Abstract:

The limitation of traditional reliability modeling methods was analyzed, the dynamic reliability modeling method based on general stochastic Petri net (GSPN) for complex system was proposed, the transitions were divided into instant transitions and timed transitions,the timed transitions were used to express the failure and maintenance process of elements.The inhibitor arc was used to express the changeover among failure states and restore states,a typical GSPN model of repairable system was established. Aiming to deal with the problem of non-exponential distribution and state explosion in complex repairable system, the Monte Carlo simulation was uesd to solve the GSPN model, by calculating the parameters such as the occurring times and sustained time of transitions and places, the dynamic reliability data of the elements and system are obtained. A hydraulic loop of digging machine was selected as an example, the feasibility of the method was validated.

Key words: dynamic reliability, general stochastic Petri net（GSPN）, timed transition, failure, Monte Carlo simulation

摘要：

分析了传统可靠性建模方法存在的缺陷，提出基于广义随机Petri网（GSPN）的复杂系统动态可靠性建模方法，将变迁分为瞬时变迁和赋时变迁，采用赋时变迁表示零部件故障及其维修过程，以禁止弧表达系统故障与修复状态的转换，建立典型可修复系统动态可靠性分析的GSPN模型。针对因非指数分布、状态空间爆炸等导致的GSPN模型求解难题，提出基于蒙特卡罗仿真的求解方法，通过对变迁发生次数及其持续时间、库所状态等参数的统计，计算系统动态可靠性指标。以挖掘机液压回路为例，验证了所提方法的可行性。

关键词: 动态可靠性, 广义随机Petri网, 赋时变迁, 故障, 蒙特卡罗仿真