A Case Study of Challenges Facing the Design of Resilient Socio-technical Systems
Introduction
In the general sense of the word, resilience refers to the ability to withstand and recover from various types of stresses. This concept has gained significant popularity in as disparate fields as psychology, engineering, and ecology (Birkland and Waterman, 2009; de Bruijne, Boin and van Eeten, 2010; Woods, Schenk and Allen, 2009). The emphasis in the field of resilience engineering lies at the abilities of individual teams and organisations to adapt and survive in the face of various disturbances. However, the resilience studied at one level depends on the influences from levels above and below (Woods, 2006). These different levels operate at different timescales, and there may be tensions and incompatibilities between them, which impacts the nature of the system’s resilience (McDonald, 2006). For this reason, it is important to investigate the relationships between the various levels of a socio-technical system, in order to understand the ways in which resilient performance is achieved. This chapter therefore aims at paying attention to the way resilience of a socio-technical system is influenced by the interplay between various stakeholders in a multi-level and multi-actor context.
As a basis for the analysis outlined in this chapter, a case study of the decision-making process at the design stage of railway tunnel projects in Sweden is presented. Decision-making in this type of projects is undertaken in settings characterised by a multiplicity of diverse actors and perspectives, where no single actor has the authority to make a final decision. In this way, the present context is characterised by some crucial differences from studies of resilience in individual teams and organisations, and the chapter aims at contributing with new insights as well as areas for further consideration to the field of resilience engineering.
The case study presented in this chapter takes as a point of departure four factors described by Woods (2003) as important challenges for managing a system’s resilience. These factors include:
• Failure to revise assessments as new evidence accumulates
• Breakdowns at the boundaries of organisational units
• Past success as a reason for confidence
• Fragmented problem-solving process that clouds the big picture
These four factors are used as a basis for the analysis presented in the subsequent sections, and in this respect, the chapter has some similarities with the study presented by Hale and Heijer (2006). Before the case study is further described, some background of the decision-making process in Swedish railway tunnel projects will be outlined.
The Decision-making Process at the Design Stage of Railway Tunnels
The case study presented in this chapter builds upon semi-structured interviews with a total of 18 respondents involved in the decision-making process at the design stage of railway tunnel projects in Sweden. Six different railway tunnel projects were included in the study, and together these projects encompass 28 tunnels, ranging in length between 180m and 8.6km. Preliminary findings from this case study have been presented in Cedergren (2011), and for a more extensive presentation of the case study, see Cedergren (2013).
Two main sets of actors are involved in the studied decision-making process. These key players are shown in Figure 8.1, which schematically outlines the decision-making process in question. The first set of actors will be referred to as the project team. The project team consists of employees from the Transport Administration, which is the national authority responsible for building and maintaining railway infrastructure (as well as road infrastructure). In addition, the project team includes various consultants, which are appointed by the Transport Administration to this type of projects, particularly for conducting risk assessments and other types of safety-related documentation. The second set of actors will be referred to as the local actors, and consists of the local building committee and the local rescue service from the municipality in which the tunnel will be constructed. The role of these players will be further described in the following sections.
Figure 8.1 Schematic outline of the main actors involved in the decision-making process at the design stage of railway tunnel projects
Results and Analysis
As described above, the decision-making process in the studied railway tunnel projects has been analysed from a resilience engineering perspective, with special consideration of the ways in which the interplay between various stakeholders in a multi-actor and multi-level context influences the system’s ability to perform resiliently. The results from this analysis are outlined in the following sections, under the four headings that were used as a point of departure.
Failure to Revise Assessments as New Evidence Accumulates
One of the key issues for decision-making at the design stage of railway tunnels relates to the provision of means of evacuation from the tunnels. While the Building Codes regulate means of evacuation for most other types of buildings in Sweden, they are not applicable to railway tunnels. Instead, another set of laws and regulations are applied to this type of constructions. According to these laws, an approved building permit is required from the local building committee before a railway tunnel can be taken into operation. This means that the local building committee (the local authority) needs to approve the design suggested by the Transport Administration (the national authority). In the studied projects, the local building committees experienced that they did not have sufficient competence on questions related to risk, safety or means of evacuation. For this reason, they usually appointed the local rescue service as their expertise on these issues. As a result, these local actors held a prominent position in the decision-making process.
Another important aspect of the laws applied in railway tunnel projects is the lack of a risk acceptance criterion for the design of railway tunnels. For this reason, the Transport Administration has issued an internal handbook for the design of railway tunnels, which builds upon a risk-based approach. According to this handbook, railway tunnels need to be equipped with various types of safety measures in such way that the overall level of risk, as estimated in risk assessments, meets the established risk acceptance criterion. One of the most important types of safety measures in railway tunnels is the evacuation exits, which are located at regular intervals inside each tunnel. These evacuation exits are also used by the rescue service as intervention points in the case of fires or other types of emergencies.
Due to the significant costs associated with each additional evacuation exit, the Transport Administration are normally reluctant to design tunnels with a larger number of exits than deemed required from their risk assessments. However, in most of the projects included in this study, the distance estimated in these risk assessments was seen as too long by the rescue services. For this reason, the rescue services rejected the risk-based approach adopted by the Transport Administration. In contrast to the risk-based approach to decision-making adopted by the Transport Administration, the rescue services adopted a more deterministic approach. In their view, the occurrence of a fire inside a tunnel was taken as the starting point for decision-making, regardless the probability of such event.
In order to minimise the walking distance inside a potentially smoke-filled tunnel, the rescue services plead for a shorter distance between the evacuation exits. The different stakeholders thus drew on different types of evidence claims for proving their point in the decision-making process. The “evidence” highlighted by the rescue services rested upon experience from rescue operations in similar environments, whereas the “evidence” presented by the Transport Administration was based upon the outcome from risk assessments and cost-benefit considerations. These divergent perspectives implied that the problems in question were looked upon in contrasting ways, and different solutions were advocated. This resulted in a decision-making situation in which none of the players were able, or at least not willing, to revise their assessments when the other player presented new evidence.
Breakdowns at the Boundaries of Organisational Units
The various types of evidence presented by the different players reflected their diverse framings of the risks associated with railway tunnels. As a result, controversies arose in several of the studied projects, particularly with regards to the distance between evacuation exits. As described above, the Transport Administration used a risk-based approach for estimating this distance. However, in most projects, the rescue services did not agree upon the distance between evacuation exits estimated by the Transport Administration. In order to intervene in the case of emergencies, they argued for a shorter distance between these evacuation exits. They also argued for a need to provide additional safety measures and rescue equipment, such as smoke ventilation and water pipe systems. Unless these additional measures were included in the design, the local authorities were not willing to approve the building permit. In this way, members of the project teams in several projects experienced that the rescue service tried to “kidnap” the building permit.
As a result of the additional demands raised by the rescue service in many projects, the Transport Administration experienced that they were trapped in a double bind. On the one hand, agreeing upon the additional demands raised by the rescue service would lead to increased costs for the project. On the other hand, rejecting the demands raised by the rescue service would imply that the project would be delayed, which also resulted in increased costs. Consequently, no matter what actions they took, it would lead to undesired outcomes.
In a similar way, the local actors experienced that they were trapped in another kind of double bind. On the one hand, approving the design suggested by the Transport Administration could potentially (in the aftermath of an accident with severe consequences) imply that they would be blamed for having approved the construction of a railway tunnel with an insufficient safety standard. On the other hand, disapproving the building permit would imply that they would be blamed for delaying the project, which typically constituted an important infrastructure investment for the region in question. In this way, no matter what actions they took, they experienced that they would be blamed for the outcome of their decision.
The double binds experienced by both of the main players involved in the decision-making process resulted in controversies and deadlocks in many projects. These breakdowns at the boundaries between the various organisations illustrate the challenges associated with decision-making in settings characterised by a multiplicity of stakeholders with diverse roles and perspectives.
Past Success as a Reason for Confidence
In order to reconcile the deadlocks emerging in several projects, and thus attaining an approved building permit, the Transport Administration agreed upon some of the demands raised by the rescue services. For example, in one of the projects the Transport Administration agreed upon the demands raised by the rescue service on providing a water pipe system. However, by agreeing upon this safety measure in this particular project, the rescue services in each succeeding project also raised demands on provision of the same type of water pipe system. In this way, each safety measure that was approved in a specific railway tunnel project gave rise to a “precedent”, that is, a decision made in one project which was used as a justification for making the same decision in future projects. As a result, the level of safety measures was gradually raised in each consecutive project.
As a reaction to the increased demands for safety measures in each new project, several members of the project teams emphasised the need to take a cost-benefit perspective on safety investments in railway tunnels. In their view, the low probability of accidents in railway tunnels did not justify the large amount of safety measures demanded by the rescue services, and the lack of previous accidents in railway tunnels was taken as a reason for confidence in this viewpoint. For these players, the establishment of precedents created frustration. In addition to the increased costs and delays associated with the establishment of precedents, the large influence from decisions reached in previous railway tunnel projects downplayed the role of the risk assessments. In this way, the risk assessments were not the only, and not even the most important, basis for decision-making. Rather, decisions were in many cases highly influenced by negotiations between the main actors involved in the decision-making process. For this reason, members of the project teams questioned the value of spending considerable resources on conducting risk assessments when these assessments, in the end, were not used as a basis for decision-making.
So far, it can be concluded that an informal decision-making process was evolving in parallel to the formal decision-making process, in which risk assessments were conducted as a basis for decisions. As a result of this informal process, precedents, negotiations, and power relations between the different players were highly influential for reaching a final decision. The next section describes the ways in which the railway system’s ability to perform resiliently was affected by these processes.
Fragmented Problem Solving Process that Clouds the Big Picture
The controversies arising in many of the railway tunnel projects arose as a result of the diverse framings held by the local and national authorities. Due to the need for an approved building permit in these projects, the viewpoint adopted by the local actors gained significant influence in the decision-making process. This viewpoint primarily focused on railway tunnels from a local perspective, that is, the way each tunnel was equipped with different types of safety measures (in order to ensure evacuation and rescue operation). However, due to the large emphasis on local aspects, limited attention was devoted to the railway system’s performance from a regional or national perspective. The bigger picture of the railway system’s functioning, including its ability to perform resiliently in the face of failures, was thus overlooked. In this way, no single actor had a coherent view of the railway system and its functioning at the overall level. This was particularly clear with regards to the choice of tunnel type and provision of means of evacuation. Different ways of building railway tunnels are associated with different solutions for means of evacuation, and two of the solutions adopted in the studied projects are illustrated in Figure 8.2.
Figure 8.2 Different tunnel types with different solutions for providing means of evacuation. A) schematically illustrates a tunnel type with two parallel railway tunnels, and B) schematically illustrates a single-line tunnel type with a dedicated parallel evacuation tunnel (walking paths for evacuation not illustrated in the figures)
Figure 8.2a shows a solution in which two parallel railway tunnels are connected with rescue tunnels at regular intervals. Figure 8.2b shows a single-line railway tunnel with a dedicated parallel evacuation tunnel. Due to technical constraints, the evacuation tunnel in this latter design had to be constructed of almost the same size as the full-sized rail tunnel. If the evacuation tunnel had been made somewhat larger, it would have been possible to prepare this tunnel for using it as a redundant railway tunnel (in effect, transforming it into the same type as shown in Figure 8.2a). According to several respondents, this solution would only imply a negligible additional cost, but provide the railway system with a significantly improved ability to continue operations in the face of failures affecting the single-line railway tunnel. However, since focus in the decision-making process was concentrated to the local features of each tunnel, consideration of the ways in which each tunnel affected the performance of the larger part of the railway system was restricted. In this way, the railway system’s ability to perform resiliently from a regional and national perspective gained limited attention in the projects.
Discussion
In this chapter, special consideration has been paid to the ways in which resilient performance of a socio-technical system is shaped by the interplay between actors with diverse roles and perspectives. As shown in the case study presented in the previous sections, the various players involved in this type of multi-actor and multi-level setting framed risk in different ways, and as a result, they drew on different evidence claims in order to influence the risk management process (cf. van Asselt and Renn, 2011). This decision-making situation, in which no single actor had superior authority, resulted in disagreements and controversies among the key players. Due to these disagreements, both of the two key players experienced that they were trapped in different types of double binds, that is, decision-making situations leading to undesired outcomes no matter what actions were taken (cf. Dekker, 2006). While the municipal actors experienced a double bind related to blame, the project team experienced that they were facing increased costs no matter what actions they took. In parallel to the formal decision-making process, in which risk assessments were conducted as a basis for decision-making, an informal process evolved. In this process, “precedents”, that is, decisions made in previous projects, played an important role. As a result, the outcome of risk assessments was downplayed, whereas greater weight was placed on efforts to reach agreements through negotiations between the different players.
While there is no doubt that studies of individual teams and organisations provide valuable insights to the field of resilience engineering, this chapter demonstrated that it is important to also take cross-organisational aspects into account (see also Mendonça, 2008). In large-scale socio-technical systems, resilience is an ability characterising the system at the macro-level. However, the macro-level abilities are created by the individual actions and decisions taken at the micro-level (cf. Vaughan, 1996). For this reason, understanding of the way that resilience of socio-technical systems is fostered requires close consideration of the cross-scale interactions between various levels (see also McDonald, 2006; Woods, 2006), that is, the link between activities at the local level, and their effects on the global level. By restricting studies to a single actor’s perspective, the ways that resilient performance of a system is affected by the interplay between various players in a horizontal as well as vertical dimension go unheeded. For this reason, the objective of this chapter was to apply a resilience engineering perspective to the analysis of a large-scale socio-technical system by adopting the four factors described by Woods (2003) as important challenges for managing a system’s resilience. While it can be concluded that these factors negatively influence the ability of a socio-technical system to perform resiliently, it is not possible from this limited case study to claim the opposite, that is, to determine that resilient performance emerges simply by reversing these factors. Rather, the main contribution of this chapter was to present additional insights into the challenges facing the design of resilient systems. In particular, the results from this study demonstrated that double binds, negotiations, precedents, and not least, power relations between the various actors, play essential roles. In order to design resilient socio-technical systems, these aspects need further consideration, and the chapter underlines the importance of paying attention to the multi-actor context in which socio-technical systems are designed and managed.
Concluding Remark
While the various activities and processes going on at the level of individual teams and organisations have gained considerable attention in the field of resilience engineering, less focus has been directed towards the effects these activities have on the potential for resilient performance of the higher levels of a socio-technical system. The case study presented in this chapter showed that the system’s ability to perform resiliently at the regional and national level was constrained by the local perspective adopted by the various stakeholders. This chapter has thus contributed by emphasising the need to take the cross-scale interactions between various levels of a socio-technical system into consideration by illustrating the ways in which these interactions poses challenges to designing resilient systems.
Commentary
This chapter shows how the basic arguments presented by Komatsubara (Chapter 7) can be applied in a more formalised manner to a concrete example. It has provided an illustration of the many dependencies and potential conflicts that exist in the real world, and that have to be taken into account when trying to manage large projects. The temporal relations, in particular, are often very important but difficult to see, especially if safety management relies on a structural description of the organisation. The overall functioning, and therefore also the resilience, should not be considered without acknowledging the intricate interplay between the multiple stakeholders that exist across levels and contexts.
References
Birkland, T.A. and Waterman, S. (2009). The Politics and Policy Challenges of Disaster Resilience. In C.P. Nemeth, E. Hollnagel, and S. Dekker (eds), Resilience Engineering Perspectives, Volume 2: Preparation and Restoration. Farnham: Ashgate Publishing Limited, 15–38.
Cedergren, A. (2011). Challenges in Designing Resilient Socio-technical Systems: A Case Study of Railway Tunnel Projects. In E. Hollnagel, E. Rigaud, and D. Besnard (eds), Proceedings of the fourth Resilience Engineering Symposium. Sophia Antipolis, France: Presses des MINES, 58–64.
Cedergren, A. (2013). Designing resilient infrastructure systems: a case study of decision-making challenges in railway tunnel projects. Journal of Risk Research, 16(5), 563–582. doi:10.1080/13669877.2012.726241
de Bruijne, M., Boin, A. and van Eeten, M. (2010). Resilience: Exploring the Concept and Its Meanings. In A. Boin, L.K. Comfort and C.C. Demchak (eds), Designing Resilience: Preparing for Extreme Events. Pittsburgh: University of Pittsburgh Press, 13–32.
Dekker, S. (2006). The Field Guide to Understanding Human Error. Aldershot: Ashgate Publishing Limited.
Hale, A. and Heijer, T. (2006). Is Resilience Really Necessary? The Case of Railways. In E. Hollnagel, D.D. Woods and N. Leveson (eds), Resilience Engineering: Concepts and Precepts. Aldershot: Ashgate Publishing Limited, 125–147.
McDonald, N. (2006). Organizational Resilience and Industrial Risk. In E. Hollnagel, D.D. Woods and N. Leveson (eds), Resilience Engineering: Concepts and Precepts (pp. 155–180). Aldershot: Ashgate Publishing Limited, 155–180.
Mendonça, D. (2008). Measures of Resilient Performance. In E. Hollnagel, C.P. Nemeth and S. Dekker (eds), Resilience Engineering Perspectives: Remaining Sensitive to the Possibility of Failure. Aldershot: Ashgate Publishing Limited, Vol. 1, 29–47.
van Asselt, M.B.A. and Renn, O. (2011). Risk governance. Journal of Risk Research, 14(4), 431–449.
Vaughan, D. (1996). The Challenger Launch Decision: Risky Technology, Culture, and Deviance at NASA. Chicago: The University of Chicago Press.
Woods, D.D. (2003). Creating Foresight: How Resilience Engineering Can Transform NASA’s Approach to Risky Decision Making. Testimony on The Future of NASA for Committee on Commerce, Science and Transportation. John McCain, Chair, October 29, 2003. Washington D.C.
Woods, D.D. (2006). Essential Characteristics of Resilience. In E. Hollnagel, D.D. Woods and N. Leveson (eds), Resilience Engineering: Concepts and Precepts. Aldershot: Ashgate Publishing Limited, 21–34.
Woods, D.D., Schenk, J. and Allen, T.T. (2009). An Initial Comparison of Selected Models of System Resilience. In C.P. Nemeth, E. Hollnagel and S. Dekker (eds), Resilience Engineering Perspectives, Volume 2: Preparation and Restoration. Farnham: Ashgate Publishing Limited, 73–94.