Combinatorial optimization approaches to the design of load shedding schemes against voltage instability

This paper proposes a methodology for the design of automatic load shedding against long-term voltage instability. In a first step, a set of training scenarios is set up, corresponding to various operating conditions and disturbances. Each scenario is analyzed to determine the minimal load shedding which stabilizes the system, with due consideration for the shedding location and delay. In a second step, the parameters of a closed-loop undervoltage load shedding scheme are determined so as to: (i) approach as closely as possible the optimal
sheddings computed in the first step, over the whole set of scenarios; (ii) stabilize the system in all the unstable scenarios and (iii) shed no load in the stable ones. The corresponding optimization problem is solved using three methods : genetic algorithms, branch and bound
approach and the so-called “sequential design”. A detailed example is given on the Hydro-Qu´ebec system in which load shedding is presently planned.