationTopic.Net Mobile