基于动态故障树的卫星姿态控制系统寿命预测方法研究

发布时间：2018-05-07 03:13

本文选题：系统寿命预测 + 动态故障树　；参考：《南京航空航天大学》2016年硕士论文

【摘要】：伴随长寿命、高可靠性航天器的发展需求,卫星的寿命预测研究成为当前研究热点。本文针对卫星姿态控制系统及其关键部件开展剩余寿命预测研究,建立了基于动态故障树的系统失效机理模型;以关键部件动量轮为例,研究了部件带工作状态切换下的寿命预测方法;针对备份部件失效概率密度描述问题,提出了改进的备份部件失效概率密度描述方法;重点研究了系统配置及工作模式与系统在轨剩余寿命的关系。以多配置下的动量轮系统为例,对不同配置、不同工作模式下系统的寿命预测技术进行了仿真验证。首先,通过对卫星姿态控制系统进行失效工作机理分析,以动态故障树作为载体建立卫星姿态控制系统的失效机理模型。为了降低建模复杂度,采用自顶向下的方法依次对卫星姿态控制系统中的子系统进行失效机理分析与建模,进而得到卫星姿态控制系统的失效模型。其次,研究了部件带工作状态切换下的寿命预测方法。考虑到基于动态故障树模型开展卫星姿态控制系统寿命预测时,需要已知部件(底事件)的失效概率密度函数,且部件在工程应用中存在多个工况。因此,以动量轮为例,在多个单一工作状态下失效概率密度函数已知的条件下,结合Nelson假设(部件的残存寿命仅依赖于已累积的失效和当前应力,而与累积方式无关)提出了部件带工作状态切换下的寿命预测方法,并以动量轮为例开展了数值仿真验证。再次,研究了系统不同配置下的寿命预测。为实现基于动态故障树的系统寿命预测分析,需要对其进行求解(即:顶事件失效概率密度函数的求解),但动态故障树模型规模比较大,直接求解比较困难,为了降低求解的复杂度,本文采用模块化的方法进行求解,静态模块和动态模块分别采用基于二元决策图(BDD)和离散时间贝叶斯网络(DTBN)的方法进行求解。在动态模块求解过程中,针对现有备份门的求解方法中存在的备份部件失效概率密度累积和不等于1的问题,提出了改进的备份部件失效概率密度描述方法,有效地解决了失效概率密度累积和问题。并将该改进方法应用于基于动态故障树的卫星姿态控制系统失效机理模型求解,实现了系统不同配置下的寿命预测,并以多种配置下的动量轮子系统为案例开展了数值仿真验证。最后,研究了系统不同工作模式下的寿命预测。通过分析系统工作模式下相关部件的工作状态及切换情形,在部件带工作状态切换下寿命预测方法的基础上,开展了基于动态故障树模型的系统不同工作模式下的寿命预测研究,并以四斜装的动量轮子系统为案例开展了数值仿真验证。
[Abstract]:With the development of long life and high reliability spacecraft, the research of satellite life prediction has become a hot topic. In this paper, the residual life prediction of satellite attitude control system and its key components are studied, and the failure mechanism model of the system based on dynamic fault tree is established, and the momentum wheel of the key component is taken as an example. In this paper, the method of predicting the life of parts with working state switching is studied, and an improved method for describing the failure probability density of backup components is proposed. The relationship between system configuration, working mode and system residual life in orbit is studied. Taking the momentum wheel system with multiple configurations as an example, the life prediction technology of the system with different configurations and different working modes is simulated and verified. Firstly, by analyzing the failure mechanism of satellite attitude control system, the failure mechanism model of satellite attitude control system is established with dynamic fault tree as the carrier. In order to reduce the modeling complexity, the failure mechanism of the satellite attitude control system is analyzed and modeled by top-down method, and the failure model of the satellite attitude control system is obtained. Secondly, the method of life prediction is studied. In order to predict the life of satellite attitude control system based on dynamic fault tree model, the failure probability density function of known components (bottom events) is required, and there are many working conditions in engineering applications. Therefore, taking the momentum wheel as an example, under the condition that the failure probability density function is known in several single working states, combined with the Nelson hypothesis, the residual life of the component depends only on the accumulated failure and the current stress. A method for predicting the life of components with switching state is proposed, and the numerical simulation of momentum wheel is carried out. Thirdly, the life prediction of different configurations of the system is studied. In order to realize the system life prediction and analysis based on dynamic fault tree, it is necessary to solve it (that is, solving the probability density function of top event failure), but the dynamic fault tree model is large in scale and difficult to solve directly. In order to reduce the complexity of the solution, the method of modularization is adopted in this paper. The static module and the dynamic module are solved by the methods based on binary decision graph (BDD) and discrete time Bayesian network (DTBN), respectively. In the process of dynamic module solving, an improved method for describing the failure probability density of backup components is proposed to solve the problem that the cumulative sum of failure probability density of backup components is not equal to 1. The problem of accumulation of failure probability density is solved effectively. The improved method is applied to solve the failure mechanism model of satellite attitude control system based on dynamic fault tree. The numerical simulation of momentum wheel system with various configurations is carried out. Finally, the life prediction of the system under different working modes is studied. Based on the analysis of the working state and switching situation of the relevant parts in the system, the method of predicting the life of the components with working state switching is presented. Based on the dynamic fault tree model, the life prediction of the system under different working modes is studied, and the numerical simulation is carried out with the momentum wheel system with four oblique loads as an example.
【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：V448.22

【参考文献】