Pump Head Guide
A comprehensive engineering guide to calculating total dynamic head (TDH) for pump selection in HVAC water systems. Covers the Darcy-Weisbach equation, local losses, static head, pump power estimation, and practical field assumptions.
What Is Pump Head?
Pump head, also called total dynamic head (TDH), is the total mechanical energy per unit weight that a pump must impart to the fluid to overcome all resistances in the piping system. It is expressed in meters of fluid column (m) and is the sum of three components:
Total Dynamic Head = Friction Loss + Local Loss + Static Head
Each component addresses a different physical phenomenon. Friction loss accounts for energy dissipated by fluid shear along straight pipe walls. Local loss accounts for energy dissipated by flow disturbances at fittings, valves, and equipment. Static head accounts for the elevation change the fluid must overcome against gravity. Understanding each term is essential for accurate pump sizing and avoiding costly oversizing or undersizing.
In HVAC practice, pump head is preferred over discharge pressure because it is independent of fluid density. A centrifugal pump operating at a fixed speed delivers the same head regardless of whether it pumps water, glycol solution, or brine — the resulting discharge pressure changes proportionally with density, but the pump curve (head vs. flow) remains valid for all fluids of similar viscosity.
Key Parameters Explained
Five parameters govern the pump head required for any hydronic system:
- Flow rate (Q) — The volume of fluid circulated per unit time, typically in m³/h or L/s. Higher flow rates increase velocity and, because friction loss scales with v², drive head requirements upward non-linearly.
- Pipe diameter (D) and length (L) — The internal diameter of the pipe determines flow velocity for a given flow rate. Smaller diameters produce higher velocities and dramatically higher friction losses. The total straight pipe length directly multiplies the friction gradient (m/100m).
- Fittings and valves count — Each elbow, tee, valve, reducer, strainer, or equipment connection introduces a localized pressure drop. In a compact mechanical room, fitting losses can exceed straight-pipe friction.
- Static elevation (Δh) — In open systems (cooling towers, open tanks), the pump must lift the water from the free surface of the source to the highest point of discharge. In closed loops, static head cancels and the pump need only overcome friction and local losses.
- Fluid density and viscosity — While head is density-independent, friction loss calculated via Darcy-Weisbach depends on the Reynolds number, which in turn depends on kinematic viscosity. Fluids with higher viscosity (e.g., glycol mixtures at low temperature) increase friction losses significantly.
How the Calculation Works
The pump head calculation proceeds in four steps:
1. Friction loss in straight pipes. The Darcy-Weisbach equation is the fundamental method:
h_f = f × (L / D) × v² / (2 × g)
where h_f is friction head loss (m), f is the Darcy friction factor (dimensionless), L is pipe length (m), D is internal diameter (m), v is mean flow velocity (m/s), and g is gravitational acceleration (9.81 m/s²). The friction factor f is obtained from the Moody chart or the Colebrook-White equation and depends on the Reynolds number and relative pipe roughness. For quick HVAC estimates, empirical friction gradients are often used: water in commercially sized steel pipe typically experiences 0.02–0.05 bar per 100 m (0.2–0.5 m head per 100 m) at design flow velocities between 1.0 and 2.5 m/s.
2. Local loss from fittings and equipment. Local losses can be estimated by two methods. The simplified approach assumes local losses equal 30% of the friction loss — a rule-of-thumb adequate for preliminary sizing when fitting details are unknown. The K-factor method sums the loss coefficient of every component:
h_m = Σ(K_i) × v² / (2 × g)
where K_i is the loss coefficient for each fitting or valve, available from manufacturer literature or industry standards (e.g., ASHRAE, Crane Technical Paper No. 410). For example, a standard 90° elbow has K ≈ 0.75, a fully open gate valve K ≈ 0.15, and a swing check valve K ≈ 2.5, all referenced to the velocity in the connecting pipe.
3. Static head. In open systems, the static head is simply the vertical distance from the free surface of the source fluid to the highest discharge point or the free surface of the receiving tank. In closed hydronic loops, the static head from the supply rise is exactly balanced by the return drop, so the net contribution is zero. This is a critical distinction — attempting to add static head in a closed-loop calculation would produce a grossly oversized pump.
4. Total head and pump power. The total dynamic head is the sum of the three components. Once TDH is known, the required pump shaft power is:
P = ρ × g × H × Q / η
where ρ is fluid density (kg/m³), H is TDH (m), Q is flow rate (m³/s), and η is pump efficiency (typically 0.65–0.85 for well-sized centrifugal pumps). The motor power must additionally account for motor efficiency and any service margin (typically 10–15%).
Friction Loss Estimation by Pipe Size
The table below shows approximate friction losses for water in Schedule 40 steel pipe at typical HVAC design velocities. These values are useful for preliminary estimates only — detailed design should use the full Darcy-Weisbach method with actual pipe roughness, fluid properties, and fitting counts.
| Nominal Pipe Size (DN) | Internal Diameter (mm) | Typical Flow (m³/h) | Velocity (m/s) | Friction Loss (m/100m) |
|---|---|---|---|---|
| DN25 (1") | 27.3 | 1.5–2.5 | 0.7–1.2 | 0.25–0.60 |
| DN40 (1½") | 41.0 | 4.0–7.0 | 0.8–1.5 | 0.20–0.55 |
| DN50 (2") | 52.5 | 7.0–12.0 | 0.9–1.6 | 0.18–0.50 |
| DN65 (2½") | 68.0 | 12.0–20.0 | 0.9–1.5 | 0.15–0.45 |
| DN80 (3") | 80.7 | 18.0–30.0 | 1.0–1.6 | 0.12–0.40 |
| DN100 (4") | 106.3 | 30.0–55.0 | 1.0–1.7 | 0.10–0.35 |
| DN125 (5") | 131.7 | 50.0–85.0 | 1.0–1.7 | 0.08–0.30 |
| DN150 (6") | 157.1 | 75.0–130.0 | 1.1–1.8 | 0.07–0.25 |
| DN200 (8") | 209.0 | 130.0–220.0 | 1.1–1.8 | 0.06–0.20 |
Note that the friction loss values assume clean water at approximately 20°C. For chilled water (4–10°C) the viscosity is slightly higher but the difference is negligible for most sizing purposes. Glycol mixtures, however, can increase friction losses by 15–40% depending on concentration and temperature, and must be factored in during final pump selection.
Common Mistakes in Pump Head Estimation
Even experienced engineers sometimes fall into these traps. Avoiding them is essential for reliable pump selection.
- Adding static head in closed loops. The most frequent error. In a closed hydronic system the static head from the supply riser is returned on the down-comer. Adding it doubles the calculated head and leads to an oversized, inefficient pump that may operate far from its best efficiency point (BEP).
- Ignoring local losses from fittings. Using only straight-pipe friction and neglecting valves, elbows, strainers, and equipment connections can under-predict the required head by 20–50%. Compact mechanical rooms with many fittings are especially vulnerable.
- Confusing nominal pipe size with internal diameter. Nominal pipe size (e.g., 2") does not equal the actual internal diameter. Schedule 40 vs. Schedule 80 pipes have different wall thicknesses and therefore different internal diameters, leading to different velocities and friction losses for the same nominal size.
- Selecting a pump without reviewing the pump curve. A pump selected at a single design point (head, flow) may perform poorly if the actual system curve differs. The selected pump should operate near its BEP at the design point and must not operate in the run-out zone or near shut-off head. Variable-speed pumps require additional checks across the speed range.
- Neglecting fluid property effects. Water at 80°C has roughly one-third the viscosity of water at 10°C, and 30% propylene glycol at 0°C is roughly 8 times more viscous than water at 20°C. Using water-at-20°C friction data for a glycol system significantly underestimates pump head.
- Using the wrong safety margin. Adding a blanket 20–30% margin without justification is common. A better approach is to understand the sources of uncertainty (pipe roughness, future fouling, unknown fitting counts) and apply a targeted margin of 10–15%, then verify that the pump curve still provides adequate head at the required flow.
- Assuming pump efficiency is fixed. Efficiency varies with flow rate along the pump curve. Using a single efficiency value from the BEP may overstate the actual efficiency at the operating point. Always read the efficiency from the pump curve at the design flow rate.
Frequently Asked Questions
What is the difference between pump head and pressure?
Pump head (meters of fluid column) and pressure (bar, psi) are related by H = P / (ρ × g). For water at room temperature, 1 bar equals approximately 10.2 meters of head. The head convention is preferred in pump engineering because it is independent of fluid density — a given pump operating at a fixed speed delivers the same head for any fluid of similar viscosity, while discharge pressure changes with density.
Why does static head cancel out in closed-loop systems?
In a closed hydronic loop — such as a chilled water system or a heating loop — the water column on the return side is at the same height as the supply side. Every meter of elevation gained on the supply riser is recovered on the return drop. The net static differential across the pump is therefore zero. The pump only needs to circulate the fluid against friction and local losses, not lift it. True open-system static head applies only when there is a free surface difference (e.g., cooling tower sump to spray nozzles, or a condensate return system).
How accurate is the 30% local-loss rule of thumb?
As a preliminary estimate, assuming local losses are 30% of straight-pipe friction losses is reasonable for typical hydronic systems with moderate fitting density. For long piping runs with few fittings (e.g., campus distribution lines), 10–15% is more appropriate. For compact equipment rooms with many valves, strainers, and fittings, local losses can reach 60% or more of friction losses. The only accurate method is to sum K-factors for each component using the K-factor method with actual velocities.
When should I use variable-speed pumps instead of fixed-speed pumps?
Variable-speed (VFD) pumps are beneficial when the system load varies significantly over time — such as in HVAC primary or secondary loops with variable flow control, or condenser water systems serving multiple chillers. The affinity laws show that pump power scales with the cube of speed, so a 20% reduction in speed reduces power consumption by roughly 50%. For constant-flow systems (e.g., primary-only designs without control valves), fixed-speed pumps are simpler and more cost-effective.
How should I account for future fouling or pipe aging in pump head estimates?
Pipe roughness increases over time due to corrosion, scaling, and biological fouling, which raises friction losses. A common approach is to add 10–15% head margin at the design flow to accommodate future fouling. Alternatively, select the pump with a slightly steeper curve so that reduced flow at increased head naturally compensates for aging without exceeding the motor power rating. Regular cleaning and water treatment can minimize the fouling rate and allow tighter initial head estimates.
This guide is for educational purposes only. Always consult a licensed professional engineer for final pump selection and system design.