Conclusions

As was mentioned in the introduction section, some of the central aims of the IRD concept have been to provide overview information, support early detection of failures and disturbances and help operators to diagnose the problem and stabilize the process. IRD displays are intended to support the development of acceptable level of situation awareness as well as collaboration and co-ordination of activities within a crew by establishing a ‘common ground’. The main design principles include display normalization, and the Dull Screen and Information Richness principles. In the following the fulfilment of the aims and the suitability of the design principles to presenting information of nuclear power production process will be discussed.

5.7.1 Usefulness of IRD design principles

One of the most salient features of the IRD displays is that they are different from more traditional type of large-screen displays. Despite of the novel form of presentation experienced operators were able to take advantage of the Fortum IRD pilot and detect deviations successfully. The Fortum IRD pilot displays were also used during the process stabilization phase along with the process monitoring system displays. In addition, they offered operators a common view to the process. Therefore, it can be said that by these means the development of situation awareness was supported. It seems to be that it is not a fatal problem for the overview displays if the presentation format is different from what the operators are used to. However, it is necessary to take care of that some consistency and familiarity with other interface elements of the CR is maintained. Because of the IRD concept’s novel features, it is clear that in order to be able to use the display in an efficient way, sufficient training is needed.

As said, display normalization is intended to support rapid visual processing and detection of failures when several normalized graphs are set along a horizontal line. Operators did make use of trend information in the simulator runs, but they, however, said that display normalization is not a very informative property, and some of them even doubted if normalization is suitable for the presentation of information of the nuclear power process. Our results, thus, do not provide much support for the usefulness of display normalization. It can be that display normalization would have been more useful in other types of scenarios in which the changes occur more slowly.

The aim of the Dull Screen principle is to make the display clearer and prevent visual noise by using specific colouring rules. The main aim in the use of the principle is to improve the visibility of alarm colours. Our results, however, suggest that the chosen set of colours make other state changes even more difficult to detect. While the visibility of alarm colours was experienced

fairly satisfactory, the colour coding for state changes in different plant instrumentation (e.g. pumps and valves) were experienced as inadequate.

Because of the lack of sufficient practice and training, the operators were not able to take use of all the information that was placed in trends and bar graphs. It is, however, probable that with more practice they would have improved in the ability to use this information. Our claim is that ‘information richness’ of graphs is one of the most useful features of the IRD concept in the nuclear field.

5.7.2 Evaluation of the Fortum pilot

Despite the rapidity and ‘ad-hocness’ of the design process, the final prototype is surprisingly mature and well-structured. This finding suggests that this kind of agile approach is well suited for the development of overview displays for industrial purposes. The approach would further benefit if it could incorporate into itself more systematic mode of operations that are based on functional modelling of the target system.

Our results suggest that the Fortum IRD pilot displays have many useful features such as presentation of history information through trend graphs, use of Gestalt grouping principles in element clustering and information richness of graphs. These features make the displays pleasant looking, and they also help operators in the identification and diagnosing of failures if they have had enough time to practise them. However, the displays have also several features that make them poorly suited to their purpose. Inconsistencies in the presentation of information (e.g. variable and inconsistent direction of reading, font size and use of grouping principles) are one of the biggest problems. The lack of labelling and lack of exact numeric information was also considered problematic.

All in all, the Fortum IRD pilot is a nice set of displays that have shown to be useful in the detection, identification and diagnosing of failure states in the nuclear power process. Considering the rapidity and spontaneity of the design process it is a respectable achievement. On the other hand, the final prototype is also some kind of ‘bastard hybrid’ of IRD displays and traditional displays based on process and instrumentation diagrams. This hybrid has its benefits, and as our results suggest, it can function both as an overview display providing useful information of the overall state of the power process and as a supplementary display that helps operators to early detect failures and problems in the power process. However, we propose that for neither of these purposes it is the best solution. IRD displays are good for rapid detection of failure states but more traditional overview displays that are based on functional modelling of the target system are more suitable for providing an overview of the state of the power process.

Hence, our suggestion is that the overview displays should be based on process architecture, and IRD principles should not be applied in their design.

On the other hand, we suggest that there could be one dedicated IRD display specially designed for rapid detection of changes in the main process parameters.

In the design of this display all the IRD principles would be followed, and it could also take use of those elements of the concept (e.g. polar star diagram) that were not used in the design of the Fortum IRD pilot. Our hypothesis is that this kind of real and genuine IRD display may be useful in the rapid detection of failures even in the nuclear field – but of course more research is needed to show that this is the case.

6. The Ecological Interface Design Experiment (2005) – Qualitative Analysis of Operator