- Are you sure we shouldn't trip the compressor?
- Who's in control?
- The whole is greater than the sum of the parts.
- Well, now we know.
Did you know...
A quick check of our database shows that the average crude head operator spends about 3.9% of their time checking email.
On-Site Plant Audit
Now Available
We will come to your location and evaluate the following variables:
- Alarms
- Displays
- Procedures
- Training
- Staffing
For more info, click here.
Control Room Design Manual
Our manual is now available via email. Click here to request a copy.
|
Dear Reader, We hope you will find our Summer newsletter to be informative and thought-provoking. Please feel free to contact us with any questions or comments you may have. |
| Considerable work is occurring in the
development of decision aids, particularly for early event
detection. The nature of these aids is such that there will be
some degree of uncertainty regarding the actions that should be
taken or the events that are implied.
Theoretical principles in situational
awareness and empirical research show that there is a need to
convey the degree of uncertainty in these probabilistic decision
aids. For example, the early event detector may show a reactor
runaway is about to occur, but a better aid also provides
operators with how soon and how sure the system is of the
assessment (i.e., how likely the event is to occur). While they used simulation of stock purchases, the implications for other systems is clear. The authors varied two factors (1) the specificity of the uncertainty (i.e., the certainty of the probability of the outcome) and (2) the presentation formats (e.g., numeric, encoded in symbols, verbal). Examples of the formats used in the study include the following:
By varying the specificity, they found that how they presented the information (numeric, linguistic, iconic) was far less important than what was being conveyed. They came to three conclusions: (1) level of specificity impacts decision performance, as more ambiguous (less specific) probabilities were almost always treated as identical despite numeric differences, (2) while all subjects had threshold probability level for making a decision, the threshold was higher with increased specificity, and (3) specificity interacted with formats, such that graphical formatting of specificity (use of symbols or icons) resulted in more conservative actions than did numeric representations. |
| Dynamic allocation of control functions
between humans and control systems is referred to as adaptive
automation. This type of automation has been shown to be
superior to fixed automation for some monitoring tasks. While
adaptive automation is superior to no automation, recent
research shows that it is best applied to low-level cognitive
tasks (Kaber, D.B., Wright, M.C., Prinzel, L.J., Clamann, M.P.
“Adaptive automation of Human-Machine System
Information-Processing Functions”, Human factors, Vol. 47, No.
4, Winter 2006, pp 730-741). In this study, adaptive automation was applied to information acquisition, information analysis, decision-making, and action implementation for an air traffic control task. The greatest gains came in automation of the action implementation (simple motor tasks). This is consistent with Beville’s anecdotal evidence that batch programs to carry out a sequence of motor tasks (e.g, putting distillation train to minimum heat and maximum cooling) can aid operators during high stress conditions. The research indicates that the actual decision to remove the heat is best made by the operator. While automation can provide great benefits to an organization in terms of safety as well as staffing, determination of what to automate and when is not always obvious. Beville can help to evaluate your current staffing and automation through our Job Sampling and Upset Response methodologies, incorporating the latest in human factors research to provide practical, real-world solutions. |
| Do you teach the whole task or parts of the
task? Do you purchase a simulator for your whole process or a
part of it? Recent research helps to answer parts of these
questions (Sohn, M., Douglass, M., Anderson, J.R.,
“Characteristics of Fluent Skills in a Complex, Dynamic
Problem-Solving Task”, Human Factors, Vol. 47, No. 4, Winter
2006, pp 741-752).
Tasks were decomposed into their different components; in this case, an identification task was decomposed into search, initiation, classification, and save sub-tasks. The authors found that component-level (or sub-task level) fluency is critical to achieve overall fluency. Also, practice showed differential improvements in performance, where practice on the cognitive components showed greater gains in overall fluency than did practice on the motor components. This research argues that the greatest performance gains come from part-task training of cognitive components. |
| In the stone ages of DCS design, text options
were few – full versus half size was about the only decision
required. But like may other issues in interface design, the
advances in technology have allowed greater flexibility, but
also more decisions, of which text size and style are only a
couple. Which is best? Under a grant from Microsoft, research
into font design and display parameters was conducted (Sheedy,
J.E., Subbaram, M.V., Zimmerman, A.B., and Hayes, J.R, “Text
Legibility and the Letter Superiority Effect”, Human factors,
Vol. 47, No. 4, Winter 2006, pp 797-815). The bottom line is that fonts should be at least nine pixels in height, in Verdana style with sub-pixel font smoothing. This is of course for viewing computer displays from a normal seated distance; a larger font would be needed if it were to be viewed from farther away. Like much research, some of the results were counter-intuitive. While two of the sans serif fonts had the best legibility (Verdana & Arial), another had the worst (Franklin), indicating that the presence of serifs (the little curls at the ends of the letters) are not a major contributor to legibility. Beville Engineering has helped numerous petrochemical companies to resolve these types of issues in the development of graphics standards. Our most recent effort has produced an electronic document for ease in maintenance and distribution. The font used? Why, Verdana of course! |
| Copyright © 2006 Beville Engineering, Inc. All Rights Reserved. |
voice: 937-434-1093
fax: 937-434-8809
web: http://www.beville.com