MONITORING INSTRUMENTS; DCS VERSUS ANALOG GAUGES

NEWSLETTER ARTICLE

It is a known fact that for holding a setpoint, distributed control systems are far superior to pneumatic control systems. In the words of one process manager, once a setpoint is selected "DCS will take you down the road" and requires few further adjustments, whereas the pneumatic control system will demand continued tweaking. Recognizing the control advantages of the DCS, shouldn't this translate to DCS operators spending less time monitoring/interacting with their control system? Some recent data suggest perhaps not.

Using data from Beville's job sampling database, a comparison was recently made between operators using pneumatic control systems and operators using distributed control systems in relation to the percentage of total time operators from each group spend monitoring/adjusting their instruments. The data for the comparison consisted of 16 four-hour job samples of pneumatic control system operators and 20 four-hour job samples of DCS operators, all taken during normal steady-state operation. The data suggest that operators using DCS spent roughly twice as much time monitoring/adjusting instruments as do their pneumatic control system counterparts. The apparent difference was found to be significant to the .05 level of confidence.

The results seem to be paradoxical; operators who used the better control system spent roughly twice as much time monitoring/adjusting their instruments. A number of plausible explanations for the data exist.

One possible explanation is due to the nature of DCS. DCS operators have only a limited view of the process, because the number of instruments that will fit on a screen is limited. For large process units with considerable instrumentation, operators must page among several displays in order to make needed comparisons between instruments. The additional time, then, is due to searching through the displays rather than reading and interpreting the instruments. Under this scenario, the DCS would, in effect, limit the rate at which the operator could assimilate process information.

Another reason which has been suggested is the newness phenomenon; because DCS's are new, operators lack a certain amount of confidence in the system and hence, monitor them more frequently. However, the DCS's in the study had been in service for at least two years, certainly enough time for such an effect to have diminished.

A third probable cause for at least a portion of the difference is likely due to the number of alarm occurrences. During normal steady state operation, processes on DCS control have nearly four times as many alarms per hour as those on pneumatic control. However, the time required to respond to an alarm is relatively short. Three additional alarms per hour would likely take an additional 2% of the operators time. The additional alarms do account for some of the difference, but by no means for all of it.

If the data are representative of the industry and for some reason it does take operators twice as long to assimilate DCS information, then it would follow that process upsets should tend to magnify the effect. To overcome possible upset response degradation, the factors which cause the effect should be identified and factored into the design. The study highlights the importance of designing DCS display systems that are efficient during both steady-state and off-normal operation. For graphic displays, rather than simply translating P&IDS directly into DCS displays, the displays should be constructed based on a hierarchy of functions which is determined through a task analysis.

Copyright 1992 Beville Operator Performance Specialists, Inc., All Rights Reserved