Matheuristic for emergency response in the electric distribution system with unforeseen islanding

Since power supply may suffer interruptions, caused by climate events or even due to equipment and human failures, these interruptions are associated with procedures commonly referred as emergency orders. These emergency orders must be attended by repair crews in an attempt to reestablish the power supply all over the system. Assuming a set of these emergency orders, the network operation center (NOC) must find a scheduling so that they are all attended by the repair crews and the power supply reestablished as soon as possible, while considering the minimization of the total service cost. Distributed generation (DG) may be one alternative for furnishing temporally power supply, such as photovoltaic and wind power generation sources. In such approach, the islanding operation confers such a complexity when considering how much time should be needed to supply the affected loads and which ones may be even selected. This study proposes a novel matheuristic for emergency response to the contingency planning problem related to electric distribution operations, considering unforeseen islanding of DG. From using a mixed integer linear programming (MILP) model, based on the emergency order scheduling problem (EOSP) with islanding operation formulations, the optimization criteria involves the minimization of the cost of non-supplied energy and the penalties related to high important customers, while respecting electrical constraints. Computational results on the IEEE 33-bus system, modified with the insertion of DGs, show the effectiveness of the proposed solution on reducing the total attendance cost when compared to a traditional scheduling solution without considering the islanding operation.