Achieving real-time economic dispatch in power networks via a saddle point design approach

Increasing distributed energy resources introduce rapid and fast fluctuations in power supply and demand. As a result, the control and economic optimization for power networks need to operate at faster time-scales for reliability and economic efficiency. In this paper, we consider the problem of achieving real-time economic dispatch for power networks with controllable loads. We first present system dynamics over the transmission network under exogenous disturbances and formulate an economic dispatch problem with power network congestion constraints. We then propose a saddle point design framework in which system dynamics and the optimization problem are combined to derive a completely distributed controller. As a result, the closed-loop system asymptotically converges to an equilibrium point where the optimization problem is solved. An example using the IEEE 14-bus network illustrates the effective performance of the control scheme.