If you type “alarm management” in a search engine, you will get hundreds of results including articles, papers, presentations and guides.
Those results may contain some useful information, but in many cases, they deal with the complete process of alarm rationalization.
The problem is that not everyone will dedicate the time and resources needed to implement an exhaustive alarm rationalization process that includes a comprehensive audit and the development of an alarm management philosophy document. This blog presents an easy and concrete method for quickly reducing the number of alarms received.
Making your bad actor list
In most cases, we can apply the Pareto principle to the number of alarms received, affirming that 20% of your alarms will generate 80% of your incoming alarms. This means that you could significantly improve your alarm management performance by focusing your efforts on one fifth of your alarms. These are your bad actors, and your first step consists in identifying them.
Some advanced alarm management software applications easily provide a bad actor list, but they are not necessarily required for this step since a simple identification can be done with your human-machine interface (HMI), Historian or Excel. The bad actors are simply the alarms that were generated the most during a given period, so you need only count the number of times each alarm became active. Once the bad actor list is created, you can move on to the analysis and solution-finding phase.
Analysis and solutions
For this step, each bad actor is reviewed to determine what action should be taken to reduce the number of instances that the alarm is triggered. Since a number of solutions are possible, we propose the following:
- Can the alarm be removed? To determine whether an alarm is necessary, ask yourself the following: What action is taken by the operator when the alarm is activated? Usually, if there is no action associated with the alarm, it can be removed, which will solve the problem.
- Is the equipment faulty? It is possible that a faulty instrument is causing multiple alarm events. If that is the case, contact maintenance about the issue and consider placing the equipment out of service or “shelving” the alarm to prevent it from being reactivated for a select period of time. Disabling the alarm in this type of situation is never recommended as it is easy to forget to re-enable it once the problem is resolved.
- Is the alarm threshold properly set? For analog signals, the alarm set point may not have been properly configured or may be set too close to the normal operating conditions. You should therefore review the alarm set point to see if it can be changed to reduce the number of alarm occurrences.
- Can I use an OFF delay? This will keep the alarm active for the duration of the OFF delay after the alarm conditions are no longer present. By keeping the alarms active, you prevent the same alarm from chattering. The OFF delay is usually preferred because it does not delay the activation of the alarm, and there is usually no risk with keeping the alarm active longer.
- Can an ON delay be used? An ON delay will raise the alarm only when the alarm conditions are continuously present for longer than the chosen delay. This can prevent fleeting alarms from being raised if their duration is shorter than the ON delay. The ON delay should not be used for a safety alarm. We must also keep in mind that the alarm will be raised only after a delay. That being said, a short delay usually does not have a negative impact on the process, but it can greatly reduce the number of alarms.
- Can we use conditioning alarms? An alarm can often be raised for no reason, such as a low-flow alarm or a pump that was turned off. You can avoid this by using conditioning alarms, where an alarm is enabled only in certain conditions, for instance, while the pump is running. This can remove many alarms that are generated during start-up and end-of-cycle sequences.
As you can see, alarm management does not have to be complicated, and significant gains can be easily achieved with the proposed method. In many cases, managing bad actors is a good starting point. It quickly produces excellent results that will motivate the team to continue its efforts and consider a more comprehensive alarm rationalization process.
MAXYM LACHANCE, P.Eng., is Engineer, Optimization and Advanced Control