基于贝叶斯推断的复杂系统可靠性分析

随着武器装备系统日益复杂,成本越来越高,大量的全面系统测试逐渐变得不可行,很难得到大量的可靠性信息,可靠性的统计分析与评估面临着挑战。提出一种基于贝叶斯推断的复杂系统可靠性分析方法,介绍贝叶斯推断以及基于贝叶斯推断的复杂可靠性分析步骤。利用事件树图对复杂系统结构进行建模,归纳可用的数据类型,对组件随机模型和先验分布进行描述。利用马尔科夫蒙特卡洛方法对模型进行求解;构建计算示例验证方法的可行性与有效性。该方法适用于不同的建模粒度,可以灵活地融合专家知识、全系统测试数据、子系统和组件级数据,一致性地估计系统、子系统和组件的可靠性参数。

RELIABILITY ANALYSIS OF COMPLEX SYSTEM BASED ON BAYESIAN INFERENCE

With the increasing complexity of weapons and equipment systems and higher costs, comprehensive system testing has become increasingly infeasible. It is difficult to obtain a large amount of reliability information. Statistical analysis and evaluations of reliability are facing challenges. In order to solve the problems, we proposed a reliability analysis method for complex systems based on Bayesian inference. Bayesian inference and complex reliability analysis steps based on Bayesian inference were introduced. Event tree diagrams were used to model the complex system structure, the available data types were summarized, and the random component model and prior distribution were described. Then, we discussed how to use the MCMC (Markov Chain Monte Carlo) method to solve the model. A calculation example was constructed to verify the feasibility and validity of the method. The proposed method can flexibly integrate component and subsystem level data, expert knowledge, and system-wide test data. It is applicable to different modeling granularity and can consistently estimate the reliability parameters of components, subsystems and systems.