White Paper

Meeting The Challenges Of Saving Energy In Water Pumping


By Tom Walski and Tony Andrews, Bentley Systems

The phrase "water-energy nexus" seems to be on the lips of virtually everyone around the water market today. It takes a lot of water to produce energy, but it also takes a lot of energy to treat and move water around – and the demand for one could soon cripple our use of the other.

Except for utilities that obtain their water from mountain reservoirs, energy consumed by pumping is one of the largest cost items in most water utility budgets, accounting for, in some cases, 95 percent of a water utility’s electrical demand. Water utilities are some of the largest users of electricity. In the United States for example, the U.S. EPA states that between 3 and 4 percent of all electricity produced is consumed by water utilities. This equates to 56 billion kilowatts costing $4 billion per annum.

Water utility operators are therefore always looking for a magic button that they can push to reduce electricity usage in their water pumping operations. Unfortunately, there are quite a few reasons why utilities don’t operate at perfect efficiency. Operators need to first identify where energy is being wasted before they can reduce their energy use and energy costs.

A utility must pump Q gallons of water against a head of h feet with a peak efficiency of e and thus there is a certain minimum amount of energy they must use. Most utilities use more than that for a variety of reasons and it therefore becomes necessary to look for wasted energy as pumps don’t know if they are wasting or using too much energy.

The usual problems that can be identified are:

  • Mechanical wear (particularly seal failures) on the pump;
  • Pumps that were poorly selected;
  • Systems that have changed so that the pump operating point moves;
  • Poor choice of pump operation, which can be a result of poor combinations or poor timing.

Each of these problems needs to be diagnosed and solved separately. Pumps in excellent condition are being operated at the wrong time, or they may have been sized correctly when new but the system has changed so much that they are now operating inefficiently. This article provides suggestions for identifying and solving these problems.

Mechanical Condition

When a pump is new, it should operate with heads and efficiencies given in the manufacturer’s pump curves. However, because a pump is basically a mechanical piece of equipment, it wears out over time depending on its operating context. What results is efficiency deteriorating due to bearing wear, impeller damage, clearances that are out of specification, corrosion in the casing and a variety of other reasons, which ultimately cause the pump to perform poorly. Measuring the pump’s flow, head, and efficiency over a wide range of conditions is an important first step in determining whether the pump’s mechanical performance has deteriorated. In some cases, cleaning and coating the casing interior or adjusting the clearances is all that it takes to restore performance. In other cases, replacing the pump with a newer pump with a higher efficiency is the most cost-effective solution.

Water utilities are beginning to turn to asset performance management methods such as reliability-centered maintenance (RCM) and condition-monitoring programs to identify failure modes, and causes and effects in order to proactively identify possible functional failures and employ a predictive strategy to optimize maintenance spend. Incorporating these techniques in a pump management strategy can improve equipment running status, reduce operating cost, extend operation life and control, and optimize cost and output over the lifetime of the asset.

Actual Operating Points

Centrifugal pumps in water systems do not produce a fixed flow, but rather the pump flow depends on the pump and the conditions of the system, which can change over time. As part of the pump testing, the pump operating points over a wide range of conditions should be measured and compared with the expected operating points. The expected operating points can be determined using tools available in some water distribution system models. The model can determine the system head curve, which embody the hydraulics of the system. The intersection of the system head curve and the actual pump head curve will yield the expected operating point. If this expected operating point does not agree with the measured operating point, the operator needs to determine the cause. Has the system changed since the pump was installed (e.g. new tanks or pipes) or has someone left an important valve closed? Figure 1 shows a typical pump head and efficiency characteristics curve.

Figure 1: Pump Operating Points

Figure 1 illustrates that the pump should be discharging 210 gallons per minute at a head of 118 feet and an efficiency of 66 percent. However, if the pump is actually operating at one of the green points, it is only operating at 61 percent efficiency and at one of the red points; it may be operating at only 42 percent or 63 percent efficiency. It is only by making these measurements and comparing them with where the pump should be operating on its curve can an operator understand that energy is being wasted and can begin to look for solutions.

Remember that a system head curve is really not a single curve but a band of curves that the pump experiences depending on water levels in tanks, demands, and the status of other pumps and control valves in the system. Figure 2 shows how the system head curve and hence the operating point of a pump can vary over the course of a day. The operating point is the intersection of the brown pump head characteristic curve with the system head curve at that time. In this example, the flow from a given pump can vary from 230 to 255 gallons per minute depending on conditions as manifested in the changing system head curves.

Figure 2: Variation in system head curve

Understanding the Energy Tariffs

Most consumers won’t buy something unless they understand the price, yet many water system operators don’t understand the tariffs under which they purchase energy. There is good reason for this, as tariff documents are often confusing, but it is worth the effort to understand them. Energy cost is in most cases not a simple case of energy used times unit price of energy. Costs are complicated by peak demand charges, time-of-day price differences and block rate pricing, all of which mean that minimizing energy use does not necessarily mean minimizing cost of energy. In particular, with peak demand charges, a single 15-minute period, where an operator makes some poor decisions, can overwhelm cost savings in other aspects of operation.

Reviewing the Energy Bill

The result of the tariff and energy usage is the bill that the water utility must pay. Many operators never see this bill, yet it provides insight into possible savings. With specialized hydraulic modeling software the operator is able to calculate what the energy bill should be and compare it with what the utility actually pays. Then using the hydraulic model, alternative operating strategies can be tested to determine the best way to operate the system.

Constant vs. Variable Speed Pumping

Variable speed pumping is often touted as a means to save energy. However, the variable speed drive introduces inefficiency and the cost of the variable speed drive is not insignificant. Variable speed pumping is most often justified in a system with no storage when a pump cannot be shut off. When a distribution network has storage, it is possible to fill the tank with a constant speed pump running at an efficient operating point and turn it off. In general, slowing down a variable speed pump reduces its efficiency, as the most efficient speed at which to run a variable speed pump is OFF. A common mistake is sizing a variable speed pump to run efficiently at peak flow only to have that pump run at a terrible efficiency 90 percent of the time. Using a hydraulic model, which can accurately calculate energy use and cost, can identify which type of pumping is best.

Pump Combinations

Pumps that may run efficiently when running alone may not run efficiently when running in parallel with other pumps. This is especially true when the pumps are not identical. When flows are low and the pumps are running at the flat portion of the system head curve most pumps work well together. As flows increase, pumps with a higher head at their best efficiency point tend to “shut off” lower head pumps. In Figure 3, with one pump running (blue head characteristic curve), the flow will be 200 gallons per minute and the efficiency will be 70 percent. However, when two pumps are running (green head characteristic curve), the flow only increases to about 270 gallons per minute and the efficiency drops to 63 percent. The cyan and brown curves represent the range of system head curves.

Figure 3: Operating points and efficiencies with one and two pumps operating

It is helpful to identify which pumps or combination of pumps work best for a given station discharge. Figure 4 shows a situation where different combinations have their peak efficiency at different flow rates for a complicated pump station. A small pump running alone may be efficient at 2,000 gallons per minute but at 12,000 gallons per minute a combination of three large pumps may be needed.

Figure 4: Efficiency curves for different pumps and pump combinations

Don’t pump to a PRV

Pressure reducing valves (PRV) waste energy as they reduce pressure to acceptable levels. Utilities certainly don’t want to pump into PRVs yet there are many situations where utilities pump to a higher pressure zone only to burn up that energy allowing the water to flow back downhill into lower zones. In some situations, operators may not realize how much energy is being wasted because the PRV is located far from the pump. In complicated systems with many pressure zones, it pays to look for these energy-wasting situations and eliminate them when feasible. In general, once water is raised to a high-pressure zone, it shouldn’t come back down, except in an emergency.

Storing water below hydraulic grade

Usually water storage tanks are placed at a level where the water level in the tank is the same as the hydraulic grade line in that pressure zone and water can freely move in and out of the tank. For a variety of reasons (e.g. zoning and lower construction costs), some tanks are constructed with their water level below that of the local hydraulic grade. This means the energy lost when filling the tank must be resupplied when pumping out of the tank. This should be avoided but when pumped storage is necessary, it is best to keep the tanks as full as feasible, while still maintaining adequate turnover in the tank. For example, in a study for the Dallas-Fort Worth Airport, Freese and Nichols, a multi-discipline consulting firm, found significant energy savings by installing an elevated tank to supplement ground storage tanks that more than paid for the tank. Freese and Nichols reported that, “The energy cost analysis provided by WaterGEMS is evidence of a positive cost-to-benefit ratio in energy savings over the life of the elevated tank.” Their estimated reduction in energy consumption was approximately 51 percent, which would save about $117,000 in energy costs per annum.

Reduce water loss

Pumping water simply to have it leak out of the system loses not only the water but the energy used to pump the water. Any reduction in leakage will be accompanied by a corresponding reduction in pumping energy.

Pump scheduling

The amount of energy needed to pump water to customers is pretty much the same whether the pumps turn on at 9 a.m. or 10 a.m., and the amount of energy saved by analyzing pump scheduling is usually marginal. However, there are some situations where the scheduling of pumping does make a difference in energy cost.

When there is time-of-day energy pricing, pumps should run as much as possible during times when the cost of energy is low. The utility is not only storing water in tanks but also the energy used to lift the water. When multiple lifts to different pressure zones are involved, the water in the highest pressure zone costs the most to pump and has the greatest need to be pumped during off-peak hours.

When long pipelines are involved with most of the energy used to overcome friction as opposed to lifting the water, it is best to pump as steadily as possible to minimize friction loss because friction loss varies as the flow squared.


So a pump that at first sight appears to have a simple function operating in a straightforward way in fact turns out to be a greedy consumer of energy, can be costly to run and has plenty of scope to be operated inefficiently. This article has attempted to show that there are numerous options to investigate and identify why a pump may be operating inefficiently. This is hopefully good news as there is plenty of scope to improve the situation. But of course, because there are many ways that energy can be wasted in water system pumping, there are many analyses that need to be conducted to determine the source of wasted energy and perform the economic analysis to determine the best solution. Using analysis tools available in hydraulic models can help in identifying problems, studying alternative solutions and quantifying the benefits of the solutions.

Tom Walski, Bentley Systems, Senior Product Manager

Tom is a veteran of the water industry. He is a widely respected expert on hydraulic modeling and has authored some of the industry’s definitive text books on how to use software to improve the design and operation of water infrastructure.

Tony Andrews, Bentley Systems Solution Executive for Water and Wastewater

Tony has worked in the water industry in Europe and North America for the past 25 years working in hydrology, GIS, hydraulic modeling and asset management for a variety of research institutions, consultancies, GIS and software companies.

Bentley Systems: http://www.bentley.com/en-US/

Image credit: "Hydraulic water pump," © 2009 Elsie esq., used under a Attribution 2.0 Generic license: http://creativecommons.org/licenses/by/2.0/