An unique approach for Voltage Stability improvement with probabilistic Neural Network

This work proposes an exclusive approach for improving voltage stability limit using a Probabilistic Neural Network (PNN) classifier resulting remedial controls available in the scenario of various contingencies. The sensitivity of the whole system is analyzed to classify the weak buses with ENVCI valuation when it approaches zero. The input to the controller, named as Voltage Stability Enhancing Neural Network (VSENN) controller for training are line flows and bus voltages near the notch point of the P-V curve and the output of the VSENN is a control variable. For various contingencies, the control action that improves the voltage profile as well as stability index is observed and trained accordingly. The trained VSENN is finally tested for its robustness to improve load margin and ENVCI with new contingencies which are not trained before. The proposed approach is established in IEEE 39-bus test system.