VFD Pump Efficiency: System Head Is the Limit, Not Hertz

A pump running at 40% of nominal frequency may show a sharp efficiency decline in the drive or motor, yet frequency alone does not decide whether the operating point saves energy. Current, thermal loss, running time, static head, and the pump-system intersection decide the result. Treat the drive, motor, pump, piping, control method, and auxiliaries as one energy chain.

Wrong fixes for low-speed efficiency

The first wrong fix is declaring 40% frequency a universal efficiency cliff. A frequency value does not identify drive input power, motor shaft speed, motor load, pump efficiency, or delivered hydraulic power. Read each quantity separately.

The second is comparing only the pump shaft-power reduction. That omits drive heat, drive cooling power, motor loss, motor cooling limits, and other auxiliaries. The opposite mistake—rejecting a drive because its percentage efficiency falls at light load—also fails. A larger percentage loss applied to a much smaller load can still produce low absolute loss.

The third is comparing variable-speed control with an idealized throttled system rather than the actual alternative. A throttle valve intentionally dissipates pressure, but a variable-speed pump must still overcome static head and minimum process pressure. Neither method wins automatically.

The fourth is keeping an oversized pump and expecting very low speed to repair the selection. A smaller pump, a small pump operating in parallel with larger units, or storage that smooths variable demand may place the equipment nearer an efficient operating region.

Energy-chain quantities and loss boundaries

The number that matters is total input energy per unit of useful hydraulic output over the operating schedule. This is heat, not logic: every watt entering the drive that does not reach the pump shaft becomes loss in the drive, motor, wiring, cooling system, or mechanical train.

Use a declared measurement boundary. For a complete installation, include drive input power and separately supplied cabinet fans or cooling equipment. If auxiliary power comes through the same measured feeder, avoid counting it twice.

P_hydraulic = rho * g * Q * H
eta_drive = P_motor_electrical / P_drive_input
eta_motor = P_shaft / P_motor_electrical
eta_pump = P_hydraulic / P_shaft
eta_wire_to_water = P_hydraulic / P_total_input

Power factor affects current and apparent-power loading; it is not an additional efficiency multiplier after real input power has been measured. Use true three-phase real-power measurement at the drive input. Current alone cannot determine kW without voltage, power factor, waveform effects, and phase information.

Low-speed thermal checks remain separate from energy calculations. Reduced motor-mounted fan speed can reduce cooling, while a variable-torque pump load often falls rapidly with speed. Compare measured current, motor temperature, drive temperature, and continuous operating time with the manufacturer’s derating and thermal data.

Pump-system operating-point physics

The pump curve and system curve intersect at the operating point. Changing speed moves the pump curve; it does not remove the system curve. A useful system model separates static and friction components:

H_system(Q) = H_static + k * Q^n

When friction dominates, reducing flow sharply reduces required head. Variable speed can then remove much of the pressure that a throttle valve would otherwise dissipate. When static head or a minimum reactor or process pressure dominates, the required head changes little as flow falls. Speed reduction reaches a floor where the pump can no longer produce the required head.

This distinction explains why pipelines, boiler-feed duties, and pressure-constrained chemical processes require individual analysis. Their suitability depends on the measured static or minimum-pressure component, not the application label. Broad claims that hydraulic-power systems inherently favor wide speed ranges require the same curve and duty-cycle test.

A flat pump characteristic or similar pump and system curve slopes can produce poor control authority near the intersection. Small speed changes may cause an undesirable operating-point response, and measurement noise or controller action can create hunting. Plot both curves over every required speed before selecting the control strategy.

Affinity-law screening at 40 percent speed

For the same centrifugal pump and impeller geometry, ideal similarity relationships provide an initial screen:

Q2 / Q1 = N2 / N1
H2 / H1 = (N2 / N1)^2
P2 / P1 = (N2 / N1)^3

If actual pump speed is 40% of the reference speed, the idealized head is 0.40^2 = 0.16, or 16%, and shaft power is 0.40^3 = 0.064, or about 6%. Flow scales to 40% only when the system permits the corresponding operating point. Static head, changing efficiency, minimum flow constraints, and control limits can invalidate a direct field prediction.

Use measured shaft speed rather than assuming electrical frequency equals pump speed. The frequency-speed relationship depends on the motor and control method, while mechanical transmission ratio and slip can also affect shaft speed.

A low drive efficiency percentage at this point may have little effect on annual energy if the absolute input power and time at that condition are small. Conversely, long operation at low hydraulic output can make fixed drive, cooling, and motor losses economically significant. Integrate kW over time rather than judging one instantaneous percentage.

Required supplier and field data

Request efficiency and loss information at the actual combinations of speed and load. A single rated-point efficiency cannot support a part-load calculation.

Ask whether quoted losses include internal fans and whether external enclosure cooling is excluded. Record the temperature and switching conditions associated with supplier curves because thermal management can change auxiliary demand.

Duty-cycle calculation procedure

1. Define required flow, required discharge pressure or head, and annual hours for each operating band. Separate scheduled demand from short transients.
2. Test whether storage can smooth production toward a constant average rate. Include tank constraints and pumping schedule in the economic comparison.
3. Build the system curve from static head plus flow-dependent loss. Add any minimum process pressure explicitly.
4. Overlay pump curves at the speeds needed to meet each duty point. Exclude points that violate the pump’s published operating constraints.
5. Read pump efficiency and calculate hydraulic and shaft power at each valid intersection.
6. Apply motor and drive part-load data at the matching speed and load. Add cooling and auxiliary power inside the selected facility boundary.
7. Calculate the existing throttling or bypass case from measured data. Use simultaneous flow, head, and true input-kW readings where practical.
8. Multiply total input kW in every band by its annual hours, then sum the bands. Add maintenance and capital costs only after the energy model is internally consistent.
9. Repeat the calculation for a smaller pump, parallel pump staging, and storage when those alternatives can cover low-demand periods.

A field heuristic has suggested keeping much of the variable duty around 70–85% of a selected design flow, considering a smaller pump when substantial time lies below 70%, and revisiting the design point when substantial time lies between 85% and 95%. Treat those percentages as a screening trigger, not an acceptance rule. The pump curve, system head, allowable operating region, and annual energy calculation make the decision; a proposed redesign targeting 95–110% of a new design rate still requires those checks.

Commissioning and verification

Verify the calculation at several stable operating points spanning the real duty range, including the lowest sustained demand and the most common operating band. Log flow, suction and discharge pressure, shaft speed or reliable speed feedback, drive input kW, relevant auxiliary kW, current, and temperatures. Allow readings to stabilize before comparing them with predicted curves.

Calculate measured hydraulic power and wire-to-water efficiency at every point. Reconcile large differences by checking instrument scaling, pressure datum, fluid density, valve position, parallel flow paths, bypass flow, and the actual pump configuration. Compare absolute loss in kW as well as efficiency percentage.

Trend temperature over the longest low-speed operating interval. A point that meets the energy target but continues heating is not a valid continuous-duty point. Confirm stable control by stepping the demand within the permitted process range and checking that flow and pressure settle without sustained hunting.

Frequently asked questions