How to Verify a Pump Curve Against Its Operating Point
This is for the engineer holding a manufacturer pump curve who wants to know whether the machine is actually running where the paperwork claims. You do it by finding the operating point symbolically, where the head the pump produces equals the head the system demands at the same flow, then comparing that predicted point to what the field instruments read. It is a paper-and-pencil check you can do before you ever touch the pump, and it tells you whether a low-flow or high-amps complaint is a real fault or just physics.
Verify Pump Curve vs Operating Point in one line: To verify a pump curve against its operating point, write the pump head as a function of flow and the system head as static lift plus a resistance term that grows with flow squared, set them equal, and solve for the flow where the two curves cross. That intersection is the only place the pump can run. Compare its predicted head, flow, and implied power to the discharge gauge, the flow meter, and the motor amps in the field; a match confirms the curve, a gap points to wear, a throttled valve, or a wrong assumption.
Write the Pump and System Heads as Functions of Flow
A pump can only ever run at one point: the flow where the head it develops exactly equals the head the piping demands. To find it on paper, express both as functions of the same variable, flow Q. The pump curve, read off the manufacturer sheet, is well approximated near the duty region as H_pump = H0 - a*Q^2, where H0 is the shutoff head at zero flow and a is a coefficient you fit from any second point on the published curve. If the sheet gives you shutoff head and one rated point (Q_r, H_r), then a = (H0 - H_r) / Q_r^2. That is the whole pump side in one expression.
The system side is the head the piping asks the pump to overcome, and it has two parts. A fixed part, the static head, is the vertical lift plus any pressure difference between source and destination that does not change with flow. A flow-dependent part, the friction head, rises with the square of flow because turbulent pipe loss scales that way. So H_sys = H_static + k*Q^2, where k rolls up the resistance of every foot of pipe, elbow, and open valve in the path. You get k from a design hydraulic calc or by back-solving from one known operating condition. Two clean quadratics, one falling and one rising, are all the algebra this check needs.
Reading the curve correctly matters here, so if the shape of that falling line is not familiar, the explainer on what a pump curve is covers how head, efficiency, and NPSH-required are stacked on the same sheet. The rising system line is treated on its own in the note on what a pump system curve is. This procedure assumes you can pull H0, one rated point, and the static head from those two sources.
Solve for the Intersection Symbolically
Set the two expressions equal and the flow falls out cleanly. H0 - a*Q^2 = H_static + k*Q^2 rearranges to Q^2 * (a + k) = H0 - H_static, so the operating flow is Q_op = sqrt( (H0 - H_static) / (a + k) ). Substitute that Q_op back into either curve to get the operating head H_op. You now have the single point (Q_op, H_op) where this specific pump, in this specific system, must sit. Every downstream check compares field data to those two numbers.
Work a concrete example so the shape is obvious. Say a pump has shutoff head H0 = 200 ft and a rated point of 400 gpm at 160 ft, so a = (200 - 160) / 400^2 = 40 / 160000 = 0.00025 ft per gpm squared. Say the system has H_static = 120 ft and a design k = 0.00015. Then Q_op = sqrt( (200 - 120) / (0.00025 + 0.00015) ) = sqrt( 80 / 0.00040 ) = sqrt(200000) = about 447 gpm, and H_op = 200 - 0.00025 * 447^2 = 200 - 50 = 150 ft. The pump should be pushing roughly 447 gpm at 150 ft of head. That predicted pair is your yardstick.
Notice what the algebra tells you before any field trip. If the static head rises, say a tank fills and back-pressure climbs, the numerator shrinks and the operating flow drops along the pump curve. If someone throttles the discharge valve, k rises, (a + k) grows, and again flow falls while head climbs toward shutoff. The symbolic form makes those cause-and-effect links explicit, so when the field reading disagrees with the prediction you already know which term to suspect.
Compare the Predicted Point to the Field Readings
Now take the prediction to the machine. Convert H_op to a pressure the discharge gauge can show: pressure in psi is roughly head in feet times specific gravity divided by 2.31, so 150 ft of water is about 65 psi at the discharge, minus the suction pressure to get the differential the pump actually adds. Read the flow meter for Q_op. Read the motor amps and compare to the power the curve implies at that point. When discharge differential, flow, and amps all land near the predicted values, the pump is on its curve and healthy, and any complaint is a system-design question, not a pump fault.
A disagreement is diagnostic, not just a failure. If measured flow and head are both low and sitting to the left of the predicted point, the system resistance is higher than assumed, usually a throttled or fouled valve, a closing check, or a plugged strainer raising k. If the pump makes less head than the curve at the measured flow, the impeller or wear rings are worn and the whole pump curve has sagged below the published one, which the note on what a pump wear ring is explains as widening internal clearance. If amps are high for the flow delivered, suspect a mechanical drag or a specific-gravity assumption that was wrong.
Because the check is only as good as the readings, verify the instruments themselves are trustworthy before you trust the conclusion. A discharge gauge that has never been field-checked can send you chasing a pump problem that is really a gauge problem, so the routine in how to field-check a pressure gauge is worth running first. With sound instruments the symbolic operating point becomes a reliable acceptance test you can repeat any time the pump is suspected of drifting off its curve.
Common Mistakes
The most common error is comparing the pump curve to a single field flow without accounting for the system curve at all, then declaring the pump bad because it is not making rated flow. A pump almost never runs at its rated point; it runs at the intersection, which is usually a different flow entirely. Always solve for where the two curves cross before judging the machine.
The second recurring mistake is mixing units and reference planes. Head is in feet of the actual fluid, not feet of water, so a heavier or lighter specific gravity shifts every pressure conversion. Suction pressure must be subtracted to get the pump differential, and a gauge reading absolute versus gauge pressure will throw the whole comparison off by an atmosphere. Pin down units and references first, then the algebra is trustworthy.
Frequently Asked Questions
Why does a pump run at the curve intersection and not at its rated point?
A pump develops whatever head the system demands at a given flow, and the system demands more head as flow rises. The only self-consistent operating point is the flow where pump head equals system head, which is the intersection of the two curves. The rated point on the datasheet is just one point the manufacturer chose to publish; the installed system almost always crosses the pump curve somewhere else, so the real duty point differs from the nameplate rating.
What does it mean if the pump makes less head than its curve at the measured flow?
The pump curve itself has sagged below the published one, which means internal wear. As wear rings and impeller clearances open up, more fluid recirculates inside the pump and less reaches the discharge at any given speed, so the whole head-versus-flow line drops. If the head deficit grows over time on a trend, that is degradation you can plan a rebuild around rather than an installation error.
Can I verify the operating point without a flow meter?
Partly. If you have a reliable discharge and suction pressure you can read the head the pump is making, then walk that head back onto the pump curve to infer the flow the machine must be passing, which is a common field trick. It is less certain than a real flow measurement because a worn pump reads off a curve you can no longer trust, but combined with motor amps it usually narrows the operating point enough to confirm or reject the prediction.
Automation services
Need help turning this into a working system?
Merobix integrates SCADA, programs Allen-Bradley and Siemens PLCs, and designs and fabricates industrial control panels.
Meeting requests are reviewed before confirmation.