Energy losses estimation tool for Low Voltage Smart grids

Abstract

The so-called 20-20-20 targets committed to by the European Union drives the need for a more efficient distribution network. The energy efficiency improvement required involves a 20% reduction of the energy consumption compared to the 1990s. For such cutback, Distribution Systems Operators are encouraged to develop the best strategies to identify and reduce power losses in their networks. This task becomes challenging in Low Voltage Distribution networks due to diversity in the feeder’s topology configuration, load distribution and the presence of renewable-based distributed generation. In this paper, a clustering-based methodology is proposed as an energy losses tool to support the energy efficiency decision-making process. A feeder’s clustering process using the K-means algorithm is carried out upon a customised network characteristics set that was previously reduced to two coordinates by applying Principal Component Analysis. The relationship between power losses and the net energy imported under the different scenarios is obtained for each feeder class identified. The data and network used in this process correspond to the roll-out deployed at the Spanish Smart Grid Demonstration Project (OSIRIS)