Under-Frequency Load Shedding Via Integer Programming

We propose and test a new systematic approach to set load shedding under-frequency relays that replaces much of the conventional simulation-based trial-and-error heuristics by a mixed-integer linear program that can then be solved by commercially available highly efficient solvers such as CPLEX. The new formulation can account for an arbitrary number of random generator outages and for the particularities of the power system such as inertia, damping and regulation parameters, and under-frequency/time limitations. The under-frequency relays are characterized by a sequence of frequency set points and allowed time spans below such levels together with the amount of load shedding needed to meet all specified time and frequency criteria. The objective function is to minimize the average amount of load shed over a set of random generator outages. Two simulation studies are presented to illustrate the use of the proposed method to obtain suitable relay settings.