Results concerning operators’ thoughts on IRD key features and Fortum pilot

The following comments are based on training sessions and discussions that were conducted after each simulator run and on the final discussion that was

conducted at the end of the test. Results of all the three crews are considered.

The operators could quite well describe what has happened in each of the simulator runs, and they quite well remembered from what sources they had detected the failure. The list of questions that were discussed in the debriefing is presented in Appendix K.

5.4.3.1 Training session: Operators’ first comments on Fortum IRD pilot The first comments concerning the central principles of IRD were somewhat reserved. One of the operators was strongly of the opinion that the normalisation principle would not work in nuclear power plants. This is due to the large number of important parameters that the NPP operators have to follow all the time. Also, there are a lot of continuously changing parameters (e.g. primary loop pressure) with many levels that all have different meaning, which makes it necessary to see the real signal all the time instead of a normalised signal. The same operator said that the LSD does not give enough information because of the lack of exact parameter values. The other operators were not as strictly towards the pilot display, but also they said that there should be more parameter values and component labels visible.

A common opinion was that the colouring is too grey, i.e. green colour is not distinguishable from the grey background, which makes it difficult to notice state changes. This is especially relevant in the case of pumps and bypasses.

Initially one of the operators said that black background would be better than grey because the operators are used to it, but later the same operator commented that the grey colour is good since it does not irritate eyes. Many operators were hoping that there would be generally more contrast and that unacknowledged alarms would be presented with flicker.

During the training session, the overall layout of the display was commented mainly positively, although at first the display looked unfamiliar. One of the operators first asked: “Is this representing our process?” Later he said that the layout resembled a lot the layout used in the current PMS-displays that were developed in 1990’s. The first impression of another operator was that it is fine that there is the same amount of space for the both sides of the plant. The lack of labels and texts made it difficult to grasp the display at first, but after seeing briefly a version of the display with labels attached and after live demonstrations, the operators said that it is quite easy to learn to know which component is which and to get used to the display.

Some ways of presenting information were not immediately clear for the operators. For example, the small symbols used in bar and trend diagrams for representing pressure, temperature, flow, and level were a bit confusing and hard to remember. One operator commented that a triangle would be an intuitive symbol for flow, but instead a square is used for flow and a triangle for temperature. Because of the density of information, the operators had difficulties in interpreting the trend diagrams containing information of valve positions.

Also, the “steam log” and “feed water log” were difficult to distinguish and their presentation could be improved with better placement and thicker lines.

The necessity of some forms of information, such as the bar diagrams around the pump symbols showing partial production, the electric board section, the bubbler, and the many pre-heater (RH) tanks, was doubted. Instead, some other forms of information that nowadays are presented on wall panels, e.g. regulation deviations, were considered very important to be shown on the large screen display. Also, some symbols were thought to be in wrong places or missing, such as the generator and transformer breakers.

5.4.3.2 Overall view of the IRD concept

Overall, operators utilized the information shown on the LSDs, and they thought that the IRD displays are useful in the detection of initial events and the first signs of failures. They also commented that the trends and bar graphs aggregate infor-mation in useful formats. Some of the positive comments are listed in Appendix L.

Some of the negative features of the Loviisa IRD pilot are listed in Appendix M.

Operators thought that LSDs are typically not continuously monitored but they are only rapidly gazed now and then. The LSDs provide the first signal, and the rest is done by using the information from other sources. For example, the stabilization of the system and diagnosing of the failure are mainly carried out by using information of desktop screens. At these later stages the IRD displays should provide information that help operators to maintain the overview of the power process and know how the things are going on in the process.

5.4.3.3 Operators’ thoughts on the key design features of the IRD-concept Display normalization

Normalization was considered to be a useful feature: operators can immediately detect if all the values are in a particular area or not. On the negative side, because of normalization information of the actual parameter values is lost, and the contact to the physical system becomes more difficult. Since the operators need this exact information in some cases, they have to learn the limits that are shown on the graph. Some operators had also some doubts about whether it is reasonable to sacrifice such a large part of the display area just to satisfy this particular aim. In addition, at lower power levels the normalization is not very useful since the graph lines do not easily stay aligned. To prevent this, it is important that exact numeric information is presented at the location of the trends and bar graphs.

Overall, presentation of history information was thought to be especially valuable. Normalized symbols without trends were also thought to be useful but the operators still considered them less valuable than trends due to the lack of history information. The time scale of the trends is perhaps too short, and

because of the short time-scale operators cannot utilize the trend information during long runs. It was asked whether the update scale could be different in different trends. However, a deeper look of this suggestion showed that it may be misleading if different trends have different update scales.

Grouping of display elements

On the IRD displays, because of element grouping, the deviations can be more easily detected based on the ‘good continuation’ grouping principle. On the negative side, since graphical elements are gathered from different parts of the display into a single place, information that is related and should be presented at the same place is sometimes scattered over the display and presented far from its real context (e.g. RA-pressure, level of pressurizer, neutron power). Therefore, it was found in the test that information that was needed was not always immediately detected, but operators had to search for the critical information for a while.

Use of colour in display design

Since the simulator was quite dark, and some contrasts between a target and the background were large, the IRD displays were quite bright. The operators thought that, because of these reasons, it would be impossible to watch the IRD screen for a long period of time, especially because they are normally sitting quite near the display screen.

The alarm colours (red and yellow) were quite well detected from the display.

Especially the red colour could be immediately noticed, and overall, operators thought that the alarm colours are well chosen. One operator, however, found out that he had some problems to notice the change of the colour from grey to yellow. The operators also thought that different states of pumps and breakers should be presented in a more conspicuous way. Neither were the colours of the surrounding frame of the valve and pump symbols seen very clearly. Because of colour contrast, the colour of the centre part of these symbols may also change a little bit, and so the meaning of the symbol may remain unclear. One solution to this is that the surrounding frame is broader, and the whole symbol is a little bit larger.

The change from one grey level to another is used to inform that a particular component is by-passed, is not connected or is out of function. Since the differences between different grey levels are quite small, these changes very easily remained unnoticed. Overall, by-passing and components that are out of function should be better presented.

The operators also complained that the logic behind the use of alarm colours was somewhat confusing. Especially, the use and function of the colour yellow remained unclear to them. Overall, the discussion of the logic of the use of alarms and alarm colours formed a big part of the debriefing discussions.

Information richness

Due to the lack or insufficiency of practice, the operators were not able to utilize all the information that was presented on mini trends and normalized symbols without trends. A common finding was that they identified a particular deviation (the change of the level of a bar), but they did not necessarily understand what its meaning was. To be able to utilize this information, more practice is, therefore, needed.

The blue dot indicating whether the device is in manual or automatic mode was considered useful but it was somewhat difficult to notice. It was discussed whether the shift supervisor is able to see it at his/her seat, and it was suggested that its size should be increased. The meaning of the horizontal dash lines that were placed in the trend graphs remained unclear to operators, and they were not able to use this information.

The arrow showing that the level was rising or falling was useful, but it was suggested that this property could be further developed: If the level is rising rapidly two arrows could be shown, and if the level is rising too rapidly, three arrows could be presented.

5.4.3.4 Presentation of information on Fortum IRD pilot displays

On the left-hand side of the display the information is read from left to right, but on the right-hand side the direction is however partly reversed, and the information is read from right to left from the condenser towards the steam generators. Some operators considered this especially problematic, since trend graphs are read from left to right even on the right-hand side of the display. It would be important that information should be presented consistently in such a way that operators need not change the direction of reading.

A lot of discussions were spinning around the question of the placement of symbols. Due to lack of space, some components are not located where they should be, and because of this, some operators sometimes had to actively search for information that was needed. For example, since the level of pressurizer is displayed on the reactor operator’s side, the turbine operator has to gaze quite far away from his location. Therefore, the turbine operators hoped that this information could be shown ‘on their side of the display’. Also, the location of the generator should be changed: it would be better if it is located to the right of the turbine along the same line. In addition, it could be better if the intermediate cooling pumps (VG-pumps) are presented where all the other pumps of the auxiliary system are shown.

Many of the pipelines and clusters of components should be reorganized. For example, all information that is related to the main steam piping (RA) or to the feed-water system (RL) should be grouped together. It was also complained that some components play a too dominant role in the display. For example, pre-heaters could be combined and presented as a single group.

Some essential information was thought to be lacking. For example, because of lack of space many important valves and pumps are not presented on the Fortum pilot display. Information of their state was, however, shown to be critical in some of the scenarios. There was also a longing for flicker information, and operators did not think that flicker is disruptive in this kind of displays. For example, it was suggested that flicker should be present until the alarm has been acknowledged. Another possibility is that the flicker is automatically stopped after a short period of time, and after that the component is still presented by using the alarm colour.

The scarcity of numerical information was complained. Even though some operators considered that the presentation of all the numerical data may make the display look cluttered, it was hoped that some numerical values could be presented at the bar graphs and at the trends. There was no general agreement on the presentation of labels and identifiers: Some operators thought that the presentation of labels makes the display look confusing, some other operators thought that labels and other identifiers are needed. Some operators, for example, complained the lack of labels at the place where the symbols and other graphical elements are located. Because of the lack of identifiers, the operators had some problems in the identification of displayed elements. Therefore, it would be useful to show the labels as long as the operators cannot identify the components without them.

On the other hand, the operators also thought that some information is useless, and could be removed. For example, there was a lively discussion of whether the electrical systems (e.g. transformer circuit-breaker, generator breaker) should be presented on the IRD display or not. Some operators thought that, except diesels, all the other components of the electrical system could be removed. Some operators hoped that all information concerning the electrical systems could be put into a separate display, whereas some operators thought that at least their grouping and arrangement should be considered more carefully.

The operators had different views on the presentation of pipelines. Some operators thought that a larger number of pipelines should be displayed. It was also complained that the pipelines are not presented very clearly, and it is somewhat difficult to see them. Some operators, however, doubted whether it is even important to see them clearly on an overview display. Arrows showing that a pipeline will continue after a break were somewhat misleading, since the arrow is typically indicating the direction of flow. The operators hoped that some of these arrows should be replaced. On the other hand, the arrows may be useful at the right-hand side of the display where the direction of reading is unconventional (i.e., from right to left). The operators also thought that the thickness of the pipelines should depend on the importance of the pipeline so that the main pipelines should be presented with a thicker line that the less important ones.

A general problem was that many of the symbols and other graphical elements were too small in size so that they are difficult to see from distance. Since the

amount of displayed information has drastically increased during the design process, the lack of space has become one of the main challenges, and many of the graphical elements and alphanumeric characters are located too near to each other. This fact also makes the information difficult to see from distance and makes the display look a bit confusing.

5.4.3.5 Development of the Fortum pilot

In the first version of the Fortum pilot only analogue measurements were shown, and only main circulation pumps and feed-water pumps were displayed. All the other information (e.g. feed-water and steam lines, valves and pumps) has been later added. A general trend was to add more information and present new valves and pumps, since operators thought it is important to present all the essential information and compensate the information shown on the panels and desks of the conventional CR. A critical question is whether this is a right way to proceed since when the amount of components displayed increases, problems with the lack of space emerges.

5.4.3.6 Usage practice of IRD displays

The operators thought the LSDs should have a general overview function: They should not only help operators to detect a failure just at the moment it is occurring but help them to get an overall view of the power process whenever it is needed. For example, they could be very useful in the shift change: The shift that is coming in to work in their shift can immediately see that everything is working well, if the colour red and yellow cannot be seen on the LSD.

It was asked whether IRD displays can support collaboration and co-operation between operators. The operators thought that since the IRD displays can provide information of what other operators are doing, they also can improve the collaboration between operators. For example, it was thought that when all the operators see the same information on a LSD, there is less need to ask other operators for help.

The operators thought that there should be no problems to use the IRD displays together with workstation screen displays because the concepts are so different that any kind of confusions are not possible. Apparently, the operators have learned to live with different types of information presentation concepts (process computer displays vs. panels and desks). It was, however, considered possible that LSDs may even disturb operators and cause trouble if the operators think that they should look at the LSD every now and then.

The IRD displays are now only aimed for watching, but they could be more interactive. A cursor could be presented on the display, and additional information could be displayed by a mouse click. For example, a code of a particular component could be presented by clicking the component with the mouse. The logic that is behind a traffic light could also be shown by this way.

In general, windowing was considered an important target of development. It may be especially useful when operators are learning to use the IRD displays.

It was also suggested that the IRD displays could help operators in the navigation through display pages. Since it has been found that operators spent a lot of time in searching for the right display, it would be nice if they could select the right display from the large screen. That is, after clicking a particular component by a mouse on the LSD, an operating window could be displayed on the workstation screen. This kind of use of the LSDs may be especially useful if the operators have to rapidly operate a particular component.

5.4.3.7 Function of IRD displays

In general, the operators thought that the IRD display should provide an

In general, the operators thought that the IRD display should provide an