AURA-NMS: Autonomous Regional Active Network Management System

Lead Research Organisation: Imperial College London
Department Name: Electrical and Electronic Engineering

Abstract

Electricity is so deeply ingrained in everyday life that when it is not available many things, essential and simply pleasurable, cease working. Finding ways to better manage faults in electricity distribution systems is a key way of improving the quality of supply offered to customers. Society also faces choices about its sources of energy and we need to find ways to remove technical barriers to the connection of small scale renewable generation without large cost penalties. The reasons that both of these tasks are difficult are (i) that the low voltage part of the distribution system was designed for simple operation without active control and (ii) the distribution system is very large and overall central control is not realistic. Solutions are also constrained by the large amount of existing equipment that is only part way through a long service life and is too expensive to replace prematurely. This project will explore a means to gradually devolving control authority from the existing central control room (which is semi-automated and semi-manual) and use a peer-to-peer network of controllers/decision-makers placed at each substation. The controllers can open and close remotely controlled switches to reallocate loads to different parts of the network and take various voltage correction actions. There is a strong need for communication to obtain feedback information and to allow a controller with only a partial view of the system to cooperate in finding an optimal set of actions to take in the event of a fault, an out-of-tolerance voltage or a generator whose output is being limited by network constraints. The project is challenging because it requires integration of research in distributed control, decision making, network analysis and communications. For this reason we have assembled a team drawn from 7 universities and 3 major international companies in the power industry.
 
Description Electricity is so deeply ingrained in everyday life that when it is not available many things, essential and simply pleasurable, cease working. Finding ways to better manage faults in electricity distribution systems is a key way of improving the quality of supply offered to customers. Society also faces choices about its sources of energy and we need to find ways to remove technical barriers to the connection of small-scale renewable generation without large cost penalties. The reasons that both of these tasks are difficult are (i) that the low voltage part of the distribution system was designed for simple operation without active control and (ii) the distribution system is very large and overall central control involves a large communications and data processing overhead. Solutions are also constrained by the large amount of existing equipment that is only part way through a long service life and is too expensive to replace prematurely. This project has explored a means to gradually devolve control authority from the existing central control room (which is semi-automated and semi-manual) by using a peer-to-peer network of controllers/decision-makers placed in substations. The controllers can open and close remotely controlled switches to reallocate loads to different parts of the network and use local small-scale generation to help regulate the voltage magnitude. An agent-based software structure has been demonstrated that can combine control actions of various algorithms in response to a network fault, an out-of-tolerance voltage or a generator whose output is being limited by network constraints. Analysis has been conducted on the communication bandwidth and protocols required to obtain feedback information and to issues control instructions. And proposals made on the communications to be adopted in future. By the end of the project, the proposed algorithms had implements on a realistic sub-station-approved computer and tested against a real-time simulation of the several network case studies. The controller had been prepared for installation in a real substation for "open-loop" testing during Autumn 2010. The project has been challenging because it required integration of research in distributed control, decision making, network analysis and communications. To achieve this we had a team drawn from 8 universities and 3 major international companies in the power industry.
Exploitation Route The project collaborators included ABB who are an equipment/software vendor for two of our other collaborators EDF Energy Networks (now UKPN) and Scottish Power Energy Networks. All of the develop algorithms were coded for and test on an ABB control platform (one that was proven for use in substations). Thus by the end of the project ABB were in a position to assess the algorithms running on their own hardware running against a real-time simulator. ABB were then able to conduct field tests with
Sectors Energy

 
Description ABB Limited 
Organisation ABB Group
Country Switzerland 
Sector Private 
Start Year 2007
 
Description EDF Energy Networks 
Organisation EDF Energy
Country United Kingdom 
Sector Private 
Start Year 2007
 
Description SP Power Systems Limited 
Organisation SP Power Systems Limited
Country United Kingdom 
Sector Private 
Start Year 2007