Increasing the Observability of Electrical Distribution Systems using Smart Meters (IOSM)

点击翻译按钮获取中文摘要

Real-time monitoring and control of distribution systems is very limited due to the lack of sensors and communication systems. Hence the distribution system can be described as under-determined with the number of measurements insufficient to make the system observable. Once the complete system state is available, then any quantity in the system can be calculated. The observability and controllability of a system are mathematical duals, which means an unobservable system cannot be fully controlled. Distributed energy resources introduce significant uncertainties and, at high penetrations, may lead to operational difficulties in a network. Therefore the provision of accurate system state information to the network operators is critical for them to operate the system in a safe, prompt, and cost-effective manner, and also to make best use of the assets. Smart metering is widely recognised as the first step towards a Smart Grid future and the UK is committed to the full deployment of smart meters by 2019. Smart meters and the associated ICT (information and communication) infrastructure can greatly improve observability. Therefore there is a need to investigate the technical feasibility and key technologies of using smart metering to increase the observability of the distribution system through state estimation techniques.The research programme is structured around three challenges: Research Challenge 1: The load demand needs to be aggregated at the MV nodes using data from smart meters connected to the low voltage (LV) nodes. A big challenge is how a state estimator deals with both various kinds of measurement errors with non-normal distribution and the influence of the measurement configuration (type, location, accuracy of measurements) effectively and provides accurate estimation on the system state.We will improve the distribution state estimation to make it robust to the influence of both the measurement error distribution and the measurement configuration of a distribution system.Research Challenge 2: Smart metering may change the behaviour of energy consumers and thus lead to more dynamic demand (e.g. load that is sensitive to price). Therefore the second challenge is how to model extremely dynamic load and to provide pseudo measurements to the state estimator under conditions of large latency or failure of the ICT infrastructure or if there are un-monitored quantities. We will provide a theoretical contribution to MV nodal load modelling through investigating a new machine learning method which is able to obtain knowledge from past experience (e.g. past smart meter data). Research Challenge 3: What and where additional real-time measurements should be placed, in addition to the smart meters, to make the estimated system states accurate enough for particular Smart Grid functions and reduce the impact of the measurement configuration. We will develop an optimal meter location method considering impact from both measurement errors and measurement configurations while minimising the extra metering cost.The research will benefit from close collaboration with national and international industrial partners, and will gain insight and make contribution to the research challenges through both theoretical study of using smart meter information to increase the observability of distribution systems, and technical demonstration via small scale test facility, i.e. the Smart Metering test rig and the Smart Grid test rig developed at Cardiff; medium scale test facility in RSE, Italy; and practical case study using a BC Hydro network. The impact on potentiol beneficiaries will be delivered through collaboration, communication, and commercialisation. We will also utilise EPSRC HubNet as a dissemination platform to facilitate a wider communication.