Modeling, Analysis And Software Design For System Testability Based On Multi-signal Flow Graph

With the development of science and technology, there appear all kinds of largescale systems and equipments with more powerful functions. Accordingly, the structureof system is more complex which leads to the difficulty of fault diagnosis and the highfees of maintenance. For the reasons above, it is necessary to conduct testability designand analysis. Lack of domestic self-developed testability design and analysis softwarehas become a major shackle which has restricted the development of the domestictestability research. Therefore, in-depth study theories of system testability and theindependent research and development of software for testability design and analysisare in great significance.Firstly, this paper studies the modeling idea of multi-signal flow graph model,illustrates the modeling steps with a concrete example; and develops an overall designscheme of the graphical modeling software for testability. By embedding the Visiodrawing interface to build a multi-signal flow diagram drawing environment, this paperdevelops a toolkit of components needed for modeling. This paper completes propertiesof the module by using the text documents attribute set, save, and outputs dependencymatrix by the use of Excel table.Secondly, this paper discusses the method to generate dependency matrix based onmulti-signal flow graph; through analysis of the Visio backstage system files obtains theadjacency matrix, then to get reachability matrix, and to complete failure matrix byextraction the corresponding fault line and corresponding test column of reachabilitymatric, unites fault and test functions to generate a functional failure correlation matrix,then combines complete failure matric and functional matrix to dependency matrix.Thirdly, taking the radar transmitter system as an example, through analysis of thedependency matrix, completes system test analysis. In static testability analysis, throughthe analysis of the dependency matrix information of obtains indicators such as faultdetection rate, fault isolation rate, undetected faults, ambiguity groups, and redundanttest will be ranked. In dynamic testability analysis, genetic algorithm is used to find theoptimal test set; AO*algorithm is used to search for and/or figures to complete thesolution of the test sequence.Finally, taking the radar transmitter system for instance, this paper introduces thetestability graphical modeling software and analysis software functions. Throughcomparing with foreign commercial software testability-TEAMS, the same resultverifies the reliability of the developed program and software.This paper realizes the important components of testability analysis software, suchas graphic user input interface, model set up, testability analysis, and results output. Thesoftware has the features of friendly interface, good accuracy, and a high degree of automation. It is considered to be comparable with the present commercial testabilitysoftware.