摘要
为了在缺乏先验数据时分析得到系统寿命,提出了一种基于蒙特卡洛树搜索(Monte Carlo Tree Search,MCTS)和系统故障演化过程(System Fault Evolution Process,SFEP)的系统寿命分析方法。方法首先明确SFEP,通过空间故障网络(Space Fault Network,SFN)将其抽象为网络拓扑形式,再转化为空间故障树,最终得到经典故障树结构。这是MCTS分析的前提条件。确定SFEP的起始元件及其寿命分布特征,获得各元件寿命样本集合。设定系统总仿真工作时间和分析时间间隔,以及系统失效度阈值,即可得到达到该阈值后的迭代次数,最终确定系统寿命。虽然存在时间间隔尺度的误差,但所构建方法可对SFEP实时分析,特别适合系统故障认知有限和无先经验情况。
To analyze the system life in the absence of prior experience data, a system life analysis method based on Monte Carlo Tree Search(MCTS) and System Fault Evolution Process(SFEP) is proposed. Firstly, SFEP is abstracted into network topology form by Space Fault Network(SFN), then transformed into a space fault tree, obtaining the classic fault tree structure. Therefore, the life of the system should be determined by the component that first fails. Because MCTS is designed for tree structure and SFEP is expressed as SFN, SFN is transformed into a space fault tree and the classical fault tree structure is obtained for analysis. This is the precondition of MCTS analysis. The initial components and their life distribution characteristics of SFEP are determined, and the life sample set of each component is obtained. By setting the total simulation working time, analysis time interval, and system fault threshold, the number of iterations after reaching the threshold can be obtained, and thus the system life can be determined. Although there are errors in the time interval scale, the proposed method can be used for real-time analysis of SFEP, especially in the case of limited system fault recognition and no prior experience. The maximum possible system life is 7 000 h when the system fault degree reaches 0.95. Since the set time interval is 100 h, the actual system life corresponding to the fault degree of 0.95 should be between 6 900 h and 7 000 h. The proposed method can also be used to analyze the system life under the influence of other factors. For example, this method can be used to determine the temperature range when the system fault degree is greater than 0.95. However, it is difficult to obtain the original data, only the component life must be determined in the design stage. Therefore, the research object is the system life.
作者
李莎莎
崔铁军
LI Sha-sha;CUI Tie-jun(School of Business Administration,Liaoning Technical University,Huludao 125105,Liaoning,China;College of Safety Science and Engineering,Liaoning Technical University,Huludao 125105,Liaoning,China)
出处
《安全与环境学报》
CAS
CSCD
北大核心
2022年第1期58-63,共6页
Journal of Safety and Environment
基金
国家自然科学基金项目(52004120)
辽宁省教育厅科学研究经费项目(LJ2020QNL018)
辽宁工程技术大学学科创新团队资助项目(LNTU20TD-31)。
