Technical and economic assessment of information and communication technologies for smart grids

Résumé

The smart grid concept represents a change of paradigm that implies the modernization of current electricity networks to obtain a sustainable power system. The first step to achieve the potential advantages of the smart grid is the successful design and implementation of a reliable, secure and cost-effective communication infrastructure. However, at the present time there is no existing standardized communication infrastructure that has been widely accepted and used to transform the current electric power grid into a smart grid. In this thesis, the performance and the impact of Information and Communication Technologies (ICT) in current smart grids applications are analysed from a novel technical and economic perspective to support decision-making in their implementation. From all the smart grid communication technologies, Power Line Communications (PLC) is considered the most cost-effective solution because it allows the reutilization of the power infrastructure as communication channel. PLC is especially used for the deployment of Advanced Metering Infrastructures (AMI), but the performance of this technology is highly affected by the local network conditions. Among the different PLC protocols used for AMI, PRIME provides the best performance under favourable conditions. For this reason, a methodology to analyse the performance of PLC PRIME networks has been developed, which is based on a simulation framework that takes into account the characteristics of the physical channel and the network behaviour. The two Key Performance Indicators (KPIs) used for the previous analysis are the number of registered nodes and the time to read all meters, which have shown the strong influence of the registration process and the user density on the communication performance. The developed methodology has been satisfactory applied to a wide range of distribution networks with a tree topology of different sizes. For this purpose, an algorithm to compute the PLC transfer function between any pair of nodes has been proposed, which is based on transmission line theory. Additionally, Reference Network Models (RNMs) have been used to obtain representative networks that can be later analysed with the proposed methodology avoiding the use of confidential data from Distribution System Operators (DSOs). The combination of the simulation framework and the RNM has resulted in a very powerful way to analyse the communication performance of PLC networks at a regional level. The added value provided by better information and communication systems in smart grids is another open issue for the smart grid development. Then, a new methodology to analyse the impact of these systems in a centralised voltage control application with On-Load Tap Changers (OLTC) and PV based on inverters has been presented. Since this kind of voltage control application aims to send set-points from the central unit to the OLTC and PV inverters to optimize the system operation, the presented methodology is focused on the impact of the accuracy of the forecast used to calculate the set-points and the time interval used to update these set-points in the PV inverters. The voltage deviation cost, the energy curtailment cost, and the energy losses cost have been chosen as economic KPIs. All the conclusions obtained in this PhD provide a significant improvement for the analysis of the performance of smart grids, and encourage the use of the developed methodologies and their extension to other communication technologies and applications.