Transmission Expansion Planning (TEP) is usually performed on a few operating situations or snapshots. In order to get a representative set of snapshots, it is necessary to select them carefully. We propose a clustering method based on the Kmeans algorithm that uses features drawn from information about system operation. Features based on price differences and non-controllable injections are considered and a small test case is proposed. We suggest replacing local features by statistical indicators over the system to reduce the clustering complexity. The obtained results show that statistical price differences can be used as a good clustering feature for snapshot selection and the error introduced in the investment solution compared to the solution without clustering is very small.
Keywords: clustering methods, renewable energy sources, statistical features, transmission expansion planning
Publicado: junio 2015.