Jody Minor believes that the downturn in the economy over the past few years will move asset reliability to the forefront in manufacturing. “When a company runs its equipment to failure, that reflects total reactive maintenance, which can generate more than just replacement costs,” says Minor, a former industrial instrumentation-and-controls and reliability specialist. He speaks from 17 years experience with Houston-based Lyondell-Bassell (www.lyondellbassell.com) at its Clinton, Iowa, ethylene and high-and-low-density polyethylene production facility.
Minor’s experience there led him to become a champion of asset management, particularly of predictive maintenance (PM) technology. “With predictive maintenance, you’re forewarned. When failure occurs, you can have a mitigation plan already in place,” says Minor, who is currently a technical consultant with Midwest Engineering Center of Emerson Process Management (www.emersonprocess.com).
However, “people are all over the map with asset reliability programs and their associated data,” Minor observes. For example, he notes that one facility’s definition of reliability and best practices often differs from another facility’s. Still, the central question remains universal: “What can I do to keep something 100 percent available? Knowing that everything you do has an associated cost.”
Though critical to an intelligent maintenance plan, Minor cautions that manufacturers should be wary of slipping into PM overload. This can happen from something as seemingly innocuous as employing a different maintenance plan for larger valves. It can also happen with the passage of time. For instance, at year zero, a company installs and learns a PM technology. Then years go by and some new technology gets installed. That often results in end-user questioning of the applicability and usefulness of the legacy PM technologies.
Despite the potential for problems, Minor maintains the importance of PM technologies to successful production operations. When asked what he considers to be an asset management system’s most important components, he replies: “Robustness and accessibility,” taking care to note that accessibility links with availability. “I want it to be able to connect to the vast majority of my assets… and it has to be easily accessible.” Otherwise, culture problems will arise if operators find the technology “cumbersome and inconvenient.”
Assessing the Technology
Minor’s asset management experience at LyondellBassell stems from his interaction in the plant with Emerson’s Asset Management Suite (AMS) Device Manager. “We first got it in 2000 as part of a control-system upgrade to our [Emerson] DeltaV distributed control system (DCS).” However, only the plant’s information technology group used it. “That happened because of a lack of knowledge in operations about what AMS could do,” he recalls.
That knowledge gap closed in 2004 when Minor tried to solve a control-room problem and discovered the asset management tool. “I talked to my boss, the reliability manager, and told him we already had this great tool and didn’t have to spend extra money.” The key aspect of AMS he liked most was how field devices could communi-cate problems through it.
For six months after discovering AMS’s potential in operations, Minor spent a few hours every week evaluating the asset management tool. “By fall of 2004, we knew we had something useful. We had a plant process-control network and I had a dedicated PC in my office to which AMS was connected.” From there, Minor could connect to all field devices in the plant: 200 Fieldbus Foundation (www.fieldbus.org) devices and about 1,600 HART Communication (www.hartcomm.com) Protocol and Fieldbus Foundation devices.
At this point, Minor was the only operations and maintenance staff member with access to the tool. Shortly afterward, however, the company leveraged existing fiber-optic cables throughout the plant, allowing the instrument shops to get dedicated PCs with access to AMS.
Once this happened, “we started turning on the alerts. Besides the predictive piece, there were diagnostics for the instrument technicians,” Minor explains. “If we had to look at something, we could view AMS first so the technicians could get an idea of what they were going to run into.”
After the initial, successful evaluation by some members of the operations staff, “we got money for extra computers and an instrument technician to help me.” But the bigger challenge was getting buy-in from operations, maintenance and instrument technicians as a whole. “Getting those stakeholders to use the tool was harder, because that was changing their culture.” This change took six months to a year, Minor recalls.
Reacting to the Data
Predictive maintenance “only works if you act on what you’ve learned,” says Minor. “If you choose to do nothing different with the information, you’re no better off than you would be using reactive maintenance.”
At Lyondell-Bassell, Minor and the others acted. “Depending on how critical the device was to operations, we set a two- to six-hour polling rate. At minimum, we looked at AMS once every day.” Whoever viewed the alarm report would determine what needed immediate attention and when operations’ assistance was needed. “Depending on what the issue was, there was a specific course of action we would follow.”
Part of the success of the asset management technology at LyondellBassell came from the attention paid to it by Minor and others. “I often looked more than once per day, because the PC in my office was connected to it.” That focus produced action-able results and bragging rights. With AMS, “I was finding problems with valves before operations found them… if they found them at all. That caught their attention.”
Using AMS helped prevent several process losses at the ethylene plant, Minor states. “There were issues identified in valves prior to upsets. We saved a day’s production per year,” he adds.
To date, LyondellBassell continues to use the PM technology it discovered al-most a decade ago.
Similar to LyondellBassell’s experience with asset management technology, a large semiconductor facility also discovered the technology’s value when it came to mainte-nance on its three large, natural-gas-fired heaters. The company serviced and maintained the heaters regularly, but maintenance sometimes noted the degree of fouling in the heaters varied from maintenance period to maintenance period. That caused variability in benefits from its existing preventive maintenance program.
Like any heat-exchange equipment, a fired heater experiences deteriorating efficiency due to factors such as fouled heat-transfer surfaces and variations in the fuel’s heating value. The company needed a mechanism to monitor the actual fouling-related degradation in efficiency. With that information, the manufacturer could adjust the preventive-maintenance cycle and save fuel.
To optimize maintenance of the fired heaters, the company wanted a PM ap-proach via condition monitoring, says Umesh K. Chitnis, operations manager for Consult IT, ABB Inc. (www.abb.com) Having used the vendor’s asset monitors on other process equipment, the semiconductor manufacturer chose ABB to provide a real-time asset-monitoring application to monitor the fouling, Chitnis says. The company also wanted to alert maintenance personnel when the fouling degradation exceeded a specified threshold.
To monitor change in the fired heaters’ efficiency, ABB’s 800xA Asset Optimiza-tion framework was selected. The 800xA system presents a common environment with a single-user interface to display and analyze data related to an asset’s health. The semiconductor company already used the 800xA DCS to instrument, control and monitor the heaters.
Based on process values available through the DCS, ABB created a mathematical model to calculate the heaters’ efficiency, Chitnis explains. “The Fired Heater Asset Monitor would calculate the real-time efficiency degradation and compare it to a base-line curve. This comparison was based on the instrumented values of the fired heater.”
Through several discussions, manufacturer and vendor validated the model. They also created a functional specification detailing inputs, outputs and the algorithm to be used in the application, Chitnis says.
Once the functional specification was implemented, the company was able to rigorously test the asset monitor using simulated data, Chitnis notes. Testing included both performance and longevity monitoring. In addition to efficiency deviations, other aspects like natural-gas consumption, monthly run-time of the heaters, the number of lockouts, as well as total heat and firing rates were also calculated—all to assist in accurate reporting of the asset.
Some of the immediate benefits the semiconductor manufacturer realized from the new asset monitor included email and/or text notifications whenever alarms trip; better coordination between operation and maintenance for predicted and recommended maintenance; and elimination of the offline use of Microsoft Excel for maintenance, as the system now automatically provides online calculations of fuel consumption, heater run-time and number of lockouts.
Visualization of all this data, via the 800xA system, represented a significant benefit for the manufacturer. “This was the first time that all relevant data was calculated in unison and reported together with a possible cause and suggested action,” Chitnis explains.
>> Podcast on Burner Management: Charlie Fialkowski, U.S. Process Safety Manager for Siemens Industry, talks about standards updates for burner management system safety. Visit http://bit.ly/awpod40