Real-Time Transient Stability Assessment of Post Fault Scenario in Power System by Using Probabilistic Neural Network

With the integration of renewable energy sources and plug-in electric vehicles, the recent modern power systems are becoming more complex and dynamic in nature and are susceptible to contingencies. It is therefore, important to monitor whether the system operates within its stability limit. In the events of stability limit violation proper control measures should be initiated to bring back the system to normal operating state. But precise real time monitoring and controlling of power system is a challenging task. Accurate monitoring of power system enables the power system engineer to take proper control measures. In this paper, a methodology is proposed to predict the transient stability status post-fault and its intimation to Power System Engineer. It will provide enough time for engineer to activate control measures. In the literature, prediction of transient stability state has been carried out by using Artificial Neural Networks. In this work, the transient stability status post-fault is predicted using Probabilistic Neural Network (PNN). The PNN has to be trained to predict unforeseen events in Power System. The 26 different contingencies are created in the IEEE 39-Bus system and post fault 6-cycle data of Rotor angle and Voltage magnitude at faulted bus are measured real time. Totally 156 data sets are generated. In that, first 5 cycle Rotor angle and Voltage magnitude are considered as input variables and sixth cycle rotor angle is considered as target variable. The total data sets are separated for training and validating of the PNN and its performance is analyzed. This proposed work will aid the power system engineer to take precautionary control actions well in advance.