An analytic alarm information method for power grid fault diagnosis

Alarm information obtained under supervisory control and data acquisition systems is the basis of building a model for power grid online fault diagnosis. The main problems restricting the online application of power grid fault diagnosis are the complex modeling process and heavy maintenance requirements. Combined with the feature of alarm information acquired from the electric power dispatching automation system, this paper proposed an analytic method for power grid alarm systems using semantic analysis. The method can automatically extract the basic logic variables from alarm information by character decomposition. Based on connection analysis, the associated relation among electrical devices, circuit breakers and alarm information was built, which was corresponded to the analytical system. According to the relay protection tripping principle and the types and properties of relay protection, the associated relation between alarm information was automatically established. This relationship was coupled with the analytic rules of the model, which realized the automatic mapping from alarm information to fault diagnosis and simplified the modeling work. Finally, based on a practical fault case that occurred in a regional power grid, the validity and practicality of the proposed method for online fault diagnosis was verified. © 2014 Chinese Society for Electrical Engineering.