Obstacle-aware optimization of offshore wind farm cable layouts

In this article, an integer linear programming model for cost minimization of cable layouts in offshore wind farms is presented. All turbines must be connected to power substations by cables. Up to a given number, turbines may be connected along a joint cable in a series circuit, and cable branching at turbine locations is possible. No two cables are allowed to cross each other. As an improvement over previously available models, the model under study enables optimal adaptation of the cable routes to obstacles. Obstacles of two different kinds are considered. First, a set of regions in which cables cannot be laid is accepted as part of the input to the model. Second, the trajectory of one cable plays the role of an obstacle to all other cables. Both obstacle types are modeled by introducing optional connection points, which, contrary to the turbines, do not have to be visited by any cable. By introducing such optional connection points at selected positions, we arrive at a model with some resemblance with the Steiner tree problem. We demonstrate that, by virtue of the optional points, the suggested model is able to identify feasible solutions to problem instances where other models fail to do so. In other instances, the model yields more cost-effective cable layouts than previously studied models do. Computational experiments with realistic wind farm instances of up to 88 turbines prove that cabling cost reductions of about $$1\%$$1% are achievable by the model.