Greenfield transmission planning for electric power systems

Transmission Expansion Planning is a problem well known by the power system planners. It is a difficult task and therefore still requires attention and researches. New challenges in the sector highlight the need of efficient methods. The integration of renewable energy on larger scales will indeed request the design of new networks to connect the remote generation sites to the consumers. The growing power consumption in some developing countries will also demand specific tools to make decisions about the installation of new transmission lines. This master's thesis aims at providing solutions to those problems. After an overview of the existing literature on the subject, it therefore develops methods to design efficiently greenfield transmission networks. The work is based on two practical cases corresponding to the two above-mentioned challenges. The first one consists in connecting wind farms in the North Sea to onshore connection points. The second one concerns the design of the grid in Burundi based on the forecasted power levels. The problems are greenfield projects for both lines and substations of the networks. They are both deterministic and do not consider operational costs. The proposed method for the former case involves a partitioning technique and a heuristic approach to provide a meaningful list of candidate substation nodes. Another approach, based on a systematic meshing of the space, is presented for comparison. A statistical analysis is also performed in this work to assess the idea of neglecting additional substation candidates by considering only the existing nodes of the grids. The results motivate this choice for the second problem whose N-1 security constraint increases the complexity. The proposed method combines a mathematical optimization method (Benders Decomposition) and heuristic techniques. It appears to outperform a purely heuristic approach based on evolutionary algorithms. The developed methods applied on these two practical cases give substantial results. This motivates the design of these specific tools for providing significant savings.