Condition monitoring programs – typically reserved for high-dollar components such as boilers and turbines – are gaining momentum for pumps and other motor-driven equipment.
Condition monitoring programs implemented as part of a larger comprehensive predictive maintenance program usually more than pay for themselves after one or two key machinery saves. Ideally, purchasing data collection equipment and investing in specialized personnel training will reap great financial benefits resulting from enhanced predictive maintenance decisions based upon machine condition.
Routine data collection leads to useful trends, which point toward imminent machine failures, leaving ample time to plan and perform corrective actions. The net result should be a decreased number of expedited emergency work orders, less lost production downtime, reduced spares inventory, and reduced maintenance spending. Too good to be true? If done the correct way, with enough focus and consistency, it can be true.
While most reliability and maintenance organizations understand the need for routine condition monitoring, many do not have the committed internal resources to successfully operate a monitoring program. The people trained to collect and analyze the machinery information are often the same people who must do the daily firefighting, handling one emergency after another. Given the choice between repairing a failed condensate pump or running data collection routes to gather information for long-term trending, the choice is usually the former. Other organizations struggle when the one person in the facility trained in vibration analysis retires or transfers to another position. And still others have difficulty maintaining the expertise often required to accurately interpret the reams of collected data.
Filling this gap, some major equipment OEMs have stepped in by offering enhanced reliability services. Creating a win-win situation between themselves and their customers, equipment OEMs are increasingly becoming more involved in the service and operation of their products by providing condition monitoring programs. Their expertise in the monitoring and evaluation of their own equipment, coupled with increasingly common partnerships with condition monitoring hardware and software vendors, has enabled equipment OEMs to extend their abilities into the condition monitoring arena. OEMs now perform tasks ranging from fully instrumenting boiler feedwater pumps during manufacture to conducting monthly data collection and analysis audits utilizing ultrasonic flow measurement, vibration and temperature analysis, oil sampling and thermography. With such capabilities, OEMs have developed the expertise to perform condition-monitoring regimes that provide plant maintenance groups with the vital machinery health information needed to make operations and maintenance decisions.
Regardless of industry, every plant contains a wide range of rotating equipment that, based upon criticality, cost and accessibility, should be considered for routine monitoring. Certainly unspared high-dollar equipment tops the list of candidates most likely requiring full instrumentation to allow continuous monitoring and protection. Much of the balance-of-plant equipment can be covered adequately with a monthly monitoring route data collection program. From boiler feedwater pumps to motors to fans and blowers, many machine trains in a plant are worth the investment in condition monitoring to help predict and prevent unexpected failure.
For instance, take a double suction cooling tower pump driven by a 600 hp motor. A routine vibration check of the machine train indicated high motor vibrations. As seen in Figure 1, the motor outboard bearing vibration spectrum indicated mechanical looseness and a high frequency spike at 56 times the running speed. The internal characteristics of the motor were unknown at the time, but the spectrum points to the possibility of cracked or loose rotor bars. An additional reading focused on the lower frequency portion of the spectrum (Figure 2) revealed pole pass frequency side bands around the running speed harmonics. The pole pass frequency is the slip frequency of the ac induction motor multiplied by the number of poles in the motor. Modulation of running speed and its harmonics by the pole pass frequency appear as side bands on the spectrum, which are a strong indicator of rotor bar problems. The plant proceeded to switch over to the spare pump and bring the equipment down in an orderly fashion. Infrared inspection and teardown of the motor revealed that there were, in fact, two broken rotor bars. Had the equipment not been monitored and analyzed properly, the motor could have eventually failed catastrophically, destroying itself and potentially damaging much of the surrounding equipment.
Temperature trending of bearings and mechanical seal flush plans, a great companion to vibration measurements, opens a whole new avenue to early problem detection. A Plan 11 seal flush plan directs flow from the pump discharge nozzle to the seal cavity through a flow-restricting orifice. During normal operation, one would expect the temperature on both sides of the orifice to be similar, indicating regular flow. During normal data collection rounds at a midwestern U.S. refinery, the discharge temperature and the flush line temperature downstream of the orifice were measured on a monthly basis. Showing little temperature differential for months, a sudden decrease in the orifice’s downstream temperature provided a clear indication of a plugged orifice. Continued operation under these conditions would eventually cause seal failure and require teardown of the pump. Maintenance to clear this condition involved a low dollar work order to break the line, clear the restriction and reassemble. Compared to the cost of pulling the pump out and replacing the seal and bearings, this example of predictive maintenance saved the operator both money and equipment downtime. As illustrated in this case, knowledge of what to trend coupled with consistent data collection provides a powerful tool in detecting and correcting abnormal performance.
Routine condition monitoring comes in many forms. At a northwest power plant, for example, a pump performance specialist conducts an annual comprehensive inspection of the plant’s boiler feedwater pumps. Using a dual channel ultrasonic flowmeter on the suction line to obtain the total flow through each pump, combined with pressure and temperature readings, the performance of the equipment is compared to its original test curve performance. Figure 3 exhibits the annual pump performance as compared to the original factory performance curve. Performance has degraded 13 percent overall from the factory performance curve, with 4 percent degradation from the prior year’s measurement. Vibration readings at the bearing locations indicate excessive vibration in the pump and the inboard end of the turbine. The vibration spectrum, shown in Figure 4, reveals a dominant 1X running speed component in the horizontal radial inboard bearing measurement. Large radial vibration at running speed frequency throughout the pump, coupled with low axial vibration, leaves imbalance as the probable cause of the high vibration. This analysis has enabled the plant to plan upcoming maintenance before performance degrades to the point of emergency shutdown and lost revenue.
In addition to all of the high-tech collection equipment, data processing software and elaborate plots, an added benefit of routine route collection is another set of experienced eyes roaming through a plant. Familiarity born from daily interaction with the same equipment can dull the senses and lull plant personnel into a false sense of security. Visiting equipment on a monthly, quarterly or even annual basis, experienced reliability personnel can notice abnormal circumstances that the maintenance staff has become accustomed to and no longer sees. Reliability personnel can notice leaking seals, loose drive belts and backed-off anchor bolts. They also can observe missing floor grates on elevated platforms and incorrectly plumbed heat exchangers. Such consistent route monitoring provides one essential element that even the most expensive online monitoring systems can’t – an up-close detailed view of the environment surrounding the equipment, and all of the associated indicators, signs and warnings.
How is the success of this type of program measured? One of the simplest ways is to track the average cost per work order of the plant’s rotating equipment. With the goal of detecting potential failures before they become catastrophic, conditioning monitoring programs enable proactive maintenance actions to be taken ahead of time to prevent failure. These steps will naturally be at a lower cost than maintenance after the equipment has partially or totally destroyed itself. Also, maintenance activities can be planned better and the number of emergency work orders should decrease.
To measure success, each machine save attributable to the condition monitoring program is logged and assigned a value by the plant reliability engineer. The value is typically the cost of the repair subtracted from the estimated maintenance cost had the equipment run to catastrophic failure. Additionally, if lost downtime or production was thwarted, those savings are also included. As seen in Figure 5, in terms of maintenance savings versus the cost of the program, payback occurred within the first month. While one can argue about the actual savings that occur from hypothetical failure avoidances, there is no question that the savings are significant. Tracking the savings in such a manner has the added benefit of justifying predictive/proactive maintenance programs to upper management.
Honeywell’s Equipment Health Monitoring detectors are designed to monitor industrial equipment for operating condition variances that can lead to breakdowns, increased maintenance costs and shutdowns.
Whether employing just one technology, such as vibration monitoring, or doing everything from thermography to ultrasonics, having experienced reliability personnel perform condition monitoring of rotating equipment on a consistent, routine basis provides measurable, tangible improvements in plant equipment reliability. p
Author – Phillip Mosher is the Lifecycle Advantage Condition Monitoring Program Manager for Flowserve Corp. He has 12 years of industry experience in rotordynamics, gas compressor seals and condition monitoring. Mosher holds bachelor’s and master’s degrees in mechanical engineering from Texas A&M University.
Condition Monitoring Tools
By: Brian K. Schimmoller, Managing Editor
Many pump failures are missed because of inadequate or nonexistent equipment monitoring practices. Such failures result in unplanned downtime, lost productivity and higher maintenance costs. A number of equipment manufacturers are expanding and enhancing their condition monitoring capabilities to help end users prevent such failures and stay competitive.
Texas Instruments Sensors & Controls is introducing a rotating equipment monitoring system called Pump SystemAlert that will use electrical and mechanical input parameters to detect the real-time machine health status of pumps and other motor-driven equipment. A significant fraction of motor-driven equipment failures are detectable through electrical monitoring since they occur inside the motors. Problems such as motor overload, phase imbalance and improper wiring can be detected via electrical characteristics. By combining electrical sensing with traditional vibration sensing, end users can obtain a more sensitive analysis of machine health status information.
“The traditional approach of employing sporadic condition monitoring with portable equipment is not only expensive, but leaves spans of time when equipment is not being monitored,” said Jim McGuinness, program manager for TI S&C. Pump SystemAlert enables existing or new sensors installed throughout a piece of rotating equipment to continuously collect and record maintenance and process data in real time. The data is compiled and stored in the control unit, allowing maintenance personnel to review equipment performance in real time or over a given period of time. Pump SystemAlert is currently in beta testing and will be released later this year.
Honeywell has released a similar set of tools, called Equipment Health Monitoring detectors, to monitor industrial equipment for operating condition variances that can lead to breakdowns, increased maintenance costs and shutdowns. These devices monitor the steady-state temperature, vibration and other characteristics of targeted equipment, alerting operators when values drift from predefined settings. Detectors are available for 10 conditions: temperature loss, fluid flow, temperature rise, mechanical noise, fluid leak, mechanical wear, available noise, vibration, tilt (slope change) and mechanical insertion.
The EHM detectors are easily installed: bolted on with two screws for most equipment or attached with cable ties for pipes. Once installed, the EHMs are connected to a 24 V DC power supply and then configured for fault monitoring. The threshold of the device is set until a red LED indicator light comes on. The user then backs off on the threshold a quarter-turn until a bright green LED light comes on. If the chosen parameter drifts outside the preset limits, the LED changes from green to red. Monitoring is reduced to a simple go/no-go digital output. The module can also be connected to an audible alarm. “An output signal from the device can be sent to a programmable logic controller if desired, but most users opt to install the device independent of the control system so they can move it around and use it on different pieces of equipment as conditions warrant,” said Leslie Neill, product manager with Honeywell.
For the power generation sector, a likely application would be to detect potential bearing failure. “Just before ball bearings seize, they turn oval and emit a characteristic rumble,” said Neill. “The EHM-D-RUMBLE detector monitors the high- and low-frequency signature from bearings to identify this rumble and prevent early failure.”
To enhance the condition monitoring capabilities of its industrial machinery, Flowserve Pumps recently formed an alliance with GE Energy to incorporate Bently Nevada machinery protection solutions into its pump products. The agreement covers Bently Nevada brand transducers (e.g., proximity probes and accelerometers), machinery protection systems and software (optimization and diagnostic software).
“A significant part of the value we add for customers is through our cooperative relationships with machinery manufacturers,” said Richard Chapman, GE Energy’s product general manager for Bently Nevada Asset Condition Monitoring. “This alliance with Flowserve Pumps tangibly benefits our mutual customers by enabling improved pump availability/reliability and lower overall lifecycle costs. Because the Bently Nevada condition monitoring products chosen by Flowserve Pumps are already so widespread in the industries we serve, customers can also look forward to easy integration.”
Advanced software and analytical systems can help end users analyze equipment data for subtle operating changes that could portend pending failure. Kansas City Power & Light (KCPL), for example, has been working with SmartSignal Corporation’s EPI*Center software solution to provide early warning on power island and balance-of-plant (BOP) equipment.
KCPL sought to improve BOP availability across its roughly 4,300 MW fossil fuel fleet, particularly among pumps. The EPI*Center solution leveraged KCPL’s investment in OSIsoft’s data historian to provide real-time analytics that evaluate process signal information in context of vibration data. The result is insight into a wider range of faults across all operating conditions.
After a data study demonstrated the ability to provide several weeks’ earlier notice of condensate pump anomalies, KCPL deployed the EPI*Center solution across a wide range of pumps – e.g., condensate, circulating water, boiler feed and heater drain pumps. Since then, the solution provided early warning of several power island and BOP equipment anomalies, including a circulating water pump issue where the cooling flow to the bearings was restricted. By finding it early, KCPL was able to fix the problem quickly and avert damage to the pump.
Power Engineering magazine recently participated in a Webcast with SmartSignal and KCPL to explore the value of condition monitoring for BOP systems. The free Webcast is archived on the Power Engineering Web site at www.power-eng.com.
Valve Shipments to Power Industry Looking Up
By William S. Sandler, Valve Manufacturers Association
The power industry continues to be one of the top five valve consumers in the United States, as well as worldwide. The Valve Manufacturers Association (VMA) outlook for 2005 puts power fifth after water/wastewater, chemical, petroleum production and petroleum refining. If one were to combine power generation with cogeneration, however, the power industry would then move up to being the third largest valve consumer. Although a variety of valve types are used in power generation, more than half of the valves used by this market are control valves.
As the accompanying bar chart indicates, power generation’s share of the total U.S. industrial valve market is predicted to increase during 2005 after declining the previous two years. In part to satisfy the valve needs for tens of thousands of MW of new gas turbine-based generating capacity, 2002 was a banner year for the power industry after seven years of continual growth in market share.
At a recent presentation to the VMA membership, several power industry experts identified pockets of growth in the United States that should present opportunities for valve manufacturers and distributors: coal plants being developed to meet the market recovery, air quality control systems being installed to comply with emission regulations, liquefied natural gas (LNG) plants being built to meet growing North American gas demand, the possible development of integrated gasification combined cycle (IGCC) facilities, and continued nuclear services.
Additionally, valves are playing a role around the world, with new power generation facilities being built in Southeast Asia and other developing nations. North America retains significant valve manufacturing capabilities, although many manufacturers are sourcing some parts overseas and are establishing a presence in high-growth countries like China through joint ventures or wholly owned subsidiaries.
William S. Sandler, CAE, is president of the Valve Manufacturers Association of America. VMA represents the interests of more than 100 valve and actuator manufacturers, which account for approximately 85 percent of total U.S. and Canadian industrial valve production. The North American valve industry supplies 40 percent of worldwide valve demand. For more information on VMA, contact Sandler at 202-331-8105 or [email protected]