Part 14(b)—Submersible Pump Design, Part 2
By Ed Butts, PE, CPI
Designing a water system using a submersible pump is not appreciably different hydraulically than doing so with a vertical turbine pump.
However, the electrical and motor cooling requirements are certainly different with voltage drop to the motor and various other factors becoming much more important. In Part 2 of this three-part series on the design of a submersible pump we will design our pump end using the hydraulic design data to fit the same sample application we previously used for a vertical turbine pump.
The final installment of this series will outline the specific electrical characteristics and factors common to submersible motors.
The Example Installation
Using our example installation in previous columns of The Water Works, we have successfully designed a sample water pumping installation for a vertical turbine pump with the following conditions:
- Conditions of Service (COS)
- Required primary design capacity: 500 GPM
- Low design capacity: 156 GPM
- Required design head (TDH): To be determined
|Components of the total dynamic head:||at 500 GPM||at 156 GPM|
|Well lift||90 feet||62 feet|
|+ Design operating pressure = 60 psig||140 feet||162 feet (at 70 psig)|
|+ Network loss = 6 psig||14 feet||7 feet (at 3 psig)|
|+ Estimated 6-inch riser pipe and check valve hf||10 feet||3 feet|
|Estimated TDH||254 feet, use 260||234 feet, use 240|
Note: The 6 inches of head has been added to both TDH figures to compensate for potential additional well lift.
- Final primary COS = 500 GPM at 260 feet TDH.
- Alternate COS = 156 GPM at 240 feet TDH.
Depth: 250 feet
Diameter: 12 feet
Casing depth: 200 feet
Well casing type: Steel
Well screen size: 12 inches T.S.
Well screen interval: 200-245 feet
Well screen slot(s): 100 slot
Static water level: 50 feet
Well seal depth and type: 100 feet/cement
Well yield and pumping level: 1000 GPM at 165 feet PWL (max.)
750 GPM at 130 feet PWL
500 GPM at 90 feet PWL
300 GPM at 72 feet PWL
Sand production: None measured at full capacity
Well alignment: 16-inch uniform offset over 250 feet in one direction
Design flow rate(s): 500 GPM (primary) 156 GPM (alternate)
Discharge pressure: typically 60 psig, Range: 35-70 psig
Network loss: 10 psig at 1000 GPM
Network loss: 6 psig at 500 GPM
Network loss: 3 psig at 200 GPM
Required fire flow: 1000 GPM at 25 psig
Distribution pipe: 8 inch PVC, I.P.S. Class 160 psi
Supplementary water: Provided from adjacent city
Storage: Provided from onsite pressure tank
Well access: Clear, off private road
Weather protection: Low temperature—10°F
Weather protection: High temperature—90°F
Onsite sanitary and storm discharge
Now that we have completed the determination of the required pump capacities and operating head, the first step in actually selecting a submersible pump end is to estimate the target horsepower required for the design conditions. From Part 1 (WWJ, January 2017):
BHP = GPM × TDH = 500 GPM × 260 feet TDH = 43.77 BHP
3960 × PE(1) 3960 × .75
(1) PE = pump efficiency [at flows < 200 GPM, use .65 (65%); at flows > 200 GPM, use .75 (75%)]
BHP = 156 GPM × 240 feet TDH = 14.54 BHP
- × .65
From the above brake horsepower (BHP) estimates, it is apparent there will be a wide disparity of required horsepower (almost 30 BHP) between the two operating points. Generally, an application that requires two operating points so far apart requires strong consideration of either using a variable frequency drive (VFD) to use the affinity laws to lower the pump and motor speed for the alternate condition, an inline control valve to regulate outlet pressure and pump discharge, or a pump with an extremely flat head-capacity (H-Q) curve.
In our example, the use of a VFD has previously been determined to be the most cost effective solution, although an inline control valve could also have been used. However, it is highly unlikely the desired horsepower of 14.5 BHP at the alternate COS would have been met as the majority of submersible units have steep operating curves, owing to multiple stages plus the pump speed of 3450 RPM.
The final option, use of a flat curve pump, would also be unlikely as a preferred choice. Again, this is due to the pump’s rotation speed and number of pump stages required, except for the possible use of a 10-inch or 12-inch-diameter pump, which would require fewer stages than a smaller unit.
Now that we have completed Step 1 from our suggested Table 1 checklist in Part 1, it is time to move into the next steps.
Step No. 2: COS to available bowl diameter
Although the hydraulic design is primarily vested in the pump’s capacity and head, the bowl diameter is also a critical factor. With a submersible pump, the bowl diameter is generally dictated by two primary conditions: the required pumping rate needed (in gallons per minute) and the limiting diameter of the well casing or wet well the bowl will be placed into (the maximum bowl outer diameter [O.D.]).
Bowl O.D. Maximum/Minimum Flow Range (GPM) at BEP
|3450 RPM||1760 RPM|
|4 inches||100/5 GPM||75/10 GPM|
|5 inches||250/30 GPM||125/30 GPM|
|6 inches||450/30 GPM||200/30 GPM|
|7 inches||650/100 GPM||400/90 GPM|
|8 inches||1200/175 GPM||900/100 GPM|
|9 inches||1600/250 GPM||1200/150 GPM|
|10 inches||1800/300 GPM||2000/250 GPM|
|11/12 inches||2000/400 GPM||3500/300 GPM|
|13/14 inches||2200/500 GPM||4000/600 GPM|
|15/16 inches||N/A||5000/1000 GPM|
|17/18 inches||N/A||7000/1500 GPM|
Table 1. Maximum and minimum design capacities at best efficiency point (enclosed or semi-open impellers).
Table 1 cites the general maximum and minimum flow rates (including speed reduced minimum flows) for various bowl diameters at their respective best efficiency point (BEP) from various manufacturers for 3520/3450 RPM (3600 RPM synchronous speed) and 1760/1725 RPM (1800 RPM synchronous speed) rotational speeds.
The maximum rated capacity for each bowl diameter and speed are based on the typically highest BEP from various manufacturers, while the minimum flow for each bowl diameter and 3600 RPM speed represents an approximate maximum pump and motor speed reduction of 40% from the practical BEP at the lowest rated flow rate for each diameter and speed.
This would approximate correction of the performance of a submersible pump and motor when used on a VFD or when used with a control valve to maintain a minimum motor speed of 36 hertz (~2100 RPM) to maintain proper motor cooling and bearing lubrication, well above the manufacturer’s typical minimum of 30 hertz (~1750 RPM).
Vertical turbine pumps (VTPs) do not generally operate with the same flow range limitations as submersible pumps and motors. Therefore, the range of allowable flow rates with a VTP is often greater than that with a comparable submersible unit. The vast majority of 6-inch and most 8-inch-diameter submersible pump motors below 100 HP operate at a two-pole speed, or 3600 RPM. Therefore, this example pump selection should also be conducted using that same speed.
Given the knowledge of the primary and secondary (alternate) design capacities (500 GPM and 156 GPM) and the well diameter (12 inches) creates a fairly easy determination of the bowl diameter. From Table 1 it is apparent either a 6-, 7-, 8-, 9-, or even a 10-inch-diameter bowl assembly at 3450 RPM will likely work for this application with a 6-inch-diameter bowl at the extreme end of its practical and efficient flow range for 500 GPM.
The minimum recommended flow for a 10-inch-diameter bowl is also above the BEP for the low flow of 156 GPM and will most likely only require two or three stages to produce the needed head, which will result in a flatter total head curve. This is generally not as desirable for use with a VFD and compromises the pump efficiency and optimum clearance inside a 12-inch-diameter well by using a 10-inch-diameter bowl. This tends to limit the best overall selection to a 7-, 8-, or 9-inch-diameter bowl.
As opposed to VTPs, submersible units are often available in two types of selection procedures:
- Building a unit through an analysis of a per-stage performance of individual stages (Figure 1), as with a VTP, and then dividing the total head required by the head per stage to determine the number of stages and horsepower needed to create an assembled pump.
- Evaluating a manufacturer or supplier’s preselected and preassembled units and then selecting a pump that comes closest to the required flow and head (Figure 2).
When using a single-stage performance curve to evaluate a potential submersible pump, always be cognizant multi-stage pumps almost always display a higher efficiency at the same operating point, impeller trim, and capacity than a single-stage unit, so an efficiency correction may be needed.
For example, the single-stage bowl assembly shown in Figure 1 is the same bowl assembly with the same impeller trim (4.875 inches) and nominal speed (3600 RPM) as the 4-stage bowl assembly in Figure 2. However, the efficiency is three points higher (77.9% vs. 74.9%) for the 4-stage bowl. This relationship holds true for both VTPs and submersible pumps.
Usually, if any correction is required for multiple stages, this is generally indicated on the pump curve itself. In many cases this type of unit is further classified by the bowl’s BEP design flow and/or motor horsepower, especially when stainless steel impellers are used. Stainless steel impellers are not as easy to trim. Therefore, knowledge of the well diameter and the approximate required horsepower will often provide a shortcut to a pump selection.
For our example, 43.77 of estimated HP translates to the probable need for a 50 HP motor. This could provide the information required to select a preassembled submersible pump with a rating of 500 GPM and a 50 HP motor. This procedure is often shown on pump selection data sheets or curves with nomenclature to indicate the bowl diameter first, followed by the pump’s rated capacity or relative rating, the number of stages, or the motor horsepower.
For example, a specific manufacturer may use a model number such as 7TLC, 7CHE, 8RJO, or 8M23. The first number (7 or 8) usually signifies the bowl’s outer (nominal) diameter. The second and/or third letter (TL, CH, RJ, or M) may designate the manufacturer’s bowl capacity or head rating, such as L for low, M for medium, or H for high. The final letter or number often describes whether the impeller is an open (using an O) or enclosed (E or C) impeller. The use of a specific number (as in 23 for 230 GPM) may indicate the bowl’s rated capacity at its BEP. In some cases, “S” is inserted into the model number to signify the unit is a submersible pump.
Finally, the number of stages and the horsepower rating is often applied to the end or sometimes as part of the model number. A complete model number for an assembled submersible pump unit, for example, may be an 8SHHE-7-100, to signify an 8-inch nominal bowl diameter submersible pump, with a high capacity and head rated enclosed impeller, equipped with seven stages, and a 100 HP rated horsepower motor. In all cases, you should verify the breakdown of a specific model number with the manufacturer as many pumps do not follow these criteria.
Occasionally, I receive a request from someone to design a submersible pump using a semi-open impeller. Although I have used this type of impeller numerous times on VTPs, I do not routinely use them on subs for several reasons.
First, since they are locked onto the pump shaft and often situated several hundred feet down a well, they cannot be adjusted to regain performance or efficiency without pulling from the well. Secondly, although semi-open impellers are often a few points higher in efficiency, they usually display more axial and radial thrust than enclosed impellers, making them undesirable for use on the lower thrust rated submersible motor.
Finally, designing an application using semi-open impellers is at best an estimation since the pump’s performance and horsepower draw is primarily a factor of the impeller’s proper running clearance from the bowl. Any variation to this clearance from the manufacturer’s published curve data will adversely impact the performance by underperforming, overperforming, over possibly overloading the motor.
The seven submersible pump ends in Table 2 represent only a fraction of those available for the primary conditions of service. However, these potential selections nonetheless represent a cross-section of the typical bowl diameters and number of stages to consider for this application.
The final determination of the selected pump must weigh several factors. Some are universal while others may be site or locality specific. And since it is fairly obvious all the selected bowls will fit inside the 12-inch well casing, this initial factor can be ignored.
The next selection criteria I generally examine is the BHP requirement and pump efficiency at the specific operating condition that will be subject to the highest use. In our example, even though the bowl assembly has been designed for a primary design condition of 500 GPM, in actuality the pump will usually operate somewhere between the two design points with the alternate COS (156 GPM) in service much more than the primary COS. The three units with the lowest BHP requirement at the alternate design condition in the table are pumps R-1, F-1, and M-1.
My next criteria, particularly since the efficiency at the primary COS is close for all units, is to evaluate the operating speed at the alternate COS. This is more important to the success of the installation than one might imagine, particularly when a VFD will be used for motor control. Most motor (and some pump) manufacturers dictate the motor speed shall not fall below 50% speed (30 hertz, or ~1750 RPM). This is to provide adequate bearing lubrication in the motor as well as maintain enough velocity past the motor for cooling. As previously stated, when feasible, I prefer to design an installation so the pump and motor will not exhibit a minimum speed below 40% of the motor’s rated speed or about 2050-2100 RPM.
Tempered with this fact, however, is the knowledge the motor must be permitted to operate at a low enough speed to facilitate a reasonable VFD operating range and proper control settings. Experience has taught me this factor works best for a multi-stage submersible unit when using a shutdown speed between 70%-90% of full load (FL) speed—a range of 75%-85% of FL motor shutdown motor speed often works the best. Obviously, all these criteria must be ascertained after a full evaluation of the pump curve (flat vs. steep) and HP at the minimum speed.
From an examination of the pumps in Table 2, it is evident all the sample pumps fulfill these desires, with pumps G-1, R-1, L-1, G-2, and F-1 fitting the best at the reduced flow of 156 GPM along with a reduced speed range between 75%-85% of FL speed.
Finally, when cost is a factor, weighing the individual options for the lowest initial and operating cost is often conducted. For this final evaluation, pumps R-1 and F-1 were both good choices, although my ultimate selection was for pump R-1 as the runout capacity (Q = 575 GPM) was lower than F-1. The BEP for the R-1 pump was also slightly to the left of the primary design point which helps to retain higher operating efficiencies at lower speeds, plus the pump was represented locally and replacement parts were more readily available.
The runout capacity of ~575 GPM is an important selection criteria for this example to avoid excessive well overdraft, especially since flows above 500 GPM will be served from a supplementary source. See Figure 3 and Figure 4 for full speed and variable speed curves for pump R-1, a 7-inch × 3-stage bowl assembly.
Now that the pump end has been selected, we generally examine any special construction or metallurgy required for the pump end. If sand or abrasives were a concern, bowl wear rings might be warranted. If the bowl’s upper stages were exposed to excessive high pressures, O-rings or gaskets may be indicated to prevent inter-stage leakage. Since most 6- and 8-inch bowl assemblies are constructed using threaded construction between stages, this would usually not be a concern unless a 10-inch or larger diameter bowl was selected and only then with pressures in excess of the manufacturer’s pressure rating.
Next, static or dynamic balancing of the impellers should be considered. On one side, this pump uses a stock pump with only three stages and fairly small diameter (4.875 inches) impellers, plus we plan to use a 7-inch-diameter bowl inside a 12-inch well casing, so balancing of the impellers is probably not needed. However, the added cost for balancing just three impellers does not generally represent a huge added cost to the bowl assembly. So if there is any concern regarding the well’s alignment, this may be a desirable option.
Finally, many firms feel all larger capacity units should be factory tested to verify performance. Although I often require factory testing for expensive or large or deep well pumping units, submersible and vertical turbine, I rarely require or recommend factory testing on smaller (<10 inches) wet ends for various reasons.
Besides the added cost (which can actually cost more than the pump itself) and the associated time delay, experience has shown conducting factory testing on smaller diameter, multi-stage pumps does not generally result in any true power savings or added assurance to the owner, especially since the selection curves from the majority of pump manufacturers have repeatedly been shown to be accurate for capacity, head, and horsepower. Also note the majority of submersible pump models between 6 and 12 inches have been tested by their manufacturer during development.
Remember, any of the above concerns will usually result in not only added cost but a delay in constructing and shipping the unit, so consider these carefully. For our example, none of these concerns are present, so the final factor would be the pump setting and drop pipe size.
Drop Pipe Size and Pump Setting
The riser drop pipe and check valve sizes depend on various factors: the desired minimum and maximum uphole velocities (critical when using a VFD); friction losses; pipe cost; well and drop cable clearance; and adequate space for any additional elements in the well/pipe annulus—sounding tubes, well or water level measurement devices, future chemical treatments. For our example installation, the pump will operate within a range of 156 GPM minimum, up to 500 GPM maximum.
Figure 5b shows flow ranges for various pipe sizes, and indicates a desired flow range of 160-700 GPM for 5-inch drop pipe. These values will maintain a minimum uphole velocity of ~2.5 feet per second (FPS) at 160 GPM to transfer any heavier solids from the well to prevent settlement onto the pump, with a maximum uphole velocity of 10 FPS at 700 GPM as an upper economic sizing. Figure 5b shows friction loss for new 5-inch steel drop pipe at 500 GPM is 4.16 feet per 100 feet of pipe with a velocity of 8.02 FPS.
Figure 6 and Figure 7 are included for those who use either PVC or flexible hose as drop pipe. We will finalize the friction loss calculation upon determining the pump setting.
The desired pump setting is truly a case-by-case determination that must be performed with full knowledge of the well casing diameter and depth, well screen upper termination depth, the reliable pumping water level (PWL), sand or abrasives pumping potential, and a reserve (safety) factor for unusual well drawdown or seasonal drafts.
Typically, I like to plan deep well installations to maintain a minimum of 10 feet of submergence over the top of the pump, not the suction or inlet, at the maximum projected pump capacity. For our example, this translates to a minimum pump setting of 110 feet (~100 feet PWL at 515 GPM + 10 feet of submergence).
In cases such as this example, when more depth is available I prefer to set the pump at least 20 feet deeper in the well to compensate for any future well decline or seasonal shifts. So, I will plan for 150 feet of drop pipe. This results in the use of seven lengths (147 feet at 21 feet per length) plus a 3-foot length for the upper well seal to provide an easy-to-remember figure for future reference.
The total friction loss would therefore be 4.16 feet/c × 1.5 (for 150 feet) = 6.24 feet + 1.60 feet for the riser check valve and miscellaneous loss = 7.84 feet total, well under our original estimate of 10 feet. The riser check valve should be placed between 5 to 20 feet above the pump, with 10 to 11 feet (the first full 21-foot joint above the pump cut and threaded in half), my usual location. This places the check valve at a sufficient distance above the pump to allow for the vertical movement of any entrapped air or vapor from the pump to the check valve and avoid air locking of the pump. It also provides two shorter pipe lengths to enable easy installation of the pump/motor and the well seal.
Also, remember in our example and any installation with a VFD, the riser check valve should be designed for continuous operation on a VFD to prevent possible valve chatter and premature failure. The proposed finished installation, along with alternate wellhead completion techniques, are shown in Figure 8.
This concludes Part 2 of the three-part series on sizing of submersible pumps for municipal, industrial, and commercial water supply. In Part 3, we will conclude with a detailed discussion of sizing and installing the submersible motor and drop cable, plus some of the pitfalls and issues associated with using this type of pump and motor.
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 firstname.lastname@example.org.