The Value of Alarm Management

In the photo you can see Ian Nimmo standing with a Polyester Fiber Bale dating back to the 80’s when this young engineer faced many uphill battles; some you all may relate to. Like most projects they come with unique challenges.

 

Ian Nimmo, president of User Centered Design Services had mastered PLCs, so programming robots and Automatic Guide Vehicles (AGV’s) was easy. He had in-depth knowledge with Real-Time Process Control Computing. While Ian had mastered proportional–integral–derivative (PID) control and Advanced Process Control (APC) he had to face some technical and political challenges as well as unforeseen encounters that were created after the automated solution was implemented.

 

“To transport and ship Polyester Fiber Bales involved gravity conveyors, which introduced blockages, breakdowns, and many variabilities requiring lots of manhandling. Customer shipping tickets were all man-made using a John Bull stamp kit and an operator reading the weigh scales. This was a perfect project for automation.  Using the Automatic Guide Vehicles, they maintained alignment for shrink-wrapping the bale and getting rid of ugly ears, which were a hazard as the edges were sharp. The new consistent speed allowed automatic labeling using a preformed printed ticket, adding the Date/Time/Consecutive number, and finally, weight read from a scale.”

 

Ian says, “The UK Government Weights & Measurers Department had an issue with traceability. After lengthy negotiations they were able to demonstrate ‘traceability’ by adding a tally role with the Date/Time/ Consecutive number and weight of the bale. The technical problems were somewhat easy to resolve. Looking back on how good the solutions were, I am proud of what we achieved, now we must look at how the automation impacted the operators. They were expected to oversee the automation and step in when the system could not correct the situation. An operator must determine what is wrong and fix the problem by taking manual control. Even today, many automated processes require operator oversight. It is a common goal to provide situation awareness to guide the operators as they monitor several processes. Over the years, we have provided operator tools to achieve this goal around the peripherals of our automated solutions. Still, in some ways,  we have failed to predict the impact of that solution on the shift workers. Additional workload, poor situation awareness, data overload, and human factors contributed to the additional workload.”

 

The sign of an exceptional alarm system is that it is a constant improvement process. For alarms to be an effective tool for the operator, the operator must be in a position to trust that the system itself is up to date. Alarm rationalization initiatives assess criteria and risk assessment for each alarm, and address alarm priorities and alarm floods to ensure that the alarm system is working for the operator, rather than against them. Ian says, “When we first started, we didn’t think through the solution. We thought the more information we could give the operator the better they would be able to detect, diagnose, and respond to an abnormal situation. More Information = Faster Operator Response. We never anticipated that we could overwhelm operators with too much information. Alarm management was not very well implemented.  Case in point, when we added the alarms, we did not do any alarm rationalization or “bad actor” resolution. Bad actors are nuisance alarms that often go unresolved and cause the alarm system confidence to fade quickly. We left it to the commissioning engineers to apply the alarm limits which were workable for the process but not useful for the operator.”

 

Ian says, “So here we are, over 30 years later and we continue to talk to operators that have major Alarm Management issues, mainly too many alarms.  We have had major disasters associated directly with bad alarm management and have the case studies that directly call out ‘alarm floods’ as a major contributor . The industry has responded by providing alarm management standards, and still, companies have not resolved these issues. We could ask, if we truly have good standards, why do we still have a problem?”

 

“Every year I see the same companies, and several new ones that have not followed the standards. I see inferior alarm philosophy documents, which no one ever follows. They do some alarm bad actor resolution, but these adjustments are temporary.  They never fix the problem for more than a few weeks. Many companies still have no idea how to group alarms into manageable amounts with proper priorities. They go back to the old school of: A high alarm as a high priority and a High-High alarm as a higher priority. Prioritized alarms should be based on a risk assessment and the consequence of a failed operator response.”

 

Ian says, “Most companies can’t pull this off because they are not convinced of the value of alarm management. They do not align upper management and lower management in a “Do It Right The First Time” methodology, so they are not prepared to invest the time and resources to treat it as a Mission Critical Project. For example, the operator has a High-Level alarm and a High-High-Level alarm. Normal operating practice, the culture today, says it is okay to allow operators to ignore the High Alarm and wait to act until the High High Alarm is activated. This type of alarm usability promotes using alarms for information instead of action. When you use alarms for information, the system fails. When you have an abnormal event that requires a quick response, operators should not be scrolling through an alarm flood to find all the action-required alarms buried in the non-action information alarms. Information alarms will slow the response time and destroy situational awareness. I have witnessed companies that have gone through rationalization and have failed to implement the new solution due to a lack of resources. The biggest issue is the absence of ownership by the operations department and leadership, which should be aligned with the fact that the Alarm System is a Mission Critical Tool that must be effective “all the time” and NOT “some of the time.“

 

Ian continues, “With any automation project, technical issues will exist but when it comes to operator oversight and the alarm system, management commitment is the more significant issue. The standards are very well written and will provide a healthy system if followed. It is crazy to think that we are still talking about alarm management today after so many standards, guidelines, white papers, and presentations on the subject.”

 

“While it is easy to point the finger at the Alarm Management System, I can tell you an overloaded Human Machine Interface will have a major impact on the number of alarms assigned to an operator. If you squeeze lots of information on a screen and provide no operator task modeling into the design, or prioritization of that information into the display, operators will require more information. It is easy to get that information by simply adding an alarm. We must stop this way of thinking and start visualizing information on the displays. Only using alarms when time is a factor and the consequence of not managing the situation has a Negative economic, environmental, safety, or financial impact and a physical action is required by the operator. Again, the industry responded by providing new Display Standards. Similarly, the standards are not being followed. Very few places have developed Level 1 overview screens. Even fewer companies provide Level 2 screens. The vast majority of control screens are just Level 3 (P&ID) screens with no attempt to put data into context. Another subject for another time, perhaps.  The bottom line here is: we need to put as much effort into operator support systems as we have into the automation solution.”