In this thesis we study the wind farm layout problem and the cable routing optimization. In particular, the aim of the first part of the thesis is to determine turbine locations to maximize the expected power output under different types of constraints and different wind scenarios. Wake effect between turbines is taken into account in the optimization. In the second part the cable routing optimization problem is addressed. Given a turbine layout and cable prices, our software is able to find the cable routing that minimizes the total cable cost, taking different constraints into account.
The solution method we propose is a Mixed Integer Linear Programming (MILP) approach with some ad-hoc heuristics. Several heuristics are proposed, implemented in C programming language, and computationally compared on both real-world and synthetic instances. Validation experiments show the effectiveness of our method.
The original results of the thesis have been considered of interest both from a company point of view (as our wind farm layout tool turned out to outperform the commercial software
that is currently used at Vattenfall), and from a research point of view, as we presented a new approach able to deal with instances with huge amount of variables (10000+ possible
turbine positions) within a few minutes on a standard PC.