A genetic algorithm for finding microgrid cable layouts

Abstract

Microgrids play a crucial role in the electrification of rural areas. Designing a microgrid comprises of multiple parts including finding suitable sites for generation units, sizing the components of the microdgrid, and determining the layout of the cables to connect the components. In this work we focus on the latter part, which we formalize as the Microgrid Cable Layout problem. In this problem we assume that the locations and sizes of the generators and consumers are already given. The goal is to find a cost-minimal cable layout that connects these locations and that is sufficient to handle the consumer demands. The cables may be of different cable types, and we may connect multiple cables not only at the locations of the generators and consumers but also at any other point. Microgrid Cable Layout is a strongly NP-hard, non-linear optimization problem. We present a hybrid algorithm for the Microgrid Cable Layout problem, which employs a genetic algorithm for optimizing the topology of the layout and a heuristic for assigning the cables to the edges of the topology. An evaluation on a set of benchmark instances indicates that this algorithm is able to find good solutions within a short amount of time. We further evaluate the performance of the algorithm in a case study on a real-world microgrid in the Democratic Republic of the Congo.