Why Throttling Changes a Pump Operating Point: A System-Curve Example

Partly closing the discharge valve on a centrifugal pump reduces the flow rate, but the reason is easy to misread: the pump itself does not change. It keeps delivering the same head-versus-flow performance described by its pump curve, where head is the pumping pressure expressed as an equivalent column of liquid. What changes is the system curve, the relationship between the head the piping requires and the flow rate. A throttling valve adds artificial friction loss, steepens that requirement curve, and shifts the operating point — the intersection of the two curves — to a lower flow and a higher head. This article explains the mechanism, walks through a worked numerical example, and summarizes what throttling costs in energy and where it applies.

The operating point: where the pump curve meets the system curve

Every pumping circuit is described by two curves. The pump curve, published by the manufacturer, gives the head a specific pump produces at each flow rate at a given speed and impeller diameter; for a centrifugal pump it typically falls gently as flow increases. The system curve gives the head the pipeline requires at each flow rate. It is the sum of two parts: static head, the fixed elevation or pressure difference the liquid must be lifted against, which is the same at every flow; and friction head, the loss in pipes, fittings and equipment, which grows roughly with the square of flow. In equation form, H = H_static + kQ², with k a resistance constant for the pipework as built.

Because the pump supplies exactly what the system absorbs, the circuit settles by itself where the two curves cross. That intersection is the operating point, also called the duty point: the flow and head the pump actually delivers. Nothing needs to be enforced — if the pump briefly delivers more than the system takes, pressure backs up and flow falls until the two match again. Anything that changes either curve therefore changes the operating point: trimming the impeller or reducing speed alters the pump curve, while pipe fouling, a partly closed valve or added hardware alters the system curve. Guidance such as the AIChE treatment of pump network performance makes constructing these two curves carefully the basis for all such predictions (AIChE CEP, January 2018).

What closing a throttling valve actually changes

A throttling — or passive — valve is one whose position is fixed by hand and whose loss coefficient does not react to other changes in the pump or system; it changes only when someone re-sets it, as the Hydraulic Institute describes in its Pumps & Systems guidance. While the setting is fixed, the valve behaves like any other fitting: an extra resistance whose loss grows with the square of flow. Closing it adds a new term to the friction part of the system curve, so the whole curve becomes steeper.

The pump curve, by contrast, is untouched. With the steeper system curve, the intersection moves: flow falls and head rises along the pump curve. In the Hydraulic Institute’s wording, the pump must now produce more head to overcome the increased system loss and settles where the new balance of flow and head is reached, at a lower flow rate. The extra head is not doing useful process work — most of it is dissipated exactly where it was created, across the valve itself.

A worked example: throttling from 100 to 80 m³/h

To make the mechanics concrete, consider the following hypothetical circuit. The numbers are constructed for illustration, not measured data.

Setup

A clean-water transfer line lifts liquid through a static head of 10 m. With the control valve fully open, friction in pipes and fittings produces a head requirement of 0.002·Q² (in metres, with Q in m³/h), so the system curve reads H = 10 + 0.002·Q². The fictional but representative pump curve reads H = 45 − 0.0015·Q².

Step 1 — the fully open duty point

Set pump head equal to system head: 45 − 0.0015·Q² = 10 + 0.002·Q². This solves to Q = 100 m³/h at H = 30 m. That is the unthrottled operating point.

Step 2 — ride back along the pump curve

Now close the valve until the flow is 80 m³/h. The pump does not choose its head independently; at 80 m³/h it produces H = 45 − 0.0015·(80)² = 35.4 m. Flow fell by 20%, yet the head the pump develops is higher than before.

Step 3 — locate the lost head

At 80 m³/h, the genuine process — lifting plus pipe friction — needs only 10 + 0.002·(80)² = 22.8 m. The difference, 35.4 − 22.8 = 12.6 m, must be absorbed by the partly closed valve. That 12.6 m serves no process purpose; it is the artificial resistance deliberately introduced to move the operating point.

Step 4 — the power consequences

Hydraulic (water) power is proportional to flow times head. At the original duty it is about 8.2 kW; at the throttled duty about 7.7 kW, of which only roughly 5.0 kW reaches the process usefully. About 2.7 kW is dissipated across the valve. In other words, roughly 36% of the head produced at the new duty is spent in the valve, and a 20% flow cut saved only about 6% of the power.

Two caveats keep the example honest. First, the figures are hydraulic power only; shaft power equals hydraulic power divided by pump efficiency, which was assumed constant even though efficiency in reality varies along the curve. Second, real pump and system curves must come from manufacturer data and measured pipework, and the valve’s own sizing determines how far it must close to reach a given flow. The purpose of these numbers is to show the mechanism, not to size hardware.

Why throttling wastes energy

The example quantifies what Pumpfocus, in its treatment of throttle control valves, calls the extra unwanted flow loss that throttle regulation always creates: shaft power must be supplied to overcome it, and in a completely loss-free regulation method the shaft power would be lower by exactly that throttle power. The same source notes that this loss, and the power it represents, are greater when the system curve is steeper for the same change in flow from Q1 to Q2 — so the deeper the throttling, the larger the share of the pump’s work converted instead to heat and vibration at the valve.

The economics matter because pumping is rarely a marginal energy user. The AIChE review reports that pumps account for a fraction of electrical power consumed at typical plant sites in the U.S. and Canada reported in the literature at between 70% and 90%, that roughly 90% of installed pumps in the chemical process industries are centrifugal, and that electric power makes up about 55% of the dominant 20-year lifecycle cost of an average pumping application. Against that background, Pumpfocus observes that although a control valve remains inexpensive to buy — and although throttle regulation is still the most common continuously variable control method — rising energy costs are making the approach uneconomic. Variable-speed drives change the pump curve itself rather than steepening the system curve, which is why they avoid most of the valve loss; that comparison is left to a separate article.

Where throttling works — and where it does not

Throttling suits centrifugal pumps precisely because their head falls as flow rises, which lets a steeper system curve pull the duty point back. Pumpfocus notes that throttle regulation cannot be used for displacement pumps at all: a positive displacement pump sets its flow by the volume it moves per stroke or revolution at a given speed, so partly closing the discharge valve barely reduces flow while the discharge pressure climbs — a combination that wastes power and stresses the pump without achieving control.

Manual throttling valves are also passive in the Hydraulic Institute’s sense: fixed in position, they do not respond to changes elsewhere in the pump or system, so a shift in process demand or suction conditions requires deliberate re-setting. Their persistence in practice owes to their low procurement cost, simplicity and fine instantaneous control — with the steady energy penalty quantified above as the trade-off.

Key points

• The operating point is where the pump curve and the system curve intersect; the pump settles there automatically.
• Throttling changes only the system curve: closing the valve adds loss, steepens the curve, and lowers the flow while the pump head rises along its own unchanged curve.
• The extra head produced under throttling is mostly dissipated in the valve itself — in the example, 12.6 m of the 35.4 m produced, about 36%.
• Throttle loss is pure waste: a loss-free alternative such as speed control would need less shaft power by exactly that amount.
• Throttling applies to centrifugal pumps, not displacement pumps, and — despite cheap hardware — grows uneconomic as energy costs rise.