Pumps & Systems, October 2007
In my August column ("Tricks to Taking Flow Measurements in the Field"), I compared the pros and cons of one of several techniques that can be used to estimate pump flow rate when troubleshooting pump operations under less-than-desirable conditions: directly measuring it with a flowmeter.
One of our readers, Don Casada of Diagnostic Solutions, LLC (Knoxville, TN), responded to that discussion with criticism of the direct measurement method from several perspectives, including the correct positioning of the external flowmeter on a pipe with appropriate requirements for a certain length of straight pipe; the absence of obstructions and bends; and similar HI guidelines.
This month, we continue our discussion by examining two other indirect methods of determining pump flow rate in the field:
-
Pressure (head) measurement
-
Power (amps) measurement
In both of these cases, you must obtain the pump performance curve for the respective application, which normally comes in combined or single-line formats, as shown in Figures 1 and 2.
Figure 1. Pump performance curve in a combined format. Source: 2004 Goulds Pumps Manual, ITT Industries
Figure 2. Pump performance curve in a single-line format. Source: 2004 Goulds Pumps Manual, ITT Industries
A combined format curve is typically available from a pump OEM generic catalog, while a single-line curve is usually supplied with specific pump quotation, or better yet, a factory-tested solution. In this example, a red triangle denotes the pump-rated point (70-gpm at 100-ft head) where a pump is expected to operate, but the actual flow is found suspect by operators.
Pressure (Head) Method
Let's assume that the discharge gauge reads 55-psig and the suction gauge reads 10-psig, thus a 45-psi pressure differential exists. This would correspond to 45 x 2.31 = 104-ft head (assuming cold water, SG = 1.0). A horizontal 104-ft head line intersects the H-Q curve (at the proper impeller diameter, which is 5.12-in here in this case) at a little less than rated flow, approximately at 60-gpm.
Power (Amps) Method
The power curve indicates approximately 3.2-hp at the rated point. Power meters (kW-meter) are rarely available, with amps and volts being more commonly displayed at the control panel. Power can be calculated from these readings, although some assumptions of the power factor and motor efficiency would be required:
BHP = (I x V x 1.73 x EFFmotorx PF) / 1000
In our example, a 5-hp 460-V motor is used and we actually read 450-V and 3.9-amps. A typical assumption of the product (EFFmotorx PF) is 0.85, although a somewhat better value can be obtained if one is willing to spend some more time on research work.
Thus, in our example:
BHP = (3.9 x 450 x 1.73 x 0.85) / 1000 = 2.6 hp
This is slightly less than the expected 3.2-hp, meaning a straight horizontal line at 2.6-hp intersects the power curve at flow approximately 50-gpm, depending how accurately you eyeball the curve.
Obviously, too many assumptions and approximations in reading curves bring bad news. However, the good news is that based on two methods, we can state that the flow appears to be somewhere between 50-gpm and 60-gpm. For many troubleshooting purposes, this answer is sufficient.
That's not all. If we also add to this information the approximately 55-gpm data that was registered from the field flowmeter using the technique we discussed (in August) for the less than optimum pipe location, our confidence of the flow actually being somewhere between 50-gpm to 60-gpm will increase even further - a very good thing.
As a note on the power method, some people feel more comfortable simply taking the ratio of actual amps to the motor nameplate (if one is still attached!) amps rating, then multiplying the result on motor rated power. In our example, if the motor rated amps were, say, 8.5-amps, and rated motor power 5-hp, we could then assume the actual power is 3.9 / 8.5 x 5 = 2.3-hp. This is close to the 2.6-hp value we derived earlier by using a power factor and motor efficiency assumption.
The power method can be applied very successfully for field troubleshooting of many pump types, but it has significant drawbacks and cannot be applied for high specific speed (Ns) pumps, such as mixed flow and vertical turbine pumps.
As HI illustrates nearby when comparing impeller profiles for various specific speed designs, pump power is not a nice continuously rising curve, as is the case for most end suction and split case pumps. Instead, the shape of the power curve can be entirely different. It can rise, drop, or stay constant with flow, even making its shape so flat that it becomes difficult to distinguish the difference for a rather wide variation of flows.
The bottom line is that each method has its own place, strength and limitations:
- The pressure (head) method is the simplest and quickest, but requires one to have a pump curve and gauges that are not broken or out of calibration. In the realities of the field, these curves are unfortunately long lost or misplaced for the old pumps, and even if they do exist, it is often impossible to know the most recent impeller diameter inside the pump after numerous prior pump repairs and modifications.
- The power (amps) method does not require one to "get dirty" around the pump replacing broken gauges, but inaccuracy of the power factor and motor efficiency is a drawback. (Reference power factor fundamentals presented by Joe Evans in "Power Factor: Electricity Behaving Badly (Part One)" (Pump Ed 101, Pumps & Systems June 2007)
- Direct flow reading is the most sure way, but most pumps do not have in-line flowmeters installed. Cutting into lines to install them is impractical and expensive. External (ultrasonic) meters are simple, but accuracy is limited due to difficulties in locating a good (HI approved) spot along the pipe of the real field installation.
Often, applying all three methods reduces the error by allowing the user to learn to intelligently interpret the reasons for the differences, be able to explain the peculiarities and inconsistencies of each method, and correct such inconsistencies by solid reason, some understanding of flow mechanics, and reasons for deviations of practice from the theory.
As always, our habit for leaving a parting thought with you: What simplifying assumptions were made in describing the pressure (head) method? What additional errors can these assumptions introduce, and to what magnitude? The first three correct answers get you a winning ticket to our next Pump School session(s). Keep on pumping!