Equipment failure can be expensive and potentially catastrophic. Unplanned production downtime, missed contract deadlines, costly machinery replacements, as well as safety problems, environmental concerns and regulatory violations are all potential consequences of a maintenance program that fails to predict and monitor equipment problems.
Instituting a predictive maintenance program using condition monitoring can help minimize unwelcome surprises. Logging live-equipment-condition data on a regular basis through a condition monitoring routine offers plants a more scientific approach to managing equipment performance. Collecting data on a regular basis enables technicians to identify trends and thus anticipate imminent machine failures. With advance notice comes sufficient time for planning corrective actions.
Although most organizations recognize the benefits of routine condition monitoring, many do not have the in-house resources to operate a monitoring program cost-effectively and efficiently. Fortunately, some OEMs are offering services and software to fill the maintenance gap, offering their end-user customers product, service and operation expertise through condition-monitoring programs. With a condition-monitoring program, OEMs can provide end-users with crucial information about the health of their machinery by performing temperature and vibration analysis, oil sampling and thermography, among other techniques.
All rotating equipment can benefit from routine monitoring. Decisions about how often and to what degree monitoring should occur depend largely on criticality, cost and accessibility. Vital, high-dollar equipment could call for full instrumentation to allow continuous monitoring and protection. Most other plant equipment can be covered adequately with a monthly data-collection monitoring program.
Temperature Trending and AnalysisEquipped with handheld data collectors, technicians can track temperatures at bearings on motors and pumps, at process equipment and on mechanical seal hardware, including seal flush, quench and heat exchanger reservoirs, to detect abnormal operating conditions at predetermined locations throughout the plant. Collected data can downloaded to asset-management software and analyzed for nonconformities or notable changes in operating conditions.
One means by which condition monitoring is helping to achieve early detection of equipment failures is temperature trending of bearings and mechanical seal flush plans. For example, this method helped save machinery at a refinery in the Midwest. On one of the refinery's pumps, an API Plan 11 seal-flush plan directs flow from the pump discharge nozzle to the seal cavity through a flow-restricting orifice. The flush line temperature downstream of the orifice and the discharge temperature were routinely measured. When operating normally, the temperature readings on both sides of the orifice were similar, indicating regular flows.
For one chemical-duty pump, after months of normal readings, the temperature differential across the orifice increased, indicating a plugged orifice. If the operation had been allowed to continue under these conditions, it would have caused seal failure and ultimately would have required equipment teardown. Regular condition monitoring allowed early detection and correction of abnormal performance, and ultimately resulted in savings for the end user.
Critical Point MonitoringSome specific conditions have been proven to lead to premature component failure. Condition monitoring can alert users to these “trigger points.” Collected data such as critical pressures, process and seal fluid flow rates, reservoir levels and pump speed are downloaded to asset-management software and analyzed for nonconformities or notable changes in operating conditions. If a trigger point condition is identified, the end-user is alerted.
Critical-point condition monitoring identified such a condition at a power plant in the Northwest. In this case, the boiler feedwater pumps at the plant received a comprehensive inspection by a pump performance specialist. To obtain readings on the total flow through each pump, technicians used a dual-channel ultrasonic flowmeter on the suction line. They then took pressure and temperature readings and compared equipment performance to the original test curve performance.
Results indicated that over the life of the pump, performance had degraded 13 percent from the original factory performance curve. Vibration readings taken at the bearing locations showed excessive vibration both in the pump and at the inboard end of the turbine. The probable cause of this high vibration was imbalance, determined from the vibration spectrum and the time waveform. Through this analysis, the plant was able to plan for maintenance before performance degraded further or the vibration led to equipment failure, emergency shutdown and lost revenue.
Critical point monitoring can be targeted to meet a plant's specific needs. For example, vibration data collection and analysis monitors how a piece of hardware responds to external forces. With this technology, bearing problems, cavitation, recirculation, misalignment, resonance and imbalance can be identified. While the OEM's on-site technicians use handheld devices to gather vibration data from the designated equipment, off-site analysts can look for vibration signatures that exceed preset alarm limits and make recommendations for reducing or eliminating the causes.
This type of vibration analysis was used to predict and prevent unexpected failure at a chemical plant in the Northeast. In this example, identification of the potential problem started with a routine vibration check of the machine train of a double-suction cooling tower pump. Unusual vibration of the pump motor was discovered. The motor outboard-bearing vibration spectrum indicated a high-frequency spike at 56X running speed as well as excessive play. At the time of the check, the motor's internal condition was unknown, but the spectrum pointed to the possibility of loose or cracked rotor bars.
The results of an additional reading focused on the lower-frequency portion of the spectrum, which showed pole pass frequency side bands around the running speed harmonics. This provided further evidence of a rotor bar issue. The plant decided to switch over to the spare pump and was able to bring the equipment down in an orderly fashion. Subsequently, a teardown of the motor revealed two broken rotor bars. If the equipment had not been monitored and analyzed, the motor could have failed catastrophically -- destroying not only itself, but also potentially damaging much of the equipment surrounding it.
Ensuring ROIHow can plants best measure the success of a routine-condition monitoring program? One of the easiest methods is to track the average cost per work order on the rotating equipment. Condition-monitoring programs are designed to detect potential failures before they become catastrophic. Clearly, using predictive maintenance to prevent failure will cost less than maintaining the equipment after it has been seriously damaged or compromises safety or regulation standards. With condition-monitoring, there should be fewer emergency work orders and maintenance activities can be better planned.
Each machine “save” that was a direct result of the condition-monitoring program can be logged by a plant reliability engineer, who assigns a monetary or other value. The value generally is determined by estimating the maintenance cost had the equipment run to catastrophic failure and subtracting the cost of the repair. If the “save” prevented losses due to downtime or lowered production output, or avoided a safety or environmental issue, those additional and significant savings also should be calculated.
Using industry averages to compare maintenance savings with the cost of a monitoring program, plants can expect payback in as little as one month. These quick returns can be used as supporting evidence when approaching upper management about incorporating condition monitoring and a predictive maintenance program into the budget. OEMs or industry partners can provide this type of supporting data.
Applying the rigor of condition monitoring in a predictive maintenance program offers a range of benefits. Combining condition monitoring and predictive maintenance with the use of asset management, other equipment management, and procurement applications and services can provide a positive return on investment.