Call Announced: February 2017




Application deadline: May 2017

Decisions to applicants: June 2017

Start/end date: June 2017 - Dec 2017

Funding available: Up to £40,000 to be divided amongst the selected proposals

Call Document

The submissions will be reviewed according to the parameters stated in the call:

  • Match to network remit

  • Engagement of early career academics

  • Realistic objectives within the constraints of funding and time

  • Multi-disciplinary and innovative science

  • Industrial support

  • Value in enhancing research activity

  • Alternative funding sources 


After a thorough review process by the Management Team and Steering Committee, the following were awarded funding:

Dr. Weisi Guo (The University of Warwick), Prof. Liz Varga (Cranfield University)

Title: Nature Inspired Routing for Resilient Networked Systems

Complex systems inevitably experience perturbations, and as interdependencies between infrastructure layers evolve, vulnerabilities to said perturbations may emerge which lead to cascading failures. The aim of this study was to analyse the resilience of the Greater London rail transport system to cascade delays and station closures during peak morning hours. Analysis was carried out via hierarchical trophic coherence analysis, which allowed analysis of network coherence and network core interconnectivity.

  Analogy between Ecology Food Web and Transport Network Energy Transfer Expressed as a Hierarchical Graph with Trophic Levels.

The study found that the network exhibits a particularly low resilience to delay cascades as a result of its incoherent structure (this is true even when only flows from urban to suburban locations are considered i.e. commuter flow through stations is approximated as largely unidirectional), but notes a high resilience to station closures resulting from its highly interconnected core. The opportunities that biological network modelling may offer to the evaluation of resilience in complex systems are clearly presented in this report.



Warren Greig (Cranfield University)

Title: Interdependencies, what do we know, how do we come to know it?

Though we understand each infrastructure layer to some extent, we do not yet possess an adequate understanding of how infrastructural systems interact i.e. we rarely recognise and model multiple infrastructures as one complex system. This has implications in our understanding of resilience risks and failure propagation (e.g. where might a perturbation located on one infrastructural layer propagate to cause failure or fault on another?). Furthermore social and economic changes can cause interdependencies to change, and as interdependencies increase, additional system vulnerabilities may be exposed. The review explores our current understanding of the interdependencies between energy, waste, water, telecommunications and transport infrastructure, via a discussion about the ways in which such interdependencies have previously been modelled, and aims to answer the question ‘How have infrastructure interdependencies been defined and modeled in the literature”?’. The author then suggests a list of changes to complex system modelling methodology that should be implemented in order to improve consistency and understanding of the field.

The author identified 62 articles associated with interdependency modelling using a carefully constructed web of science search string. Though all authors tended to define at least one unique interdependency type, only 5 were frequently reoccurred – Physical (40 times), Cyber (24), Geographic (40), Logical (24), Functional & Spatial (11). Papers focused overwhelmingly on Energy (51), telecoms (31), transport (27), and to some extent on waste (12) and water (12). There was some consideration of soft infrastructure (8). Operational state of the infrastructure modeled varied between articles, though the effects of repair (12) and system stress or disruption (10) on resilience were somewhat common. Types of failure investigated varied between reports, though cascading failure (19) was the most frequently examined.

The author highlights the fact that Interdependency is a field with no standard practice, and that a common framework and language need be developed to ensure effective collaboration and rapid progress. It is also noted that there are very clear gaps in the study and modelling of human and behavioral interaction effects on infrastructures, and that studies considering this type of interaction will likely need to involve social science tools, such as surveys. However, the author believes that integrating existing social sciences research into existing interdependency models may be feasible, and that this may be an interesting area to examine.


Dr Chao Long, Dr Lee Thomas, Prof. Jianzhong Wu (Cardiff University)

Title: Feasibility of applying blockchain and smart contract technology to distribution grid management in the GB power system

The authors simulate the interaction of agents with smart contracts, and assess the impact this could have on the state of the distribution network. 3 agents (each operating from an individual laptop) trade their energy on a private blockchain, and energy traded is transferred across a simulated distribution grid. The power time series produced can be used as input to a power flow solver which contains a model of the distribution network, and the network state information acquired from power flow studies can be used to determine additional trading rules that must be integrated into smart contracts to maintain acceptable grid conditions and functioning.



The authors were successful in constructing the required trading environment, and thus the feasibility of such a trading mechanism has been proven. The authors intend to develop the simulation such that the results of the power flow are used to adjust the terms of the smart contracts, so that energy trading does not cause distribution network violations. The study has attracted the interest of several DNOs, who have provided 150,000 in funding for related work. The work has promising implications for the development and of peer to peer (P2P) trading schemes, and presents the beginnings of an investigation into how such schemes may need to be regulated to ensure network reliability.


Sathsara Abeysinghe, Meysam Qadrdan (Cardiff University)

Title: Generating random-realistic topologies for electricity distribution networks 

To ensure the future resilience of the power distribution grid, we must understand the impacts that new technologies and behaviours may have on it. However, in order to extensively prove resilience, we must perform simulations using many variations of network topology as different topologies exhibit very different behaviors under stress. As real network data is limited and often difficult to obtain, it is helpful to be able to generate a large database of representative network topologies. In this study, the authors obtained GIS data detailing the topological and electrical properties of UK distribution networks owned by western power grid, and used this to determine how the properties substation count (ground and pole mounted, 11 kV/400 V and 33 kV/11 kV), line length (underground and overhead) and network box counting dimension, relate statistically to population density. An algorithm was developed which uses population density as an input to generate representative network topologies.



The authors were able to generate networks with similar number of 11kV subs as in real networks, and with comparable distribution. The authors hope to further develop the algorithm to better simulate cable path lengths between network nodes, overall cable length per unit area, and network component and cable ratings. The further development could allow power network researchers access to a virtually limitless pool of topologies, rather than the limited pool of test cases that are currently available, and thus improve our ability to test complex systems under uncertainty.


Giuliano Punzo (University of Sheffield)

Title: Compartmental analysis of Infrastructures

Draws the analogy between compartmentalization studies in immunology and failure in complex engineering systems; an otherwise healthy component in certain systems fails under stress if overloaded as a result of the failure of other components to which it is in some way connected, in the same way that a healthy individual without immunization becomes more susceptible to infection if spatially and socially connected to a greater number of infected individuals. The author uses the compartmentalization analogy to examine the evolution of transport disruption and ageing of transport nodes in airline and small world networks.



The analysis suggests that synchronised failure can result from infrastructure ageing, and that this effect may be mitigated through heterogeneous distribution of node degrees e.g. the presence of dominant hubs appears to reduce the probability of synchronised failure. Furthermore, frequent network rerouting may also increase the probability of synchronised failure. 



The author notes that a directed ring network stabilises to give equal occupancy and infection rates at each node, whilst a heterogeneous network results in unequal occupancy at each node, but universally low chance of infection from neighbouring nodes in a 4 node system. The project provides an insight into how appropriate use of epidemiological models may aid our understanding and predictive power with regards to complex engineering system resilience.

Dr. Sabato Manfredi (University of Naples),  Prof. Vito Latora (Queen Mary, University of London)

Prof. David Angeli (Imperial College London)

Title: Dynamics and Resilience of Multilayer Cyber-Physical Social Systems

The understanding of complex socio-technical and cyber physical systems relies on our ability to model their different layers, and the interactions between them. These can be represented as the physical, information, and decision layers. By coupling these layers, it is possible to obtain models of Cyber-Physical systems (CPS), Cyber Social Systems (CSS) and Cyber Physical and Social Systems (CPSS). This feasibility study presents abstract models of CPS, CSS and CPSS systems, describing them analytically. Numerical simulations for a CPSS model are then presented. These give insight into the systems resilience, and the dependence of this on the systems physical configuration. The work is developed through 4 research outputs delivering:

1. Analysis and dynamic modelling of the social and cyber-physical (engineering) system layer.
2. Modelling of opinion dynamics (i.e. diffusive model) at the social layer and study of the onset of a unique, or cluster of opinion equilibrium. Investigation of the effect of network layer topological feature on the opinion equilibrium.
3. Investigation of decentralised optimising control strategies for CPSS. In this set-up, the social users are modeled as agents taking utility functions that interact with an engineering layer characterized by specific constraints in terms of energy and service rate.
4. Analysis of dynamic CPSS performance under finite node capacity and link velocity variations. Investigation of interplay between CPSS performance and topological features.



The authors were successful in modelling the attainment of consensus in the multi-agent networked system (in which agents possess individual characteristics) under weak connectivity conditions. The investigation into opinion dynamics through mutual interactions on a finite number of alternatives (using a nonlinear network model) showed that convergence to consensus or to clustered opinions is dependent on the directed or indirect characteristics of the interaction graph. A distributed flow control law to deal with the congestion and energy autonomy problem in energy harvesting for wireless sensor networks was proposed, and the dynamics of a multi-layer transport network with finite buffer size was analysed. Optimal characteristics in terms of journey times and packet loss (as a function of the buffer size and the speed of the two layers) were shown to be related to the structure of the network through betweeness centrality.



Share the NEWS