MPC-based Load Shedding for Long-term Voltage Stability Enhancement Using Trajectory Sensitivities

This paper presents model predictive control method for load shedding based on trajectory sensitivities. According to the basic principle of model predictive control, aiming at minimizing the quadric performance index which comprehensively considers both the control cost and the deviation between bus voltage predictive trajectories and their reference values, the receding optimization model is described as an optimal control model with constraints consisting of differential-algebraic equations in continuous and discrete time. The trajectory sensitivities between bus voltages and control variables are computed by means of trajectory sensitivities of hybrid systems and quasi-steady-state simulation, and hence the optimal control model for emergency voltage is simplified into a quadratic programming model with which independent variables are increments of control variables. This both enhances the computational efficiency and achieves better control performance. Results on New England 10-machine 39-bus system demonstrate feasibility and effectiveness of the proposed method for enhancing long-term voltage stability of power systems.