Version classiqueVersion mobile

Proceedings of the fourth Resilience Engineering Symposium

 | 
Erik Hollnagel
, 
Éric Rigaud
, 
Denis Besnard

From technical resilience toward urban services resilience

Serge Lhomme, Marie Toubin, Damien Serre, Youssef Diab et Richard Laganier

Résumé

Urban resilience will be the next objective of many countries if they want to overcome the worthening disasters and the growing global competition. But the specificities of urban systems, at both technical and organisationnal level requires new methods to tackle complex behaviours and interdependencies. Indeed, technical networks are considered here as the backbone of the city, providing energy, water and other services to every point of the territory. Thus, they need to be resilient in order to ensure a smooth functionning and the continuity of service in case of disturbance. Inspired from safety methods, failures and disturbances of technical networks can be thoroughly assessed and geolocated with a dedicated tool. Then the operability of the method is ensured by a collaborative work between network managers and city services. Based on the scientific assessment of failures and interdependencies, a shared diagnosis highlights the interactions between actors' decisions and fosters an integrated planning in order to improve urban resilience.

Note de l’auteur

Note portant sur l’auteur1

Note portant sur l’auteur2

Note portant sur l’auteur3

Note portant sur l’auteur4

Note portant sur l’auteur5

Texte intégral

1 Introduction

1The year 2007 marks a turning point in history: half of the world population now lives in cities [UN-Habitat, 2007]. Moreover, the total urban population is expected to double from two to four billions over the next 30 to 35 years [United Nations, 2006]. This growing rate is equivalent to the creation of a new city of one million inhabitants every week, and this during the next four decades [Flood resilience Group]. So, this urban development coupled with technical failures has increased risk and corresponding challenges to urban risk management [Ashley et al., 2007]. For instance, volcanic eruptions, earthquakes, hurricanes, cyclones, tornados, tsunamis, fires, floods, landslides, avalanches and flash floods cause colossal loss of life and economic damage every year in the urban environment. Despite efforts made to tackle hazards, we often observe hazard defense failures leading to finally increased risk in protected area. That is why hazard management is not sufficient to manage urban risk. These circumstances oblige to manage urban risk by integrating new concepts like urban resilience. Thus, disaster management has been moving away from solely emergency response, initiated during and after a hazardous event, toward mitigation and preparedness, initiated before an event, in order to reduce impacts more effectively [Wilhite, 1987; Hooke, 2000; Fontaine & Steinemann, 2009]. First, concerning urban risk management, vulnerability concept has been developed and secondly the concept of resilience has been integrated. « At the urban scale, for example, vulnerability, seen as an internal risk factor, must be related not only to exposure of the material context or to the physical susceptibility of the exposed elements but also to the social frailties and lack of resilience of the prone communities » [Cardona, 2003]. Therefore, one of the objectives in order to decrease urban vulnerability is to improve urban resilience. A research prospectus concerning urban resilience has precisely been submitted by the Resilience Alliance. « The Urban Resilience program will focus research on the major challenges facing urban systems and the landscapes they comprise. The same questions arise for urban as for regional social-ecological systems: how much and which kinds of disturbances can urban areas absorb without shifting to alternative less desirable system regimes? ». This new paradigm in urban risk management needs new methods to shift from conceptual analysis toward operational tools.

2In the first part, we will discuss the interest of methods stemming from safety engineering in network management and in particular their application in resilience management. Then we will emphasize the role of these assessment methods in decision-making and demonstrate that they need collaborative approaches to be implemented efficiently by the local authority in order to improve the global resilience of urban services [Pinel, 2009].

2 The use of Safety methods to improve networks resilience

3For a critical infrastructure, getting dysfunctional is a phenomenon that transcends by far the failure of any, even major, single component. The often incomprehensible cause of system crashes stems from the inherent features of the critical infrastructures: they are multicomponent systems, prone to cooperative behaviour, and typically responding in a non-linear fashion to stimuli and perturbations. Then we propose the use of safety methods to study impacts of hazards on network infrastructures and to study interdependencies between different networks. Based on a functional analysis of the network infrastructures, FMEA (Failure Mode and Effect Analysis) was carried out for each network. Then, using these FMEA, failure scenarios can be designed. Indeed, we design network systems failure scenarios by linking failure causes to failure modes, and then to failure effects. In this way, the failure mechanisms are modelled as series of functional failures representing the relevant physical processes taking place within the system and leading to loss or deterioration of functions. Taking into account an important components number, a tool has been developed in order to automate scenarios design.

2.1 The methodology

4New paradigm in urban risk management needs new methods to shift from conceptual analysis toward operational tools. In this way, we first study city’s specificities in order to have a better understanding of its functioning. Power and influence of a city are based on its ability to provide a safe environment to its citizens and an efficient support to its activities. Urbanity is then linked with resources accumulation enabled by economic activities that depend on powerful networks. Indeed, as illustrated in the literature, a city appears as a set of components interconnected by networks. In this urban system, networks affect the well-being of the people and the smooth functioning of services and, more generally, of economical activities. For instance, over 19 billions tons of freight valued at $13 trillion dollars was moved through the U.S. multimodal transportation system during 2002 [U.S. Department of Transportation, 2006]. So, evaluating network infrastructures for potential vulnerabilities is an important component of strategic planning, particularly in the context of managing and mitigating service disruptions [Murray et al., 2008]. Yet, multiple networks that innervate the city are particularly sensitive to hazard, through their structures and geographic constraints. There is a need to understand how networked systems are resilient because societal functions are highly dependent on networked systems and the operability of these systems can be vulnerable to disasters. The first step for that is to study their failure modes and effects, in particular interdependencies between different networks thanks to risk analysis.

5It appears that risk analysis is not a discipline but rather an engineer activity who tries to assess the risk. In fact, there are several approaches to fulfill this activity, which are clearly identified and formalized in the industry (nuclear plant, aeronautic…). They can be gathered into two families [Zwingelstein, 1995]: internal methods and external methods. Internal methods are based on detailed knowledge of the system functioning that is why these methods are relevant to study complex infrastructures like networks. Functional modeling is one of the two main internal methods.

6The principle of functional modeling is to study the interactions between components of a system and its environment in order to establish a link between the functions failure, their causes and effects. There are various techniques for functional modeling systems: Analysis of Failure Modes and Effects Analysis (FMEA), FMEA completed by a criticality analysis (FMECA), the methods of the Tree or Consequences Events... Functional modeling allows better understanding of how the system operates and that’s why it allows better understanding of failure mechanisms. Thus it is possible to produce failure scenario and to take into account complex interdependencies thanks to these methods [Lhomme et al., 2010].

7For this study, we mainly used FMEA. FMEA is a procedure to identify component failures which have significant consequences affecting the system operations in the application considered. FMEA only provides qualitative analysis. Firstly, FMEA requires breaking down the system into components (structural analysis). Secondly, it’s necessary to identify the functional structure of the system and how the components contribute to functions. Then FMEA requires defining failure modes of each component and finishing perform analysis for each failure mode of each component and recording results in table.

8After compiling the FMEA data, we can determine the most important failure modes of the systems, their causes and their effects. So, using the FMEA, the failure mechanism model had been defined, and failure scenarios had been designed thanks to event trees. The event trees analysis was developed in early 1970 for risk assessment of nuclear power plants. Here we are just using the method without quantitative aspects, but this model highlights domino effects induced by networks failure.

9We design network systems failure scenarios by linking failure causes to failure modes, and then to failure effects. In this way, the failure mechanisms are modeled as series of functional failures representing the relevant physical processes taking place within the system and leading to loss or deterioration of functions. So it’s possible to produce failure scenarios using this specific methodology.

2.2 Results

10The methodology, presented above, allows producing networks failure scenarios. Nevertheless, it is only possible to produce few scenarios. Maybe these scenarios are the most important and plausible, thanks to a good knowledge and expertise of the networks, but we can't take advantage of the overall FMEA analysis. For these reasons, a tool is needed to automate these scenarios and take advantage of the FMEA analysis.

11The tool responds to three main objectives. The first objective is to allow visualization and update of the FMEA. The second objective is to design failure scenarios. The third objective is to analyze the results and to allow an overall understanding of interdependent networks failure modes thanks to diagram representation of the results. In order to produce failure scenarios we implement FMEA in database and use the user interface to display results.

12Here, analyzing the results in order to validate the methodology developed is not a trivial issue. Indeed, the tool was designed because there are too much scenarios and it was impossible to produce all these scenarios manually. So it is also impossible to validate all the scenarios manually. The solution is to use the general results produce by the tool and to compare these results with general results based on feedbacks. The study concludes that electrical network is the most aggressive networks (Figure 1) and that some components are very problematic. Indeed, numerous scenarios involve these specific components. These components are not the origin or the final effect of these scenarios, but components which spread disturbance effect of a component. Thus these components disturb components which can not directly be impacted by an origin component (pump stations, road…). These conclusions are quite similar with general results based on feedbacks. That is why, even if it is impossible to validate all the scenario, the methodology seems to be valid.

Figure 1. Component failure effects (orange) or causes (blues) diagram

3 A necessary shift toward network managers

13The identification and assessment of the critical components of technical networks is the first step of the resilience improvement. Then, it is necessary to cope with the disturbances and implement measures to improve its resilience. Based on expertise and scientific methods, the acknowledgement of critical points is facilitated and should be accepted by all stakeholders. For the second step, the issue is to have those network managers share their diagnosis and collaborate in order to tackle the interdependencies between networks. Mutual learning and collaborative approaches are one way of ensuring the shared diagnosis and the efficiency of the subsequent measures.

3.1 Collaborative approaches and their interest for networks resilience

14Collaborative approaches appeared first in the environment management field, where the multiplicity of stakeholders and the complexity of issues could not be tackled by a single person any more [Olsson, Folke et al., 2004]. Several ways of having experts, local authorities, private interests and populations work together have been developed in the last decade, with more or less success. The objective is to have a common understanding of the issues of a project in order to facilitate a solution that satisfied the broadest interest possible. In resilience management, the framework is very similar, though the aim is quite different.

15The urban context gathers many different actors, with their particular power scope, their objectives and constraints. Networked services are ruled by one or several actors, being public or private, and they usually don’t collaborate with each other. Each sector is able to identify its own vulnerabilities and the use of FMEA enables the manager to know its dependencies to other networks. But, if interdependencies are not completely shared by all stakeholders, the legitimacy and honesty of the diagnosis can easily be questioned by the others. The acknowledgement is facilitated by a collaborative work and a mutual sharing in knowledge where confidence and collaboration are essential.

3.2 Tools and methods to enable a shared diagnosis of networks resilience

16Several tools are developed in collaborative approaches [Walker, Carpenter et al., 2002], but they have been designed to tackle issues of environment management and are specific of a local context. Moreover, the use of complex tools such as multi-agent systems or multi-criteria decision-making may not be appropriate for our purpose. Among all the collaborative processes, we are positioned here in the field of collaboration and not participation (which implies populations and private interests as well) and our purpose is problem identification (on the contrary to problem solving where the issue is already known). Indeed, the collaborative tool is to be used by experts and non-experts in order to ensure everyone has the same level of information and the same understanding of a situation [Ridder, Mostert et al., 2005]. The need for a mediator is then highlighted and the collaborative process needs a framework to be carried out efficiently. Sociology of translation is one way of improving mutual learning and it could be the basic principle leading to the development of a collaborative tool adapted to resilience management. The issue is to find a balance between expertise and decision-making.

17The question of expertise is indeed central in our research. When it is considered by some actors as the only valuable knowledge, it can also easily be contested by the others. Scientific hypothesis or parameters are as many questionable choices that can hamper mutual knowledge. Then FMEA should be carried out jointly by experts from each sector and the local actors. Another source of knowledge identified here is field knowledge, usually coming from local actors [Healy, 2009]. The confrontation of both expertises is not a difficulty for resilience assessment; they have to get rich from one another and give food for the shared diagnosis. Managing this mutual learning is one of the main issues of this research.

4 Conclusion

18With these two levels of expertise and the involvement of the corresponding stakeholders, resilience improvement should be effective. Indeed, when supported by scientific methods already acknowledged in industry, the assessment of critical points is necessarily admitted by network managers, all the more that they are involved in the process. The efficiency will rely on the quality of the tools used to model the networks behaviour and disturbance. In order to provide the more detailed information, a GIS software is developed and will be the basis for the negotiation between network managers who will have to plan and coordinate their actions for network resilience.

Bibliographie

References

Ashley, R., Blanksby, J., Chapman, J. & Zhou J. J. (2007). Towards Integrated Approaches to Reduced Flood Risk in Urban Areas, in R. Ashley, S. Garvin, E. Pasche, A. Vassilopoulos, C. Zevenbergen, Advances in Urban Flood Management, pp. 415- 432.

Cardona, O. D., (2003). The Need for Rethinking the Concepts of Vulnerability and Risk from Holistic Perspective: A Necessary Review and Criticism for Effective Risk Management, in Bankoff, G. et Frerks, D. H., Mapping Vulnerability: Disasters, Development and People, Earthscan Publishers, Londres.

Fontaine, M. F., Steineman, A. C. (2009). Assessing vulnerability to natural hazard: Impact-based mathod and application to drought in Washington State Natural Hazards Review (10), pp. 11-18.

Healy, S. (2009). Toward an epistemology of public participation. Journal of Environmental Management, 90 (4): pp. 1644-1654.

Hooke, W. H. (2000). “U.S. participation in international decade for natural disaster reduction.” Natural Hazards Review (11), pp. 2–9.

Lhomme, S., Serre, D., Diab, Y., Laganier, R. (2010). “Resilience: From ecology concept to urban application. A GIS for urban flood resilience” 25th International Conference on Environmental Science and Technology 2010, July 28-31, Environmental Conference Program, American Academy of Sciences, Houston 10p.

Murray, A. T., Matisziw T. C., Grubesic, T. (2008). “A methodological overview of network vulnerability analysis”, Growth and Change, Vol. 39 No.4, pp. 573-592.

Olsson, P., Folke, C., et al. (2004). "Adaptive Comanagement for Building Resilience in Social–Ecological Systems." Environmental Management, 34 (1): pp. 75-90.

Pinel, W. (2009). La résilience organisationnelle : concepts et activités de formation Centre Risque et Performance, Maîtrise en sciences appliquées, 142 p.

Ridder, D., Mostert, E., et al. (2005). Apprendre ensemble pour gérer ensemble -améliorer la participation à la gestion de l'eau. Osnabrück, Allemagne, HarmoniCOP, Université d'Osnabrück, Institut de recherche en systèmes environnementaux: 112 p.

United Nations, (2006). “World Population Prospects: The 2008 Revision » http://www.un.org/esa/population/publications/wpp2006/wpp2006.htm

United Nations Habitat, (2007). “State of the World’s Cities 2006-07 » http://www.unhabitat.org/pmss/listItemDetails.aspx?publicationID=2101, 108 p.

U.S. Department of Transportation, (2006). “Freight on america: a new national picture” http://www.bts.gov/publications/freight_in_america/pdf/entire.pdf

Walker, B., Carpenter, S., et al. (2002). "Resilience management in social-ecological systems: a working hypothesis for a participatory approach." Conservation Ecology, 6 (1): 17 p.

Wilhite, D. A. (1987). “The role of the gouvernement in planning for drought: where do we go from here”, Planning for drought: toward in societal vulnerability, Wilhite an Easterlin (eds.), pp. 425-444.

Zwingelstein, G. (1995). « Diagnostic des défaillances: théorie et pratique pour les systèmes industriels », Traité des nouvelles technologies, série Diagnostic et maintenance, Hermès, Paris.

Notes

1 Université Paris Diderot-Paris 7, UMR PRODIG, Paris, France Université Paris Diderot-Paris 7, UMR PRODIG, Paris, France Serge.lhomme@eivp-paris.fr

2 Université Paris Diderot-Paris 7, UMR PRODIG, Paris, France Université Paris Diderot-Paris 7, UMR PRODIG, Paris, France Marie.toubin@eivp-paris.fr

3 Université Paris-Est, Ecole des Ingénieurs de la Ville de Paris - EIVP, Paris, France

4 Université Paris-Est, Ecole des Ingénieurs de la Ville de Paris - EIVP, Paris, France Université Paris-Est, LEESU département génie urbain, Marne-la-Vallée, France

5 Université Paris Diderot-Paris 7, UMR PRODIG, Paris, France

Table des illustrations

Légende Figure 1. Component failure effects (orange) or causes (blues) diagram
URL http://books.openedition.org/pressesmines/docannexe/image/1043/img-1.jpg
Fichier image/jpeg, 295k

Le texte et les autres éléments (illustrations, fichiers annexes importés) sont sous Licence OpenEdition Books, sauf mention contraire.

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search