HMIs: Industry Must Do Better

How many times have you been in this situation: The fire alarm goes off in a public building, and nobody makes a move to leave? We're conditioned to assume that it's a false alarm, so we're slow to react, hoping that we don't really have to stop what we're doing and head outside.

This was the scene at a presentation this morning at the Yokogawa Users Group Conference in New Orleans, and it was the perfect prop for Ian Nimmo as he tried to finish up his presentation about all the things that are wrong with process human-machine interfaces (HMIs) these days. High on his list was how heavily HMIs rely on alarms, making the operator system reactive rather than proactive.

But as demonstrated so aptly by the audience's reluctance to get to its collective feet as the fire alarm went off, alarms are often ignored, as people assume they're just nuisance alarms. "Some common things you'll see is their alarm system overwhelmed the operator and was totally useless for prioritizing," Nimmo said. "The HMI was ineffective and didn't give the operator a good situational awareness."

Nimmo, president and founder of User Centered Design Services, continually came back to the point of situational awareness as the main goal of an HMI. Recent major accidents in the process industry—such as at BP's Texas City refinery or Texaco's Pembroke facility—were caused by a lack of situational awareness, Nimmo said. "How much can a true high-performance HMI impact your plant?" Nimmo asked. "If you don't have it, we've seen the result in these major accidents."

A Tale of Two Operators

Nimmo presented two types of operators as examples of how the process industry has gotten away from good operator interfaces rather than improving. The first operator, an old hand preparing to retire, is the go-to guy when information is needed. "He started back in the 1950s," Nimmo described. "His automation was made up of pneumatics and a physical hard panel. He started on one plant and learned everything about that plant and became the expert." Above all, he had good situational awareness, with a good plan in his head about how the whole system works.

The other (younger) operator came to the position with a similar education and work background, but his knowledge of the system and how he reacts to abnormal situations isn't nearly as good. That's not because the younger generation has a poor work ethic, Niimo said. Instead, his poorer performance is because of the tools he's been given.

Operators who interact with the plant primarily by reacting to alarms is a poor one, and is a major problem with how today's HMIs work, Nimmo said. "If you have reactive operators, we have an issue. I want to see proactive operators. The guy on the right [the younger operator] has been brought up to believe that the automation system is doing its job, and he sits around until an alarm goes off, then he does his job. Not good."

Nimmo likened the situation to what it would be like if an airplane flew by alarms. He presented a picture in which the pilot reacts when the alarm tells him he's flown too high, then reacts again when it tells him he's flown too low, and so on. "It wouldn't be very comfortable, would it?"

Of course, the cost of all these alarms in process automation isn't just a loss of comfort. "We see a big impact because we allow our processors to run by alarms," Nimmo said. "It's not just the lost production. Equipment reliability is reduced. If we want reliable equipment, we have to be running with no alarms."

A Better Way

On the positive side, a true high-performance HMI can also have an incredible impact your plant. Operators who are able to catch a problem quickly—before there's an alarm—and bring it back to normal can keep equipment running much more reliably. "The longer you stay in abnormal, the more it costs you, and the more it impacts the reliability of your equipment and the quality of your product," Nimmo said.

A better HMI, with fewer alarms, will allow the operator to be more proactive. Nimmo contends that the industry can do much better in this area. He made several recommendations to improve HMIs:

• Improve color schemes, including removing blue type on black backgrounds ("How ever did we do mistakes like that?") and avoiding colors like red and green, which are difficult for the color-impaired.
• Return more trending information to the HMI, giving operators a look into not just what's going on now, but also the past and the future. "I need to know that past, present and future for good situational awareness," Nimmo said. "A trend was very good at doing that."
• Eliminate the pages and pages of alarms.
• Improve navigation, letting operators see more on the first page.
• Make more use of video feedback from around the plant, particularly with the affordable price of fiber-optics today.
• Give historical data to the operators.
• Provide other relevant information to operators, such as historical data, maintenance projections, weather forecasts and news.

Addressing these issues can provide significant improvements in operations, Nimmo said. He cited sample benefits that include a 5x increase in the operator's ability to detect an abnormal situation before an alarm comes in; a 37% improvement in responding to abnormal situations; and a 41% reduction in the time it takes to complete abnormal situation tasks.

"We've got to get the tools right down to the people who can make the biggest impact: the process control operators," Nimmo urged. "The process control operators are the saviors of our business. They will make a profit or lose a profit depending on the tools you give them."

By the way, the fire alarm? It was false. "Thanks for the abnormal situation," Nimmo quipped.