Decentralized frequency control of power systems with deep penetration of wind-based generation

The introduction of the highly variable and uncertain renewable resources into the power grid is calling for more control and regulation of the power system dynamics. In particular, the automatic generation control, which is responsible for maintaining the nominal system frequency and the scheduled real power interchange, needs to be modified to include unmodeled system dynamics and to account for disturbances from renewable resources. In this thesis, we work on a nonlinear differential algebraic model of the power system which takes into account the effect of the power network and includes wind power injections. We then propose two decentralized controllers that each would stabilize the system frequency and power interchange. The first controller is based on linear quadratic (LQ) optimal control followed by an optimization algorithm to increase the sparsity of the feedback gains. The other controller is designed using the theory of overlapping control and the inclusion principle. Each controller is applied separately on a 3-machine 6-bus 2-wind turbine nonlinear model, and the simulation is carried out using Simulink. A power flow program is run at each automatic generation control (AGC) cycle to update the power flow variables. Results show that we can design decentralized controllers for each control area that can successfully track the desired frequency regardless of the disturbances associated with wind-based generation. Furthermore, we show that the performance of these controllers is comparable to that of a centralized controller