为弥补传统测试性推理方法通用性差的缺陷,研究了基于图论的通用测试性推理技术,构建了一种通用测试性推理模型。其中,图论内核抽象了框图模型、信息流模型、多信号流图模型、混合诊断模型、贝叶斯网络模型和AI-ESTATE中的信息元素,融合了Floyd-Warshall算法、FDI算法和高阶关联关系推理算法,为测试性推理的通用化、标准化和并行化提供了一种新的解决方案。结合某装备测控系统的信号形成组合进行了推理方法验证,证明了方法的有效性和先进性。

In order to make up the flaw of low universality of traditional testability inference methods, in the paper, a general purpose testability inference technology was given based on graph theory. A graph theory core was implemented to abstract all testability information from frame model, information flow model, multi-signal flow model, hybrid diagnostic model, Bayesian model and artificial intelligence exchange and service related with all test environments (AI-ESTATE), and melt general-purpose algorithms such as Floyd-Warshall algorithms, fault detection and isolation (FDI) algorithms and high order relation inference algorithms. It has provided a new solution to general-purpose, standard and concurrent testability inference for equipment. Graph theory based general-purpose testability inference technology is shown to be effective and advanced by the example of the testability inference for the certain equipment's signal generation unit.