These are alarms that initially come in, but for some reason the operator cannot remove them. They are usually classed as stale if they have been on the alarm page continuously for more than 24 hours. This could be for one of a number of reasons.
This is probably the most common standing alarm and is caused by an instrument not reading property. The instrument may have failed completely, giving no reading at all (generating a bad value alarm), it may just give a false value. I wrote a previous post on ways control systems can lie to you. In this case the operator will not be able to clear the alarm. It doesn't matter what the operator does to the process, if the instrument isn't reading the true process condition, it will stay in alarm.
The operator should ensure maintenance are aware and get the instrument fixed while using any other instruments available to infer the true process condition. If the true process condition can't be inferred, then it may be necessary to shutdown the plant until the original instrument has been fixed.
Not Always Relevant
There are some cases where you only need the alarm to be active under certain circumstances.
For example, an alarm to tell the operator that the pump has stopped. Except it is part of a duty/standby pair, and one pump will always be offline. As a result, there will always be a pump that is off, and so always have one alarm active.
Another case is that you may only need an alarm when a specific procedure is getting carried out. Examples of this include a startup condition or when regenerating a drier or catalyst.
The solution to this is to all logic that can be more clever with the alarm.
- Check if the partner pump was is in service.
- Check control valve positions to determine if a procedure is being carried out or the equipment is in service.
Alarms used for Alerts
I partially covered this in yesterday's post. An alert is a method of drawing the operators attention to something that has occured. However, unlike an alarm, it does not require a response on the part of the operator. If there is no response required, the alarm will probably just stay in place.
As I said yesterday, it is often easier to setup an alarm than it is to generate an alert.
What is the Problem?
So why do we care about stale alarms? Why have I spent a significant amount of time trying to get rid of them?
It should be pretty obvious why the faulty instrument alarms are bad. If the instrument is faulty, you are operating slightly blind. Like having a mark on your car windscreen. One or two marks are annoying but probably not too much of a concern. There is plenty of other information on the plant that you should be able to know what state the plant is in. As more and more instruments fail, the metaphorical windscreen gets more and more marks on it until you can't see where you are going. The plant management need to determine how comfortable they are with the current number of faulty instruments and if they need to bring forward planned maintenance to repair them.
The other types of stale alarms, the ones that are not always relevant or that are used as alerts, are also a problem. Even though they are no longer bothering the operator after they hit silence, the remain on the alarm summary page. Remember that the definition of an alarm is that it requires a timely response by the operator. If alarms are always on the screen, the operator gets used to seeing them. If the alarm is genuinely triggered at the same time as a different alarm, there is a risk of missing the familiar one and them when they actually need to do something about it.
I have also seen that once an alarm got fixed, the operators were so used to seeing it that they just ignored the alarm when it reappeared the following month. Time needs to be set aside to ensure that alarms are kept in a good state. The only alarms that should be on the alarm summary are ones that require a response.
It should be the operator's, not maintenance department's todo list.Go Top