Load Pickup and Feeder Reconfiguration of Distribution Networks

Abstract

The power outage of the distribution networks resulting from natural disasters could bring catastrophic damage to the system as well as millions of financial loss to the people. Restoring distribution networks after these events can help reduce the consequences and therefore have been drawing much attention since 1980s. This thesis focuses on the load pickup problem and the feeder reconfiguration problem in the restoration process. For a single feeder distribution network, system operators try to pick up as many loads as possible after a distribution network power outage. However, the load pickup problem is a combinatorial optimization problem, the computing time of which could be very long for large scale distribution networks. We model it with an ILP formulation and propose an approximation algorithm by casting it into a variant of knapsack problem. Both algorithms are tested and compared intensively with many cases on standard IEEE test feeders. For the multiple feeders interconnected with normally open switches, system operators can alter the system topology by opening and closing switches and shift the loads to pick up them. In this thesis, we formulate the feeder reconfiguration problem as an ILP problem and then perform tests to find out different corner cases that would help us to examine the performance of the ILP solution approach.