• distributed fault diagnosis;
  • multimachine power systems;
  • robustness;
  • fault detectability;
  • fault isolability;
  • adaptive estimation;
  • sensor bias


This paper presents a distributed sensor fault detection and isolation (FDI) scheme for multimachine power systems. Each generator is interconnected with other generators through a transmission network, where the interactions between directly interconnected generators are nonlinear. In the distributed FDI scheme, a local FDI component is designed for each generator excitation system in the power system based on local measurements and certain communicated information from other FDI components associated with generators that are directly interconnected to the local generator. In each FDI component, adaptive thresholds for distributed FDI are derived, ensuring robustness with respect to nonlinear interconnection and unstructured modeling uncertainty under certain conditions. Furthermore, the fault detectability and isolability properties are investigated, characterizing the class of sensor faults that are detectable and isolable by the distributed FDI method. In addition, the stability and learning capability of the local adaptive fault isolation estimators designed for each generator is derived. A simulation example of a two-machine power system is used to illustrate the effectiveness of the proposed method. Copyright © 2014 John Wiley & Sons, Ltd.