A tractable two-step MILP-QCP approach to on-line thermal constraint management in large radial active distribution systems

This paper focuses on centralized thermal overload management in active radial distribution systems which accommodate a significant amount of distributed generation (DG). The approach looks for minimizing the amount of dispatchable DG units curtailment to remove thermal overload while considering the possibility to use remotely controlled switches (RCSs) so as to reduce the amount of curtailed generation. This task is formulated as a mixed integer quadratically constrained programming (MIQCP) problem. In order to break-down the onerous computational burden of this MIQCP problem to runtimes compatible with real-time application to large distribution networks, a tractable two-step approach is proposed. The proposed approach consists in decomposing the original MIQCP problem into a mixed integer linear programming (MILP) approximation and a quadratically constrained programming (QCP) problem. We prove the interest and feasibility of the proposed approach, on a snapshot basis, in three distribution grid models of 34, 137 and 1089 nodes, respectively. Results show that the proposed approach scales much better with the problem size than the original MIQCP approach and provides solutions of comparable quality.