确定系统级测试性参数的广义随机Petri网模型

测试性设计是近年来发展起来的一门新兴学科 ,对提高复杂系统的可靠性、可维修性和可用性具有重要的意义。在测试性设计的过程中 ,合理地选择和确定测试性参数是实现费效比最优设计的关键之一 ,但目前还没有有效的方法能合理地确定每个参数的门限值或目标值。基于广义随机Petri网的原理 ,将测试看作设备整个生命周期内可靠性、维修性活动的一个有机组成部分 ,建立了系统的测试性模型。采用数值分析的方法得出系统稳态可用度与系统测试性参数之间的关系曲线 ,作为确定测试性参数的依据 ;当系统需待定的参数比较多时 ,将模型转化为一个区间线性方程 ,有效地解决了计算量大的问题。这一方法对解决复杂系统的测试性设计问题具有一定的借鉴意义

Generalized Stochastic Petri Network for Defining the System-Level Testability Parameters and Figure of Merit

Testability design is a promising technique for improving reliability, maintainability and availability of a complex system. In the design process, selection and definition of proper testability parameters and figure of merit is a key to optimal cost-effectiveness ratio. However, there isn't yet an effective method to define the proper limit values of testability parameters due to immaturity of testability technique. In this paper, a generalized stochastic Petri network (GSPN)-based approach is proposed for defining such parameters. Tests are regarded as an important part of maintenance activities in the whole life cycle of equipment. Thus, GSPN-based testability model is constructed, on which the relation curves of stable availability and testability merits are gained by utilizing numerical simulation. When more than one parameters need to be qualified, the model is transformed to an interval linear equation, and the parameters can be defined by using Hasnen's extended algorithm.