A decomposition method for analog fault location

In this paper, fault location in large analog networks by decomposition method is generalized to include subnetworks not explicitly testable. Assume that the network topology and nominal values of network components are known and the network-undertest is partitioned into subnetworks once for all. The decomposition nodes could be either the accessible nodes whose nodal voltages can be measured or the inaccessible nodes whose nodal voltages under faulty condition can be computed by a new method proposed in this paper. The new method reduces the test requirements for the number of accessible nodes and increases the flexibility of decomposition. Location of faulty subnetworks and subsequent location of faulty components are implemented based on checking consistency of the KCL equations for the decomposition nodes and using ambiguity group location techniques. This method can be applied to linear or non-linear networks, and is particularly effective for the large scale analog networks. An example circuit is provided to illustrate the efficiency of the proposed method.