PROFILINGBLACKOUTS · Cascading failures in electrical networks: stochastic analysis and distributed prevention methods
FP7 — People (Marie Curie Actions)
- Duration
- 2011-09-01 → 2013-08-31
- EU contribution
- €208,593
- Participants
- 1
- Scheme
- MC-IEF
Lines connect the coordinator with its partners.
Results in brief
Cascading failures in electrical networks: stochastic analysis and distributed prevention methods
The need to reduce CO2 emissions is fundamentally changing the way the electricity grid will be operated and designed. Variable renewable generation sources such as wind and solar are less predictable and less controllable than conventional generators. In order to cost effectively integrate these sources it will be critical to adopt a range of smart grid technologies with the objective to enhance the utilisation of electricity system infrastructure. The potential benefits of this smart grid transformation are very significant and well-studied, but it is not clear what this will mean for the reliability performance of the future electricity supply systems. The PROFILINGBLACKOUTS project aims to investigate and develop novel advanced statistical methods and tools to assess the resilience and improve the reliability performance of the electricity grid, with a particular focus on high-impact low probability events. A prime example of such events are cascading outages, where an initial outage triggers follow-up outages thus potentially leading to large-scale blackouts. The likelihood of such events occurring is vanishingly small so these events are easily overlooked. However, the impact is so significant that the possibility of such an event must not be ignored. The research has led to the development of an efficient model for cascading outages on large random networks. The model provides insight into the fundamental parameters that affect the resilience of the realistic large-scale electricity grid. The resilience is quantified through conducting a large number of computer simulations, and the results are analysed with a novel data analysis method that accurately accounts for statistical fluctuations. This leads to results with strict reliability bounds. In addition, the research project has resulted in a novel stochastic control method for domestic refrigerators that can support real time balancing of demand of supply. It enables individual appliances to take actions that collectively stabilise the electricity grid, thus significantly increasing the grid’s resilience at a very competitive cost.
Data: CORDIS, © European Union
Project objective
Electricity networks are designed to cope with a set of credible contingencies, usually involving the loss of a single piece of equipment. However, the interconnectedness of the network makes it hard to predict what happens when an unexpected contingency occurs. Sometimes this leads to an overload in one of the lines, causing it to be taken out of service, which may trigger other overloads. In recent years, such cascading failures have lead to a number of large scale blackouts.This project’s first aim is to get a deeper understanding of the fundamental vulnerability of electrical energy systems to experience large-scale cascading failures after the loss of very few components. This is done by generating large numbers of generic networks with given connectivity profiles that satisfy Kirchhoff’s circuit laws and are robust against ‘simple’ contingencies. For these networks, a risk profile is computed that indicates the frequency of outages as a function of their magnitude. These will give insights into the factors influencing a network’s vulnerability. Particular attention will be paid to connectivity, margins and the size distribution of generators.Building on this knowledge, it will be investigated whether the nodes in the network can identify and share information on network stress with their neighbours and whether this information can be used to prevent large-scale cascades. For example, a node may shed some load preventively or exercise a demand reduction contract in order to locally decrease the stress on the network. A few of these schemes will be constructed and it will be investigated how this affects the risk profile. The results will be compared with each other and the business-as-usual scenario with an aim to inform the design of real-world protection schemes.""
Original text from CORDIS.
Participants
- IMPERIAL COLLEGE OF SCIENCE TECHNOLOGY AND MEDICINE · LondonCoordinatorUnited Kingdom
Links
Data: CORDIS, © European Union