top of page

Hydronic Pipe Loss at Low Flow: Recheck Reynolds Number Before Squaring the Flow Ratio

Oct 20
5 min read
A small-bore hydronic test loop with a circulation pump, valves and measurement instruments.

A low-flow pressure reading can look too high when it is compared with a design-point value scaled by the square of flow. That comparison assumes the relevant resistance coefficient stays approximately constant. In a small hydronic branch containing a relatively viscous fluid, reducing flow may put the straight-pipe section into laminar flow, where the pressure-loss relationship is different.


Before diagnosing a blockage or changing a pump-control setting, check Reynolds number at the measured condition. This article uses an original straight-pipe example to show why half flow need not mean one-quarter pressure loss. It does not establish a minimum permissible flow for a coil, chiller or complete hydronic circuit.


1. State the assumptions behind the square-law curve


For a fixed circular pipe, Darcy-Weisbach gives friction pressure drop Δp = fD(L/D)ρv²/2. Length L and internal diameter D are geometric inputs; density ρ and Darcy friction factor fD depend on the condition. Velocity v is volume flow divided by actual internal area.


If geometry, density and friction factor remain effectively constant, pressure drop scales with flow squared. That approximation is often useful for a bounded turbulent-flow comparison. It is not an identity valid for every temperature, fluid and flow rate. Even in turbulent flow, the friction factor can change with Reynolds number and relative roughness.


KSB’s head-loss reference distinguishes the friction relationships for different flow conditions. The immediate review question is therefore not whether the original design pressure was correct, but whether the same scaling model still represents the new condition.


2. Calculate Reynolds number using the actual fluid


For a circular pipe, Re = ρvD/μ, where μ is dynamic viscosity in Pa·s. Equivalently, Re = vD/ν when kinematic viscosity ν is in m²/s. Do not enter centipoise directly as Pa·s: 1 cP equals 0.001 Pa·s.


Caleffi’s hydronic heat-transfer guidance uses below 2,300 as the laminar screening range and above 4,000 as turbulent, with an uncertain transition region between them. These are screening ranges for the stated round-pipe problem, not sharp universal equipment acceptance limits.


Assume a 60.0 m straight circular pipe with internal diameter 0.0200 m. The illustrative Newtonian fluid has density 1,030 kg/m³ and dynamic viscosity 0.00600 Pa·s at the reviewed condition. These assumed properties are not assigned to a glycol concentration or commercial product. For a project, use the actual supplier’s properties at the relevant temperature. The related glycol design review connects those properties to thermal flow and equipment selection.


At 0.120 L/s, volume flow is 0.000120 m³/s. Area is π × 0.0200²/4 = 0.00031416 m², so velocity is 0.3820 m/s. Reynolds number is 1,030 × 0.3820 × 0.0200/0.00600 = 1,311. The straight-pipe example is comfortably inside the laminar screening range.


3. Derive the correct laminar loss relationship


For steady, fully developed laminar flow of a Newtonian fluid in a circular pipe, the Darcy friction factor is fD = 64/Re. Substituting that expression into Darcy-Weisbach gives Δp = 32μLv/D². Replacing velocity by 4Q/(πD²) gives Δp = 128μLQ/(πD⁴).


At fixed temperature, fluid and geometry, pressure drop is therefore proportional to volume flow Q, not Q². The friction factor increases as flow falls, offsetting one power of velocity in the Darcy-Weisbach expression. This is a change in the physical model, not an extra allowance added to a square-law answer.


Original comparison of laminar pipe pressure loss against incorrect fixed flow-squared scaling. At 0.12 litres per second pressure loss is 11.00 kPa and Reynolds number 1311; at half flow the correct pressure loss is 5.50 kPa, not 2.75 kPa.

Original fully developed laminar straight-pipe calculation: 60 m length, 20 mm internal diameter, density 1,030 kg/m³ and dynamic viscosity 0.006 Pa·s. No fittings, heat exchangers or whole-system acceptance are represented.


For the stated 0.120 L/s example, fD = 64/1,311.44 = 0.04880. Direct calculation gives Δp = 128 × 0.00600 × 60.0 × 0.000120/(π × 0.0200⁴) = 11,000.8 Pa, or approximately 11.00 kPa. The Darcy-Weisbach form gives the same answer.


4. Halve the flow and expose the scaling error


At 0.060 L/s, velocity becomes 0.1910 m/s and Reynolds number becomes approximately 656. The flow remains laminar. The correct straight-pipe pressure loss becomes 5.50 kPa, exactly half the original value under the constant-property assumptions.


A fixed square-law calculation instead gives 11.00 × (0.060/0.120)² = 2.75 kPa. That prediction is half the correct low-flow loss. Conversely, a genuine 5.50 kPa measurement would appear twice as high as the mistaken prediction, inviting an unnecessary search for fouling or an undocumented restriction.


Neither comparison includes instrument uncertainty. At low differential pressure, the uncertainty of the measurement can become a large fraction of the reading. Confirm synchronized flow, pressure and temperature data before drawing conclusions from either a modeled discrepancy or an apparently good match.


5. Find where the simple model stops applying


Rearranging the Reynolds expression gives Q = ReπDμ/(4ρ). For the assumed pipe and fluid, Re = 2,300 occurs at about 0.210 L/s, while Re = 4,000 occurs at about 0.366 L/s. These values identify a transition region for review; they are not approved operating limits.


For example, doubling the original flow to 0.240 L/s gives Re ≈ 2,623. Do not extrapolate the laminar equation and report its result as a verified pressure loss at that condition. The calculation has crossed into the uncertain transition range and needs a suitable model and engineering judgment.


AFT’s pipe-pressure-drop documentation describes separate flow-regime equations and configurable transition handling. A smooth software curve through that region reflects a modeling choice; it does not remove the physical uncertainty. Record the method rather than hiding it behind the plotted line.


6. Temperature and geometry can change the conclusion


If viscosity falls from 0.00600 to 0.00300 Pa·s at unchanged density, diameter and 0.120 L/s flow, Reynolds number doubles to about 2,623. The original laminar pressure-loss calculation is no longer safely transferable. Obtain the actual temperature-dependent properties rather than treating concentration alone as a complete fluid description.


Internal diameter also matters strongly. Within the fully developed laminar model, pressure loss varies with D to the negative fourth power at fixed flow. A nominal size is insufficient if the actual wall thickness, lining or product bore differs. Recalculate Reynolds number whenever the bore changes; do not apply the diameter correction while assuming the regime is unchanged.


7. Rebuild the circuit component by component


The derived linear relationship belongs to the stated straight-pipe section. It must not be applied automatically to valves, strainers, coils, evaporators or an entire pump curve. Use appropriate manufacturer data and fluid corrections for those components, and keep any static head separate in an open system.


A low straight-pipe loss also does not prove acceptable heat transfer. Verify equipment flow limits and thermal performance independently. The useful review record retains actual bore, fluid properties, temperature, Reynolds number, chosen loss model and the range over which the model will be used.


For the surrounding piping, pumping and plant-design workflow, explore the paid Chilled Water System Design Handbook. Keep this regime check beside the wider calculation so a convenient flow-squared shortcut does not become the basis for a mistaken diagnosis or control change.



Comments

Rated 0 out of 5 stars.
No ratings yet

Add a rating
bottom of page