HubNet: Research Leadership and Networking for Energy Networks (Extension)

点击翻译按钮获取中文摘要

The pace of increase in renewable energy use in UK electricity is impressive with 19% of electricity sourced from renewables in 2014 (DUKES 2015) and half of that being produced by wind turbines. Use of large quantities of variable generation sources has already caused changes to the way electricity networks are designed and operated but as penetration pushes onward to 30% and thence to 50%, it becomes significantly more challenging to maintain the stability and availability that people have come to expect. It is widely said that the resolution of this problem is offered by a smart grid but despite much discussion of smart grids there is still a lack for detailed specification of the intelligent and automated functions required to operate network of the future..An example smart grid feature is that, to a much greater extent than today, demand needs to be influenced to follow available generation. With the advent of half-hourly resolution of electricity bills, facilitated by smart meters, there is a mechanism in place to reward customers who make their consumption flexible and responsive to changes in electricity availability and price. In this example of "smart" technology there are very many questions over how exactly this might work, how people would engage with this technology and how coordination of millions of individual actions will add up to stable operation of the grid.We set out to answer two broad questions. First, how exactly should new computational techniques help mange the future grid? Specifically how can we use decentralised control to perform demand-supply balancing (and voltage control, etc.) at a local level and facilitate coordination and collaboration between decentralised controllers. Also, as part of that, how do advances in data analytics help gain useful control information from the masses of data being produced in the system. Second, in this more complex system, how do we analyse the risk of a system failure and design in resilience to mitigate risk. Traditionally we have had good models of the probability of the failure of, say, a transformer. How can we analyse and design for the highly variable behaviour of customers' responses to short term electricity prices and the short term fluctuation of PV and wind energy. New design principles are needed to ensure continued high reliability of supply. Being smart about the use of decentralised control and the harnessing of newly available data is crucial to ensuring we provide an electricity system able to incorporate large amounts of renewable energy without burdening us with undue cost or degrading performance.Alongside these specific technical ambitions, HubNet maintains its role in bringing together the many academic groups who can contribute, not just electrical power engineering but also computational sciences, energy economics and behavioural modelling. Further, our role is to be the conduit for the academic community to converse with the network operators and equipment vendors so as to keep our research relevant to the sector's major challenges and provide a route to demonstrate and deploy our findings. We will continue to help identify the key research questions for the community to address and we will continue to use our findings as evidence to support energy policy and regulation.