Power Grid Fault Diagnosis Based on Random Forest

With increasing power grid architecture and operation level, the demand for power supply reliability will increase accordingly. How to establish an efficient and automated power grid fault diagnosis strategy under increasingly complex operating conditions, shorten the time of power failure, and reduce the economic loss caused by power failure. It is the basic problem of power grid dispatching operation research and application. Rule-based power grid fault diagnosis needs to introduce a large number of protection and device action rules when modeling, and the modeling process is difficult. To solve the above problems, this paper takes the fault event description words in the alarm information as the research object, studies the distribution characteristics of fault description words in different types of faults, establishes the alarm information vector space model, completes the text vectorization of alarm information, and proposes a power grid fault diagnosis method based on random forest (RF). The example simulation shows that the model can diagnose the fault type accurately.