General discussion

The discussion section is mainly concentrated on the evaluation of the Fortum IRD pilot displays that are based on the IRD concept. In addition to that, since we also interviewed designers and participated in the design workshops during 2008, we will comment on the design process as well. We have divided the discussion section into two main sections: First we present comments on the IRD concept in general, and after that we overview and present critical comments on the Fortum IRD pilot (see Appendix L).

5.6.1 Evaluation of the IRD concept 5.6.1.1 Display normalization

A lot of comments were received regarding the graphical presentation of infor-mation. For example, normalization of information was critically commented by several operators. The main idea of the original IRD concept is that when the scaled trends and bar graphs are put along a line, even very small deviations are easily detected making possible the detection of failures before the alarm is

dis-played. Comparison of aligned mini-trends and bar graphs is made possible through display normalization. In general, the value of the display normalization was considered to be quite small, and some operators thought that the normal-ized graphs are nearly useless if the exact numerical values are not continuously present to complement visually presented graphs. The operators did not report that they would have used this property (i.e. comparison of aligned graphs), since in test scenarios other types of information could be used in the detection of the failure. Therefore, our test results do not provide much evidence about the usefulness of normalization. This is a big deficiency since display normalization is one of the most important properties of the IRD concept. It should be consid-ered if the strict placement of display elements along the horizontal/and vertical lines is reasonable on the expense of the plant architecture familiar to the opera-tors. However, the operators still thought that providing history information through trend graphs was positive and arrows showing the direction of change after the violation of an alarm limit were considered useful. It is possible that display normalization shows its usefulness and importance in other types of test scenarios. Different types of test scenarios are, thus, needed in which the conse-quences of failures emerge more slowly. However, since a quite broad and com-prehensive set of scenarios were chosen for our simulator tests, it can be at least concluded that there are quite many failure and problem states in the nuclear field in which display normalization is not needed and is not helpful in the iden-tification and stabilization of the failure.

5.6.1.2 Use of colour

The application of the Dull Screen principle is one of the key design features of the IRD concept. Concerning the chosen colours, the operators’ opinions differed quite much: Some operators thought that the Dull Screen with the grey background is a good choice, especially because in the future also the background of the workstation display screens will be grey; other operators, however, thought that the grey background is not suitable, for example because of the grey background other shades of grey are not distinct enough. The main aim in the use of ‘greyish’ Dull Screen principle is to improve the visibility of the alarm colours. The used alarm colours were mostly experienced as satisfactory even though the visibility of the yellow colour appear to be more sensitive to the quality of the display technology.

Our results, however, suggest that the chosen set of colours makes other state changes even more difficult to detect. While the visibility of alarm colours was good, the colour coding for state changes in different plant instrumentation (e.g.

pumps and valves) was experienced as inadequate. The operators considered that it is essential to perceive information of the resources in use, which is necessary in order to be able to create an overview of the situation. The green colour was considered most problematic, since it has been widely used and it has several meanings. This fact decreased its value in the elicitation of attention. The main

problem was that the colours green and dark grey are difficult to distinguish from the grey background when looking far away, and in this way some of the important changes or deviations from the normal state do not catch operators’

attention. Some operators also had difficulties in distinguishing the green colour from yellow. This was especially problematic in the case of pump symbols in which the coloured frame around changed the hue of the central part of the symbol, because of colour contrast. The operators also had problems in noticing by-passes and components that are not connected, since they were not able to distinguish different shades of grey from each other. To summarize, some positive characteristics emerged when following the Dull Screen principle, but also many comments provided suggestions that the principles concerning the use of colour have to be further developed.

5.6.1.3 Information richness

‘Information rich’ trends and graphs were considered useful, but due to the lack of practice the operators were not able to utilize all the information that was presented on them. It is possible that with practice they could better utilize these information clusters. But it seems to be that the symbols in these graphs are too small, and they are located too near to each other so that even after practise it may be difficult to identify the cluster and interpret its total meaning.

5.6.2 Fortum IRD pilot

The Fortum IRD pilot displays are aimed for an overview display for 100%

power and for rapid detection of failure states. It is, one of the central tenets that the IRD displays should help users to detect deviations from normal. Even though this is important, it is necessary that these kinds of displays should also serve other purposes. For example, they should be useful when diagnosing failures or trying to stabilize the system. In fact, IRD displays should also function as overview displays that help operators to maintain accurate situation awareness. However, it seems to be that in order to attain situation awareness and perform the operational tasks that follow from the rapid detection of failure states more detailed information is needed. It is not clear whether and to what degree the Fortum IRD pilot displays are suitable for these purposes. We claim that in the nuclear field overview displays should be based on process architecture and on the functional analysis of the target system, not only on a particular design principle. Therefore, the critical question is whether it is reasonable to sacrifice several screens for this purpose only if these displays are nearly useless at other plant states. The apparent answer according to operators is no if their content is fixed. The answer, however, would be different if it could be possible to change their content according to the plant state so that some parts

of the IRD displays change as the plant state is changing. The participating operators also thought that this kind of change should occur automatically.

Some operators critically commented the way the process is displayed on the Fortum IRD pilot, the fact that on the left-hand side of the diagram the information is read from left to right, on the right-hand side the direction is, however, partly reversed, and the information is read from right to left. Overall, it seemed to be that, in the long run, the operators will have no problems to get familiar to this characteristic, but in the beginning it may seem to be a complication.

The participating operators criticized that many of the alphanumerical characters, symbols and other graphical elements were too small in size so that they had problems to identify them from the distance. This is a real problem, since operators do not normally have a possibility to walk closer to see what is displayed on the screen. Neither have they any reason to move closer especially because the IRD displays are only for reading, not for operating. An additional problem was the inconsistency of element size: For example, the size of letters could vary from one part of the display to another.

Even though the Fortum IRD pilot displays look different from more traditional overview displays, in general, the operators had no prejudices against the new displays. The experienced operators were able to utilize Fortum IRD display despite of the fact that the presentation format differs from what they have used to. However, some inconsistencies between traditional overview displays and IRD displays disturbed them to some extent. One of the main nuisances was that the symbols of components that belong together are not located near to each other on the display but they are dispersed over the display.

Since the elements are lacking labels and other identifiers the operators had serious problems in the identification of displayed elements. Because of the short training time, the operators also had some problems to remember to what different symbols (i.e., dot, line, triangle and diamond) were referring to.

Even though some of the pipelines are presented on the display, they may be more confusing than they are helping since only fragments of them are displayed. A general hope was that pipelines that in the reality are different in size should be also presented with different-sized lines. The operators also had problems to understand the arrows located at the end of the pipelines.

The final version of the Fortum IRD displays looks quite different from the first one, since a lot of information has been added on the display during the design process. Even though the amount of information has increased, the operators still made suggestions of components they would like to see on the display. For example, some operators hoped that information of detached components could be seen on the display. Contrary to that, some operators also mentioned components information of which could be removed from the display (e.g. information of electrical systems).

One operator counted on how many trend graphs is needed if a complete presentation of all the necessary information is desired: Since the reactor operator is monitoring about a hundred parameters and the turbine operator is

monitoring about three hundred parameters, about four hundred trends should be displayed on the overview display if a complete presentation is required. Since the presentation of this amount of trend graphs is impossible, the amount has to be lowered, and only the most important trends should be displayed. Screening is, therefore, needed to select the most important variables. Our opinion is that this screening and selection should not be done in a spontaneous and ad hoc manner but should be based on careful functional modelling of the target system.

The design process based on the rapid prototyping methodology was shown to be useful. It may also be the only possible way to develop a testable prototype especially if there is a lack of time and resources. Some method based on the functional analysis of the NPP process would be preferable, but it may not be feasible in this kind of project.

All the stakeholders had an important role to play in the design process. Since the IFE designers knew the IRD concept, their task was to design the Loviisa pilot displays in such a way that they support the detection of failures. The Fortum designers participated in the implementation of the designed solution, and they also functioned as mediators between the IFE designers and the operator designers that participated in the design work. Since the three operator designers were experts of the nuclear domain, their function was to provide domain expertise in what information should be presented in the overview display and by which way it should be presented.

Many of the good and bad features of the Fortum IRD pilot were added during the design process, and they are not based on the IRD concept as such. For example, the Fortum IRD pilot displays are filled with a lot of process information which make them suitable as overview displays that support the acquisition of accurate level of situation awareness. Originally, the IRD displays were, however, aimed for rapid detection of failures, and their principal aim – according to the developers – was not to function as overview displays. Since a lot of information is displayed on the Fortum IRD pilot displays, they are also suitable for stabilization and diagnosing of failures which was not the original intent.

The final prototype has thus weaknesses that cannot be blamed on the IRD concept. For example, it is not caused by the IRD concept that the displays were considered a bit cluttered and confusing. As the designers said, these defects are mainly caused by the way the prototype was designed: It is characteristic to the rapid prototyping that the development process proceeds in a quite spontaneous and ad hoc manner. The designers also complained about the hastiness of the design process which may also have made the final prototype look a bit unfinished. In addition, due to occasional malfunction of the simulator, the operators had problems during the simulator runs that may affect their evaluations even though these problems cannot be blamed on Loviisa IRD pilot displays.