## Problem Description

In the preventative security constrained optimal power flow (PSCOPF) problem selected for this competition, the objective is to provide a generation dispatch at the least cost to meet system load (demand) in the base case and in all credible contingencies cases, such that power flow obeys the physics of a power grid and satisfies system's limits.

While the underlying optimization problem includes many variables, the key operational decision is the selection of generation “dispatch,” i.e., the generating plant’s control set points of active and reactive power output and terminal voltage.

In a contingency case, a failure or series of failures will likely change total power balance in the system from the base case. A power imbalance results in deviations of the interconnection frequency from its nominal value. The first system response to support frequency (after the system inertial response) is the speed droop control (a.k.a. governor response or primary regulation). Many generators in AC power systems are equipped with speed governors, which are automatic devices that change the real power output to arrest and oppose significant frequency variations; however, the active power generation adjustments that result from governors’ actions are not sufficient to completely restore frequency to its nominal value after a disturbance. The full recovery is achieved by further adjustments made by automatic generation control (AGC), contingency reserve activation, and other control actions. The speed droop control is normally seen in the system within a few seconds after a disturbance, and reaches its maximum within a few tens of seconds after a single disturbance; therefore, the power flow resulting from contingencies is best described initially by a model reflecting speed droop controls on generators (governor power flow). The role of PSCOPF during the initial post-disturbance system state is to eliminate system violations that are not feasible or not acceptable even for a short time. The governor power flow equations are used as constraints for contingencies in this optimization. Note that the objective function representing system operating cost is assumed to depend only on the set point of generator active power output from the base case and does not consider the cost of deviation induced by speed droop control. This assumption may be justified in a market environment by observing that the compensation paid to a generator will typically be based only on the commanded power set point. We will use this model in the initial phases of competition (Phase 0 and Phase 1).

Usually, frequency is not reflected in power flow formulations. This means that instead of a more accurate power balance model that includes frequency, generation, load, and power losses as they depend on frequency variations, PSCOPF is a simplified model that includes only the balance between generation, load, and power losses. This model implies that the system imbalance caused by disturbance and power losses is allocated to generators based on their participation factors/distribution coefficients, which may reflect their speed droop characteristics.

For more information, please download a detailed version of the Problem Description or a Slide Summary.