【摘要】 本文提出了数字系统故障自动诊断的一个新的理论模型,此模型中的测试记录符合模2运算。文中讨论了模2模型故障症候的充分和必要条件,证明了任何适合此模型的系统都是在故障单元少于半数情况下一步故障可诊断的。进而指出,确定最大可能相容故障集与解模2线性规划问题是等价的。在此基础上,给出了一个求相容故障集的有效算法以及它的APL程序,并且论证了此算法的时间和空间的复杂性都是线性的。最后,将此模型推广到了随机故障诊断。事实表明此模型与其它模型相比有较多的优点。更多还原

【Abstract】 In this paper a new theoretic model of automatic fault diagnosis for the digital systems is proposed, in which the set of test outcomes tallies with the modulo-two operation. The necessary and sufficient conditions of fault syndrome in this model have been discussed. It has been proved that any digital system which is suited to modulo-two model has the one-step fault diagnosability, if the number of fault units is less than the half of total units. It is shown that Determing the most likely consistent fault sets is equivalent to solving a modulo-two linear programming problem. Based on this analysis, the author has gotten an effective method and its APL program for correctly ditect consistent fault sets, and proved that the space-time complexity of this algorithm is linear. Finaly, this model has been extended to probabilistic fault diagnosis. In fact, there are some advantages in modulo-two model over the others.更多还原