Do I really need this alarm?
Have you ever reviewed the list of standing alarms in your plant? Do you see alarms that always seem to be cycling? Are there multiple alarms that have indicated an alarm state for an extended period of time? Have operators silenced or disabled the alarms that are constantly going off? I have found numerous power plants that have forced off or silenced alarms due to the repeated activation just so they can put more focus on the 300 other active alarms during normal plant operations. A review of the standing alarms often leaves you wondering which alarms the operator notices when overwhelmed with so many. So how do you make sure you have the right alarms in place?
A periodic review of your alarm philosophy should be conducted (there are several third party software packages available to analyze your alarms and provide very useful data), or your plant maintenance employees can collect and analyze the alarm logs. There are also several systems integrators that can help with this process by utilizing their experience to help formulate solutions on a case-by-case basis. As your controls age and equipment is replaced, your alarming requirements may change and maintenance of your control system code needs to be performed.
A common occurrence is an alarm being driven by code that is no longer being utilized. For example replacing an old flue gas draft system that utilizes fixed speed fans with a new system that uses variable speed fans on a boiler. You are able to repurpose some of the I/O for starting and stopping the fans, but the main portion of the code that controls the fans is very different. Instead of completely removing the old code, the I/O is disconnected and the old code continues executing in the background. So any alarm that was active in that portion of the code remains active unless the code is deleted. I have seen these alarms show up at the bottom of alarm screens with dates going back several months or years depending on the last time the controls were cleared. My personal favorites are the diagnostic alarms that go back to when the plant was commissioned that no one took the time to address.
Another occurrence is the alarm that is always sounding. These are usually ignored by the operator due to the high frequency that it occurs. For the sanity of the operator, you should review how often an alarm occurs during a specified period and modify the alarm criteria to reduce the frequency to only sound for a true alarm condition. An example of this would be a high or low drum level alarm on a boiler. The alarm point is based on a fixed level and the turbulence in the drum can cause the alarm to activate repeatedly when operating close to the alarm point. This can be cleared up by adding a small deadband that is slightly greater than the noise or a filter with a time constant of a couple seconds to dampen the signal. I have found some alarms that have been triggered more than ten thousand times in twenty-four hour period due to operation right at the threshold of an alarm point and the noise would trigger the alarm. Another is a level switch alarm that also causes a sump pump to operate. The pump operates based on level switches, which also triggers the alarms, so during normal operation the alarm sounds every time the pump starts and stops. This alarm can be conditioned so that it only alarms if the pump does not start/stop depending on the required pump state.
According to ISA 18.2 standard, the definition of an alarm is “An audible or visible means of indicating to the operator an equipment malfunction, process deviation or abnormal condition requiring a response”. Without an alarm management guideline, commitment from management to implement alarm management and to maintain the alarm system, the control system will be an obstacle for the control room operator to prevent undesirable consequences. ISA 18.2 standard was established in 2009 in recognition that ineffective alarm systems have often been cited as contributing factors in incident investigations after major incidents. The issue that operating plants have is that alarms start to be ignored because the operators are accustomed to certain alarms always being active, so the alarm state becomes the normal. Usually, alarm conditions can be conditioned to activate for their desired reason, but not so frequently that they become a nuisance. It just takes time to communicate with the operators to find out what alarms are a nuisance and evaluate each alarm on a case by case basis to properly manage them. Where do you begin? For a large plant reviewing the entire scope of alarms may seem like a daunting task, but you can break it apart by systems and slowly work through the issues. Internally or externally there are the costs of operator’s time, engineers, maintenance personnel, safety personnel, or subject matter experts. The final solution is up to plant management to commit to changing the culture that alarm management shall be a priority. Your operator will thank you for it.
This post was written by Mike Robb. Mike is a senior engineer at MAVERICK Technologies, a leading automation solutions provider offering industrial automation, strategic manufacturing, and enterprise integration services for the process industries. MAVERICK delivers expertise and consulting in a wide variety of areas including industrial automation controls, distributed control systems, manufacturing execution systems, operational strategy, business process optimization and more.
MAVERICK Technologies is a CSIA member as of 3/20/2015