Keywords

10.1 Introduction

Supply chain safety and quality assurance aims to proactively create and maintain prerequisites for nuclear safety in supply chains. An important task is being aware of the supply chain structure and how it affects nuclear safety. In megaprojects such as nuclear new builds, this is not trivial because the number of suppliers can be very high. For instance, approximately 2000 subcontractors were involved in the Finnish Olkiluoto 3 nuclear power plant construction project, reaching up to five tiers at the construction site [1].

In this chapter, the authors describe how a network visualisation method was developed and used to support supply chain quality and safety assurance of a nuclear power plant construction project. The benefits and limitations of applying this type of visualised representation of the supply chain for safety practice are discussed.

10.2 Visualising Safety and Network Visualisation

In safety science and practice, visualisations serve many purposes. At least four types of safety visualisations can be distinguished. Conceptual safety visualisations present some aspect of the concept of safety in a visual manner. They aim to answer the question “what is safety”. Such visualisations are often educational and include a theory of how accidents occur, or what phenomena can influence safety. Examples of famous conceptual visualisations include Reason’s Swiss cheese model [2] and Rasmussen’s sociotechnical risk management and migration models [3].

Data visualisations convey safety-related data to end-users for easier recognition of patterns, for summarisation, or for economical communication of the data. They aim to answer the question “is it safe”. Examples of safety-related data visualisations include conventional charts (e.g., bar, pie or line charts) of adverse outcomes, and their trends.

Visual tools help make sense of safety-related information. They involve user interaction, including inputting, organising and analysing data visually. Such tools aim to answer the questions “how does this relate to safety” or “is this safety”. Examples of visual tools include the accident analysis methods ACCIMAP, FRAM and bowties.

Visualisations can also be used to communicate safety-related phenomenon in a dramatic and vivid manner, aiming to influence the viewers through creating affective responses. They answer the (possibly unasked) question of “what is safety” by explaining “this is safety” or “this is not safety” through means of narrative and dramaturgy. Examples of visual dramatisations include posters, movies, videos and websites with graphic content of accidents or their causation.

This chapter focuses on one type of visualisation process, the visualisation of networks. Network analysis examines the relationships between entities (e.g., friendship, communication or acquaintance networks). Although perhaps most commonly used for social networks, network analysis is not limited to social entities or phenomena, but it can be used with any relational data. In the context of supply chains, network analysis has been used for modelling contractual relationships, material flows, communication of instructions between the companies, performance incentives, etc.

Visualisation is an integral part of network analysis, because it facilitates, for instance, the detection of interactions and emergent patterns, and understanding the overall structure of the network. The most common visual representations of networks are node-link diagrams. Node-link diagrams consist of nodes, links that connect the nodes, and a layout (incl. node positioning and link routing). To improve the readability or emphasise some aspects of the diagram, various metrics are calculated based on network topology or the underlying data and are mapped to visual parameters of the diagram (e.g., node sizes and colours, and link colours and widths).

10.3 Supply Chain Network Visualisation Method

Fennovoima (the future operator of Hanhikivi-1 nuclear power plant) has granted an EPC (engineering, procurement, and construction) contract for a complete turnkey delivery of the power plant to the plant supplier, who in turn has made several contracts with vendors. According to Finnish legislation and regulatory requirements, the licensee is responsible for ensuring the safety of the nuclear power plant in all its life cycle phases [4]. In the context of supply chain quality and safety assurance, one of the implications of this requirement is that the licensee must have an overview of the status of the supply chain. The supply chain network visualisation method was developed as a partial solution to this issue. Its purpose was to help make sense of the project’s contractual structure and support quality and safety assurance activities.

Before the network visualisation, information regarding the contractual relationships was in spreadsheets. The spreadsheets were rather complicated and hard to make sense of due to the sheer number of contracts and companies involved. Simplified visualisations that only described the most important top suppliers were also available, but they only contained a small fraction of the whole supply chain. An overall visualisation was not a high-priority task because each technical discipline was well aware of the companies that were directly connected to their job. However, for supply chain quality and safety assurance, a holistic perspective that takes the whole supply chain in account is necessary to understand how companies interact, where they are located in relation to other companies, and how they contribute to the overall safety of the construction project (and ultimately, the nuclear power plant). Network visualisation was chosen to provide this overview due to the following reasons:

  • It provides an overview of contractual structure of the construction project.

  • It provides a great deal of information at a glance without having to access the raw data (incl. safety classifications, contract grades and contract expiration dates).

During the years 2017–2019, the supply chain network visualisation has been produced six times (see Fig. 10.1 for example graph). In each update, new data was added and improvements were made to its visual design. It has evolved into a highly customised graph in response to practitioner needs. After two years of utilisation and development, the supply chain network visualisation method has established itself at Fennovoima. It has been incorporated into management system procedures as one of the methods periodically utilised for gaining an overview of the supply chain. The visualisation is still in continuous development.

Fig. 10.1
An illustration of supply chain network visualization. It has 3 suppliers A, B, and C with a maximum contract to supplier A and a minimum to supplier C. Supplier C has a maximum number of safety class 1 contracts. And all have both incoming and outgoing contracts.

Extract of the supply chain network visualisation (all company names have been removed and a few random modifications have been made to ensure confidentiality). Nodes indicate suppliers and links indicate contracts. Colours are mapped to contract safety classifications, e.g., red arcs refer to safety class 1 contracts (incl. reactor pressure vessel or primary circuit components). Node shapes indicate the structural positions of suppliers (left side indicates number of incoming contracts and right side outgoing contracts)

Full size image

10.4 Application

The supply chain network visualisation has been applied as a decision support tool in defining the extent of supply chain quality and safety assurance activities, especially in the field of safety culture. As the future operator of the Hanhikivi-1 power plant, Fennovoima is responsible for assuring safety culture during construction. Fennovoima’s safety culture assurance activities include monitoring activities such as audits, facilitation activities such as trainings and collaborative activities such as work groups with suppliers (for further details, see [5]).

In the nuclear industry, graded approach is applied to ensure that the application of requirements and the stringency of control measures is commensurate with nuclear safety significance. Fennovoima has developed a specific graded approach for supply chain safety culture assurance (SCA grade) to define safety culture assurance activities for a given supplier or sub-supplier on a four-level scale (A-D).

Fennovoima identified the need for safety culture specific grading when the supply chain grew in size and the supplier and the sub-suppliers signed multiple safety-classified contracts. Actors in the supply chain became more distant (due to increase in tier length) and their significance and interrelations became more difficult to comprehend. To maintain focus on significant suppliers, SCA grading process was developed as part of a wider development of supply chain safety culture assurance. SCA grades are reviewed and updated on an annual basis (for further details, see [5]). The initial determination of the SCA grade is based on the safety classification of the contract and the initial grade given by Fennovoima’s supply chain management. Additional factors influencing the SCA grade include the type of contract and the position of the company in the project network. The supply chain network visualisation contributes to the grading process by identifying the position and role of each supplier in the overall project network and in relation to other suppliers. This approach borrows from social network analysis. The visualisation provides a way of positioning the supplier in the Hanhikivi-1 project and has already on a couple of occasions acted as evidence for raising the SCA grade of a particular supplier.

For example, the visualisation shows that Supplier A (Fig. 10.1) has multiple incoming safety-classified contracts (marked as red, orange and blue), but no outgoing ones. Hence, Supplier A acts as a key node in terms of delivering safety–critical services to multiple other companies. This position means that there are several companies already auditing and monitoring Supplier A, and providing information regarding its status. That is, a lot of information on this supplier is probably already available. This may include audit and inspection reports, observations, and other documented data. Consequently, Fennovoima needs fewer monitoring activities of its own. For example, Fennovoima does not necessarily need to do safety culture audit to Supplier A (or audit frequency can be decreased), if the customers of Supplier A have well-functioning auditing programmes. However, Fennovoima would first need to verify the capability of the customers of Supplier A to produce usable and reliable information, for instance by observing their audits, crosschecking their findings, or reviewing their assessment or inspection processes.

The position of Supplier A also suggests that there is a risk of common cause failure, because one supplier delivers to multiple customers. If this supplier fails to deliver an acceptable product, it can have widespread effects on the project. This has implications of quality and safety assurance of Supplier A: it might be necessary to investigate, what is the capability of Supplier A to manage multiple deliveries, with potentially differing, overlapping and contradicting requirements.

Supplier B (Fig. 10.1) is in a different position in the network. This supplier has a vast number of contracts with varying safety significance going out, and only one safety-classified contract coming in. It represents a key node in terms of oversight and contract management activities, including the distribution of requirements further down the supply chain and ensuring that sub-suppliers understand and apply them. Being in such a position suggests that Supplier B has an effect on many companies and needs to have a highly developed supply chain management practices of its own. However, many contracts of Supplier B are not categorised nuclear safety-significant (marked as green in Fig. 10.1).

The implications of the Supplier B position of the network include that there might not be as much documented information available from other companies in the network. However, due to the vast size of its supply chain, there exists a lot of knowledge in the sub-suppliers about Supplier B. This means that supply chain quality and safety assurance activities towards Supplier B (and other companies in similar positions in the network) are based on Fennovoima’s own data collection and generally ensuring continuous and close collaboration practices with this supplier and its sub-suppliers.

Supplier C (Fig. 10.1) represents yet another type of position in the network. It has incoming as well as outgoing contracts. Supplier C sets requirements to others and delivers services based on requirements set by someone else. That is, its customers monitor it, but it also has the responsibility to establish and monitor its own supply chain. From Fennovoima’s perspective, this company is not only responsible for quality products, but also for other suppliers. This calls for an assurance approach combining those described for Supplier A and Supplier B.

One of the main insights of the network visualisation has been in illustrating the networked nature of the project supply chain, and that suppliers in different positions require different supply chain quality and safety assurance approaches. In principle—and in hindsight—many of these observations could be deduced from the supply chain contract register spreadsheet without visual aids. However, the visualised representation of the supply chain proved helpful in orienting the supply chain quality and safety assurance activities to consider the positional relations between the suppliers, which is something that the tabular data was not able to do. Spreadsheet data is still needed when details or specific contracts need to be reviewed—as of now, the visualisation is too coarse a method for examining details. This suggests that it should be considered as a complementary tool among other tools in making sense of the supply chain.

10.5 Discussion

Experts from various disciplines at Fennovoima have communicated and presented the supply chain network visualisation in various events and meetings, ranging from top management meetings to nuclear safety committees and regulator inspections. The visualisation has been used to communicate the overall structure of Hanhikivi-1 project network, or to illustrate the reasoning behind supply chain assurance decisions. Overall, the reception and feedback has been very positive. The visualisation has been described as providing a good overview, or as offering a holistic picture of the supply chain in a simple way.

What does the supply chain network visualisation tell that non-visual data cannot? The application of the visualisation indicated that there are clear benefits to using visualised as opposed to tabular or textual descriptions of the supply chain.

First, it is an intuitive way of analysing complex data and phenomena. The visualisation helped identify patterns in the contractual data and connections between different suppliers, helping experts determine the suppliers’ roles in the project network and consequently support designing quality and safety assurance strategies.

Second, it serves as an economical communicational aid for situations where time and efficiency is of the essence. Visual inspection of the diagram combined with a few examples of the different nodes gives the viewers an overview as they are able to see the entire network at a glance. Describing the basic functionality (e.g., colours, node sizes) of the graph, the viewers learn to read it quickly, which is not easily achieved with complex spreadsheets. This benefit was evident in many top-level meetings, where only a short time window was available for presenting. Using the visualisation, experts can present massive amounts of information.

Third, it is relatively easy to return to in later communications because the viewers are already familiar with the visualisation. Hence, it provides a memorable reference point for supply chain-related discussions or for sense making.

The visualisation might not be always useful for everyone. The end user’s familiarity with the supply chain influences how they perceive the visualisation. For example, supply chain experts who know the underlying data intimately, and know how to read the supply chain contract register effectively, are likely to benefit less from the visualisation in terms of understanding how different companies relate to each other, or where they are visually located in the network because they already intuitively know this. This might not be the case for other experts or the management, who probably only know the suppliers most relevant to their tasks, but not the big picture. For them, the visualisation provides an easy-to-approach overview of the supply chain, something that tabular contract register is not able to provide.

While most viewers of the visualisation perceived it to be “interesting” and thought-provoking, its meaning or relevance—especially its relation to nuclear safety—was sometimes hard to grasp. In its current form, the only explicitly safety-related data included in the visualisation is the safety classification of contracts. Other safety-related information inferred from the visualisation relies on the experts’ interpretation. For example, social network metaphor was applied to interpret the visualisation when assigning SCA grades, as described in the previous section. Adding more data points, especially ones that relate to safety (e.g., audit or inspection findings, etc.) may be a potential approach to make the visualisation more readily interpretable for end-users and to provide a more complete overview of the status of the supply chain. However, there are major drawbacks to adding more data. Increased visual clutter is the most critical one. Even in its current state the visualisation can be very hard to read in some areas, despite the efforts to make it clear. Trade-off between readability and amount of safety-related information included in the visualisation must be successfully managed. This means finding the answers to the following questions:

  • What is the minimum amount of safety-related information that is required for the visualisation to be useful to safety practitioners?

  • What is the minimum level of readability required that the visualisation would still make sense to the end-users?

Another solution to making the connection between the network visualisation and safety more evident is integrating a conceptualisation of safety into the diagram itself. That is, combining a data visualisation with a conceptual visualisation. To the authors’ knowledge, there are no established (visual) safety models that explicate the connection between safety and supplier roles in contractual networks. For instance, Reason’s Swiss cheese model is quite clearly focused on the operations within a single organisation and describing different types of barriers, and while Rasmussen’s sociotechnical risk management model does, in principle, include external actors such as regulators or the government as part of the sociotechnical system, it still does not specifically address supplier organisations nor their roles. Neither model readily addresses how contractual structures (or other network phenomena) influence safety. This may suggest that there is a need for a completely new type of safety model, or a creative variation of an existing one.

A potential drawback of integrating a safety model with actual data is that the viewers may become anchored to this particular representation of data and perceive that the data and the safety model are inherently linked. The risk is that such a visualisation might become treated as an end-all solution to making sense of safety in the supply chain. In actuality, a visualisation (or in this case, a visual safety model) only projects the data in one way. Data in itself can have any number of projections. Similar process occurs implicitly when the viewers of the current version of the network visualisation attempt to relate the visualisation to safety by applying their mental models of safety. Therefore, integrating conceptual models of safety in the visualisation requires care from its developers, and informed practitioners who know the assumptions underlying the model and are able to avoid the safety model having too much (or unwanted) influence on their thinking or decision-making.

The supply chain network visualisation was sometimes observed to induce affective responses on experts or decision-makers: the colourful visualisation was perceived as attractive, or its complexity was perceived as shocking. Integrating (or at least considering) an explicit safety model may help better manage such effects. It is important that visual dramaturgy directs attention to the right things, and to the ones most relevant to safety. Ideally, an engaging visualisation makes the various stakeholders more aware of the importance of managing quality and safety in the supply chain and the challenges it involves, and be more motivated and committed towards solving them. One of the risks is that the visualisation conveys the supply chain as too complex in a too simple way. That is, the viewer might only remember the visual complexity or its (potentially attractive) visual appearance (i.e., the things that caused the affective response), and not the safety-related insights embedded in the visualisation (cf. picture superiority effect). Therefore, knowledge in designing visual narratives or generally visual storytelling is important to understand what kind of responses the visualisation creates in viewers to manage its dramatic impact.

10.6 Conclusions

In this chapter, the authors described a supply chain network visualisation method, its background and an example of its application in a nuclear power plant construction project. The method was developed as a solution to better make sense of how a project network creates preconditions for safety. Experiences showed that the visual representation of the supply chain helped uncover such insights of the supply chain that probably would have remained hidden if relying on tabular data only. These insights were applied in designing supply chain quality and safety assurance activities. Further development needs were also identified, especially the development of safety models that explicitly address the safety significance of various interaction phenomena in safety–critical networks.