Do your console operators have a hard time identifying alarms? This report explains why.
Dr. Nichol Lee and Dr. John Casali from Virginia Tech University evaluated current alarm sounds use in control rooms. Sound measurements were taken at two refinery control rooms using different distributed control systems. The results indicated that some of the alarm sounds could be difficult to detect for those with age induced hearing loss. A follow-on was determined to be useful for creating a means to select optimal alarm sounds.
At this pilot stage (i.e. phase 1), the VT-ASL team would visit 2 refineries in Corpus Christi, TX, and endeavor to experience (auditorially), measure (acoustically), and capture (digital recordings) all alarms that are deployed in the control rooms. This will enable both the acoustical spectral measurements, likely in one-third octave bands (or more narrow bands if tonal), for each alarm, as well as high-fidelity digital recordings, to be obtained. The researchers will also endeavor to experience and understand the nature of the control room environment and operations therein; for instance, the presence of multiple consoles with overlapping alarms, the directionality of the acoustical environment, the reverberance of the environment, barriers that block or divert sound, and any potential for masking of alarms by noise or even speech communications or coded alarms that may be broadcast over a PA or annunciator system. If possible, the team will also do a cursory interview of a few operators to gain their perspective on the alarms, and any issues associated with their perception and use of the current alarms. This operator interview effort will be helpful in formulating plans for a more complete Task Analysis, as noted above, which would then be conducted in a follow-on project to this pilot effort, if deemed appropriate.
What are the optimal sound characteristics for console alarms in multi-console control rooms?
Distributed control systems (DCS) use audible tones to alert the operator to the actuation of alarm(s) and convey the urgency needed in response (alarm priority). Typically three separate tones are utilized that can be selected from a variety of options. For control rooms where there are multiple consoles, different tones are used on adjacent consoles so that operators are not confused or distracted by alarm actuation on a nearby console. Selection of the “different” tone is done on a subjective basis.
Modern DCS use wave files to create alarm tones. Research in audible displays should be able to provide guidance on the optimal set of tonal characteristics that can be used for alerting operators. However, there is a gap between researchers knowledgeable on audible displays and individuals knowledgeable on process control environments.
After the site visit, researchers will develop a brief report that entails a list of alarms and antecedent events which elicit their deployment, and operator perspectives of the current alarm situation which are obtained via interviews. The major focus of the outcomes of this pilot (Phase 1) effort will be development of a research plan for follow-on research (Phase 2), which will include task analysis and in-lab experimentation to understand the strengths and weaknesses of the current system. Findings from the Phase 2 research will determine the need for a Phase 3 research effort, where either modification or redesign of the alarm system along with a validation experiment will occur. The final deliverable from a Phase 3 effort would be a guidance document for refinery control room auditory alarm design and selection.
To recapitulate, the final report for this pilot project will include: 1) the results of the control room visits and measurements, including a list of alarms, antecedent events that deploy alarms, and operator perspectives obtained via interviews; 2) a description of the recommendations for follow-on research for Phase 2 as described above, including a research plan outline for that Phase, based on Phase 1.