1

Industrial Doctorate Centre in Systems

Research Papers for the Engineering doctorate in Systems Annual Conference 2009/10 Cohort 1 (2006-07 start)

The UK Housing Market

Lucy Allan, DRTS Ltd. [email protected] Alan Champneys, Bristol University , [email protected] Patrick Godfrey, Bristol University, [email protected]

Introduction

In 1995 a major earthquake devastated the operations of Kobe port in Japan. It had been thought that the design of the port was adequate but events proved otherwise. Remediation scenarios after the disaster were focused on local and micro-scale solutions, e.g. the analysis of how sand behaved under seismic loading, and future- proofing followed a similar mind-set. However, the problem was both simpler and at the same time more difficult to defend Fig. 1 Perimeter wall displacement in Kobe Port against and to resolve; locally sourced sand had been used for filling material, with unintended consequences, as it didn‟t behave well under the strong seismic loading of the earthquake. Also, crucially, old seismic design codes hadn‟t been updated.

All theories tend to simplify and engineers have to make assumptions out of necessity, but often more than warranted by context, i.e. the port was economically vital to the community and country. An alternative approach is to use holistic systems thinking. This requires whole- life thinking, i.e. consideration of the implications of every decision, and recognition that what happens in one part of a system affects every other part. This can then lead to the creation of a wider design space in which more holistically conceived solutions can be formulated to any given problem. Ports have to achieve a harmonic, complex balance between the local community, the environmental integrity of their processes, and progress economically. It is not possible to design part of the port system in isolation without considering the problem and solution as a whole.

There are many definitions of a complex system however the most useful general description, sees a complex system as composed of interconnected parts that as a whole exhibit one or more properties not obvious from the properties of the individual parts. The

2 components that tend to define a complex system are the scale, nestedness, feedback loops, nonlinearity, emergence, hysteresis, and the degree of connectivity or coupling. Kobe is such a complex system and closer to home, so too is the housing market.

The UK housing market presents a unique opportunity for the study of a complex system, there is widely available public data, property and housing is central to most of our lives, the property market is viewed as a bellwether for the wider state of the economy, the location and price of property has many interesting social connotations, and recently we have become aware of housing and energy use in particular as a contributor to the levels of CO2 in the atmosphere.

Theory & Methods

There are many approaches to modelling complexity, (Forrester, 1961; McCulloch & Pitts, 1943; Weiner 1946; McCarthy, 1956; Wasserman et al. 1994). The approach developed by Checkland in the 1970s, termed the Soft Systems Methodology, (SSM), see Fig. 2. Checkland has many useful features. A real- world „rich picture‟ is central to the problem solving process. This model is compared to the real-world problem situation through a structured debate. Fig. 2 Soft Systems Methodology, Checkland.

It would appear that the engineers and planners for Kobe failed to think holistically of the „hard‟ and „soft‟ system elements, and failed to establish a rich debate. Models of the perimeter walls were analysed for their stability under seismic loading, water pressure, soil backfill loading, and vessel cargo loads, and how vessels would navigate the port. However, they failed to draw the problem boundary wide enough and incorporate into their models „soft‟ issues. They broke the Kobe port problem into separate parts for analysis and didn‟t see the interconnectivity of the risks and emergent behaviours brought about by using local sand, or how would nine months port downtime in Japan‟s second largest port and the world‟s sixth largest hub affect food and water supplies to central Japan, the Japanese stock market, and even worldwide cargo movement and supply.

And so we must ask ourselves if current econometric models for the UK housing market are adequate, or whether in fact uncertainties and dynamic relationships and interactions in an evolving, unruly and open-ended system have been dangerously simplified? Also we must ask if the economists in their desire for order through regression analysis are not somewhere missing the point, and that we must look at the data provided, all the data, in an open- minded way. And so we turn to Checkland and SSM to help generate the framework for rigorous debate.

3

Generally there are two approaches to be chosen for abstraction of real-world systems. The first is where the outputs are derived from the models, e.g. through regression approaches or more dynamically through agent-based modelling. The second is via a data driven approach where Bayesian and neural networks are examples. One quick and easy way, yet also robust and powerful modelling technique is through concept mapping by expert opinion – a similar approach to mind mapping but with slightly more rigor. The approach adopted in this research is to generate concept maps from data which can then be compared with the expert mappings of assumed reality, and thereby force debate. The value here is in the discussion and debate rather than in the underlying technologies.

Results & Discussion

We may consider mind maps as fairly simplistic but with many nodes, i.e. the resulting image may require far more skilled interpretation. Fig. 3 is an example of what a concept map of the housing data looks like*. Here more than 400 housing data variables have been fed into the map builder. Analysing the network more closely it is possible to see that the majority of the variables are connected to some degree and some are not connected at all. The two blue variables represent the most connected Fig. 3 Concept map of housing market data variables in the system and it is possible to think of them as „hubs‟ in the system.

When we examine the connectivity of the blue variables we see that the first is UK GDP, see Fig. 4. Without any knowledge of the market whatsoever it is possible to arrive at a very logical conclusion that the UK economy influences the housing market. Much more surprising and less intuitively it is possible to see that the second main hub is linked to housing benefit, i.e. the number of housing benefit claimants in the UK – see Fig. 5. This second result is a consequence of the housing market rather than a cause. However, here we are reminded of some of the potential for unintended outcomes from mortgage rates, easy lending and market dynamics, with an impact on real people‟s lives, unemployment, homelessness and the potential drain on local resources.

Fig.4 UK GDP Fig. 5 Claimants Housing Benefits

4

In this technique the data can speak for itself and real debate can be engaged to ensure models are realistic, whole, rational, and where the potential for emergent behaviour is considered. For new systems such as Kobe port, data from Monte Carlo simulations can be used for data generation and then expert models challenged.

Biography

[email protected]

Lucy Allan has a first class honours for her MEng in Civil Engineering with Spanish, from The University of Sheffield; she was a SET awards finalist in 2005; she won a prestigious scholarship from the Japan Society for the promotion of science in 2006, working in Japan for a year with leading institutions; she has also spent a year at the University Polytechnic de Catalunya (UPC) in Barcelona, researching issues on Environmental Engineering, Noise Pollution and Water Treatment. She is fluent in Spanish, and conversant in Japanese.

She joined DRTS Ltd from a research position at the University of Sheffield where she was studying the impact of earthquakes on ports and harbours. The focus of her work for DRTS is to develop systems, services and new technology for the development of sustainable infrastructure systems. DRTS are already working on tools for the aviation and automotive industries and although the civil construction industry has a major impact on all our lives, there are few mainstream systems and procedures for managing the thru-life impact of structures with respect to people, planet and profit.

Lucy chose the EngD because both for the challenging research opportunity, and the opportunity to develop her commercial skills. This intersection between new technologies and industrial systems makes for a heady mix that is both exciting and challenging. It offers unsurpassed opportunities to provide new and creative ideas and it is challenging because to take ideas and turn them into reality is not easy, usually requiring a detailed knowledge of both the research domain and industry‟s needs.

5

A Systems Approach to Organisational Learning

Neil Carhart

University of Bristol/British Energy – part of EDF Energy [email protected]

Supervised by: Dr John May, SSRC, University of Bristol Dr Sally Heslop, Civil Engineering, University of Bristol Dr Maria Priest, British Energy – Part of EDF Energy

Within any industrial organisation the safety of its employees, the public and the environment is of utmost importance. Reliability and efficiency are also important as they can have a significant impact on commercial and economic operability. Organisations therefore strive to maximise safety and reliability. This is especially the case in the civil nuclear industry. When unfavourable issues emerge it is in the organisation and industry‟s interest to understand why they happened in order to prevent them from reoccurring or to manage them more effectively if they do. Many organisations and industries share operating experience in a formalised way in order to improve their learning and hence their safety and efficiency. Within the nuclear industry there are around 800-1000 items of operating experience circulated every year (Weil and Apostolakis, 2001). When significant events happen they are often followed by high profile investigations. Both these investigations and the operating experience will draw conclusions about why the events happened and recommend learning points and corrective actions.

There are those that believe these events cannot be avoided as with Normal Accident Theory (Perrow, 1984) and those that believe they can, High Reliability Theory (for specifically the nuclear industry see Bourrier, 1996). Despite these opposing views it is inescapable that the operating experience and event reports often draw the same conclusions and recommend the same corrective actions. As Pidgeon (1997) says, we must recognise this and turn our attentions to the “deep-seated organizational preconditions and patterns which have, time and time again, been shown to pre-date catastrophic events” (p1).

In the independent review of the loss of the RAF Nimrod Aircraft in Afghanistan in 2006 (Haddon-Cave QC, 2009) it is noted that there are “uncanny, and worrying, parallels between the organisational causes of the loss of Nimrod XV230 and the organisational causes of the loss of the NASA Space Shuttle “Columbia” in 2003” (para 17.1, p448). As Haddon-Cave goes on to point out of course “There were many „echoes‟ of the Challenger accident in Columbia” (para 17.7, p449). He also draws parallels with the sinking of the Herald of Free Enterprise in 1987, the Marchioness Disaster in 1989 and the BP Texas City oil refinery explosion in 2005. He draws these parallels to highlight the importance of the fundamental organisational causes which can lead to disasters, but it also raises the question of whether organisations have been effectively learning from these events at all. In addition to this a study by Taylor and Rycraft (2004) and subsequently van Wijk et al. (2008) has drawn common learning points from ten different disasters including the Longford Gas Plant Explosion in Australia in 1998, the Hatfield Railway Accident in the UK in 2000 and the corrosion of the pressure vessel at Davis Besse nuclear power station in the USA in 2003.

With the notion of common underlying causes in mind, it is hypothesised that these events continue to happen as a result of the ways in which the events are investigated, the results are communicated and the lessons are learnt. The limitations of the existing accident investigation tools have previously been documented and System Dynamics (SD) has been proposed as a complementary way of looking at events from a systems perspective to address some of these shortcomings (see Carhart, 2009). The most widely used accident models and analysis tools such as the „Swiss-cheese‟ barrier model and Event and Causal Factors Charting are based on a linear reductionist philosophy, yet are being applied in complex socio-technical systems shaped by feedback and interdependence. While they may identify causal factors, they do not reveal the underlying structure of causality. SD provides

6 an alternative view of the causality by explicitly modelling the structure of the feedback processes which control the behaviour of the system. Here the focus is on the advantages such an approach could have on learning and moving proactively forward from these events.

The majority of individuals within the organisation are given facts and actions developed from the analysis but removed of context, and are expected to integrate this learning into their work. Adopting a systems approach may provide the organisation with the context for the corrective actions. This could increase the effectiveness of the wider organisational learning. It is also possible to develop more generic structures (at a particular level of abstraction) which allow learning to be shared in a meaningful way between industries.

Studies have shown frustration from employees over the lack of transparency of the causes of events they are told about (Huber et al., 2009). The organisational learning process within the nuclear industry has been said to concentrate too much on local issues without “deep learning” of the causes, which it has been suggested can be obtained through systems models (Carroll et al., 2002). In discussing an approach to an effective incident learning system Cooke and Rohleder (2006) reiterate the need to go beyond the identification of „root causes‟ to analyse causal structures while Woods and Cook‟s (2002) nine steps to move forward from error urges searching for underlying patterns and taming complexity though feedback. Argyris (1977) postulated that most organisations spot an error and correct it without questioning the underlying objectives or reasoning. He advocated questioning these underlying assumptions, but also warned of the dangers of dogmatically held mental models abstracted from generalisations. There is potential for this with any model including system dynamics based representations of accident causality.

Forrester, the originator of SD discussed its use in education (1992) saying “Students are stuffed with facts without having a frame of reference for making those facts relevant to the complexities of life” (p.5), a criticism that could be levelled at some organisational learning systems. The creation of causal loops has been used to investigate the benefits of non-linear thinking in schools (Plate, 2010) on the basis that thinking in casual chains can restrict understanding. SD has also been shown to aid understanding in a wider context, such as for water management issues in a case in Las Vegas (Stave, 2003). In this example a model was shown to members of the local community who suggested different strategies and variables to seed the model with. The discussion moved from placing blame for water mismanagement to understanding how the system worked. This understanding of the system can create resilience to the emergence of unwanted events. It has been suggested (Hollnagel et al., 2006) that events emerge when a system is temporarily unable to cope with its inherent complexity and resilience in this context is its ability to adapt and prevent failures. Understanding the underlying structure promotes adaptability as opposed to linear models lessons and instructions that only prevent the same or similar events occurring.

As an example The Nimrod Review contains 84 recommendations. Recommendation 22.3 states; ““Risk Cases” should henceforth be drawn up and maintained in-house by the Regulator/Services and not outsourced to Industry. All Safety Cases which are currently being managed or drawn up by Industry should be re-named and brought in house”. This recommendation is based on the analysis of the event presented in the 587 page document, however it is likely to reach key individuals as a simplified lesson or action, removed of context and reasoning. If the recommendation is given on its own then the transparency behind it is lost along with any chance of „deep learning‟ and the criticisms presented in the studies mentioned above remain valid.

Figure 1 attempts to capture the reasoning behind Recommendation 22.3 as contained in the Review. The assumptions and reasoning behind the recommendation are exposed in a way that demonstrates the complexity and interaction of the issues involved. This may promote a deeper understanding of the underlying causality. The assumptions and reasoning may be wrong or alternative solutions to the problems may be identified if one understands and questions what is really happening. This is only one

7

view, based on a specific case. The connection A→B should be taken as a causal tendency rather than deterministically and A may not be the sole variable influencing B. There are eight reinforcing feedback loops highlighted in the diagram. A brief description of the behaviours as observed or postulated by the Review and described by these loops is given below (Table 1). In addition to this there are two further important issues; 1. The loss of in-house capability can reduce the ability of the organisation to fulfil its required role as intelligent customer, to check the outsourced work and spot errors 2. The reliance on outsourcing reduces the involvement of those who actually work with and operate the technology. The safety case is written without their practical expertise, issues may be missed and the end product incomplete and distrusted Table 1 - Loop Descriptions

Figure 1 – System Dynamics Representation of Nimrod Recommendation 22.3 Loop Description 1 A decrease in in-house capability results in less employees involved in safety case preparation, reinforcing the erosion of understanding, and the decrease in in-house capability 2 A reduction in in-house capability results in more outsourcing, reducing the involvement of employees and their chance to improve their understanding and retain knowledge in-house as above 3 A decrease in employee involvement reduces their sense of ownership, their commitment to the arguments in the safety case and their understanding of it, hence a reduction in in-house capabilities 4 The habitual use of outsourcing decreases the confidence of the in-house employees, leading to increased reliance on contractors 5 Outsourcing, it is suggested in the Review (22.7), leads to an increase in the quantity of paperwork which can erode understanding of the core information 6 Increased outsourcing leads to a drain of qualified and experienced personnel from in-house to the contracting organisations, increasing turnover and decreasing in-house capability. 7 Over time increased reliance on outsourcing work can actually result in costs higher than if the work had been carried out in house, in time this can affect staffing levels. 8 As above, outsourcing can increase costs which can affect the amount of training provided to in-house personnel.

Erosion of in-house capability is a legitimate concern for any organisation that relies on outsourcing work to contractors. The loops in this model should serve as a warning of the mechanisms behind such potential unfavourable consequences. With this in mind it could be argued that the recommendation above is too generic in its reductionist reasoning of a fairly complex issue. There are other ways the problems could be addressed without removing external companies altogether. For example a corrective balancing loop could encourage closer collaboration with contract partners, embedding them within the organisation, aligning their cultures and priorities and possibly enhancing in-house capabilities through shared learning.

Conclusions There is a need for complementary non-linear, systems based accident models and tools for use within complex socio-technical systems. There is a need for further investigation into the problems with organisational learning which have allowed the underlying causes of many unwanted events to go unresolved. Further investigation is also required into the potential for a systems approach to address these issues. Rather than events being used solely to gain experience of what went wrong in order to learn which actions to modify, they can also be

8 used as an opportunity to analyse and learn how the system behaves. Tools such as System Dynamics and causal loop diagramming provide models which better reflect the complex interactions involved, exposing and communicating the underlying causality, delivering context to recommendations, allowing for „deep learning‟ of causes and the promotion of resilient systems and organisations.

References ARGYRIS, C. 1977. Organizational learning and management information systems. Accounting, Organizations and Society, 2, 113-123. BOURRIER, M. 1996. Organizing Maintenance Work At Two American Nuclear Power Plants. Journal of Contingencies and Crisis Management, 4, 104-112. CARHART, N. J. 2009. Investigating the Potential use of System Dynamics as a Tool for Event Analysis in the Nuclear Industry. The 4th IET International Conference on System Safety 2009. London. CARROLL, J. S., RUDOLPH, J. W. & HATAKENAKA, S. 2002. Learning from experience in high-hazard organizations. Research in Organizational Behavior, 24, 87-137. COOKE, D. L. & ROHLEDER, T. R. 2006. Learning from incidents: from normal accidents to high reliability. System Dynamics Review, 22, 213-239. FORRESTER, J. W. 1992. System Dynamics and Learner-Centred-Learning in Kindergarten through 12th Grade Education. HADDON-CAVE QC, C. 2009. The Nimrod Review - An independent review into the broader issues surrounding the loss of the RAF Nimrod MR2 Aircraft XV230 in Afghanistan in 2006. London. HOLLNAGEL, E., WOODS, D. D. & LEVESON, N. (eds.) 2006. Resilience Engineering - Concepts and Precepts, Aldershot: Ashgate. HUBER, S., WIJGERDEN, I. V., WITT, A. D. & DEKKER, S. W. A. 2009. Learning from organizational incidents: Resilience engineering for high-risk process environments. Process Safety Progress, 28, 90-95. PERROW, C. 1984. Normal Accidents, New York, Basic Books. PIDGEON, N. 1997. The Limits to Safety? Culture, Politics, Learning and Man-Made Disasters. Journal of Contingencies and Crisis Management, 5, 1-14. PLATE, R. 2010. Assessing individuals' understanding of nonlinear causal structures in complex systems. System Dynamics Review, 26, 19-33. STAVE, K. A. 2003. A system dynamics model to facilitate public understanding of water management options in Las Vegas, Nevada. Journal of Environmental Management, 67, 303-313. TAYLOR, R. H. & RYCRAFT, H. S. 2004. "Learning from disasters". IAEA Conference on Topical Issues in Nuclear Instalation Safety. Beijing, China. VAN WIJK, L. G. A., TAYLOR, R. H. & MAY, J. H. R. 2008. "Cultural and Organisational Factors Leading to Major Events". Topsafe. Dubrovnik, Croatia. WEIL, R. & APOSTOLAKIS, G. E. 2001. A methodology for the prioritization of operating experience in nuclear power plants. Reliability Engineering & System Safety, 74, 23-42. WOODS, D. D. & COOK, R. I. 2002. Nine Steps to Move Forward from Error. Cognition, Technology & Work, 4, 137-144.

Biography

[email protected]

Neil Carhart has been working on a project with British Energy – part of EDF energy, who operate eight nuclear power stations in the UK. The project aims to develop innovative systems solutions to the issues surrounding unplanned generation losses, in order to make the stations safer and more reliable

EngD Research Title: The Prediction and Management of Significant Unplanned Generation Loss Events in Civil Nuclear Power

9

The research has looked at previous high-profile loss events in many different industries to understand why they happened. Currently the most frequently used tools for investigating and learning from these events are based on linear, reductionist models of causality. It has been suggested that complex socio-technical systems exhibit non-linear behaviour, governed by feedback and are often incompatible with these approaches. The project has implemented tools such as System Dynamics to look at how such events develop and emerge from these complex socio-technical systems as well as how they can be proactively prevented and more effectively controlled and managed if they do occur.

Publications:

Carhart, N. J. (2009), “Investigating the Potential Use of System Dynamics as a Tool for Event Analysis in the Nuclear Industry”, The 4th IET International Conference on System Safety 2009, London, UK

Carhart, N.J., Yearworth, M. (2010), "The Use of Group Model Building for Analysing Event Causality within the Nuclear Industry", The 28th International System Dynamics Conference, Seoul, South Korea.

The Emergence of Agile Development Principles in the World's First Personal Rapid Transit (PRT) Project – Towards Agile Systems Engineering Processes

Nick Davenport, ULTra PRT (formerly ATS); Unit B3 Ashville Park, Short Way, Thornbury, BS35 3UU; [email protected] or [email protected]

Introduction Advanced Transport Systems Ltd (ATS) have designed an entirely new form of public transport, utilizing new advances in computing technology and pushing the application of computer science and mathematical thinking to the forefront of transport development. The success of ATS – and success of PRT for that matter – remains contingent on ATS‟s ability to implement the first ever PRT system at Heathrow Terminal 5 and showcase it to the rest of the world. This has been regarded as a complex, high-pressure and high-risk undertaking, and it has been suggested by commentators and experienced members of ATS that “no company would normally attempt a project of this scale with this level of resource”. So how was ATS able to develop and deliver this system – normally thought to take over one hundred engineers, managers and executives – with only ten people?

The answer lies in ATS‟s ability to balance predictive, “plan-driven” development methods with more adaptive, “agile” methods. Agile software development methodologies have become more prevalent in IT companies seeking to develop software in high-risk, fast- changing commercial environments (West, 2010). These methods help manage risk, uncertainty and unforeseen change via early and frequent releases of working software, collaborative development in small teams, customer involvement, and adaptive processes (Cockburn, 2001; Highsmith, 2002; Boehm, 2002, 2003; Schwaber, 2004; Augustine, 2005). These are a significant departure from conventional, plan-driven methods which attempt to foresee change through upfront requirements gathering, analysis and design, and manage large organisations through extensive process planning (Boehm, 2003). However, it has been shown that there is a need to balance agile methods with plan-driven methods in order to provide both high flexibility and high assurance (Boehm, 2002, 2003).

10

Increasing attention has been given to the use of agile methods for developing systems of a general nature (Turner, 2007; Smith, 2007, 2009). Here, projects with high levels of uncertainty and high rates of unforeseen change are expected to benefit from iterative design, frequent integration, and active involvement of stakeholders to establish, prioritize and verify requirements as they are better understood. However, plan-driven methods are widely regarded as essential for systems that need upfront architecture to support hardware integration and early safety assurance. Here, it is proposed that general system development projects will benefit from combined use of agile and plan-driven methods in achieving both high flexibility and high assurance.

This paper seeks to understand how elements of plan-driven and agile methods have been combined to suit certain characteristics of the Heathrow PRT project, providing evidence to support the application of agile methods to general systems development projects. Here we use the same characteristics which Boehm (2003) uses to differentiate between the two methods. These characteristics are  Application characteristics, including primary project goals, project size, and application environment;  Management characteristics, including customer relations, planning and control, and project communications.  Technical characteristics, including approaches to requirements definition, development and test.  Organisational characteristics, including customer characteristics, developer characteristics, and organizational culture.

Balancing Agile Methods and Plan-Driven Methods in the Heathrow PRT Project Application characteristics Plan-driven methods were necessary for developing safety-critical features and providing continual safety verification of the system. However, many aspects of the system were subject to ongoing, iterative development with continuous integration and testing. Safety- critical aspects were therefore bounded to make safety provable, and uphold the system's safety case throughout development.

Right from the start, ATS recognized that this project was a novel system requiring a new safety regulation, and engaged in discussion with the regulator to develop this. Later on, under a new regulatory regime, ATS worked with an independent Safety Verification Team to devise a safety verification method appropriate for safety assurance of the PRT system. While this process relied on comprehensive documentation, it also relied on a degree of trust between ATS and SVT that could only exist through sustained collaborative development.

Management characteristics The customer's approach to engaging with ATS was inherently plan-driven, using contracts as a basis for relations. Having validated critical aspects of the system through successful demonstrations of a prototype, the customer was confident that they and ATS would be able to foresee the remaining issues and work through them in advance, formalizing plans and specifications into a contractual agreement. Upfront planning was seen as necessary for pricing the contract, agreeing milestones and payments in advance, and, generally speaking, to avoid an "open-ended" project. In practice, plan-driven methods suffered from continually evolving requirements, developments, testing, and unforeseen changes surrounding a series of integration difficulties. ATS therefore took on a more agile approach to managing development activities, focusing planning efforts on particular testing/demonstration goals in the short-term, whilst using high-level plans to forecast and agree key project milestones with the customer. This approach allowed for a degree of autonomy among development teams and also allowed operational requirements to be incorporated late in the process.

11

Documented plans and specifications were subject to continual modification to a point where they could not be maintained given the limited resource. Developers therefore tended to operate on the basis of shared tacit knowledge rather than documented knowledge.

Technical characteristics ATS's risk-driven approach to defining requirements provided a degree of agility in their development process. Safety requirements and requirements relating to the core enabling systems were defined from the beginning, and were subject to rigorous change control. The extensive use of software throughout the system allowed requirements relating to non-critical features such as human interfaces to be set later on and evolve as developers' understanding of the requirements matured.

The core system architecture was designed around the needs of a small organisation, allowing individuals to own a subsystem and own their own processes. Interfaces were made extremely simple and were tightly controlled to allow for a degree of autonomy and agility within each area of subsystem development. Later in the process, management began to control the configuration of the system in an effort to achieve frequent and early system-level integration, allowing for an agile approach to system-level development. The use of software has enhanced plan-driven methods. Early consideration of software in the design process allowed developers to better integrate COTS technology that was not designed specifically for use under PRT operating conditions.

The use of software has also facilitated agile-like development approaches. Software has allowed developers to mitigate unforeseen changes and COTS incompatibility issues. Software‟s agile development processes have been used for integrating hardware – considered exclusive to plan-driven methods – via the use of simulated and hardware-in-the- loop environments. The system's extensive use of software allowed for ongoing development and frequent and early integration, allowing ATS to provide continual validation of operational capability during system integration.

Organisational characteristics The Heathrow PRT project saw a clash between ATS's R&D-based culture, which thrived on informal and emergent working practices, and the customer's plan-driven culture, which thrived on upfront planning and process maturity. ATS management were able to adapt their management style to cope with developers' inherent working practice whilst ensuring that high level plans were established and maintained with the customer. The Heathrow PRT project relied on exceptionally capable individuals, who brought experience and expertise from a variety of engineering disciplines. With no direct PRT track record prior to this project, ATS relied on the expertise and experience of their personnel to instil trust in their ability to deliver the Heathrow PRT system.

Conclusions ATS were able to combine elements of both agile and plan-driven methods in providing high assurance, yet sufficient flexibility to deal with rapid, unforeseen change. Agile methods are seen to apply to general systems engineering projects which rely on a substantial amount of new software development. Agile methods exploit software-intensive systems and small teams, allowing developers to operate on the basis of shared tacit knowledge and use documentation only where necessary. Plan-driven methods are seen as necessary for designing and building hardware subsystems, however agile methods may be used together with completely simulated, or hardware-in-the-loop (HIL) environments, to better integrate these subsystems. Completely simulated environments allow for early validation of operational capability, and HIL environments allow for more controlled and effective integration of hardware, exploiting software's automated build and test capability. Also, software provides the necessary flexibility to deal with late-in-the-process hardware incompatibility issues.

12

For novel, safety-critical systems, upfront architecture is necessary for demonstrating safety- critical functionality early, and is necessary for bounding safety-critical aspects from any fast- evolving aspects of the system; here, general system development projects must “plan to be agile”. For a novel system development project, there will be little process maturity, so organisations will have to rely on the prior expertise and track record as well as a degree of customer participation to instil trust in their ability to deliver the project. Here independent experts must be used where the customer lacks experience or expertise. Organisations can improve their “agility” by collaborating with relevant experts and representatives throughout the entire development process.

To promote the use of agile methods, hardware should be “off the shelf where possible”. This focuses the organisation on rapid integration and software development activities. To cope with the emergent, experimental nature of a novel system development, the management style must shift towards rolling-wave planning, and ensure that development iterations are trained on the evolving needs of stakeholders. For a novel system, stakeholders must accept that the development process is not wholly plan-driven. However, high-level planning is necessary for setting payment schedules and projecting completion dates. Here, the management of novel system development projects must combine agile methods with plan-driven methods to be successful.

REFERENCES AUGUSTINE, S., PAYNE, B., SENCINDIVER, F. & WOODCOCK, S. (2005) Agile Project Management: Steering from the Edges. Communications of the ACM, 48. BOEHM, B. (2002) Get ready for agile methods, with care. Computer, 35, 64-69. BOEHM, B. (2003) Balancing Agility and Discipline: A Guide for the Perplexed. COCKBURN, A. (2001) Agile Software Development, Addison-Wesley Professional. HIGHSMITH, J. (2002) Agile Software Development Ecosystems. SCHWABER, K. (2004) Agile Project Management With Scrum, Microsoft Press. SMITH, P. G. (2007) Flexible Product Development: Building Agility for Changing Markets, Jossey- Bass A Wiley SMITH, P. G. (2009) Flexible Product Development for a Turbulent World - Is "Agile" NPD the Answer? PDMA Visions Magazine. TURNER, R. (2007) Towards Agile Systems Engineering Processes. CrossTalk, Journal of Defense Software Engineering. WEST, D. & GRANT, T. (2010) Agile Development:Mainstream Adoption Has Changed Agility -- Trends In Real-World Adoption Of Agile Methods. Forrester Research.

Biography [email protected]; [email protected]

Nick Davenport studied in the first cohort of Engineering Design at the University of Bristol, before joining the EngD in Systems programme there in late 2006. His research interests include mathematical modelling and simulation, project and programme management, and knowledge management.

Nick has led early concept and feasibility studies for several PRT applications in the UK, US and UAE. He has worked with a wide range of prospective clients and stakeholders to develop their understanding of the capability of PRT and the necessary development

13 process. He has also managed technical development activities for the World‟s first commercial PRT system at London Heathrow Airport.

EngD research title: Improving the Development and Delivery of Future PRT Systems

A Personal Rapid Transit (PRT) system uses automatically-driven small vehicles to carry individuals or small groups non-stop between any pair of stations on a dedicated guideway network. All stations are offline, and so there are no intermediate stops. Vehicles are coordinated autonomously in response to – and even in anticipation of – passengers arriving at each station, promoting immediate, on-demand service. PRT was initially conceived in the 1950s in response to the need to move commuters in areas where the densities were too low to pay for the construction of a conventional metro system. The idea was that automated guidance would allow small vehicles to run with very short headways, delivering line capacity equivalent to that of a metro, however with the added benefits of operating on a network, collecting, distributing and circulating commuters to and from lower density areas. ULTraPRT (the company) is the first to succeed in developing a commercially viable PRT system, and has implemented the World‟s first system, ULTra, at London Heathrow Airport. ULTra provides numerous opportunities for site owners, operators and developers of airports, corporate and educational campuses, through to city-wide applications, however there are challenges that ULTraPRT will face as leaders of this kind of technology.

The ULTra system in particular is extremely flexible in its design, deployment and operation, and presents a virtually unmanageable design trade space if viewed from a conventional transit planner‟s point of view. This however can be overcome with the use of custom simulation design tools to provide a tight integration between stakeholder needs, system design, system performance, and the business case. Here, Nick‟s research centres on how these elements can be managed together throughout early stages of a PRT programme, focussing on the modelling and simulation developments necessary to support this process. This research has motivated the need for rapid, collaborative development with stakeholders at all levels in the supply chain, promoting flexibility in the early design process and avoiding costly rework later on.

ULTraPRT will have to manage the many risks associated with delivering projects where there is a new technology and a new capability. Much of this capability is determined by software, and so PRT programmes may benefit substantially from the use of agile principles to manage risk and maximize stakeholder value throughout the programme. Here, Nick‟s research considers at the use of incremental and agile lifecycles to support the delivery of PRT applications, exploiting the software-intensive nature of ULTra.

Publications: Lees-Miller, J.D. Hammersley, J.C. Davenport, N. (2009). “Ride Sharing in Personal Rapid Transit Capacity Planning”, Proc of 12th International Conference on Automated People Movers, Atlanta,Georgia, USA.

14

Davenport, N. Wilson, R.E. (2008), “PRT Performance Modelling with Discrete-Time Queues”, Proc of 1965th UTSG, Jan 2008, Southampton, UK.

Can Capability Planning learn from economic theory?

Kate M. Gill Group Leader, Naval Systems Department Defence Science and Technology Laboratory (Dstl), Portsmouth, England, [email protected]

Global Economic Setting. Global economic growth has been driven by globalisation over the last 30 or more years, generating networks of connections and interdependencies between the major economic powers that are unprecedented in extent and pervasiveness (Development, Concepts and Doctrine Centre (DCDC), 2009). The economic landscape has evolved rapidly with the demise of centrally planned economies such as the Soviet Union, the rise of Asian economies, particularly China, which has embraced a market aware philosophy, and the maturation of the European Union as a cohesive economic market. These changes have created a multi-faceted economic landscape that is intimately interconnected and influential.

One feature of this economic globalisation will be the continuing internationalisation of markets for goods, capital, services and labour, which integrates geographically dispersed sets of Customer and Suppliers. Markets that have existed to date as isolated communities will find their boundaries changing, and merging with other entities, with the associated merging of trading practices, rules and expectations. This reduction in barriers will be an impetus for accelerating economic growth, but will also be a source of risk, as local markets become increasingly exposed to destabilising fluctuations in the wider global economy. The “24/7” nature of the global market, enabled by mass communication of data, means goods and services can be exchanged at a rate that would not have been forecast years ago.

So what makes a sustainable global supplier? Defence planning is a long term activity, and the success or failure of the endeavour can in part be related to the stability of the aims and contributory components. If you use the principal-agent construct (Stiglitz and Weiss, 1981; Grossman and Stiglitz, 1980) shown in Figure 1, the global defence sustainability problem using two agents, Defence Customers and Defence Suppliers, can be examined.

Figure 1: Basic idea of Agency Theory (P: Principal, A: Agent) using Defence Players

Nobel Economics. Akerlof, Spence and Stiglitz (1970) jointly won a Nobel prize for their work on the contributory factors that can lead to the economic effect of market efficiency. Akerlof et al. identified a number of factors, but the one that has received enduring recognition is their assessment of market information. Their proposal was that in any market situation, there is a finite value that can be placed on the truth, and by extrapolation, the value of truthful information. If both sides of a transaction have “information symmetry”, then the resulting transaction is efficient in a holistic, global sense and benefits both parties. If one side of the transaction has more valuable information than the other, for example, the true price of production, this is termed as “information asymmetry”. The transaction has the potential to be inefficient, and will result in either more benefit to one party than the other, or, if the transaction does not take place, more “dis”-benefit to one party or the other. In any transaction, potential and reality should be recognised; information asymmetry does not guarantee an inefficient transaction. In this research, we are focusing on the sustainability of

15 the two Defence agents, and by extrapolation, the long term efficiency forecast of the players. Let us assume, that Defence Suppliers have full visibility of the cost and quality of their products or services; and that Defence Customers have a capability requirement that may, or may not, be achieved using the supplier‟s product or service. The Defence Customer has full visibility of their capability need, and how they plan to achieve the capability effect, using the suppliers product combined with other existing capabilities.

The Capability Planning Construct Viewing the capability planning problem from an economic angle, a global economic trade space that has finite boundaries and operates in a world of truth can be considered. In this situation, all information is available, and there is no asymmetry. In economic terms, this world of truth is defined as an efficient market with all Customers paying for items that they want, and paying the exact price that it takes to make the item from a Supplier; in this world, the Defence Supplier and Defence Customer are both sustainable. Explained another way, the Customer is receiving what is needed at a fair price and a sale takes place that will keep the Supplier in business, through life. Examining the sale in more detail, we can learn a great deal about both sides of the transaction. In this efficient global market, the price of an item provides information on its value to the Customer, and also its equivalence in the hierarchy of needs of the Customer. Using an example, a customer has £100k. Depending on the relative needs of the customer (which only they know for sure), they can chose a variety of routes for using the £100k. Their decision provides information not only on what they decided to do, but also what they decided not to do, which is equally informative.

Option A – The Customer decides to spend the £100k on purchasing a particular ship. This provides the information that the Customer has a need that can be fulfilled by the advertised abilities of a £100k ship. However, it also provides information that the Customer needs the abilities of that particular £100k ship, and that the capability value it represents is £100k in the global market. There is also the information that the Customer believes that the capability offered by the ship is at a higher priority than any other capability that may also be needed. This alternative capability need may not be achieved fully by the £100k, but could represent a partial capability, but even this partial capability is a lower priority.

Option B – The Customer decides to “do nothing”1 . This provides the information that the customer has a need that cannot be fulfilled by the advertised abilities of the £100k ship. However, it also provides information that the customer feels either that they will not pay £100k to fulfil the need, but may be looking for a £50k option or a £150k option. The Customer may also feel that £100k is not the global rate for fulfilling that need. Finally, the Customer is also supplying the answer to the question: “how much value do you place on not having this capability?”

If we assume that the customer prefers Option A, the Customer is indicating not only that they prefer to spend the £100k on a ship rather than staff recruitment or improved organisational but that, in the world of truth, the Customer feels that they are paying a fair price for the ship, which will maximise the market efficiency. Taken to the next level, if the Customer is presented with a range of ships that represent various capability levels, the selection represents the value placed on the incremental difference between the preferred choice and that of its nearest neighbours. In an ideal world, this value can be defined in tangible terms by building up a picture of the market, building up from its “marginal” capability effects. Margins and marginal effects are important in an efficient market as they represent the service or effect that defines the entry point to a market. Marginality is a

1 One of the UK Acquisition process criteria is to present a business case for the “Do Nothing” option at the Initial and Main Gate reviews. In the defence market, this does not necessarily mean catastrophic considerations, but could be a case of examining the “Do Nothing” option. “Do nothing” could be a review of whether something is really needed (in light of the externality effects that have taken place, or are predicted). UK Acquisition Operating Framework (v1.1.3 Dec 2009). Risk Management: Business Case Approvals. Retrieved January 18, 2010 from http://www.aof.mod.uk/aofcontent/tactical/risk/content/buscase.htm

16 measurement of how easy it is to enter a market (Akerlof et al, 1970). High marginality means that there is significant discrepancy between the “have” and the “have not”, and having adds significant status, ability or, in the defence world, capability. Low marginality enables easy entry in the market.

In the defence capability domain, some capability requirements have high marginality; as an example, the capability to sail under the polar ice caps. To achieve this capability effect, and using the UK Defence Lines of Development (DLOD) construct, you would need:  Training – a crew trained in underwater operation and endurance;  Equipment – a submarine or suitable submersible capable of under ice operation;  Personnel – a group of people with diverse skills e.g. mechanics, navigators, catering staff, maintainers;  Information – accurate navigational charts and knowledge of the currents and environment;  Doctrine and Policy – the ability to access the area, in this case, access to international waters;  Organisation – a hierarchy of people with appropriate roles and responsibilities;  Infrastructure – ability to communicate and access pertinent information to weather, movement of other assets;  Logistics – ability to supply and support for the duration of the mission.

The above list is hard to achieve for a standard individual, and even for some defence forces. Just taking one of the points, there are a finite number of operational submarines, commercial or military, at any one time. The marginality of this capability is therefore high. Taking another example, the capability to transport a box from A to B, to achieve this capability effect, and using the UK DLOD construct, you would need:-  Training – someone or something trained in manual handling;  Equipment – a person or a person with a trolley (if applicable);  Personnel – one person;  Information – instructions on the location and destination required;  Doctrine and Policy – permission to move the box (if applicable);  Organisation – authority to move the box (if applicable);  Infrastructure – instructions to move the box from a suitable source;  Logistics – a box to move.

Conclusions In Capability Planning, recognising what the capability transaction represents in terms of information is key to an efficient transaction. Information is passed Customer to Supplier, Supplier to Customer, and from both parties to the global market.

The Capability need, and its associated market marginality that can fulfil that need, is a clear indicator of price, and cost, to both parties.

Capability Planning can use the construct of marginality to categorise transactions, and to ensure that the Customer and the Supplier gain, and maintain a sustainable position. In short term transactions, the Capability Management cycle (planning, delivery, generation and operational) can be assessed for efficiency. In longer term transactions, the marginality effects are harder to track, and predict, but can still be used as an assessment of efficiency.

References Akerlof, G.A., Spence, A. M and Stiglitz, J.E. (1970). "The Market for 'Lemons': Quality Uncertainty and the Market Mechanism.” Quarterly Journal of Economics. Vol. 84, No. 3. The MIT Press, pp 488–500. Development, Concepts and Doctrine Centre (DCDC) (2009). Dimensions - Research Areas. In DCDC Strategic Trends Programme: Global Strategic Trends, Edition 4 V3, 16 July 2009. UK: Crown Copyright, pp 87-158. Grossman, S.J. and Stiglitz, J.E. (1980) “On the Impossibility of Informationally Efficient markets”, The American Economic Review, American Economic Association ,Vol. 70, No. 3 (June), pp 393-408.

17

Stiglitz, J.E. and Weiss, A. (1981). “Credit Rationing in Markets with Imperfect Information.” The American Economic Review, Vol. 71, Issue 3 (June). American Economic Association, pp 393-410.

Biography [email protected]

Kate Gill joined the EngD in Systems programme in 2007, after several years away from education and academia. Dstl were launching a programme of personnel development in Systems Skills and Systems Thinking, and the EngD in Systems was a great way to achieve personal development whilst maintaining industrial contribution. At the start of the programme, Kate was part of an Open Systems team of ten engineers and scientist delivering to the maritime defence domain. Kate is a full time EngD Research Engineer based full time at her work base in Portsmouth. During the last four years, Kate‟s career has developed and she is now a Group Leader in the same organisation with a staff of over 60. Balancing the EngD, daily work and a young family has been a challenge, but she is still going – and is hoping to complete within the deadline of January 2011.

EngD research title: Generic Process for the Delivery of Balanced Defence Capability

In the last five years, the MoD approach to the provision of capability has changed radically and comprehensively. At the start of 2006, The Defence Industrial Strategy (DIS) White Paper2 was presented to Parliament as a wide reaching discussion paper on the defence domain and how it should operate. In this paper, the MOD had reviewed their approach to delivery of defence product, and also reviewed their differing roles within the process. Historically, the MoD had only seen its role as „customer‟ and so was driven by time, cost and quality (performance). Therefore, in its assessment of quality, it had only concentrated on a small part of the lifecycle of the programme and therefore had a misconception of quality. As a result of the DIS white paper, the MoD undertook a review of procurement, and the terms of assessment of project success.

This review of quality covered a wide range of factors, and concluded with a radical restructuring of Defence Procurement that changed the organisation, processes and lifecycle assessment. These changes came into effect on 2nd April 2007.

In order to deliver this through life assessment, the entire MOD procurement infrastructure has been reorganised with clear lines of accountability and delivery. The role of the IPTs has been expanded to cover the whole life of the project and not just procurement, through the entire defence acquisition „CADMID‟ lifecycle. This remit includes all aspects of its interaction within the defence domain and UK industry. From April 2007, all the CADMID defence lifecycle major review points will include an assessment of the Defence Lines of Development (DLOD) 3 which were introduced as criteria for assessing the quality of a project. These DLOD were introduced alongside cost and time scales, to ensure a more robust review of quality. The DLOD approach should ensure balance across the project, and refocus the attention away from equipment provision.

Implementation of the DLODs assessment criteria will impact the defence procurement process, and my EngD in Systems is aimed at mapping the processes to gain the best „balance of investment‟ to the provision of defence capability. It is hoped that this work will support the continuing evolution of defence acquisition.

2 Defence Industrial Strategy Defence White Paper, Presented to Parliament by the Secretary of State for Defence, December 2005. 3 Acquisition Handbook ‘A guide to acquiring Defence Capability’, Edition 6, October 2005, Section #1 (http://www.ams.mod.uk/content/handbook)

18

Analysing Sustainable Development Social Structures Using Social Network Analysis

Nicholas Meese1, Chris McMahon2 and Juani Swart1 Email: [email protected]

1. School of Management, University of Bath, UK. 2. School of Mechanical Engineering, University of Bath, UK.

Abstract Our understanding of sustainable development (SD) issues is growing at an increasing rate. As such, it is imperative that we share this understanding with like-minded peers to ensure the latest SD thinking is being effectively utilised. This paper reports on part of a programme of work to provide knowledge and information support on SD issues to engineers in a large international engineering consultancy. As formal hierarchical structures rarely reflect how things get done in organisations, we investigate informal social structures in relation to SD topics. Using social network analysis (SNA), a systematic technique that aids the study of social structures, we examine three distinct skill groups. Members of these groups were asked to complete an online survey, detailing their SD oriented relationships with organisational peers. Common mathematical SNA techniques are then applied to reveal structural characteristics. In addition, contextual information is used to provide a richer picture of the studied structures. Findings show that, although all groups have central connecting members, they are clearly decentralised in form; i.e. they do not reflect their hierarchical pretences. Differences between the study groups are explored, with group size and markets served predominantly influencing the characteristics of SD relationships. 1. Introduction Sustainable development (SD) is increasingly recognised as an approach to balancing social, economic and environmental factors. As such, SD philosophies and practices are rapidly evolving. This means that new ideas and innovation can quickly become outdated, requiring organisations to share their knowledge more effectively to enable employees to respond and adapt to shifting SD requirements and capabilities.

The work presented in this paper aims to understand the degree of SD related connectivity between individuals within three skill groups. Members of these groups were asked to complete an online survey, detailing their SD oriented relationships with organisational peers. Using common social network analysis (SNA) techniques, gathered data were analysed to help achieve the following goals. First, we wanted to make visible the social structures facilitating SD issue resolution in project delivery. This would act as a communication tool and a way of visually interpreting the data and could be used for feedback and discussion. Second, we wanted to determine the level of cross-boundary interaction to help us assess a network‟s or sub-network‟s isolation. Third, we sought to identify the strength of relationships between members to allow us to recognise key players within the network. Finally, we wished to submit a set of recommendations that would bolster the studied networks for enhanced SD performance. 2. Social Networks, SNA and Sustainable Development As social animals, humans network (Parkhe et al., 2006). According to Cross et al. (2002), “one of the most consistent findings in the social science literature is that who you know often has a great deal to do with what you come to know.” As such, relationships between network members are often more reflective of how work gets delivered rather than formal relationships within organisational hierarchy (Cross et al., 2002, Brown and Duguid, 2000). Yet it is because of their informal nature that they are often starved of resources and management attention (Cross and Prusak, 2002). Social networks ideally connect individuals across functional and geographic boundaries in a dynamic and as-required basis (Smith and

19

McKeen, 2007). These network structures often promote organisational agility and innovation, enabling organisations to respond to changing organisational and market needs. SNA is the study of social structure that consists of actors (nodes) and a set of relationships connecting pairs of actors (Carrasco and Miller, 2009). This approach provides a rich and systematic means of assessing informal networks by mapping and analysing relationships among people, teams, departments or even entire organisations (Cross et al., 2001). Using mathematical techniques, it is possible to assess the social structure between a set of entities, which in turn can help to recognise intellectual capital, socialize new members and enhance organizational learning (Chan and Liebowitz, 2006). SNA has been applied in sociology, anthropology, information systems, organisational behaviour, and many other disciplines (Liebowitz, 2005). From an organisational stand point, SNA has been applied to various issues, including: supporting partnership and alliances; assessing strategy execution; integrating networks across core processes; developing communities of practice; personal networks (Cheuk, 2007).

Forum for the Future, a leading non-profit UK based organisation with a mission to promote SD, has adopted the following definition of SD: „a dynamic process which enables all people to realise their potential and improve their quality of life in ways which simultaneously protect and enhance the Earth‟s life support systems‟ (FftF, 2010). Whilst understanding of SD rapidly evolves, KS helps reduce inherent complexity by disseminating relevant knowledge to relevant parties, leading to an increased rate of learning and innovation. 3. Method Hatala‟s (2006) eight step SNA framework was applied in this exercise. A positional strategy was used to define three functionally bound populations (Scott, 2000). These populations, known as skill groups (SG), were selected based on their perceived KS performance according to the organisation‟s staff survey results in 2008. Gathered relationships identify which peers the respondent discusses SD related issues. As each group provides distinct services, a bespoke SD definition was drafted for each group and distributed in collaboration with the group‟s leader. For each stated SD relationship respondents were asked four questions relating to: how frequently the respondent acquires SD related information or advice from each contact; how aware the respondent is of their contact‟s knowledge and skills; the accessibility of their contacts; and how well their contacts engage with their on matters related to SD (Cross and Parker, 2004).

An online survey tool was used to gather the network data. Survey errors were addressed to encourage the required high response rate, typically between 65% and 90% (Stork and Richards, 1991), needed to map a social network. Populations, labelled SG1, SG2 and SG3, each consist of 41, 120 and 66 members respectively. The average response rate was 78%. Actor attribute information was extracted from the organisation‟s personnel database. Microsoft Excel and UCINET (Borgatti et al., 2002) were used to analyse the collected data. Cross-boundary interaction is analysed by looking at each contacts‟ characteristics. Functional and geographical boundaries are analysed in relation to all stated contacts. Centrality and substructure indices are used to describe the social structure within the studied groups. Centrality seeks to quantify an actor‟s prominence within a network by summarising the ties between all actors (Knoke and Yang, 2008). This concept provides insight into power, stratification, ranking, and inequality in a network (Chan and Liebowitz, 2006). Substructures are the study of exceptionally densely directly tied actors (Brandes, 2008, Knoke and Yang, 2008), whom often exhibit their own cultural norms, values and behaviours (Scott, 2000). 4. Findings The sociograms for the three groups studied are shown in Figure 1. Cross-functional analysis of the groups reveal that SG3 is the most insular group studied, with 39.0% of stated contacts residing outside of its immediate group; SG1 and SG2 have 59.3% and 60.5% respectively. Cross-geographical analysis shows that SG3 possesses the least cross-

20 boundary interaction with only 18.1% of stated contacts residing in different offices; SG1 and SG2 had more than double with 44.9% and 41.8% respectively.

Figure 1. Sociograms of the three groups studied (from left to right: SG 1, 2 and 3)

Centrality indices indicate that SG2 has 12 (10.0%) members with more than four references, whereas SG1 and SG3 have 7 (17.1%) and 6 (9.1%) members respectively. All groups contained members who seek knowledge from other members but are not referenced by other members: SG1=8 (19.5%); SG2=16 (13.3%); SG3=12 (18.2%). Closeness centrality shows that SG2 has the closest actors (xˉ =1.50), with SG1 and SG3 having more distributed networks (xˉ =13.01 and 12.59 respectively). However, although SG2 has the closest actors it is the least dense (0.02); SG1 has the highest density (0.07), with SG3 sitting in the middle (0.04). Substructure analysis reveals that, with the exception of isolates, SG3 is the only group to be internally cohesive; SG1 consists of two disconnected substructures, SG2 is most disconnected consisting of five.

Relationship analysis suggests that nearly 30% of SG2 contacts are asked for SD related information on a weekly basis; SG1 has the least maintained SD relationships – nearly half of all contacts are asked for SD information or advice less than once a quarter. All group respondents indicated that they have a good understanding of their contacts‟ SD related knowledge and skills, with an overall average of 80.1% „agree‟ or „strongly agree‟ responses. The data implies that SG2 and SG3 contacts are highly accessible for SD information and advice, with over 85% of contacts falling into the „agree‟ or „strongly agree‟ categories. SG1 was also strong in this area, but received a „neutral‟ percentage score of 23.6%. Finally, SG3 ranked highest in the engagement category, with contacts receiving 97.7% in the „agree‟ or „strongly agree‟ categories. SG1 and SG2 also achieved high percentages, receiving 76.4% and 86.0% respectively. 5. Discussion and Conclusions The findings show the insular nature of SG3. This is believed to be the result of the markets the group serves. SG3 delivers projects to clients who are risk averse and often desire high levels of confidentiality. As such, they prefer work to be carried out in an orthodox manner. Thus, SG3 members are unlikely to seek support across boundaries due to the prescriptive nature of their work. SG1 and SG2, on the other hand, serve markets and clients that are less risk averse, encouraging greater innovation and opportunity recognition in the majority of their projects.

The number of key players in each group reflects the decentralised structure within all groups. This implies that the groups are not dependent on a small number of individuals (which can easily create bottlenecks); the accessibility and engagement scores reinforce this by indicating that members often obtain useful responses within reasonable timeframes. This is encouraging as the majority of relationships are maintained on a monthly or quarterly basis. Thus it seems that members are happy to share SD knowledge.

There is a concern that a high number of members who seek knowledge are not recognised by their peers; something that does not resonate with the perceived high awareness of peers‟ knowledge and skills. The organisation has already attempted to overcome this by implementing a skills search system. However, it seems the system is not fully utilised and further efforts should be made to connect isolated members and structures, whilst

21 recognising members with SD interests. Although members believe that they posses a good understanding of their contacts‟ SD knowledge and skills, they are unaware of other group peers‟ within their vicinity who also possess useful SD knowledge and skills, thus creating pockets of knowledge.

In conclusion, this paper provided an overview of a research exercise that applied SNA techniques to understand SD knowledge sharing in three groups. Several considerations emanated from the findings. First, markets and clients served by a group can directly affect its social structure. Second, relationships do not necessarily need to be maintained by frequent contact. Third, opportunities exist for increasing awareness of SD knowledge and skills outside members‟ network. The study is not without limitations; it was conducted with a single company and it would be dangerous to generalise these findings to other contexts without further research. References BORGATTI, S.P., EVERETT, M.G. AND FREEMAN, L.C., 2002. UCINET for Windows: Software for Social Network Analysis. Harvard, MA: Analytic Technologies. BRANDES, U., 2008. Social Network Analysis and Visualization. IEEE Signal Processing Magazine, November, pp.147-151. BROWN, J.S. and DUGUID, P., 2000, The Social Life of Information, Harvard Bus. School Press, Cambridge, MA CARRASCO, J.A. AND MILLER, E.J., 2009. The social dimension in action: A multilevel, personal networks model of social activity frequency between individuals, Transportation Research, 43, pp.90-104. CHAN, K. AND LIEBOWITZ, J., 2006. The synergy of social network analysis and knowledge mapping: a case study. Int. J. Management and Decision Making, 7 (1), pp.19-35. CHEUK, B., 2007. Social Networking Analysis: Its application to facilitate knowledge transfer. Business Information Review, 24 (3), pp.170-176. CROSS, R., BORGATTI, S. AND PARKER, A., 2002. Making Invisible Work Visible: Using Social Network Analysis to Support Strategic Collaboration. California Management Review, 44 (2), pp.25-46. CROSS, R. L. & PARKER, A., 2004. The hidden power of social networks: Understanding how work really gets done in organizations. Harvard Business School Press. CROSS, R., PARKER, A., PRUSAK, L. AND BORGATTI, S.P., 2001. Knowing what we know: Supporting knowledge creation and sharing in social networks. Organizational Dynamics, 30(2), pp100-120. CROSS, R. AND PRUSAK, L., 2002. The People Who Make Organizations Go – or Stop, Harvard Business Review, June, pp.104-112. FFTF, 2010. What is sustainable development? [Accessed 7 Jan 2010]; Available from: http://www.forumforthefuture.org/what-is-sd. HATALA, J., 2006. Social Network Analysis in Human Resource Development: A New Methodology. Human Resource Development Review 2006, 5 (1), pp.45-71. KNOKE, D. & YANG, S., 2008. Social network analysis. Sage Publications. LIEBOWITZ, J., 2005. Linking social network analysis with the analytic hierarchy process for knowledge mapping in organizations. Journal of Knowledge Management, 9 (1), pp.76-86. PARKHE, A., WASSERMAN, S. AND RALSTON, D.A., 2006. New Frontiers in Network Theory Development. Academy of Management Review, 31 (3), pp.560-568. SCOTT, J., 2000. Social Network Analysis: A Handbook. 2nd Ed. Sage Publications. SMITH, H.A. AND MCKEEN, J.D., 2007. Developments In Practice Xxvi: Social Networks: Knowledge Management‟s “Killer App”?, Communications of the Association for Information Systems, 19, pp.611-621. STORK, D. AND RICHARDS, W.D., 1992. Nonrespondents in Communication Network Studies: Problems and Possibilities. Group & Organization Management, 17 (2), pp.193.

22

Biography [email protected] Nicholas Meese, having completed BSc Computer Science (Hons), Cardiff University (2001-04), MSc Engineering Management, Aston University (2004-05), been a Research Assistant at SEIC, Loughborough University (2005-07) Nicholas embarked on his EngD in Systems in 2007.

Currently Nicholas works with the Halcrow Group Ltd., focusing on social and technical knowledge sharing practices. EngD Research Title Knowledge sharing for sustainable development.

Publications Meese, N. C., Swart, J., Vidgen, R., Powell, P., McMahon, C., (2010). “Addressing Data Collection Problems in Web Mediated Surveys”, Proc. of 30th Computers and Information in Engineering Conference (part of the ASME International Design Engineering Technical Conferences), Montreal, Canada.

Meese, N. C., McMahon, C., Swart, J., Vidgen, R., (2010). “Providing knowledge and information support to engineers for enhanced sustainable development practice”, Proc. 7th International Conference on Product Lifecycle Management, Bremen, Germany.

HalSTAR: systems engineering for a sustainable built environment

O.J.D. Pearce ([email protected]) Halcrow Group Ltd, 1 The Square, Temple Quay, Bristol. BS1 6DG EngD Systems Centre, University of Bristol, Queen's Building, University Walk, Bristol BS8 1TR

Introduction Sustainable development (SD) is complex, subjective and value laden, and must be treated as such. It involves the integration of multiple disciplines in balancing a wide range of issues, with different stakeholder requirements and priorities in every context. The construction industry has a major role to play in achieving SD – the built environment that we create has a significant influence on the structures and systems within which everyone operates. The core objective of the research has been to enhance the sustainability of Halcrow‟s projects. With projects in about 70 countries across the world, the needs and values of different clients and stakeholders can vary significantly. Addressing sustainability in such a wide range of contexts can be highly complex, but this diversity also presents potential advantages to multidisciplinary construction consultancies, with the opportunity to drive step change by evolving the holistic understanding necessary to achieve effective, integrated sustainable developments.

Theory and Methods or background of research The most widely quoted definition of SD is that framed by the Brundtland Commission: development that "meets the needs of the present without compromising the ability of future generations to meet their own needs.” [UN, 1987]. Whilst laudable in intent and underpinnings, the „Brundtland definition‟ has been described as being sufficiently vague to justify a wide range of positions, without really challenging the existing market economic

23 paradigm, or „business as usual‟ to any significant extent [Ross, 2009]. Halcrow has adopted Forum for the Future‟s 5 capitals model at the core of its strategic framework for sustainability [FftF, 2009], based on the concept of capital stocks from which benefits flow, all of which depend on natural capital, the stock of which is not inexhaustible.

There are many different methods for appraising sustainability. In the UK, these range from voluntary rating schemes such as BREEAM [BREEAM, 2009] and CEEQUAL [CEEQUAL, 2009], through a range of public and private sector „sustainability toolkits‟, to statutory processes such as Sustainability Appraisal (SA). Most methods were designed with a context in mind and can be very prescriptive as a result, using a predefined checklist of requirements against which performance is measured. Different methods define different sets of criteria and often use inbuilt weightings to produce an overall score; they imply an objective, generally applicable definition of sustainability.

However, interpretations are rooted in our perspectives and value systems, and it is often difficult to formulate and justify apparently simple decisions – what is a “good” solution? Either decision criteria do not adequately address the issues, or the way in which they address them is not acceptable to key stakeholders. Irrational trade-offs are often made in order to gain credits, key issues are often missed, and sustainability becomes an add-on to the development process, rather than a potential source of added value. In addition, as most infrastructure systems are coupled, addressing issues in isolation not only results in missed opportunities but also has the potential to generate unintended consequences. A sustainable approach to construction has to be grounded in the unique project contexts that result from different client-led, local, national and international influences and priorities. This context implied that grounded theory would be an appropriate methodology – helping to understand the “meanings and concepts used by social actors in real settings‟‟ [Gephart, 2004]. The iterative interaction of sustainability issues results in inherent uncertainties associated with decisions, with potentially significant consequences. Such “complex adaptive systems” have several key features that must be understood in order to achieve effective sustainable solutions: they are nested, subject to adaptation, and exhibit emergence. Sustainability is an emergent property of complex adaptive systems. By improving our understanding of complex systems we should be able to improve sustainability [Godfrey, 2006].

Results & Discussion The wide range of perspectives of what is involved in achieving SD makes a holistic framework within which stakeholder requirements can be managed essential. The Halcrow Sustainability Toolkit And Rating system (HalSTAR) is designed to provide a truly holistic methodology, incorporating many features of existing approaches whilst mitigating their disadvantages. It provides a common framework to assess and manage sustainability, integrated with a generic decision support method. HalSTAR is based on a systems model of SD (Figure 1), which represents a balance between a range of needs (capitals), for a nested system of stakeholders, throughout the lifecycle of a project or process.

Figure 1 Systems model of sustainability

24

Based on Halcrow‟s strategic approach, the HalSTAR sustainability wheel (Figure 2) is grounded in the definitions implicit in over 400 existing approaches, including assessment methods, indicator sets, legislation, planning policies, corporate responsibility reports, and the requirements of key stakeholder groups. The “issues framework” (Figure 3 3) adds another dimension of the systems model. Issues in the diagram are associated with stakeholders on different scales: the client, project, end users, and local, regional, and global impacts. The resultant database of generic requirements contains approximately 840 sub-issues, with about 4200 qualitative criteria and 2000 indicators. The grounded approach results in a world wide „shared view‟ of sustainable development that connects high level policy with project level practice. The framework provides a comprehensive but highly adaptable means of aligning relevant requirements, enabling the context-driven appraisal of all of the key factors affecting the sustainability of a project or Figure 2- The HalSTAR sustainability wheel programme.

Relevant issues are managed throughout the project lifecycle using an iterative loop (Figure 44) of scoping, assessment, review and action. This is a bespoke process; the elements used, the extent of their use, and the number of iterations will depend on project needs. The scoping stage involves establishing which issues are relevant, through engagement with the client, stakeholders, and project team; the HalSTAR framework is designed to align their concerns and requirements from the outset. Weightings developed through consultation attain a measure of objectivity by triangulating multiple subjective opinions. Context specific performance measures are developed on the basis of scoping information. Appraisals are bespoke - impacts are assessed in the level of detail required, and results are conveyed to decision makers and stakeholders at an appropriate resolution for their needs. Once results have been reported, follow-up actions are agreed with the client and project team. Upon completion, post occupancy evaluation, feedback on client and stakeholder Figure 3- HalSTAR issues framework with inset satisfaction, and monitoring of resultant performance can be used to ensure feedback from real world results is captured and compared to predictions.

The methodology has been applied with public and private sector clients on projects ranging from eco-cities and mixed use developments, to flood defence schemes and renewable energy projects, and has also been used to inform office relocation and ongoing operations within Halcrow. It is adaptable to any given context (including projects, products, processes and organisations) and is designed to mitigate risks and add value in that context. Potential applications range from simple options appraisals and monitoring frameworks to lifecycle asset management, corporate responsibility reporting, and infrastructure programmes. On

25 some projects, the framework may simply be used to scan for unintended consequences, or to carry out a gap- analysis on identified issues. It is also being increasingly used to identify issues and priorities during bidding processes, and in building strategic relationships with clients - helping to understand their needs.

Conclusions HalSTAR provides a comprehensive yet context-specific means of Figure 4 - The basic HalSTAR process efficient, effective management and proactive enhancement of sustainable projects and programmes. It ensures that sustainability, rather than being an add-on to the development process, becomes a potential source of added value for clients by ensuring all issues are considered while focussing on the key impacts and opportunities; reducing risks of fines or delays. Through consideration of stakeholder needs and values in the development of context-specific performance measures and appraisal criteria, progress can be tracked to ensure stakeholder concerns are addressed effectively, helping to define priorities and enhancing clients‟ reputations.

Acknowledgements Many thanks to all of the colleagues at Halcrow and the EngD Centre whose reflections, support and inspiration have contributed to the development of this research. Particular thanks to my supervisors - Antony Bursey, Prof. Sally Heslop and Prof. John Davis - and to Nick Murry, Andrew Kluth, and Tim Broyd for their support within Halcrow. Thanks also to the EPSRC for their funding, and to Halcrow‟s Innovation Fund for helping to develop the approach into viable software.

References BREEAM (2009) http://www.breeam.org/ CEEQUAL (2009) http://www.ceequal.org FftF (2009) Models of sustainable development, http://www.forumforthefuture.org.uk/aboutus/sdtools_page398.aspx Gephart, R. P. (2004) Qualitative research and the Academy of Management Journal. Academy of Management Journal 4, 47, 454-462. Godfrey, P. (2006) Systems thinking approach to sustainable innovation. Joint International Conference on Construction Culture, Innovation, and Management. Ross, A. (2009) Modern interpretations of sustainable development. Journal of Law and Society 1, 36, 32-54. UN (1987) Report of the World Commission on Environment and Development, United Nations

Biography [email protected]

Ollie Pearce, a Research Engineer at the University of Bristol. Having completed a MEng in Civil Engineering in 2004, Ollie joined Halcrow Group Ltd., working as a structural engineer for two years before commencing the EngD programme as part of a collaborative project between Halcrow and the EngD centre focussing on systems approaches for sustainable

26 development. Ollie is the author and Project Manager for the Halcrow Sustainability Toolkit and Rating System (HalSTAR) – a tool for the holistic appraisal and improvement of sustainable development issues in context. Committed to the application of sustainable principles, with an extensive understanding of sustainable construction, and has acted in an advisory role on sustainability issues in numerous projects.

Ollie’s Research Title is Systems Thinking for a Sustainable Built Environment.

The research project specifically relates to the creation of methods for assessing and developing knowledge on the sustainability of Halcrow projects and processes. The project has primarily involved the development and application of HalSTAR. The HalSTAR system works to improve sustainability on projects and programmes by providing a common framework to guide decision-making and appraisal within unique contexts and priorities. Issues and impacts are assessed in relation to the client, end users, region and wider environment in the level of detail required. Assessments are bespoke, guided by HalSTAR‟s holistic framework while providing flexibility to ensure a client-orientated, context-driven approach. The process is designed to add value by improving life-cycle efficiency (e.g. through energy savings), reducing operating costs, and can reduce risks associated with planning and regulatory compliance by aligning and managing relevant requirements throughout the project, flagging up issues early-on.

ANALYSING RISKS IN A NOVEL TRANSPORT SYSTEM

Alan Peters1,2, Torquil Ross-Martin2, John May1 and Alin Achim1

1 University of Bristol, UK 2 ULTraPRT Ltd, Bristol, UK Email: Alan.Peters.bristol.ac.uk

Abstract ULTra is a new personal rapid transit (PRT) system. As safety is of paramount importance, a comprehensive safety case has been developed. As part of this safety case a quantified risk model has been constructed. This model describes all foreseen hazards to passengers and thirds parties resulting from operating the transport system. It also includes the mitigations that reduce the risk. The risk model provides a powerful starting point for understanding some of the complexities within the system. It has already been used to inform decisions on safety versus availability.

A method of analysing the model to identify priorities in the continuous effort to reduce risk is proposed. A number of recommendations for the future use of the model are made:

1. The model should continue to be flexible to allow it to grow with increased information and changing beliefs. 2. It is necessary to create a list of variables in the model that can be updated on the basis of system operations. 3. Good communications with the operations team needs to be maintained.

1) INTRODUCTION ULTraPRT Ltd have developed ULTra, a new personal rapid transit (PRT) system. Personal rapid transit, a developing form of transport, is designed to provide non-stop urban travel using small driverless vehicles. As a concept, PRT has been around since the 1950's but it

27 is only recently that systems true to the core principles of small vehicles and non-stop travel have been developed. The ULTra system is at the forefront of this new development wave.

As with any novel transport design, safety is of paramount importance and therefore a comprehensive safety case has been developed. Two important ideas are central to the safety approach taken with ULTra. Firstly, safety needs to be designed into a system from the start of the concept phase. Secondly, when it comes to assessment of risk, the system needs to be considered as a whole, as it can have emergent properties not seen in the component subsystems. These two ideas are further discussed in Leveson (2003).

Risk analysis techniques are well developed in various transport industries (Luxhoq, 2003; Ale, 2000; Evans, 1994). Fault trees, used to determine the probability of a hazard, are commonly used as part of this risk analysis process. Constructed correctly, they can identify weak areas of the system design that other analytical methods may have overlooked. As part of the ULTra safety case, a quantified risk model has been created.

The model consists of sixty fault trees with two main constructs: nodes that represent faults or external events (represented in faults / year) and nodes representing the system's response to an event (represented as unavailability / demand). These nodes are joined together by a system of logic gates. This frequency architecture is then multiplied by severity estimations to obtain the risk of hazards to passengers and third parties in the environment of the system. The model was developed with the assistance of independent experts and the assumptions were reviewed and refined with an independent safety review team. Wherever appropriate the assumptions are based on data from related transport, e.g. car speed and pedestrian injury data to support the estimate of the level of harm that might result from a collision between an ULTra vehicle and a trespasser on the system. However, it is recognised that in the absence of any directly related historical data the estimates of the frequency of faults and external events, the alertness of the controller and the harm that could befall passengers or the public can only be based on an informed judgement at this stage. For this reason, a prudent approach to the frequency of events, faults and the level of harm has been adopted.

The risk model provides a powerful tool for understanding some of the complexities within the system and it has already been used to inform decisions on safety. Additionally, it has been proposed that it could aid decision making in aspects of system availability. This research aims to enhance the understanding of the risk model and therefore the system it represents. This aim can be represented by three objectives: i) The formulation of a decision-making structure based on risk ii) An understanding of limitations of traditional risk analysis techniques iii) Development of an efficient proposal for continuously updating the risk model

2) USING RISK TO AID DECISION MAKING Risk analysis is often looked at negatively, however it can suggest improvements in system availability and therefore be a constructive tool. A key step in the analysis of the risk model is a sensitivity analysis. This describes the effect on the overall risk if one of the inputs is altered. It can be used to suggest where engineering or operations effort should be concentrated to reduce the risk.

In this work, the hazard probabilities and relevant consequences were inputted in to the Fault Tree+ software package which allows the automation of the sensitivity analysis to be set up. This has the advantage that when the model changes, the sensitivity analysis can be easily reproduced. The result of the sensitivity analysis is a list of parameters that the overall

28 risk is the most sensitivity to. The significant parameters are those with a high sensitivity where there is low confidence that the assumptions are not either accurate or cautious.

It is proposed that the priority of engineering / operations effort can be found using the following formula:

PriorityValue = (Sensitivity)((MaxConfid+1) - Confidence) (Controllability) where Sensitivity is a % figure generated by the analysis. Confidence is an integer in the range 1 to 5 inclusive, with 5 being the most confident and 1 the least. MaxConfid is the maximum of the confidence scale (in this case 5). Controllability is 1 for controllable and 0 for uncontrollable. For example, number of merges in a guideway design is uncontrollable once the guideway has been constructed. The higher the PriorityValue the more an investigation into the parameter is required.

Furthermore, by changing the consequences to reflect, say, hours lost per failure, the model and analysis can be used to improve system efficiency as well as reduce passenger risk.

3) UNDERSTANDING AND OVERCOMING LIMITATIONS The quantified risk model is a traditional tool but there is increased discussion of the limitations of such techniques. Firstly, there is a tendency to assume it is possible to identify all the various ways in which a system can fail. This is particularly significant when there are people involved in a system who can 'fail' in a number of complex and unforeseen ways. Secondly, the model is only valid if the input data is regularly updated. This is particularly relevant in the case of ULTra as the system will provide the world's first operations data on PRT systems and therefore updating the model on the basis of the new data is crucial.

The more involved hazard scenarios (such as loss of separation between vehicles) may benefit from a more holistic view. System dynamics (SD) has been proposed as a method to enable greater understanding of complex scenarios (Carhart, 2009). SD models are usually constructed from a Casual Loop Diagram. This diagram visually represents the positive and negative affects of each parameter on each other. It is proposed that sections of the risk model involving a large degree of human interaction, may benefit from the addition of a static system dynamic model.

Keeping the model up to date and valid is important as the data from the operating system is likely to come quickly and in large quantities. To achieve this process efficiently, a number of recommendations are suggested:

 The model should continue to be flexible to allow it to grow with increased information and changing beliefs. This can be enabled by having well-defined Fault trees with consistent numbering systems. The continued use of the specially developed FaultTree+ software is also considered important.  It is necessary to create a list of variables in the model that can be updated on the basis of system operations. These are likely to be those systems that are regularly tested and events that are predicted to happen frequently.  Good communications with the operations team needs to be maintained as they will be responsible for much of the data gathering (which will allow the model to be updated).

4) CONCLUSIONS This paper has described a risk model that has been used to analyse levels of risk in a novel transport system. It has suggested a method of analysing the model to identify priorities in the constant effort to reduce safety risk. It further makes a number of recommendations for maintaining the model and gaining the maximum amount of benefit from the model.

29

Future work is required in the development of system dynamic models. This may help to identify common cause failures. A multivariate sensitivity analysis would further increase understanding of the model and the system it represents.

Acknowledgements This work was supported by the Systems Centre and the EPSRC funded Industrial Doctorate Centre in Systems (Grant EP/G037353/1) and ULTraPRT Ltd.

References ALE, B. J. M. & PIERS, M. (2000) The assessment and management of third party risk around a major airport. Journal of Hazardous Materials, 71, 1-16. CARHART, N. (2009) Investigating the potential use of system dynamics as a tool for event analysis in the nuclear industry. The 4th IET International Conference on System Safety 2009. EVANS, A. W. (1994) Evaluating public transport and road safety measures. Accident Analysis and Prevention, 26, 411-428. LEVENSON, N. (2003) White paper on Approaches to Safety Engineering. LONG, R. A. Beauty and the Beast- The Use and Abuse of the Fault Tree as a Tool. LUXHOJ, T. (2003) Probabilistic Causal Analysis for System Safety Risk Assessments in Commercial Air Transport. Proceedings of the Workshop on Investigating and Reporting of Incidents and Accidents (IRIA}.

Biography

[email protected]

Alan Peters, a Research Engineer at the University of Bristol, is a member of the first Cohort due to complete the EngD in Systems in the autumn of 2010.

Alan joined the EngD in Systems programme in 2006, after completing a MEng in Knowledge Engineering (Department of Engineering Mathematics, University of Bristol). He works with ULTraPRT Ltd (formally Advanced Transport Systems Ltd) and his current research interests include risk modelling and sensor systems for a driverless transport system.

Alan’s research title is The Safety of Personal Rapid Transit Systems.

Personal Rapid Transit (PRT) is a new mode of transport designed to provide a viable alternative to the private car that is more sustainable, safer and more pleasant to travel in. PRT uses small driverless autonomous vehicles, travelling on segregated guideways, to provide an on-demand, non-stop service in urban environments. Currently a first system is being tested at Heathrow Airport. The long-term PRT vision is to have large complex networks in urban areas such as city centres. Routing algorithms will optimise the movement of vehicles to maximise passenger capacity using less substantial architecture than alternative transport forms. PRT, or a similar system, is necessary to reduce traffic congestion in large cities.

The EngD research has centred upon analysing and improving safety in PRT systems. This work includes risk modelling as part of the system safety case, design of an occupancy detection system, design of a visual method for navigation and design of a wireless collision avoidance system.

30

Framing the Debate about Sustainable Scales for Water Systems in the UK

C.M. Way*, S.E. Heslop*, D.B. Martinson**, R.C. Cooke*** * Department of Civil Engineering, University of Bristol, University Walk, Bristol, BS8 1TR, UK (E-mail: [email protected]) ** Department of Civil Engineering, University of Portsmouth, Portland Street, Portsmouth PO1 3AH, UK ***Buro Happold, Camden Mill, Bath, BA2 3DQ

Abstract There is currently debate around what is the most appropriate scale of water system in the UK context. There has been an increasing level of enthusiasm for decentralised and small scale approaches such as rainwater harvesting to complement the traditional “mains” supply and sewerage systems. However, with additional systems come additional impacts; environmental, social and technical, which are not necessarily negated by the reduced burden on the existing infrastructure. Before entering into an adversarial comparison of the “business as usual” versus “alternative” strategies, there needs to be an improved understanding of the problem situation. This paper seeks to frame the debate by defining the root problem and exploring the complex web of issues between the two.

INTRODUCTION Water systems in the UK are currently undergoing a phase of scrutiny not seen since the privatisation of the water industry in 1989. Small or micro-scale water systems are growing in popularity, but there are tensions between these building-level systems and the macro „mains‟ systems. There are questions being asked about the environmental credentials of, for example, rainwater harvesting (Clarke et al. 2009, Thornton 2008), as additional systems raise the environmental impact of the building, while not necessarily reducing the environmental impact of the mains networks, which typically remain in place. This paper outlines what is driving change in the building industry and water industry, before considering what the problem situation is. This highlights some of the key issues and tensions which need to be resolved, accommodated or exploited in order to move forwards.

CURRENT STATE This section is focused on identifying the root cause that is driving change. Why are there moves to change the existing system, which on the whole provides a good service at a price the majority of customers find acceptable (MORI 2002, WaterUK 2010a)? The standpoints of the building industry and the water industry are taken, as this provides interesting insights into demand and supply management issues.

Drivers for change in the building industry 1. Pressure from building rating systems Building rating systems such as The Code for Sustainable Homes (CLG 2008) and the Building Research Establishment Environmental Assessment Method (BREEAM 2009) have supported and driven recent enthusiasm for alternative small scale (up to hundreds of homes) water systems like rainwater harvesting (RWH) and sustainable drainage schemes. This can be seen in the rapid growth of the UK market for RWH systems from around £1M to £10M in the past 7 years (Johnen 2010). These rating systems require significant water savings from a baseline case, which at higher levels can only reasonably be achieved through use of alternative water sources for lower grade uses such as toilet flushing. This has led to RWH becoming part of the suite of technology indicators of an eco-development. Along with elements such as green roofs, timber construction, and solar-thermal panels, it is becoming intrinsically connected with the idea of a „green building‟ (Cutler 2009).

2. Interest in community solutions across all sectors

31

Community developments such as the new Ecotowns (CLG 2010) are seeking community scale energy technology solutions such as combined heat and power and district heating, and there is pressure to consider the applicability of similar scale water systems.

3. Increasing environmental consciousness End users are beginning to look at how to reduce water use, if only, for those on meters, for financial reasons. Those not on meters may be seeking to reduce use as part of a wider environmental consciousness (Stern 2005).

Drivers for change in the water industry 1. Pressure to reduce resource use in the macro system. There is pressure to reduce energy use (carbon emissions), water wastage and cost. The water industry‟s energy use and greenhouse gas emissions are increasing, largely in response to increased demand and quality requirements (WaterUK 2009). Initiatives like the Carbon Reduction Commitment and Climate Change Act (UK Govt 2008), mean the carbon emissions of the water industry are coming under scrutiny. Reducing water wastage through leakage is a perennial problem, and will increasingly need addressing as the asset base ages. Financially, there is also pressure to do more with less (WaterUK 2009), as OFWAT (the Water Services Regulation Authority) has a remit to keep prices „affordable‟.

2. Pressure to respond to changing conditions. The weather is becoming increasingly variable due to the effects of climate change. The water industry is at the forefront of dealing with these effects, as they are felt predominantly through impacts on the hydrological cycle and extreme weather events such as droughts and floods are becoming increasingly common (EA2009, Water UK 2008). These all change the flows the existing infrastructure was designed for, and strains its capacity and resilience. As previously mentioned the physical asset base is aging, and much of it is coming to the end of its useful life. This means decisions need to be made whether to regenerate, remove or replace it (EA 2009, WaterUK 2009).

3. Pressure to respond to changing demands. The EU Water Framework Directive (EU 2000) is requiring stringent standards of discharge to the aquatic environment, and meeting these requires modifications to the current system. This means financial expenditure, higher energy use, and sometimes additional facilities. Water quantity demands are also changing, in terms of absolute quantity and distribution. In recent years there has been less demand from industry, and more from the domestic sector (DEFRA 2008). The population is not only growing, but aging and migrating within the UK, and household size is decreasing with the rise of single occupancy dwellings (CRC 2008).

4. Requirement for advance planning Currently water companies have to follow a cyclic five yearly asset management process, which sets up strong, hard-to-change frameworks once they are agreed. Water companies also have to set out statutory 25 year resources management plans (DEFRA 2008) meaning that decisions are being made with long term impacts, so there is pressure to get it right now.

WHAT IS THE PROBLEM? The building industry has traditionally been characterised by demand-side approaches to resource management. Supply-side management has typically not been considered since the majority of people became connected to the “mains” supply and sewerage systems. This makes exploiting sources such as rain for lower-grade uses within buildings somewhat new territory for the industry.

32

The water industry has traditionally been supply focussed, following a predict-and-provide model. It is only in recent times that demand management has been considered, in part to alleviate pressure on the infrastructure.

The two groups are coming into each others traditional areas of expertise. Utility companies are no longer the only legitimate managers of energy, water and waste. In the building industry, organisations such as housing associations and local authorities are negotiating and delivering new energy water and waste services for their tenants (Van Vliet et al 2005). This creates a complex arena for regulation and decision making, and “normal” approaches may no longer be viable.

There is a feel that household and community scale (several hundred homes) systems may have benefits for reducing overall energy usage and lowering carbon emissions (EA 2008) but it is still very much a „feel‟. The water industry is making moves to respond to the pressures previously outlined (DEFRA 2008, EA 2009, WaterUK 2010a), which raises the question of whether this undermines the argument for small scale systems being “better” from an environmental perspective. The role of these systems in future resource management is not clear, and there needs to be an improved understanding of when different scales make sense from a demand and supply perspective. In order to frame this, the key tensions are highlighted.

KEY TENSIONS The historic focus of the water industry has been on providing safe, plentiful, affordable water (EA 2009). There is now ever increasing pressure to additionally consider carbon impacts and stricter discharge qualities. This is tied in with a broader move towards carbon accounting, which regulation is still trying to accommodate. Regulation has traditionally been divided between economics (OFWAT) and quality (Drinking Water Inspectorate, Environment Agency). How the balancing will be done between cost/ quality/ quantity/ carbon without conflict is not clear within existing legislation and funding mechanisms.

The complexities and uncertainties associated with this sort of environmental problem make scientific input and technical knowledge a cornerstone of the decision making process (Larson et al 2009, UKWIR 2004). There is a need to quantify the impacts of various types of water system, but research and development has been the big casualty of privatisation as water companies focus more on short term returns (EA 2009).

There is a deeply rooted historic precedent for our current water systems and attitude towards their management and consumption. Centralised water supply systems grew out of a need to provide for increasingly localised and large populations. These became so large that the structures and practices demanded were beyond the capacity of individuals, leading to municipal authorities taking responsibility. This meant public policies developed a paternalistic character, a ‟leave it to the experts‟ approach, and removed the burden of responsibility from the end user. However, people have become comfortable not being responsible for water, and will tend to distance themselves from water scarcity problems (Askew & McGuirk 2004). This leads to tension when there are attempts to return some of this responsibility. Reports often talk of engaging with the consumer or the public, and assigning greater responsibility to them for the protection of the resource, suggesting co- management between consumers and providers (EU 2000, Van Vliet et al 2005, EA 2009), but do they actually want it? And what if they don‟t?

There is tension in how far the public should be involved in decision making. The general public is known to have a relatively low awareness and concern for their water and sewerage services (MORI 2002) and public attitudes can limit the range of possible management choices. People rarely make rational decisions (Reed 2009, Sutherland 2007, Routhe et al 2005), so at what point do experts need to dictate what is best for the greater good? This is

33 a sensitive area, as people become fiercely protective over things that may affect their homes (As typified by the well-worn saying “An Englishman‟s home is his castle”).

There is a social reluctance to change habits and routines, which combined with the physical inertia of systems, means that change happens slowly and incrementally, and can be understood as a process of reconfiguration as old and new interact (Van Vliet et al 2005). This ties in to the idea of progressive retro-fitting; systems need intergenerational compatibility. However, there is pressure to change quickly, stressing the various social and physical systems.

Water is one of the most highly regulated industries in the UK, and this policy structure can be very constrictive and forms significant regulatory resistance. „It is not a question of technical capability, but rather economic and regulatory incentives and constraints‟ (EA 2009). UKWIR (2004) considers environmental legislation to be the key driver, even going so far as to say „it must take precedence in situations where there is a potential conflict between complying with legislation and achieving a sustainable solution‟. There is also a call for greater competition in the water industry (Cave 2009), but this is almost impossible with it being such a tightly controlled natural monopoly (most consumers do not have a choice of supplier, and it would be unfeasible to duplicate the existing infrastructure).

CONCLUSIONS It is not apparent that the interface between the building industry (demand management focus), and the water industry (supply biased approach) is currently being addressed. They are both engaged with different aspects of water management, and seeing the problems from different viewpoints. The debate around the application of small scale systems has brought this to light as they are located in this gap, and the roles and responsibilities of each side are not always clear. This disconnect needs to be resolved and mutually supportive pathways need to be developed to effectively tackle the issue of water stress in the UK.

Primarily there needs to be clarity in what needs to be achieved, and why. Traditionally, the building industry has dealt with buildings, the energy industry with energy provision, the water industry with supply and disposal of water and so on. Moving from this clearly delineated approach to an integrated resource network will inevitably be a wicked, messy problem (Jackson 1993).

REFERENCES Askew L.E., McGuirk P.M. (2004) Watering the suburbs: distinction, conformity and the suburban garden. Australian Geographer 35(1), 17-37 BREEAM. (2009) Breeam: The Environmental Assessment Method for Buildings around the World. available at: www.breeam.org (accessed 25th September 2009) Cave M. (2009) Independent Review of Competition and Innovation in Water Markets: Final Report. Department for Environment Food and Rural Affairs. London. Clarke, A., Grant, N. and Thornton, J. (2009) Quantifying the Energy and Carbon Effects of Water Saving. Environment Agency and Energy Saving Trust. CLG. (2008) Greener Homes for the Future. Communities and Local Government. London. CLG (2010) Eco-towns, Communities and Local Government website, Available at: www.communities.gov.uk/housing/housingsupply/ecotowns/ (accessed March 5th 2010) Cutler J. (2009) Water Efficiency. Contract Journal, 466(6713), 17. CRC (2008) State of the Countryside 2008, Commission for Rural Communities, Cheltenham DEFRA (2008) Future Water - the Government's Water Strategy for England. Department for Environment Food and Rural Affairs. London. EA (2009a) Evidence: A Low Carbon Water Industry in 2050, Report SC070010/R3. Environment Agency, Bristol. EA (2009b), Quantifying the energy and carbon effects of water saving - full technical report. Elemental Solutions, Hereford EU. (2000) Water Framework Directive. 2000/60/EC. European Parliament and the Council of the European Union. Jackson, M. C. (2003) Systems Thinking, Creative Holism for Managers. John Wiley & Sons. Chichester. Johnen L, 2010, personal communication

34

Larson K.L., White D.D., Gober P., Harlan S. & Wutich A. (2009) Divergent Perspectives on Water Resource Sustainability in a Public-Policy-Science Context. Environmental Science and Policy, doi:10.1016/j.envsci.2009.07.012 MORI (2002) The 2004 Periodic Review: Research into customers’ views, August Reed J. (2009) In search of the one litre drop. Water and Effluent Treatment News, March, 8 Routhe A. Jones R.E. Feldman D.L. (2005) Using theory to understand public support for collective actions that impact the environment: alieviating water supply problems n a non-arid biome. Social Science Quarterly. 85(4), 874-897 Stern P. C., (2005) Understanding Individuals’ Environmentally Significant Behavior. Environmental Law Reporter, News and Analysis, 35(11). Sutherland S. (2007) Irrationality, Pinter & Martin Ltd Thornton, J. (2008) Rainwater Harvesting Systems, Are They a Green Solution to Water Shortages? Green Building Magazine. Spring Edition, 40 - 43. UK Govt (2008) Climate Change Act. Hansard vol. 483 Col. 855 UKWIR (2004), Sustainable Waste Water Treatment Works for Small Communities, Volume 1: Sustainability and the Water Industry. Report 04/WW/04/9, UK Water Industry Research Limited, London Van Vliet B., Chappells H. & Shove E. (2005) Infrastructures of Consumption, Environmental Innovation in the Utility Industries. Earthscan Publications Ltd., London WaterUK (2008) How the water industry is adapting to climate change. WaterUK Briefing, December WaterUK (2009) Sustainability Indicators 2008/2009. London WaterUK (2010) Building a New Water Business Model Available at: www.water.org.uk/home/news/press-releases (accessed 16/03/2010)

Biography [email protected]

Celia Way worked at Buro Happold following graduation, and shortly afterwards joined the EngD program at Bristol University. Within Buro Happold, Celia is a member of the Sustainability and Alternative Technologies Group, which provides consulting services related to sustainable development and alternative energy, water and waste technologies for buildings. Her current research is focused on the application of decentralised water systems in the UK.

EngD research title: The Application of Decentralised Water Systems in the UK

There is currently debate around what scale of water system is „better‟ in the UK context. There has been an increasing level of enthusiasm for decentralised and small scale approaches such as rainwater harvesting to complement the traditional „mains‟ supply. However, with additional systems come additional impacts; environmental, technical and social, which are not necessarily negated by the effect of drawing less water from the existing infrastructure. Before considering the balance of these impacts to determine which strategy is „better‟, there needs to be improved understanding of what to measure or compare and why. This EngD research seeks to frame the debate, and investigate the complexity of the arguments, in part through interviews with experts in the field.

35

Towards a Cyclic Systems Model of Technology Development

Ian Watson1,3, Edward P. Goddard2,3, and Katherine H. V. Fulcher2,3 1Frazer-Nash Consultancy, Dorking, RH4 1HJ, UK 2Frazer-Nash Consultancy, Bristol, BS1 5TE 3The Systems Centre, University of Bristol, BS8 1TR, UK INTRODUCTION This paper advocates a systems approach to understanding the process of developing and investing in new technology. The process of investment has hitherto been studied in isolation from the processes of technology development - which creates value for the technology developer - and market introduction - which realises that value to the technology developer, and in turn to the investor. This paper seeks to address this by developing an outline systems model of the process of investing in technology.

APPROACH The work is founded on a literature review of each of the venture capital, technology development and marketing and innovation literatures. Checkland‟s [1] Soft Systems Methodology was used to determine a structure for the model, an analogy was developed to introduce the logic which links each of the principal parts of the model and this was refined into an initial systems model which is presented for discussion. The literature to date has been written to satisfy the demands of the individual stakeholders in the system. The venture capital literature takes an investor‟s standpoint. Similarly both the technology development and marketing and innovation literatures are composed from the standpoint of the company in receipt of the investment, seeking to develop and market the technology. Each area of the literature contains scant links to the others. By analysis of the literature a set of requirements for a whole-system model to represent the process of new technology commercialisation were derived. These requirements are:  The three principal actors in the system (investor, technology developer and the customer) must be represented.  The model must represent the process undertaken by each actor.  The model must show the steady advance in technical maturity of the product.  The model must represent the interaction between market and technology needed to develop a commercial product.  The model must show how the process of technology development generates the investor‟s return on their investment.  The model must be flexible, to accommodate investment at any stage of technical development, and exit at any future stage as appropriate. From this, the system of investing in technology has been conceptualised by three interconnected sub-systems operating concurrently.

AN ANALOGY In order to explain how the interaction of the three sub-systems creates value, we present an analogy drawing on the plant cycle of growth. Here, the technology is the plant, from seed to tree and the organisation in which it exists is the soil. The model is comprised of six stages and is shown in figure 1.

TOWARDS A SYSTEMS MODEL Following Trott & Hartmann [8], a cyclic model was developed, consisting of three principal streams of activity; - investment, technology development, and marketing. These capture the interests of the three principal actors in the system – the investor (who carries out the investment activities); the technology developer (who carries out the technology

36 development and marketing activities) and the customer (who is engaged by the marketing activities). The cycle shown in figure 2 has four principal phases, which seek to align the venture capital process [4] with the staged-gate technology development approach [6]. The first phase seeks to identify the opportunity for investment. Market research is being carried out by the developer to understand the likely value of a proposed technology development to the customer. Meanwhile, the investor is searching for investment opportunities. The two actors first interact at the end of phase 1, typically when a technology developer presents a business plan to the potential investor. Phase 2 concerns the assessment and quantification of the opportunity on the part of both parties. The investor will screen the business plan against their basic investment criteria [2,3], and if they are met will proceed to detailed evaluation of the investment opportunity. This will almost always involve assessing the technology itself, as well as the suitability of the organisation which proposes to Fig 1: Analogy of Technology Commercialisation develop it to undertake such a task. During this time the technology developer may undertake organisational planning activities and refine the market research to develop an initial market segmentation. If the investment proceeds, the phase concludes with the maturity of the technology understood by both the investor and the technology developer, a commercial plan in place, and financial closure of the investment. The third phase represents the technology development process, from the point of investment until the technology is ready for market. It is in this phase that value is created. The capital invested is spent on activities which seek to mature the technology. If these are successful, then the value of the technology increases as it comes closer to market [7]. The broadening of the arrows in the model seeks to represent both the increasing effort spent on each phase in the system, and the corresponding increase in the value of the technology under development.

This stage usually concludes with a Fig 2: A Systems Model of Technology Commercialisation technology in the range TRL 7/8 [5];

37 being as developed as is possible before a commercial launch of the technology. At this stage the initial marketing strategy is finalised. The final phase is that in which the technology developer realises the value of the technology. This is usually achieved through a commercial launch, producing and selling items to businesses or consumers. The initial marketing strategy is implemented, and an ongoing process of monitoring and control is begun. This is often the stage when an investor will seek to realise the value of their investment by cashing out. Investors must sell their claim on the ownership of the technology developer in order to repay the capital to those who have invested in their funds. The investor‟s return depends on the market perceiving this claim as more valuable after the technology development activity than it was before the work was undertaken. From the standpoint of a particular technology, phase 4 can be viewed to continue for its entire product lifecycle. However from the point of view of the technology developer it is often at this stage that the cycle loops back and begins again. The resources used to develop the technology are no longer needed to support it in production. Product updates will require further research and development work; spin-offs and complements may require developing, or other markets may emerge which require the technology to be adapted. The first phases of the cycle may not be undertaken explicitly for future iterations of technology development; however few companies will proceed, even if they themselves are acting as investor, without at least an outline business plan, assessment of the suitability of the technology, and a formal decision to proceed.

VIII. CONCLUSION This paper presents a systems model of the technology commercialisation process, taking into account the interdependency of investment, technology development and marketing activities. It builds on the staged-gate technology development process by identifying the investment and marketing activities necessary to enable to technology to develop to maturity. The model is conceived to represent a wide range of technology developments and funding processes, from internal investment within a large corporate organisation, through corporate venturing to external venture capital models. The model is cyclic, emphasising the continuous nature of technology development as a process of ongoing value creation, and investment as the process which seeds this cycle and harvests the reward at a later time.

ACKNOWLEDGMENT Ian Watson, Edward Goddard and Katherine Fulcher are registered on an Engineering Doctorate Programme at the University of Bristol; financial support from the UK Engineering and Physical Sciences Research Council (EPSRC) is gratefully acknowledged.

REFERENCES [1] P. Checkland, “Towards a systems-based methodology for real-world problem solving,” Journal of Systems Engineering, vol. 3(2), pp. 87–116, 1972. [2] I. C. Macmillan, L. Zemann, and P. N. Subbanarasimha, “Criteria distinguishing successful from unsuccessful ventures in the venture screening process,” Journal of Business Venturing, vol. 2, pp. 123–137, 1987. [3] I.C. MacMillan, R. Siegel, and P. Subbanarasimha, “Criteria used by venture capitalists to evaluate new ventures proposals,” Journal of Business Venturing, vol. 1, pp. 119–128, 1985. [4] T. Tyebjee and A. Bruno., “A model of venture capitalist investment activity,” Management Science, vol. September, pp. 1051–1066, 1984. [5] J. C. Mankins, “Technology readiness levels,” Advanced Concepts Office, Office of Space Access and Technology NASA, Tech. Rep., 1995. [6] R. G Cooper, S. J. Edgett, and E. J. Kleinschmidt, “Optimizing the stage-gate process what best practice companies are doing,” Research Technology Management (Industrial Research Institute, Inc.), vol. 45, no. 5, 2002.

38

[7] P. Bigwood, M, Managing the New Technology Exploitation Process. Industrial Research Institute Inc., 2004. [8] P. Trott and D. Hartmann, “Why „open innovation‟ is old wine in new bottles,” International Journal of Innovation Management, vol. 13, no. 4, pp. 715–736, December 2009.

Biography

[email protected]

Ian Watson studied Engineering, Economics & Management at the University of Oxford, before joining the EngD in Systems programme at the University of Bristol in 2007. His research interests are the management of technology, decision sciences and investment appraisal methods.

Ian is also a consulting professional, with over five years‟ experience working with over 13 marine energy developers on behalf of Frazer-Nash consultancy. He has led technical due diligence assignments on behalf of major utility companies and has advised a wide range of technology developers, from major blue-chips to university based start-ups developing wave and tidal energy devices.

EngD research title: Improving Investment Decision Making in Complex Engineering

From wind turbines to aircraft carriers, many investment decisions are ultimately founded on consideration of complex technological systems. In these instances, it follows that the characteristics of the investment decision follow from the performance and associated uncertainty of the technical system.

If the system is sufficiently well defined to encompass not just the „hard‟ technical system, but also the „soft,‟ people-centred system in which it sits, then a bottom-up approach to investment appraisal can be developed. Ian‟s research centres on folding in the investment aspects into the overall system, moving from the philosophy of appraising a technical system based on key system performance parameters to modelling the system as part of the enterprise process, taking into account the strategic aims and objectives of the business.

Ian hopes that such an approach will allow earlier, more informed investment to be undertaken, by better characterising the nature and impact of risks which arise from the project under consideration. Such an approach could greatly benefit any organisation which invests money, time or resources in complex technological systems.

Development of Customer Focused Ice Pigging Equipment for the Food and Pharmaceutical Industries

E.A. Ainslie1, D.G. Ash1, J. Chalmers2, T.J. Deans1, A.N. Leiper1, M. Herbert1, G.L.Quarini1

1. Faculty of Engineering, University of Bristol, Queen's Building 2. ABB Ltd, Daresbury Park, Daresbury, Warrington, WA4 4BT

Introduction Ice Pigging is a patented process (Quarini 2001, 2004) which can be used for various duct cleaning demands. The process involves pumping a crushed ice slurry into and along a pipeline, where the slurry exhibits both solid and liquid behaviour, depending upon the geometry of the pipe, the flow conditions and the properties of the ice slurry. The slurry

39 presents an elevated viscosity when compared to water, thus providing a larger shearing force at the pipe wall for the same velocity of flow (Shire, 2005). In addition the ice pig maintains plug, flow even under challenging geometries, such as changes in diameter, 90° and 180° bends (Shire, 2006) and various heat exchangers (Shire et al., 2009). The plug flow behaviour can be exploited to clear pipelines of remaining product, such as (Shire et al., 2005), with minimal contamination, resulting in improved levels of product recovery, when compared to water.

Background Water Industry Extensive testing has been carried out in the water industry, over the previous three years (Evans, 2007, Quarini et al, 2010), with over 30Km of live water mains having been pigged out in order to reduce the levels of sediment, and hence both the average turbidity levels and the risk of customers complaints due to discolouration events. This has resulted in the process being licensed to Agbar Environment and rolled out across the UK. The application of ice pigging to the water industry represents the simplest situation, with one relatively straight pipe being cleaned per job, typically with no changes in diameter, except at the entrance and exit. As a result early equipment developed for the manufacture, storage and dispensing of the ice slurry was heavily focused on the technological aspects of the process, requiring skilled operators who understood the process, which due to the simplicity of the situation was acceptable. In addition as the pipe line in question was different every time, the equipment had to be correctly assembled for each job, which also required a skilled operator.

Food and Pharmaceutical Industry When ice pigging was first investigated the food and pharmaceutical industries were identified as particularly high value areas (Quarini, 2002). The manufacture of food stuffs and consumer goods, typically result in several batch changes per day, with a significant volume of product being lost in the effluent between batch changes. It is this product recovery which the ice pig could improve, resulting in more product being produced per batch and reducing the cost of treating effluent from the factory, due to the lower product content. Despite successful initial testing, where the process provided high levels of product recovery (Quarini et al., 2007) there were several issues surrounding the operation of the process, and equipment being used, which have lead to only one system currently being installed in a food factory situation.

Analysis of Current Situation

The system which was used for factory trials was analysed on several different levels: the basic hardware, the interface between ice pigging hardware and the interface with the end user (Figure 1). The first two investigations lead to numerous material improvements to the equipment, however the third stage leads to some interesting and fundamental problems, which may explain the minimal uptake of ice pigging within the mentioned industries: • High level of user understanding required for successful operation • Multiple operators required • Integration into factory system requires extensive customer input • Minimal system protection • Appears to be custom built trial kit • Lesser known brand components • Academic involvement not associated with reliability

Once again these could separated into factors related to the actual operation and performance of the ice pigging equipment, the first four, and the customers perception of the ice pigging equipment, the last three.

40

Figure 1 - Current system interfaces and proposed system interfaces

Proposed Solution

In order to improve both the system integration and the customer perception a relationship with an industry leading automation company, ABB, was developed. This provided the knowledge and equipment to standardise an ice pigging product such that integration into any factory system was easily possible. In order to address the knowledge and decision making load which was previously placed upon the local operator a Programmable Logic Control was added which could automate many regular operations and move the operator decision making to a higher, task based level. Combined with the PLC was an operator interface which provides clear access to the task based system, whilst also providing the operator with improved feedback concerning current tasks being performed (Figure 1). This also allows for improved reliability of both the process and the equipment performing the process. The PLC can be used to ensure that the equipment is always operating within give bounds, reducing damage which could occur due to operator error.

Future Development

The proposed improved system has now reached the commissioning and testing stage with the following key aspects to be tested:

• Do the system safeguards prevent the ice pigging rig being damaged? • Do the safeguards provide adequate protection to the factory production facilities? • Has the process been simplified to an attainable level? • Does the operator learn about the process from the feedback provided? • Can the operator successfully operate the machine without significant prior instruction? Conclusions

The ice pigging process has been proven to provide significant customer benefit, across a range of industries, however is only currently being exploited to any level in the water supply area. To improve uptake across other sectors, primarily food and pharmaceuticals, improvements to the equipment are required. The changes which are thought to provide the largest benefit include: improved factory integration capability; improved safety systems to protect both equipment and user; simplification of demand on local user.

References

Evans, T. S., 2007. Technical aspects of pipeline pigging with flowing ice slurries. PhD Thesis, University of Bristol. Quarini, G. L., 2001. Cleaning and separation in conduits. Quarini, G. L., 2002. Ice-pigging to reduce and remove fouling and to achieve clean-in-place. Appl. Therm. Eng. 22, 747–753. Quarini, G. L., 2004. Methods of cleaning, clearing and separation in conduits.

41

Quarini, G. L., Evans,T. S., Shire, G. S. F., 2007. An environmentally friendly method of cleaning complex geometry food manufacturing equipment and ducts with maximum product recovery, using ice slurries. Food Mf. Eff. 1(2), 15–27. Quarini, G. L., Ainslie, E.A., Andrews, S., Ash, D., Deans, T., Herbert, M., Haskins, N., Norton, G., Rhys, D., Smith, M., 2010. Investigation and development of an innovative pigging technique for the water-supply industry. Proc. IMechE, Part E: J. Process Mechanical Engineering 224, 79 – 89. Shire, S., Quarini, G. L., Ayala, R. S., 2005. Experimental investigation of the mixing behaviour of pumpable ice slurries and ice pigs in pipe flow. Proc. IMechE, Part E: J. Process Mechanical Engineering 219, 301–309. Shire, S., 2006. Behaviour of ice pigging slurries. PhD Thesis, University of Bristol. Shire, G. S. F., Quarini, G. L., Rhys, T. D. L., Evans, T. S., 2008. The anomalous pressure drop behaviour of ice slurries flowing through constrictions. Int. J. Multiphase Flow 34, 510–515.

42

Industrial Doctorate Centre in Systems

In early 2006 the University of Bristol, in collaboration with the University of Bath, was awarded £3.4M EPSRC funding to set up the EngD Centre in Systems. The Centre received an immediate high level of interest both from companies and potential Research Engineers (REs). In its first three years of operation, 31 EngD projects were established, across the full range of disciplines, sponsored by more than 20 different companies. Our industrial partners have been enthusiastic and proactive participants in the Centre, a factor that has contributed strongly to its success. In April 2009 the Centre was awarded a further £5.3M from EPSRC to establish a new Industrial Doctorate Centre (IDC) in Systems. The first year (2009-2010) of the IDC has brought on board a further 18 REs – the highest rate of recruitment across all IDCs, and a strong indicator that the IDC in Systems, as with its predecessor EngD Centre in Systems, is truly meeting industrial needs for holistic Systems Engineers, for which there is a substantial unsatisfied demand. Each of Bristol and Bath Universities is pursuing an integrated strategy to establish itself as an internationally competitive, research-intensive university, committed to the transfer of knowledge and the provision of a world-class education. One of the key elements of this strategy is to expand and enhance postgraduate education and research activity. The IDC in Systems, hosted by the Systems Centre at the University of Bristol, is a flagship for this development. The Universities, industry partners and many professional engineers have invested significantly in the Centre. The „learning‟ that is emerging is influencing Faculty strategies and, in turn, postgraduate and undergraduate education. The following quotes from our many letters of support for the IDC in Systems articulate the industrial and national significance of Systems Engineering and Systems Thinking: “The IDC for Systems will have enormous benefit to industry in providing access to the latest techniques, methodologies and processes required to deliver effective world-class complex integrated systems. This will revolutionise the UK manufacturing sector ….” – RAEng “(Dstl) have recently introduced an accelerated development scheme for identified staff to enhance their systems skills. The successful implementation and delivery of this scheme is one of our corporate targets. “…. such high-level skills are needed and in short supply. I believe that generally better application of the systems approach will be of wide benefit across the whole of the UK.” – Rolls-Royce “Clearly we believe and demand from our clients supports this, that UK has a current and growing need for the highest competence in systems engineering.” – Frazer-Nash (Part of Babcock Group) We wish to thank all the organisations that are supporting the Centre. Our industrial partners contribute to the Strategic Development of the Centre and collaborate on research.

43

INTRODUCTION TO THE EngD

The Engineering Doctorate (EngD) is an alternative to the traditional PhD for individuals keen to pursue a career in industry. A 4-year programme combines PhD-level research with taught courses, and Research Engineers (EngD „students‟) spend about 75% of their time working directly with a company. EngD programmes are attracting top quality people and many with already significant industrial experience. Research Engineers (REs) are required to meet not only the University‟s high academic entry requirements, but also to undergo stringent selection by the sponsoring company. The Systems EngD also requires individuals to be, or have the potential to become, „systems thinkers‟.

The sponsoring company proposes the EngD „project‟, and gains from having a talented and highly motivated individual pursuing business case-based research identified as important to the company‟s future. The RE should be treated for all *intents and purposes* as an employee, with the company committing in advance to releasing the individual to attend course modules, and to supporting their research. Existing company employees are able to join the programme as „employed‟ REs and this provides a means for the company to develop (and/or retain) their very best people. The 2-year part- time MRes in Systems provides a pathway to the EngD for companies unable to commit fully to the 4- year EngD programme.

The EPSRC supported Industrial Doctorate Centre in Systems awards stipends (£15,090 pa in 2010- 11) and fees (£3,440 pa in 2010-11) to selected projects and REs; also generous support towards, eg, conference attendance. The Centre assigns academic supervisors to support the RE and their project, as well as to ensure sufficient academic rigour. Companies are required to „top up‟ the RE‟s stipend by a minimum £3,500 pa, and to pay a contribution to the Centre‟s operating costs. The overall cost to a company is typically £10,000-12,000 pa for larger companies.

EngD projects should be regarded by HMIR as research *by definition*, hence the company should be eligible for R&D tax credits in respect of all costs associated with the EngD project. These costs may go well beyond the contribution the company makes to the RE‟s „training‟, but this is for the company itself to explore; the Centre can provide only limited guidance on this.

Useful sources of information

EPSRC‟s website provides useful information on the Industrial Doctorate Centres: http://www.epsrc.ac.uk/funding/students/centres/Pages/indd.aspx

EngD Stakeholders Survey. An assessment of the relevance of the EngD scheme to its key stakeholders, including views of current and past REs, academics and industry: http://www.epsrc.ac.uk/pubs/reports/Documents/EngDReviewReport.pdf

R&D tax credits: http://www.hmrc.gov.uk/randd/index.htm

Dr. Oksana Kasyutich Systems centre manager [email protected] 0117 3315421

Dr. Peter Ereaut Business Manager Systems Centre [email protected] 0117 9289003

44

Learning together

Industrial Doctorate Centre in Systems www.bristol.ac.uk/eng-systems-centre/idc

University of Bristol University of Bath Engineering Faculty, Queen’s Building School of Management University Walk Bath BA2 7AY Bristol BS8 1TR