Grasp Maintenance’s Bottom Line

Effective predictive maintenance demands more than technology.

By C. Kenna Amos, contributing editor

Share Print Related RSS
Page 2 of 2 1 | 2 Next » View on one page


Rosemount now is running field tests of high-speed sensors (22-Hz data collection) for EED. Trials include for detecting: flooding and damage to internals in distillation columns; flame instability and burner problems in fired heaters; catalyst transfer issues in fluid catalytic cracking units; fouling in heat-exchangers; and presence of wet gas in compressor suction lines.

The Abnormal Situation Management Consortium, Phoenix, Ariz., cites several situations in which early information has provided significant benefits. In one case a frozen differential pressure cell that could have led to column flooding was detected six hours before the first alarm would have been triggered, saving a plant from a major upset and a large financial hit.

Indeed, EED’s value isn’t only in warning of developing problems but in helping track down their root causes such as failed sensors, plugged lines and sticking valves, adds Steve Johansen, global services marketing manager for advanced solutions at Honeywell.

“The multivariate statistical analysis and principal component analysis methods within Honeywell’s EED toolkit are generic in the sense that they can be used to analyze and correlate very small (a single piece of equipment) to very large (an entire process) data sets,” he says.

Process-oriented technology can detect oscillation in pumps, notes Moin Shaikh, marketing manager for process control systems at Siemens Energy & Automation, Spring House, Pa. It looks at variation in pump out-flow, assessing whether valves are performing optimally or not. “Based on that, you can set predictive-maintenance conditions to schedule mainte-nance,” Shaikh adds.

Sometimes, though, it’s necessary to have smart instrumentation that trips an alarm to alert operators that a certain piece of equipment has reached a critical maintenance level, he notes.

“Online remote diagnosis with real-time data and data validation — with efficient logic analysis tools and effective work processes — are keys for successful predictive-maintenance program at any site,” says Amit Ajmeri, consultant for asset-management solutions with Yokogawa Corpora-tion of America, Sugar Land, Texas.

Eyes in the Field
But people will continue to play a crucial role. “You need them to see the obvious things,” Kant says. Honeywell agrees. Its chemical plants (within Honeywell Specialty Materials) employ maintenance-excellence programs that rely on operator field surveillance. Checking pumps for noise may alert operators to a problem “normally days, if not weeks, before failure,” notes Winters.

Brady advocates operator-driven reliability (ODR), a preventive/predictive maintenance mix that focuses on having reliable eyes in the field. “During shifts, the operations guys have always done rounds with check sheets.” Now, though, devices such as industrial personal digital assistants (PDA) allow data “to get into a database people can view. If a trend becomes high, the maintenance guys get notified,” he says.

While operators are making visual inspections they also can generate, not just record, data, e.g., on vibration, thanks to wireless sensors (Figure 1). Brady says such devices will assume an ever-increasing role primarily because of their reliability and safety, especially with hazardous environments.

Figure 1
Tall order no longer. Wireless sensor means that equipment location doesn’t hamper monitoring.
Source: SKF


Regardless, “it’s good for operators to be out there. It’s seeing the machine, hearing the machine, smelling the machine,” Brady adds. “Nobody can replace a person.”

This role for operators makes the need for cooperation with maintenance all the more important, Kant emphasizes. “Because of a lot of these advanced-monitoring techniques, it’s the operating people who get the first warning or signal. Therefore, they need to call maintenance right away.” Kincer warns, “If operations and maintenance aren’t talking, then there are problems.”

Pay attention to another people issue — staff expertise, counsels Johansen. With people who are really knowledgeable about specific equipment increasingly scarce at sites, it may make sense to bring in an authority to help with some PdM issues.

Based on Honeywell’s experiences and those of its end-users, Winters lists three criteria for excellent maintenance programs. First, get the basics right and use the right technology. He suggests a good field-surveillance program with mobile hand-held devices — Honeywell uses its own Intela-Trac PKS devices at its facilities — to collect information about an asset so those data can be trended.

Next, monitor other relevant parameters such as vibration, lubricant condition and temperatures that can be sensed through thermography. “Typically, these technologies can be used to predict failures,” he says. Consider how wireless technology can help, he adds. “We recently completed a project on an oil tanker. We put wireless infrastructure inside the ship to monitor the pumps.”

Finally, adopt technologies that give data useful for first-principles or empirical EED modeling. “That’s a little bit more specialized. It’s not employed as broadly as the others. It’s a technology we use in certain instances to monitor important events.” For example, at one of its chemical plants, Honeywell uses an IntelaTrac device with EED, he says.

The Culture Question
The availability of field-intelligence technologies and the quest to minimize shutdowns and downtime certainly are strong drivers for PdM, says Kant. However, he adds, success also depends upon top-management buy-in, proper motivation for excellence and company culture. “Culture is where the actual execution of predictive maintenance takes place. It is a change of thinking.” All are needed, he stresses.


“When top management deployed such PdM strategy, but did not have full support from maintenance as well as not having the strategy implemented as part of the work process, that had a high probability of failure,” cautions Ajmeri.

Brady emphasizes that besides technology to enable a successful program and a process to get it done, you need right attitudes to make the other two click (Figure 2). “Where do we see things fail? Where the culture’s not there. Unless the culture’s there, it [the maintenance program] doesn’t go anywhere.”

Figure 2
Necessary Elements: Success with a maintenance program requires proper tools, attitude and procedures.
Source: SKF


This demands a change in thinking about maintenance, says Lee. “We have to elevate the image of maintenance systems.” Deafness by top management no longer works — and shouldn’t, he stresses. After all, as Kincer states, “Predictive maintenance is a matter of competitiveness.” Indeed, notes Kant, top management that does buy into PdM sees its real value: “They love it because it goes to reliability and, thus, the bottom line.”

Page 2 of 2 1 | 2 Next » View on one page
Share Print Reprints Permissions

What are your comments?

Join the discussion today. Login Here.

Comments

No one has commented on this page yet.

RSS feed for comments on this page | RSS feed for all comments