Engineering of Water Systems

Part 13—Vertical Turbine Pump Design (3)

By Ed Butts, PE, CPI

In this installment of The Water Works we will expand upon our example pump selection with some additional technical considerations, as well as examine which option for flow control is best—a variable frequency drive or an inline control valve.

As a reminder, this example was based on the initial use of a vertical turbine pump (VTP) with 150 feet of 6-inch × 1- inch column and lineshaft setting, installed in a 250-foot-deep well and capable of supplying water to a new subdivision with a projected average-day demand of between 35,000-40,000 gallons per day and a maximum day demand estimated between 90,000-95,000 GPD.

In this case, this translates to three separate design flows: 156 GPM for the calculated minimum flow for average day consumption, 392 GPM for the system’s maximum day demand, and 500 GPM for the combined duty of irrigation demand coincident with the maximum day demand.

This will help the designer and client make informed judgments on a potential purchase well before it is actually made.

The majority of each separate demand is estimated to occur over a period of just four to five hours per day. We will detail much of the remaining well pump selection criteria, including the economic analysis prepared to justify the pump selection along with the flow control method.

Two additional points must be made. Usually on most water system designs the source pump (well or booster pump) capacity is based on a singular primary condition of service (COS) since interim storage or additional sources are generally planned to be available to supply any relative shortage of production. However, on a water system design where there is little or no storage and a single source planned such as this example, I try to efficiently plan the well and pump, considering the full range of demands, and design the pump around the mid-range flow with the pump hopefully capable of extending out to the maximum demand and also working comfortably at the lowest projected flow rate.

In this example, I determined designing the VTP for the maximum flow rate of 500 GPM was the most practical. Finally, even though we have tried to design the pump to be as efficient as possible at all three duty points, the reality is water systems do not operate on a specific timetable with predictable flow rates throughout a given day. The fact of the matter is virtually all potable water systems, including this example, can conceivably need to provide anywhere between an almost zero flow rate (usually coinciding with the demand that occurs in the middle of the early morning hours, 2-4 a.m.) up to the maximum demand that can occur during the peak water demand period while grounds irrigation is also required.

In these cases, the system designer must carefully evaluate each individual water system demand and decide if a means of flow control or water storage equipped with booster pumping is required to provide water during the low or high demand periods or if the pump can safely operate at the projected low flows. In most cases some type of interim low-flow provisions must be incorporated into the system design, such as a second pump not only to save wear and tear on the primary well pump and related components, but also to save the higher power costs associated with this type of relatively inefficient operation.

In our example, where the water system will be connected to a supplementary and backup water supply, we have determined the well pump should be designed for all of the projected system demand conditions, although we also know the water system will at times only require a demand of 20 GPM or less. For this reason, as a prudent designer, we have informed our client that an additional device and a moderate volume of water storage, likely in the form of a hydropneumatic pressure vessel, is warranted to be added to the installation to protect the pump and system against these low flows and extreme pump cycling conditions.

The two choices we are examining for flow control for our imaginary client are a variable frequency drive (VFD) or a control valve (CV). A thorough analysis and review of both types of flow control devices were outlined in the October and November 2003 issues of Water Well Journal (“Control Valves vs. Variable Frequency Drives: Facts and Fallacies”). Please refer to these two columns for a full discussion of the backgrounds and associated pros and cons of each method.

The Example Design

In the last installment of The Water Works, from a careful and complete review of four potential vertical turbine pumps, we selected Pump #1 as our well pump. The efficiency of our selected pump is indicated on the pump curve in Figure 1 at the design point (84%) and at each 10% decline of efficiency across the curve. (80% at ~360 GPM; 70% at ~260 GPM; 60% at ~230 GPM; and 50% at ~160 GPM). The second pump curve (Figure 2) also shows the same full speed (60 Hz) conditions but includes a cross-hatched window and system head line to indicate the proposed area of pump performance from the maximum operating head of 260 feet down to the projected minimum system head of 190 feet (comprised of the minimum system pressure of 60 psi [140 feet] + the minimum well lift of 50 feet from static).

Figure 1. Full speed pump curve.

This also corresponds to a pump speed range between a full speed of 1760 RPM down to the projected minimum pump rotation speed of ~1466 RPM with use on a VFD at ~50 Hz. Input horsepower is also shown on each frequency line to indicate the horsepower draw at each speed and performance data point. Figure 3 shows the full speed portion of the curve that indicates a potential total dynamic head (TDH) between the design head of 260 feet up to the maximum head of 350 feet that will be encountered during operation with a control valve, with the horsepower draw at selected points along the curve.

Our well pump, whether on a VFD or with a CV, must be able to deliver any flow rate required within these multiple head conditions. To fully evaluate the best choice for our installation, we used the life cycle cost method. This method is one of many universally used when an engineering analysis is needed to compare options based on various factors—different projected lives, different purchase costs, different maintenance and salvage values, and other different factors like input horsepower and pump efficiency.

It directly compares the projected total costs for an appliance or device based on the total costs accrued over its expected lifetime. The lowest total cost for the stated lifetime of the system or component is generally the preferred option and the choice usually selected.

For our analysis we were able to compare the costs of operating the well pump at a reduced speed vs. operation against a control valve. The fundamental difference between the two types of devices is the total head. When using a VFD, the performance of the pump closely mirrors the affinity laws which state the capacity varies directly with the speed, or Q = RPM2 head varies with the square of the speed, or

So for our example on a VFD, if the original capacity at 1760 RPM was 500 GPM at 260 feet at 39 BHP (brake horsepower), lowering the pump and motor speed to 1400 RPM will produce:

Revised head will be:

Therefore, the corrected pump conditions for 1400 RPM based on the original COS would be: 398 GPM at 164.5 feet TDH at 19.63 BHP. It is important to note this reflects the reduced horsepower the bowl itself will consume and would most likely not be indicative of the entire pump including the motor excitation loss, lineshaft friction, and thrust bearing loss.

The pump efficiency also varies slightly with the speed, but usually not to the same degree as the other parameters. It is obvious from this example that lowering the pump speed by using a variable speed controller can have a significant impact on the operational parameters of the pump. Modifying the pump’s performance by using an inline control valve, though, essentially modifies the pump’s output capacity and discharge head through throttling alone. It also lowers the horsepower from this same dynamic change, primarily in the capacity, although the total output head from the pump does not appreciably vary since the pressure is simply moved to a higher value on the upstream side of the valve.

Thus, using our earlier example: with the original conditions of 500 GPM at 260 feet TDH at 39 BHP, by installing an inline throttling valve to maintain a discharge pressure of 60-70 PSI and assuming the flow is lowered to 398 GPM (to match the 1400 RPM VFD output): From the pump curve: 398 GPM = ~285 feet TDH at 80.5% efficiency.

Revised horsepower is:

In this example it is important to note the general retention in horsepower for the control valve example is mainly due to the type of pump curve (a “steep” rather than a “flat” curve), which simply transfers much of the loss capacity to the higher head in the horsepower equation. The life cycle cost analysis takes all these factors into consideration as the revised efficiency and head is factored at the average daily demand of 156 GPM.

Critical Speed Issues

As with any spinning rotary machine (rotordynamics), sustained operation at a selected reduced or increased speed (RPM) can lead to problems with critical speed. Fundamentally, critical speed is defined as a rotational speed where the natural frequency of the support structure and rotating (rotor) elements (impellers, lineshaft, etc.) approaches or matches the frequency drive operating range. frequency of the exciting forces resulting from the rotational speed of the unit at the same time.

This results in a condition referred to as resonance. Prolonged operation of a rotary machine in resonance can rapidly lead to severe vibration, noise, and undulating forces that can literally cause destruction of the pump in just a few minutes.

These natural frequencies of vibration typically occur in rotor torsional or lateral vibration or structure lateral vibration. While the rotor generally consists of the pump itself, the structure often includes the pump’s foundation, mounting method, piping and piping supports, driver, and shafting. Since pumps are designed for smooth operation at the best efficiency point (BEP), as the pump moves away from the BEP, issues with critical speed can be expected.

There are generally three critical speeds within a given 0- 1800 RPM speed range installation, often referred to as multiples— the first critical speed, second critical speed, and third critical speed. Each critical speed is unique to the installation, including factoring of the lineshaft length, bearing spacings, shaft diameter, load fixity, pump support means, and the actual rotational speed at the moment.

Generally, pumps that operate at reduced speed on a VFD pass through critical speed ranges during acceleration or deacceleration, although this type of pumping design could conceivably operate for an extended interval in any one of these ranges, particularly at low flows where the pump will be operating far from its BEP.

Although this could also become an issue with a submersible pump, the potential for problems by running a VFD on a vertical turbine pump are usually far greater due to the presence of long and relatively small diameter lineshaft found on a VTP.

For our example VTP installation, the VFD corrected pump curve indicates the lowest projected speed our installation should incur is about 1466 RPM (~50 Hz, or a ~17% speed difference). This relatively small change in overall speed is common for reasonably steep curves and advantageous as far as critical speed is concerned. We should only be concerned with the first critical speed.

Most problems with vertical turbine pumps running on VFDs occur when the pump is required to operate for a sustained period at an efficiency at or below 50% of the BEP. Our example pump is expected to operate no lower than around 67% of the BEP, so a critical speed issue is not expected to be a problem. However, each installation is different, owing to the individual mass and potential damping that occurs with the installation itself. Thus every designer is cautioned to evaluate each VTP on a VFD installation separately and incorporate the provisions necessary to avoid critical speed concerns.

Most current production models of VFDs for pump applications are now equipped with adjustable speed rejection bands, where the installer can program the rejection of specific speed ranges into the unit—meaning the pump and motor will not be permitted to operate for an extended period of time within that specific speed range. All concerns related to critical speeds can be addressed by the pump’s supplier or manufacturer at the time of the order placement and designed “out of the system.”

In addition to the critical speed concern, a further examination of our example indicates reconsideration of the lineshaft diameter is also warranted. Although all the design criteria indicate use of 1-inch lineshaft is acceptable, further investigation and a designer’s instinct also shows that the installation would be better served if the lineshaft size (diameter) was increased to 1.25 inches.

This increase in size is performed at a nominal increase in cost and only requires .45 more horsepower for the entire installation, or a total motor load of 40.78 HP—well under the maximum allowable level of 46 HP. The additional strength and rigidity afforded from upsizing the lineshaft to 1.25 inches will pay marked dividends throughout the projected service life of the installation, especially while operating on a VFD.

This drives to the heart of my earlier statement to always go back and review the design factors and always use your experience and judgment when designing water systems and well pumps. Your initial design decision may not be the best or final decision.

Any concern with the alternate use of a control valve involves the potential for higher than desired horsepower draw at low flow, valve cavitation, and thermal heating of the water behind the control valve. In our example, the system design is predicated on the use of interim pressurized storage to avoid continuous operation through the use of either device, so these issues are not expected to be of concern.

Life Cycle Cost Analyses

A life cycle cost (LCC) analysis includes the following elements:

LCC = CIPC + CIN + CE + COL + CM + CDT + CENV +/- CDD where:

CIPC = Initial purchase cost (system or element)

CIN = Installation and commissioning costs

CE = Energy cost to operate

COL = Ongoing operating labor costs

CM = Maintenance costs; labor and materials

CDT = Equipment downtime and loss of production

CENV = Environmental costs

CDD = Decommissioning, disposal, and salvage costs.

Fortunately, I have access to several economic evaluation software programs and spreadsheets to assist with determining the best economic options. In this case I used an electronic spreadsheet.

During the separate life cycle analyses, we included the following factors:

VFD installed costs: 40 HP, 460 VAC, VFD ($5000) + installation and wiring costs ($3400) = $8400.

CV installed costs: 6-inch CV ($1200) + installation costs ($2200) = $3400 (difference over a motor starter).

Pump costs (for pump option #1): $15,900: 10-inch × 7 stage VTP set at 150 feet on 6-inch × 1-inch water lubricated 40 HP motor.

Pump and motor installation costs: $2000 for all pump options (included in installation cost estimates).

Maintenance costs: $1000 per year for all options. No salvage value or decommissioning costs assumed. O

perational hours per day: Assumed five hours per day of operation at reduced flow. This roughly equates to the average daily demand of 35,000-40,000 GPD at 156 GPM.

Power costs (per kilowatt hour): $0.12—no other fees or charges added.

During the individual life cycle cost analysis, the reviews were based on three separate life spans for each device:

In each example, use of a variable frequency drive over a control valve provided the lowest life cycle cost. Once again, it is important to note these relationships applied to this example only and may or may not also apply to other life cycle cost scenarios. Each prospective installation must be evaluated individually and by using the pertinent factors fairly and objectively.

A more efficient pump will generally recover a cost difference over a less expensive pump in enough time of operation from the difference in energy costs alone. The life cycle costs for Pumps #1, #2, and #4 are reflected in Table 1 with VFD operation (using an adjusted TDH) and were based on a 20- year service life. This life cycle cost analysis was conducted to assist with the selection of Pump #1 over the other three possible units. The spreadsheet results are shown in the table.

In a potential installation such as this example, with a 10- inch bowl diameter VTP installed in a 12-inch well casing, there are limited opportunities to use a two-pump system where a smaller pump could be used for the lower demands and as a backup to the larger pump due to size and physical restraints.

Any serious consideration of reverting to no less than a two-pump system with a VTP would require serious consideration of a moderate to large volume atmospheric water storage vessel with one or two booster pumps at a potential cost exceeding $50,000 with the well pump feeding directly into the tank, which may or may not be considered by the client. This type of scenario would generally be met with far more client interest and feasibility if the installation was lower in cost and was performed using two submersible pumps with provisions to allow online swapping of pumps. We will discuss this type of option in the submersible pump example in this series.


Performing an economic analysis—a life cycle cost (shown here), equivalent uniform annual cost, capitalized cost, or 64 ? October 2016 WWJ Control valve Variable frequency drive LCC LCC 10-year life $84,807.27 $80,173.48 (lowest LCC) 15-year life $125,510.91 $116,060.23 (lowest LCC) 20-year life $166,214.55 $151,946.97 (lowest LCC) WATER WORKS from page 62 present worth evaluation—for prospective purchases is a logical and important analysis to conduct for an entire water system design or virtually any separate component.

Beyond the obviously important technical and installation factors (design conditions, pump setting, bowl diameter, pump operational and cycling parameters, pump curve shape), evaluating the potential payback and investment recovery for the system and individual components such as comparing the use of a VFD to a control valve is a critical design function to perform. This will help the designer and client make informed judgments on a potential purchase well before it is actually made.

The process itself is not hard to do. It may take additional time and effort to conduct these evaluations—but so too is providing the client with the most efficient and cost effective system.

As I said, there are numerous spreadsheets and computer software programs, as well as various manual calculation methods, now available to help with performing these vital economic determinations and payback analyses.

This type of analysis can be useful to the people paying the bill regardless of the system component—whether it is the well pump itself, choosing between water storage and multiple pumps, horsepower options, or deciding between a variable frequency drive or a control valve. This type of economic analysis is used for the basic premise of whether to use one device over another for strictly the initial purchase price and power cost reasons. Other considerations should also be included in the decision-making process for using either a VFD on a VTP or submersible pump or for using a submersible pump for a deep well pumping installation over a VTP.

Most of them are technical and related to motor life, electrical losses in the motor, and possible electrical currents through the motor bearings.

We will cover many of these, including modifying the system design to include a pair of submersible pumping units from the well, in the coming installments of The Water Works.

Until next time, keep them pumping!

Ed Butts, PE, CPI, is the chief engineer at 4B Engineering & Consulting, Salem, Oregon. He has more than 40 years of experience in the water well business, specializing in engineering and business management. He can be reached at