It is estimated that pumps account for 7 percent of the total maintenance cost of a plant or refinery, and pump failures are responsible for 0.2 percent of lost production. Figure 3 illustrates the typical causes of pump failures and the effects failures have on the process, pump equipment, the environment and the business itself. These costs can be significantly reduced if the pumps are equipped with wireless online condition monitoring.

This article discusses how wireless sensors can be installed on pumps to reduce failures, cut maintenance costs and improve reliability.

Manual rounds versus online monitoring

Basic pump maintenance is usually 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 rounds are not always sufficient to identify performance degradations in time to take action. A pump can experience problems within a few days after a manual check, and those problems may not be seen until the next round, perhaps six months later.

A few pumps deemed as critical to operation may have online condition-based monitoring in place. In a typical oil refinery, these may account for approximately 10 percent of pumps. That leaves about 90 percent of pumps subject to manual rounds, with varying degrees of frequency.

Historically, the expense of installing wired online monitoring systems has prevented online condition monitoring from being expanded beyond the most critical pumps. The challenges with retrofitting wired 4-20 milliamp (mA) or fieldbus solutions are their complexity, installation difficulty and significant cost. Many pumps are located in hazardous or difficult-to-reach areas. In many of these cases, using wired transmitters may be too expensive and therefore not feasible.

The chief advantage of wireless systems is that they are easy to install virtually anywhere in an efficient, timely and cost-effective manner. By using predictive maintenance techniques with data gathered from wireless sensors, customers can minimize pump failures.

The top ways to minimize pump failures are to use wireless sensors and predictive maintenance to monitor pump seals, cavitation and vibration.

Pump seal monitoring

wireless sensors

Figure 2. API Piping Plan 52 unpressurized auxiliary seal flush system for online measurements (On top is a pressure transmitter, and on bottom is a level transmitter.)

The latest edition of American Petroleum Institute (API) Standard 682 now shows a preference for level and pressure transmitters instead of switches to detect level or pressure alarms. The use of transmitters provides an improved view of the pump seal flush reservoirs. A level signal also allows for monitoring the rate of change of a reservoir level for earlier indication of potential seal failure.

Figure 2 shows a wireless monitoring system consistent with API Piping Plan 52 for an unpressurized seal flush system commonly used for piping systems transporting hazardous products. A rising pressure measurement indicates a leak from the process to the buffer fluid when the pumped process fluid vaporizes at atmospheric pressure. Similarly, a rising level indicates a leak from the process to the buffer fluid, if the pumped fluid remains in the liquid phase at atmospheric pressure. A slowly decreasing level is normal, but a sudden increase in the rate of level change indicates the buffer fluid is leaking across the outboard seal to the environment.

Local pressure gauges are easily replaced with indicating wireless pressure transmitters to provide continuous online monitoring and give early indication of a leak. The seal flush reservoirs must be refilled periodically. Continuous wireless monitoring of the reservoir level allows the user to identify when and where flush fluid should be replenished. In addition, using the continuous level indication from a transmitter, an online monitoring system can monitor the rate of change on the level measurement to alert operations or maintenance when the fluid depletes faster than normal. This gives operations more time to switch to a spare pump and turn the problem pump over to maintenance personnel for repair.

Cavitation monitoring

Cavitation can damage pump components, disrupt flow, accelerate bearing wear and lead to seal failure. For high head multistage pumps that cannot tolerate cavitation even for a brief time, users must monitor discharge flow (downstream of the illustration in Figure 1) and pressure, level in the suction vessel, and the differential pressure across the suction strainer (upstream of the illustration in Figure 1) to help prevent cavitation from occurring.

Although cavitation often happens when pumps operate outside their design ranges, it can also be caused by intermittent pump suction or discharge restrictions. Damage can occur before manual rounds discover problems, but can be detected sooner by online monitoring of the pump discharge pressure (see Figure 1).

Some advanced pump monitoring solutions monitor discharge pressure fluctuations to give pre-cavitation alerts when cavitation is suspected. When increases in pressure fluctuations are correlated with vibration measurements, the pump is likely experiencing cavitation.

For high-head multistage pumps that can be damaged with even brief periods of cavitation, the pump monitoring regimen should also include level measurement in the suction vessel, differential pressure across the pump suction strainer, and the integrity of any automated isolation valves around the pump suction and discharge that may inadvertently move and cause blockage that will lead to cavitation.

Vibration monitoring

Figure 5. Traditional vibration measurement (third column) shows that vibration alone cannot give early warning. Adding advanced software allows for the detection of small changes in vibration and predicts imminent failure at points A and B.

Figure 5. Traditional vibration measurement (third column) shows that vibration alone cannot give early warning. Adding advanced software allows for the detection of small changes in vibration and predicts imminent failure at points A and B.

Vibration monitoring detects many common causes of pump failure, such as worn bearings, worn shaft couplings, misalignment, impeller damage, cavitation, and foundation or frame faults. Increasing vibration commonly leads to seal or other failures.

Vibration monitoring can be used to detect several root causes of pump degradation. However, traditional vibration measurement monitoring 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 (see Figure 5). Specialized vibration measurement software detects changes.

New technologies are available to measure high frequency impacting faults, such as cases where metal comes into contact with metal, to give early warning of rolling element bearing faults. Peak impacting measurement technology can detect bearing failure with substantially more than 10 percent life remaining, providing more time for preventive action.

Analyzing conditions

Table 1 provides a summary of the most frequent fault conditions of pumps, and the measurements that can give early warning or indication of failure: seal fluid level and pressure, pump discharge pressure, and peak impacting measurements. Vibration and temperature indicate failure and are useful for root cause analysis. Modern pump health monitoring solutions observe the overall health of a pump using a combination of both process and machine measurements.

Process data includes discharge and suction pressure, strainer differential pressure, seal oil pressure, and seal oil level. These process measurements are combined with machine measurements of vibration, peak impacting, pump bearing temperature, pump speed and run/stop indication.

This data is used to give the earliest warning, such as pre-cavitation alerts, that can be easily understood by operators and maintenance personnel without the need for machinery experts to interpret the data. Some advanced pump health monitoring solutions also offer pre-engineered algorithms that work on a combination of process and equipment data to report asset health.

Wireless pump monitoring solutions should be customized to a particular process environment and pump design. They are fully scalable, ranging from simple, condition-based monitoring on a single variable to advanced alarming across a range of pump states and operating ranges.

The challenge with condition monitoring is providing for the diverse needs of both operations and reliability teams with the correct data to enable strategic interpretation and actionable information. Operators may require alarming on serious conditions to prevent fast-occurring and severe pump damage. Meanwhile, reliability teams often have enough time for longer term planning.

Wireless technology a key enabler

Every measurement solution described above can be accomplished with wired or wireless transmitters, but the cost and installation advantages of wireless makes them the better choice in many applications.

Figure 4. This pump health monitoring system uses pressure, hydrostatic level and vibration wireless transmitters.

Figure 4. This pump health monitoring system uses pressure, hydrostatic level and vibration wireless transmitters.

Wireless transmitters with built-in power modules require no wired infrastructure, open input/output (I/O) points on the control and monitoring system, or local power supply, so they can be installed in locations far away from a process unit’s wired signals. They also do not require the same supporting infrastructure as wired devices, and wireless transmitters with power modules can operate safely for years in hazardous areas. Installation, therefore, is simple. A typical wireless transmitter can be installed, configured and commissioned into a control and monitoring system in a few hours, as opposed to days or even weeks for its wired equivalent.

A complete pump health monitoring system (see Figure 4) can pay for itself in a matter of months. At one 250,000 barrel-per-day refinery, pump monitoring systems were installed on 80 pumps throughout the complex. The annual savings were 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 valued at $912,000.

Summary

Serious issues caused by pump failures are costly, but they can be avoided simply with the right maintenance at the right time to prevent hazardous leaks, fires, expensive repairs, process upsets and downtime. They cannot, however, be stopped with preventive maintenance and manual rounds alone. Online condition-based monitoring is required, although it previously has been limited to critical pumps in most facilities.

Predictive maintenance solutions using wireless technologies are changing the landscape of pump health monitoring, making online condition-based monitoring available to virtually any dual seal pump in an industrial plant or refinery.

Brian Atkinson is a senior application development consultant for Emerson Process Management. He is responsible for driving integrated product and technology roadmaps and leading new product introduction efforts. He has more than 15 years of experience in automation and control, with extensive experience in the oil and gas, refining, and pharmaceutical industries. He has a bachelor’s degree in chemical engineering from the University of Delaware.