Industrial processes across the globe require pumps to operate reliably and efficiently. The latest pump designs and coating technologies offer significant improvements in the long term performance of industrial pumps. By minimizing the effects of corrosion and erosion, users can enhance productivity and reduce running costs. Continued research into the processes that degrade pump performance is being matched by the development of better application techniques for protective coatings. By gaining a better understanding of both the pumping process and the factors that affect it, end users can make significant improvements in their maintenance strategies. Almost every industrial process involving liquids will include a pump at some point. From deep sea oil and gas to DNA sequencing, pumps are required to perform a vast range of tasks. However, no matter what the design or the size of the pump, central to every application is reliability and efficiency – minimizing down time and running costs is essential to modern industry.

For those working with large industrial pumps, often operating in harsh environmental conditions, maintaining pump performance in the face of a continuous threat from corrosion and erosion can be a particular challenge. With increased knowledge of these processes and the techniques used to tackle them, it is possible to implement a more cost effective pump refurbishment program. Corrosion is commonly defined as a chemical reaction between the component surface and the reacting fluid passing through a pump. In general a distinction is drawn between general or uniform corrosion and localized corrosion like pitting and crevice corrosion. Non-stainless materials suffer mainly from uniform corrosion whereas metals forming oxide layers that adhere to and passivate the surface are prone to localized corrosion. Flow accelerated corrosion (FAC) describes the removal of the protective oxide layer on a metal. The speed of this process is affected by the oxygen content, the flow velocity and, to some extent, the chloride content. The formation of a calcareous layer due to high carbonate hardness of the water reduces or even prevents FAC.



The influence of oxygen can be seen in the following example: Water with an oxygen content of less than 20 ppb (parts per billion) and a flow velocity around 15 m/s will typically see a corrosion rate around 0.01 mm/year. However, increased oxygen content can see the corrosion rate rise to several mm/year, which will present a significant challenge to the process. Fortunately FAC only poses a real issue for low carbon steels and cast iron. Increasing the chromium content or using stainless steel will largely eliminate the vulnerability to flow accelerated corrosion. Pumps that are used to transfer fluids containing abrasive substances, such as sand, can experience significant levels of erosion, especially in areas with high flow velocities. This can be seen in the oil and gas industry where injection pumps are employed to force water back into the oil field and thus maintain the pressure which is needed to lift the oil to the surface. The entrained sand particles act as an abrasive and the high working pressures only serve to compound the issue.

In operating conditions where both erosion and corrosion are present, the degradation mechanism can become very complex and depends on the type of substrate and the fluid chemistry. Corrosion may create oxide layers with low adherence to the substrate which is prone to erosion, or erosion may damage the passive layer, leading to an activation of the surface which accelerates corrosion. In this case surface protection regimes are often the best and sole option. Most commonly seen on the pump impeller, cavitation is caused by a pressure difference, either on the pump body or the impeller. A sudden pressure drop in the fluid causes the liquid to flash to vapor when the local pressure falls below the saturation pressure for the fluid being pumped. Any vapor bubbles formed by the pressure drop are swept along the impeller vanes by the flow of the fluid. When the bubbles enter a region where the local pressure is greater than saturation pressure, the vapor bubbles abruptly collapse, creating a shockwave that, over time, can cause significant damage to the impeller and/or pump housing.


For pump manufacturers, the key is to mitigate the corrosion problems by using the most appropriate base material in the construction of the pump. For applications where the use of carbon steel or cast iron is preferred due to cost reasons, the corrosion rate can be estimated very accurately. Based on the mutually accepted corrosion rate per year, the service life of the pump can be anticipated and factored into the maintenance costs of the application. If the expected corrosion rate is not acceptable the pump materials have to be upgraded to stainless steels which leads to higher costs. In cases where this cost increase is prohibitive, the alternative is to use advanced coatings that can be tailored to suit each application. If stainless steel is selected for an application, the expected service life is much longer, in some cases infinite. However, this is only true as long as the appropriate stainless steel grade has been chosen for the specific application, it has been produced carefully and is used within the agreed fluid specifications. Special care is required as soon as particles are introduced into the fluid.

In this case even stainless steel becomes susceptible to corrosion due to the passive layer being damaged and the base material becoming activated, which then starts to corrode. Normally the passive layer can be re-established, but if the chloride content is too high or the pH level is too low, the material may remain in an active state and the corrosion continues. Another frequent cause of corrosion in stainless steel pumps are stagnant conditions caused by process interruptions or intermittent operation. A further threat for stainless steel is chlorine, which is used to combat biological growth in the pump or the connected pipelines. Low level concentrations, around 2 ppm, will have little impact on stainless steel, but it is important to understand how and where the chlorine is introduced into the water flow, to avoid spot concentrations that will damage the protective layer. Unexpected corrosion can easily negate the anticipated improvement in durability of stainless steel compared to the much cheaper carbon steel variant.


Water hammer occurs when the flowrate of fluid in the pipe changes rapidly. It is also known as “surge flow”. It can cause very high pressures in pipes, very high forces on pipe supports, and even sudden reversals of flow. It can cause burst pipes, damaged supports and pipe racks, and leakage at joints. Water hammer can occur for any fluid, in any pipe, but the severity varies depending on the detailed conditions of the fluid and the pipe. It usually occurs in liquids, but it can occur in gases. It can cause pipes to burst and structures to collapse. This article will describe the conditions most likely to lead to water hammer problems and the issues that pump and pipe designers and operators can face. It also outlines some of the ways to resolve the problems.

Increased pressure occurs every time a fluid is accelerated or retarded by pump condition changes or valve position changes. Normally this pressure is small and the rate of change is gradual, and water hammer is practically undetectable. However under some circumstances, the pressure created can be many tens of bars, and forces on supports can be many tonnes, exceeding their specifications. In pipe bridges, collateral damage can occur. The risk to safety, assets and environment are obvious. Slight water hammer can be detected by pipe movements, banging noises and pulsing flows. Serious water hammer gives the same effects but these might be large enough to cause serious damage, and might only occur once! Pipe systems that show the characteristics that can lead to serious hammer should be analysed by computer software, especially if hazardous chemicals are being carried in them. Its presence can also be revealed sometimes by unexpected opening of relief valves.



Water hammer is a shock wave passing down the pipe as a result of a sudden flowrate change. The most common cause is a valve closing too quickly, or a pump tripping or starting up suddenly. This causes a shock wave which starts at the valve or pump and passes along the pipe, changing the fluid velocity as it goes. This is the cause of high pressure. If the wave is sharp and it passes through pipe bends, the pressure step change can cause out-of-balance forces which move the pipe. This might cause the pipe to move off its supports or transmit the force to its anchors. The pressure wave can travel through pumps, damaging the impellor and drive. Water hammer can also be caused by cavitation due to the pressure dropping below the vapour pressure, and then the bubbles collapsing as the pressure swings back up. This can happen after a valve or downstream of a pump. As the valve closes or the pump trips, the pressure downstream can fall to a level that the fluid boils, creating a vapour cavity. This suction can cause the liquid to flow backwards and the cavity collapses as it approaches the closed valve or stopped pump. When it collides with the valve or pump, a severe hammer can occur.

The closure of non-return valves can also cause water hammer. Some systems are very prone to this, and the use of a simple swing check valve can give severe water hammer. Some companies manufacture non-return valves which minimise water hammer caused by their operation. The formation of cavities in the high points of pipes due to exceeding the barometric height of vertical legs can also cause water hammer as flow is restarted. It’s not possible to give simple, infallible rules for spotting water hammer potential. Computer programs exist that allow pipe systems to be modelled and any potential for water hammer problems to be revealed. In experienced hands they can also be used to find the best solution to any such problems. Simple hecks can be done by hand calculation, and some vendors have nomographs to help predict hammer and design suitable alleviators. However most systems need good computer software to do this accurately.


Centrifugal pump systems with flow throttled by a control valve are ubiquitous at process plants. Unfortunately, the majority of pumps are oversized, which both poses problems regarding controllability [1] and wastes energy. Pump oversizing occurs for two major reasons. First, a system’s pressure drop must be calculated rather early in planning using estimates for pipework and its fittings. Unfortunately, values taken from the literature, for example of flow resistance coefficients of valves, may vary widely. Consequently, dynamic pressure losses are somewhat shaky. Static pressure differences usually are better defined. Hence, planning engineers must add some safety margins. Correcting a pressure drop calculation much later when pipework has been designed does not really solve this problem. Some margin also might be added to the rated flow. Second, the pump’s manufacturer adds safety margins, too, to ensure the pump meets guaranteed performance, in particular if the liquid handled is not water. So, it is nearly impossible to avoid installation of oversized pumps. Hence, the only realistic option is to adjust a pump based on experience once the plant is operating.

The vast majority of pumps in the process industries are centrifugal radial ones, mostly of moderate power consumption. For example, a 2010 survey of around 30,000 pumps by the Fraunhofer Institute in Germany revealed that about three-quarters require less than 15 kW and that most of these have installed spares. A centrifugal pump can be adjusted to some extent to a changed operating duty point by trimming its impeller’s diameter. Manufacturers use this method to achieve the operating point needed by the customer; it also might be appropriate later whenever necessary. In multistage pumps, a plain disk could replace an impeller.

When designing a plant, the traditional method to size a control valve consists of allocating to it 25–50% of the dynamic pressure drop of the corresponding flow path or, alternatively, an absolute pressure drop; its opening is fixed to, say, between 60% and 75%. This original design opening constitutes a2 and thus allows calculating ∆HL; also the original pressure drop may be used directly. Nevertheless, to determine the maximum avoidable pressure loss — and thus maximum potential energy savings —the control valve should be as wide open as possible while still achieving good controllability. The value of opening a2 can be found by using the configuration shown in Figure 1,while maintaining a constant flow rate (monitored, e.g., by a clamp-on flow meter)and applying the following procedure: Throttle manual valve V2; the control valve V1 will open to compensate for the reduced pressure at its outlet port. (In other words, the pressure loss of control valve V1 shifts partially to valve V2 and, as long as the flow rate does not decrease, the sum of the pressure losses of V1 and V2 remains constant, i.e., the operating duty point does not change.)Repeat this procedure several times in the field until controllability deteriorates. Select the opening just before this happens asa2 [2].

An H-Q diagram (Figure 2) provides a clearer picture. Here, D reflects the original operating duty point at design flow rate QD, and HP the original pump curve of the oversized pump. R indicates a hypothetical duty point determined by subtracting the avoidable pressure loss, ∆HL, calculated via Eq. 2. Hence, R can be used to define the diameter of the trimmed impeller, which either is inferred from manufacturer’s data or found by applying a method that is well known from the literature [3]. In some cases, installing a new pump might be preferable. The pump curve satisfying R is marked HP,R. Deleting the 2nd term containing a2 in Eq.1 yields the absolute pressure drop the control valve exerts; this is the pressure loss avoidable through replacing a throttling valve assembly as final control element by a pump with a variable speed drive.

Plant managers may have some reservations about operating a less powerful pump having a trimmed impeller. Having an installed spare pump with the original size impeller, as often is the case, should alleviate these concerns. Performing the task of modifying a pump in an existing facility demands cooperation from the plant’s operations staff and instrument and maintenance engineers as well as from process specialists. Coordinated teamwork is crucial.
Most chemical plants are working to become more energy efficient. Companies are implementing energy management software, installing occupancy sensors throughout plants to help lower electricity bills, and even changing times of operation to use less power at peak load to avoid the associated higher rates. One of the best ways to save energy is to focus on motor-driven pumps. Pumps consume more energy in chemical plants than any other category or type of rotating equipment. The average annual spending on pump maintenance and operations is approximately 50% greater than that of any other rotating machine, according to a recent study by the FiveTwelve Group. Companies that operate large numbers of pumps usually recognize the high energy costs as well as the impact pumps have on reliability and process control. However, too many organizations focus on these factors separately when, in fact, they are closely linked. 

Recent report on the use of motor efficiency technologies by the U.S. Department of Energy’s Industrial Technologies Program (ITP) contained an in-depth analysis of energy use and savings potential by market segment and industry. The report identified centrifugal pumps as the largest consumers of motor energy (Figure 1). Also, among all rotating assets in the plant, process pumps had the highest overall potential for electrical energy savings. Separate Finnish Research Center study of centrifugal pump performance found that the average pumping efficiency was less than 40% for the 1,690 pumps reviewed in 20 different plants across all market segments. That study also revealed that 10% of the pumps were operating at less than 10% hydraulic efficiency. Considering this sizable efficiency loss, you can expect that from 10% to 20% of the pumps in any continuous process plant are candidates for optimization. More than likely, the real number is much higher.

Today, companies increasingly are relying on lifecycle costing (LCC) for selecting an optimal solution to create economic and environmental value over the life of a system. Using a lifecycle-cost perspective during initial system design will minimize operating costs and maximize reliability. For pump systems, using LCC makes particular sense because the initial purchase price typically represents only about 10% of long-term costs (Figure 2).

A LCC analysis assesses the cost of purchasing, installing, operating, maintaining and disposing all the system’s components. Determining the LCC of a system involves using a methodology to identify and quantify all the components of the LCC equation. For instance, the equation provided in the Hydraulic Institute’s “Pump Life Cycle Costs: A Guide to LCC Analysis for Pumping Systems” includes terms for initial cost or purchase price (e.g., the pump, pipe, auxiliary equipment); installation and commissioning costs (including training); energy costs (predicted for entire system, including controls); operating costs (labor man-hours for normal system supervision); maintenance costs (e.g., parts, tools, labor man-hours); downtime costs (loss of production); environmental costs (leakage losses and permit violations); and decommissioning costs (disassembly and disposal). Energy consumption is a major element in pump lifecycle costs. Because excess energy consumption leads to higher maintenance costs, these two elements combined typically dominate total lifecycle cost. Thus, it’s important to determine the current cost of energy and the expected annual escalation in energy prices over the system’s projected life, along with labor and material costs for maintenance.

In this multi-part series, we will investigate several aspects of centrifugal pump efficiency. First, I will define efficiency and give some examples. Next, I will examine some of the design criteria that ultimately dictate the efficiency exhibited by a particular pump. I will also try to make that somewhat nebulous quantity, known as specific speed, more meaningful. I will illustrate its effect on the shape of a pump’s performance curve and overall pump efficiency. 

Next, I will explain the contributions of individual pump components to a pump’s overall efficiency and show why the combined efficiency of a pump and its driver is the product, not the average, of the two efficiencies. How pump efficiency can be preserved by changing impeller speed rather than reducing it diameter will also be examined. Then I will compare the value of peak efficiency versus the breadth of efficiency over a range of flow. The discussion will end with the importance, or sometimes unimportance, of efficiency as it relates to a particular application or process.

When we speak of the efficiency of any machine, we are simply referring to how well it can convert one form of energy to another. If one unit of energy is supplied to a machine and its output, in the same units of measure, is one-half unit, its efficiency is 50 percent. As simple as this may seem, it can still get a bit complex because the units used by our English system of measurement can be quite different for each form of energy. Fortunately, the use of constants brings equivalency to these otherwise diverse quantities.

Common example of such a machine is the heat engine, which uses energy in the form of heat to produce mechanical energy. This family includes many members, but the internal combustion engine is one with which we are all familiar. Although this machine is an integral part of our everyday lives, its effectiveness in converting energy is far less than we might expect. The efficiency of the typical automobile engine is around 20 percent. To put it another way, 80 percent of the heat energy in a gallon of gasoline does no useful work. Although gas mileage has increased somewhat over the years, that increase has as much to do with increased mechanical efficiency as increased engine efficiency itself. Diesel engines do a better job but still max out around 40 percent. This increase is due, primarily, to its higher compression ratio and the fact that the fuel, under high pressure, is injected directly into the cylinder.
In a condition-based maintenance environment, the decision to change out a hydraulic pump or motor is usually based on remaining bearing life or deteriorating efficiency, whichever occurs first.Despite recent advances in predictive maintenance technologies, the maintenance professional’s ability to determine the remaining bearing life of a pump or motor, with a high degree of accuracy, remains elusive.

Deteriorating efficiency on the other hand is easy to detect, because it typically shows itself through increased cycle times. In other words, the machine slows down. When this occurs, quantification of the efficiency loss isn’t always necessary. If the machine slows to the point where its cycle time is unacceptably slow, the pump or motor is replaced. End of story.

In certain situations, however, it can be helpful, even necessary, to quantify the pump or motor’s actual efficiency and compare it to the component’s native efficiency. For this, an understanding of hydraulic pump and motor efficiency ratings is essential. There are three categories of efficiency used to describe hydraulic pumps (and motors): volumetric efficiency, mechanical/hydraulic efficiency and overall efficiency.

Volumetric efficiency is determined by dividing the actual flow delivered by a pump at a given pressure by its theoretical flow. Theoretical flow is calculated by multiplying the pump’s displacement per revolution by its driven speed. So if the pump has a displacement of 100 cc/rev and is being driven at 1000 RPM, its theoretical flow is 100 liters/minute.