Diagnosing vulnerability, emergent phenomena
and volatility in man-made networks
The aim of this project was to analyse real-world infrastructure systems with a view to aiding day-to-day and emergency planning of
critical —primarily energy— European infrastructures. We applied the techniques of complexity theory to real-world networks. In order to
do so we investigated networks (both physical and service) that make up the main elements of functional interconnected networks. The
work involved project partners which included infrastructure owners and governmental decision making bodies. This approach enabled
access to knowledge sets that formed the basis of the project case studies.
Thus, the project assembled, developed and applied complementary mathematical methods to analyse large, man-made multi-element
infrastructure systems that exhibit, so-called, complex behaviour.
This will assist in the development of civil emergency preparedness strategies as well as in the general long-term planning of energy
infrastructure programmes. In order to address such interdisciplinary topics, researchers working in a wide variety of fields brought
insights to problems that, in complex systems terms, have a common qualitative nature; thus enabling a macroscopic overview of the
complex behaviour of key infrastructures.
Adopted approach
Main achievements
The project has so far achieved results in several key areas:
- Applying mathematical analysis to networks using real world data and understanding the problems and opportunities this raises.
- Mathematically analysing the electricity spot market data, and developing new insights as a result.
- Building useful new data sets, particularly around European electricity and gas networks, offering a range of new insights around
flows and vitality.
- Developing and implementing mathematical methods to couple heterogeneous networks (both functionally and topologically)
and understand how failures percolate through such systems when exposed to natural hazards.
A number of specific research results have also been produced:
- The core of the project turned out to be the problem of eliciting data in a form that was suitable for analysis. The JRC (Joint
Research Centre), with the help of industrial, governmental and commercial stakeholders and the other project partners have
developed good collaborations and made progress in obtaining integrated graphical analysis at both continental, country and
local area level within cities. The problem is finding not only the topological properties of a network but also the importance of
the nodes and links according to the various criteria of importance. The extensive data mining that we have undertaken has, for
example, enabled us (a) to consider error-tolerance of gas transmission networks and network redundancy in greater detail than
ever before, (b) to investigate the potential impact of seismic activity to disrupt the electricity and gas supply across Europe and
the potential societal disruptions (i.e., lack of electricity supply to citizens) that can accrue from this and other natural hazards
(c) to consider the critical infrastructure aspects of road networks in major cities and attempt to obtain describing motives for
physical processes such as traffic flows.
- The project has been host to extensive work on the wind energy characteristics across Europe and its implication for national
and EC policy for our future electricity supplies, and the associated questions of likely efficiency, which have improved our
understanding of wind energy potential and the necessity for this to be integrated into the national and super-national energy
networks. This study poses political implications and could raise important strategy questions in Europe. For example, the selforganising approach around deploying energy networks is not good enough in the case of wind energy. It almost certainly requires
an interventionist policy from the EC to oversee adequate international transmission of excess energy production, imposing a
more strategic super-national perspective on the energy operators. This research calls for a more strategic top down view, like a
highway system, because decisions by local governments do not take the broader “wide-network” implications into consideration.
© European Union, 2011. This document should not be considered as representative of the Commission’s official position.
The aim of the project in the sense above was more than substantially accomplished. Certainly the project partners did analyse realworld networks in collaboration with stakeholder organizations who are involved in emergency planning of critical infrastructures. To
these data were applied a number of techniques from complex systems analysis including, graph theory, time series analysis, recurrence
quantification analysis, dynamical systems theory and others. In some cases we have merged some applications from differing fields by
exploiting the varied expertise of the project partners, for example we applied spectral analysis measures usually used in earthquake
engineering to develop a hierarchy for weighted graphs.
- We have developed a cascade breakdown model which has produced initial output of breakdown scenarios. The early version,
while showing toy model cascading breakdown for the Hungarian power grid and resulting `blackouts’, has some unrealistic
close-down assumptions on redistribution following overloading of lines. The more sophisticated version involves an element of
linear programming optimization using the Kirchhoff laws. This follows a much more realistic decision-making process. It is also
underpinned by linear modelling which makes the approach scalable to continental levels.
- The first information that became available in the MANMADE project that we were able to use was Nordic electricity spot price
data. This examination was led by the Queen Mary and Westfield College of London (QMUL) and the University Carlo Cattaneo
of Varese (LIUC). The investigation was initially independent and then collaborative, and resulted in several joint papers.
- LIUC have led and published extensively on the supply chain behaviour and its dependence on electricity disruption. Their work
has been concentrated on the problems of amplification and distortion of demand in a supply chain. By eliminating or controlling
this effect, it is possible to increase efficiency and cut redundancy. The main model that has been considered is a single-product
serial supply chain with discontinuities in order supply. Effective measures of supply have also been considered. There has been
joint work with the Macedonian Academy of Sciences and Arts of Skopje, the Collegium of Budapest and JRC on this.
- Theory has been developed on, so-called, influence spreading in complex networks which is applicable to examples such as
Power Blackouts. The approach is a stochastic dynamic model which treats the network on two levels: at the network level –
each node is treated as an active entity called site and, at the local level with a Markov chain representing the possible states
of each site. The general influence model is a network of interacting Markov chains and can be used for finding optimal ways
for spreading influence, and/or preventing influence from spreading. This has been applied to modal analysis and vulnerability
of complex networks: the case of segmentation of the EU power grid; vulnerability vs. reliability – definitions, investigating
vulnerability with nodes and edge removal under random (errors) and preferential (attacks), in particular modal weight based
attacks and comparison to the recent results.
- Network algorithms have been developed which produce graphs which are acutely prone to cascade breakdown. The networks
are designed to breakdown catastrophically by cascade breakdowns. These networks can be characterized in terms of nodedegree distribution. The ability to characterize these graphs gives the potential for finding a measure of how much real networks
differ from these extreme constructs. This would be a further tool in characterising or measuring the vulnerability of real networks
to cascade breakdown. It is quite well established that overloading transmission lines in transport or energy networks can trigger
a cascade of failures resulting in a critical breakdown. A future aim of this research is to investigate the connectivity created
by these graph building scenarios and the nature of the dynamics at the root of such behaviour, and ultimately to propose
countermeasures which can be employed to prevent such catastrophes.
- Load and fault-tolerant backbones of the trans-European gas pipeline network have been investigated. Combining topological
data with information on inter-country flows, we estimate the global load of the network and its tolerance to failures. To do
this, we apply two complementary methods generalized from the betweenness centrality and the maximum flow. We find that
the gas pipeline network has grown to satisfy a dual purpose. On one hand, the major pipelines are crossed by a large number
of shortest paths thereby increasing the efficiency of the network; on the other hand, a nonoperational pipeline causes only a
minimal impact on network capacity, implying that the network is error tolerant. These findings suggest that the trans-European
gas pipeline network is robust, i.e., error tolerant to failures of high load links.
Contract
MANMADE
Coordinator
Queen Mary and Westfield College (QMUL), Great Britain
EC-DG JRC - Commission of the European Communities, Directorate General JRC, Europa
COLB - Collegium Budapest Egyesület, Hungary
Partners
MASA - Macedonian Academy of Sciences and Arts, Macedonia
LIUC - Università Carlo Cattaneo LIUC, Italy
EC-contribution
896.010,97 €
Full partner and project information available on http://cordis.europa.eu/fp6/projects.htm
The coordinator provided text and pictures for the factsheet and his copyright is acknowledged
http://ec.europa.eu/research