0710_WindUpButton
0710_WindUpButton
0710_WindUpButton
0710_WindUpButton
0710_WindUpButton

Basic Skills Needed to Combat Alarm Woes

Oct. 2, 2007
If we don’t have the foundation right, why are we surprised that operators are getting overwhelmed by alarms?
By Keith Larson, VP Content, Putman Media

For all the new wonders that the first distributed control systems made possible some 30 years ago, the proliferation of process alarms—and the resultant need for alarm management discipline—wasn’t one of them.

The move to distributed control systems made it easy (and inexpensive) to add multiple alarms to virtually every loop and measurement point, even as the operators’ panel boards, with their de facto virtue of effectively displaying process trends, were removed. The ongoing erosion of plant personnel skills means that many existing control loops remain ill-maintained, further exacerbating the alarm management problem.

Such is the soapbox from which Ian Nimmo, president of User Centered Design Services, spoke in his keynote address to Control’s AutomationXchange strategic sales summit in August. A longtime veteran of ICI, Honeywell and the Abnormal Situation Management consortium, Nimmo knows of what he speaks when he decries the decline in engineering discipline that has led to alarm proliferation. It’s a critical situation that he fears will only begin to be addressed when we’re hit with another disaster on the scale of Texas City or Bhopal.

Nimmo cites recent plant surveys showing that a full 40% of control loops in the average chemical plant or refinery are not operating correctly. Typically only one-third is okay, one-third is not operating properly and the other third is in manual. “That’s terrible!” he exclaimed. “If we don’t have the foundation right, how can we expect to do anything in advanced control, and why are we surprised that operators are getting overwhelmed by alarms? Processes aren’t in control and money is going out the door.”

Nimmo said that in the area of loop-tuning, for example, self-tuning algorithms have helped some, but not enough to make up for plants’ declining skill sets. “Once loop-tuning was done by engineers, then by operators and now most often by trained technicians,” he said. But to what end? A show of hands among the process automation professionals assembled indicated that “it remains a huge issue still.”

The situation is representative of an industry-wide slackening of engineering discipline and a pervasive neglect of the basics, Nimmo said. He cited plant after plant that has dramatically reduced the numbers of alarms their operators are presented with by first concentrating on getting the basics right. “Let’s look at this whole area of signal conditioning. Can we validate that sensors are working correctly? Are we filtering properly? Attention to basics resulted in many bad actors disappearing and reduction in alarm management problems,” he said.

Poor maintenance of basic process control systems is resulting in lots of variability, decreased quality and yield and increased operating costs. “And if you don’t believe this impacts plant reliability, you’re sadly mistaken,” he said.

“The performance of a basic closed-loop control system is often overlooked because people do not appreciate that its maintenance is an important contributor to plant performance and energy efficiency,” Nimmo added.

Very little attention is being applied to this area, and it’s a big loss because it’s one of the few things you can do that doesn’t require a shutdown.”

Just how big a carrot are we talking about here? How about a 5% to 15% reduction in energy costs, throughput increases on the order of 2% to 5%, yield improvements of 5% to 10% and reductions in quality problems of 25% to 50%?  All by going back to solve sensors problems, filter signals correctly and tune control loops.

“The returns are phenomenal, so why aren’t we doing it?” Nimmo continued. “We’re going after things that are so complex and unachievable when the basics are right in front of us.” Nimmo even related the story of one plant where alarms had been reduced the point that when they had even a single alarm, it was discussed at the Monday staff meeting.

And while Nimmo is bullish on the potential for more sophisticated tools such as dynamic alarming, he said, “We can’t do anything until we have a solid foundation to work from, or all the data will be clouded by bad actors and duplicate alarms.”

Nimmo’s concluding rallying cry: “It’s time to step up to the plate and show that we’re engineers!”

So, what’s stopping you?