Pumps account for an estimated 7 percent of maintenance costs of a plant or refinery, and pump failures are responsible for 0.2 percent of lost production. While a pump failure in a refinery may only affect one part of a process, pump failures in an oil field can shut down a well or pipeline. According to one estimate, the optimal run life of an electrical submersible pump is five years, but the average run life is only 1.2 years. Electrical submersible pumps are used for 60 percent of oil and gas production. Most are run to failure, then replaced. The average downtime for replacing a pump is one week. Figure 1 shows the typical causes of pump failures and the resulting impacts. These avoidable costs could be significantly reduced if unmonitored pumps had online condition monitoring.
Pump maintenance can be performed either through manual rounds or online condition-based monitoring. The goal of both is to prevent failures that require expensive repairs and cause process slowdowns or shutdowns. However, manual monitoring is typically labor-intensive and often cannot identify degrading performance in time for plant personnel to take action.
Historically, the expense of installing dedicated online monitoring systems has prevented them from being used on anything beyond the most critical pumps. But with the relative ease of adding online pump condition monitoring with today’s wireless sensor technology, online monitoring can be done on:
- pumps with repeat failures
- pumps without spares
- pumps that can cause a fire or environmental incident
- pumps that can lead to a significant process disturbance, process shutdown or slowdown
Many pump failures can be predicted using condition-based monitoring techniques, predictive technologies and reliability-centered maintenance best practices.
Pump Seal Monitoring
Dual mechanical seals are typically used in applications that require a more effective seal than packing or single mechanical seals can provide. These finely machined surfaces have one side affixed to the pump case and one to the shaft. The space between each seal is filled with a seal flush fluid to provide lubrication and remove heat. Contaminants can quickly degrade a seal, but with proper flushing, seals can last thousands of hours.
Auxiliary seal flush systems can be pressurized or unpressurized. Unpressurized seal flush systems are designed to leak process fluid into the seal fluid, referred to as a buffer fluid, which can then be removed through the outboard seal. Dual mechanical seals with pressurized seal flush systems obviate the need to remove this buffer fluid.
Figure 2 shows a pump monitoring system consistent with American Petroleum Institute (API) piping plan 52 for an unpressurized seal flush system. This arrangement is common with hazardous products because it provides a clean buffer fluid. A rising pressure measurement indicates a leak from the process to the buffer fluid, caused when the pumped process fluid vaporizes at atmospheric pressure.
Similarly, a rising level indicates a leak from the process to the buffer fluid. A slowly decreasing level is normal, but a sudden increase in the rate of level change indicates that the buffer fluid is leaking across the outboard seal to the environment.
API prefers continuous online measurements instead of local gauges, and pressure gauges can be easily replaced with wireless pressure gauges or transmitters to provide continuous online monitoring and to give an early indication of a leak. The latest version of API Standard 682 (fourth edition) also prefers hydrostatic level technology for continuous level measurement instead of switches to provide the signal for the low level alarm. This level-measuring device can be monitored by the control system or an online predictive asset health monitoring system, each of which can interpret rising level and provide actionable information.
Guided Wave Radar (GWR) technology can be used to measure reservoir level with accuracies of up to ± 0.2 inches (5 millimeters), which can be achieved over the level span. With GWR, the sensor can be mounted in the center of the reservoir, allowing measurement along the whole length of the reservoir, even down into the cooling coils, below the lower tap of the level instrument as currently shown in the standard.
Cavitation monitoring is needed on high-head, multistage pumps because they cannot tolerate this condition, even for a brief time. Although cavitation often happens when pumps operate outside of their design ranges, it can also be caused by intermittent pump suction or discharge restrictions. Damage can occur before manual rounds discover the problem but can be detected sooner by monitoring the pump discharge pressure for fluctuations (see Figure 3).
Some advanced pump monitoring solutions track discharge pressure fluctuations to give pre-cavitation alerts. When these increases in pressure fluctuation are correlated with increasing peak impacting measurements, they indicate the pump is likely experiencing cavitation. Cavitation detection through pressure and vibration monitoring shows a more holistic view of pump health that can be achieved through a combination of process and equipment measurements.
However, for high-head, multistage pumps that can be damaged by even brief periods of cavitation, monitoring should also include anything that could inadvertently move and cause blockage that would lead to cavitation. This includes level measurement in the suction vessel, differential pressure measurement across the pump suction strainer and the integrity of the automated isolation valves around the pump suction and discharge.
Vibration monitoring detects many common causes of pump failure. Excessive motor and pump vibration can be caused by a failing concrete foundation or metal frame, shaft misalignment, impeller damage, pump or motor bearing wear, and/or coupling wear and cavitation. Increasing vibration commonly leads to seal failure and can result in expensive repairs, process upsets, reduced throughput, fines if hazardous material is leaked, and fire if the leaked material is flammable.
Online vibration monitoring has been successful in detecting several root causes of pump degradation. However, traditional overall vibration measurement alone has limitations and may not provide information early enough to avoid failure. For example, bearings with a life of 100 percent display the same overall vibration as those with 10 percent of remaining life, and little change is indicated at even 1 percent of remaining life. In a typical application, pump and shaft vibration can detect bearing failure, but it cannot provide an early warning (see Figure 4).
New technologies can measure high frequency impacting faults, such as cases where metal comes into contact with metal, to give early warning of rolling element bearing faults. This peak impacting measurement technology, measured as Gs of acceleration, can detect bearing failure with more than 10 percent life remaining, providing much more time for preventive action. Advanced monitoring systems can detect a significant change in values and predict approaching bearing failure.
Every measurement solution previously described can be accomplished with wired or wireless transmitters. The challenges with retrofitting wired 4-20 mA or FOUNDATION Fieldbus solutions are complexity, installation difficulty and significant cost. Many pumps are located in hazardous or difficult-to-reach areas. In many instances, wired transmitters may be too expensive to be feasible.
Wireless technology has advanced to meet the rigors of the industrial sector and has enabled operators to economically expand the number of online measurements. The chief advantage of wireless systems is the ease of installation virtually anywhere in an efficient, timely and cost-effective manner.
Battery-powered transmitters require no wired infrastructure, open input/output (I/O) points at the control system or local power supply, so they can be installed in locations far from a process unit’s wired signal termination points. They also do not require the same supporting infrastructure as wired devices, and battery-operated transmitters can operate safely for years in hazardous areas.
Also, installation is simple; a typical wireless transmitter can be installed, configured and commissioned into a control system in a matter of a few hours, as opposed to days or even weeks for its wired equivalent.
Financial Benefits of Pump Health Monitoring
A complete pump health monitoring system can pay for itself in a matter of months. At one 250,000-barrel-per-day refinery, for example, pump monitoring systems were installed on 80 pumps throughout the complex.
The annual savings was more than $1.2 million after implementing the pump monitoring solution, resulting in a payback period of less than six months.
The savings came from decreased maintenance costs of $360,000 and fewer losses from process shutdowns because of failed pumps, which were conservatively valued at $912,000.
At the Eagle Ford shale play in Texas, failed pumps were costing an operator $3 million a month in lost production and well workover costs. After employing pump diagnostics and analytics to identify problems, the operator saved $14 million over a three-year period.
Predictive maintenance using online condition monitoring reduces maintenance costs and catastrophic pump failures associated with infrequent manual rounds by enabling proactive pump repair instead of reactive replacement. It also prevents unnecessary preventive maintenance activities, saving man hours that can be spent on pumps and other assets that need attention.
A critical factor in a highly effective pump monitoring program may hinge on the choice of software application used to analyze pump health from the measured variables and to provide timely, meaningful actionable information. Wireless sensors are often the lowest-cost solution, and they can be installed in hours, with payback periods measured in months.