Flow Rate, Velocity & Pipe Diameter Calculator
Three ways to read the same relationship : a flow passing through a pipe's cross-section moves through it at a certain velocity. Knowing two of these three quantities — flow rate, velocity, diameter — gives the third. The most useful case for sizing a new pipe : start from the target flow rate and a recommended velocity to find the minimum inside diameter.
| Line type | Recommended velocity |
|---|---|
| Suction (before the pump) | 0.5 to 1.5 m/s |
| Discharge (after the pump) | 2.0 to 6.0 m/s |
| Return (to the reservoir) | 1.5 to 4.0 m/s |
Calculator
How flow rate, velocity and diameter are calculated
Q = v × A · A = π × D² / 4
The flow rate through a cross-section is the product of velocity and the area of that section. For a circular pipe, the area depends only on the inside diameter : at a fixed flow rate, shrinking the diameter raises the velocity — and with it the friction head losses, which is why it pays to aim for a recommended velocity rather than the smallest diameter that fits.
An asterisk on first use points to the glossary.
Why suction, discharge and return want different speeds
The oil is not in the same predicament on all three lines, even at the same flow rate*. On the suction side the pump pulls; it does not push.
Available head is the hydrostatic column in the tank, plus atmosphere, minus strainer losses. If the speed goes up, inlet pressure collapses. The fluid releases air. The pump cavitates. Suction stays slow for that reason.
Day-to-day hydraulics aims for a mean velocity* of roughly 0.5–1.2 m/s on suction, 3–6 m/s on discharge and 2–3 m/s on return. Rules of thumb, not a standard.
The table at the top of this page is a slightly wider envelope (up to 1.5 m/s on suction, 2–6 on discharge, 1.5–4 on return), and it covers less fussy circuits. A piston pump belongs at the low end of the suction band. Discharge is already under pressure, so the oil will not boil, and a smaller tube is the ordinary choice: it costs less, bends more easily, and weighs less on a machine. The cap is noise, heat, and head loss eating the pressure the cylinder was supposed to see.
Return sits in the middle: a few bar of filter back-pressure, perhaps, and not a vacuum.
Too fast, and the return filter loads up. The tank froths. Too slow, and the hose is needlessly fat. The compromise on most machines is 2–3 m/s.
Below about 0.3–0.4 m/s in a horizontal run, chips, mill scale and free water have time to settle. The tube becomes a settling tank.
The next peak flow picks that slug up and sends it into a servo or a spool. Oil that crawls cools oddly — too much or too little, depending on the room. An oversized return looks generous on paper. In the field the oil arrives tepid at the filter, and the clogging indicators stay quiet until they do not. A diameter “well above” the band is not free: weight, bend radius, price, and a circuit that never flushes itself clean.
On suction, too slow is almost never a hydraulic problem. It is a packaging problem.
An 80 mm tube for 30 L/min will not bend under a bonnet. The suction band starts at 0.5 m/s. Going under that “to be safe”, with no other reason, does not improve the inlet. The pipe is only harder to route.
Velocity too high: noise, erosion, losses, cavitation
Twice the speed, twice the trouble? Friction losses do not scale that way.
They scale roughly with velocity squared once the flow is turbulent — which it usually is. Going from 3 to 6 m/s is about four times the straight-pipe ΔP. On a long hose the pump sees another 20 bar. That pressure never arrives at the cylinder.
Noise follows. Oil tearing through a poorly faired fitting, a tight elbow or a half-open valve whistles — a nuisance in a workshop, a regulation in a cab.
Wall erosion is quieter. At high speed the grit the filter missed sandblasts the bore. Bends take the worst of it. Two seasons later a “clean” hose weeps at the braid.
On suction, too fast has a name: cavitation*. A bore that is only “a bit tight” is already a pump that gets replaced.
Local pressure drops below vapour pressure, or below the oil's air-release point. Bubbles form. They collapse in the first pump stage. The gears pit, with a gravelly rattle. Volumetric efficiency collapses.
High suction velocity is not cured by raising pump speed. More speed means more flow, which means still more velocity in the same bore: the remedy is a fatter inlet, a larger strainer, a closer tank, not a faster pump.
Inside diameter vs DN / NPS
The calculator returns an internal diameter* (ID): the D in Q = v × π × D2 / 4, the hole the oil actually uses. Nominal diameter* — DN, and NPS in inch-based catalogues — is only a label, and DN 20 does not mean a 20.0 mm bore.
Q = v × A · A = π × D2 / 4
On steel tube, DN follows ISO 6708. Wall thickness (series, schedule) eats into the outside diameter. A light-series DN 20 may have a 22 mm bore. A heavy schedule leaves quite a bit less.
Hose is sold in dash sizes* (–8, –12) or in inches of ID. –12 is ¾ in, about 19 mm inside, not 20.
Ordering “DN 16 because the sum came to 16.3 mm”, without the supplier's bore table, is a classic miss. The actual bore has to be checked, especially on thick-wall hose and on stainless with a heavy schedule. DN matches the fittings. Flow is calculated on the inside.
Why an undersized diameter costs more to run
A tight diameter does not show up on assembly day. It shows up on the electricity bill, or on a cylinder that no longer makes force once the hose is rolled out. Every extra metre of undersize burns pressure into heat. The pump paid for that pressure — the pump power calculator is where that cost is reckoned — and the actuator never gets it back.
Line losses (Darcy–Weisbach) go as v2 / D. One step down in diameter, at constant flow, raises v (area falls with D2) and the 1/D term, and the two stack.
Three metres of –8 hose instead of –10, at 50 L/min, can cost several bar. Cycles per hour, and the motor's kWh, multiply that gap. With a diameter in hand, the head-loss calculator checks ΔP on the real run — length, roughness, bends. Sizing to the smallest D that “meets” the velocity, and stopping there, is a half job.
The D from this calculator is the straight-pipe bore. Every fitting, every elbow, every pressure tap is a singular loss, often dearer than several metres of line.
A sharp 90°, an adaptor that necks DN 20 down to DN 12 “to fit the manifold”, or a ball valve half shut: local velocity explodes even when the tube itself is the right size. K-factor tables sit on the head-loss calculation. Before the next dash size goes on the order, the bends are worth counting. Four 90s on 1.5 m of hose often beat an extra metre of straight length. A generous radius and a straight fitting do more for ΔP than the next DN.
On suction, reductions stay away from the pump flange: a straight run of at least two diameters, a generously sized strainer, and no reduced-bore isolation valve. A pump does not swallow a restriction, even one marked “temporary, just for the job”.
Hose versus rigid tube
Same ID, two different lives. Rigid tube (steel, stainless) is smooth, holds pressure, and stays put.
It is bent and clamped, and it can be sized a shade tighter, because roughness is low and no braid adds wall thickness. Bend radius is not negotiable. Too tight, and the tube ovals. The local bore shrinks, and the speed that was calculated is no longer the speed in the tube.
Hose moves. It grows a little under pressure, it shakes, and it curves.
The inner tube (rubber, thermoplastic) is rougher than drawn steel. Crimped fittings are local restrictions. Practice sits at the low end of the velocity band, or one dash size up. On suction the risk is specific: hose collapse, from vacuum plus a tight bend.
Long straight runs go in rigid tube: cheaper, and the loss is lower. Hose stays at the joints, on pumps that move, and at service points.
The two diameter logics do not mix. A DN 12 tube is not a –12 hose. The bore has to be read for each. The labels come from different catalogues and do not stand in for that reading.
Instantaneous vs average flow: a moving cylinder
A positive-displacement pump at constant speed delivers a more or less flat flow. A cylinder does not. During the stroke, Q = rod speed × area. Against the stop the flow falls to leakage. On restart, an accumulator or oil compressibility puts a spike on top.
Q = rod speed × area
Sizing the line on the machine's average flow (“40 L/min typical”) undersizes it on the peak.
The cylinder calculator gives rod speed for a given flow. Inverted, that is the flow to plan for a cycle time. That flow, not the daily average, is the one used when D is chosen. On a machine with several actuators the worst case is “two functions together”, not the biggest one on its own.
Return flow is not always the same as extend flow. Annular area on the rod side is smaller: the cylinder retracts faster, and the cap end dumps more oil.
A return line sized on pump flow can be tight when the rod comes home. Both areas need a figure. Two pipes of the same D split the flow evenly only if the paths match. A poorly placed tee, or a hose a bit longer on one side, and one branch takes 70 % of the flow.
That branch leaves the velocity band. The other silts up.
This calculator cannot see the split: it assumes one tube, one Q, one D. For a splitter feeding two cylinders, each branch is sized on its own peak flow, not on half the pump flow. The common trunk is sized on the sum. Same idea as incoming and outgoing cables in an electrical panel.
Flow-divider plates even the split, at the cost of extra ΔP. With one fitted, the loss is no longer that of the bare tube.
Velocity is worked out again downstream of the divider. The figure used is the real flow in each branch. A 50/50 guess is not that flow. The band is read on the branch that carries the oil.
Reynolds number and flow regime
The chosen speed fixes the regime. Reynolds number* Re = v × D / ν, and ν is the oil's kinematic viscosity*.
Re = v × D / ν
Laminar sits at Re ≲ 2,300. Turbulent sits at Re ≳ 4,000. Between the two is a messy band nobody designs for. A common shop oil at 40 °C is ISO VG 46, with ν ≈ 46 mm2/s. In a 20 mm tube at 4 m/s, Re ≈ 1,740 — still laminar, or on the edge.
Warm the oil to 60 °C and ν drops. Re roughly doubles, and the flow tips into turbulent.
Losses jump without the diameter having moved. A circuit that “was fine cold” whistles at the end of the shift. Laminar flow is the quieter of the two regimes. It wants a fat diameter, a thicker oil, or a lower speed.
Suction chases that laminar regime on purpose. Discharge rarely can: the tube would be huge.
Velocity is not only a cavitation check. It chooses the loss formula used next. The laminar formula follows v. The turbulent one follows v2.
Worked example: 50 L/min discharge line
A small power pack delivers 50 L/min. Discharge is aimed at 4 m/s, the middle of the 3–6 m/s band.
A = Q / v · D = √(4 × A / π)
- Q = 50 L/min = 8.33 × 10−4 m³/s.
- Area needed: A = Q / v = 2.08 × 10−4 m².
- Inside diameter: D = √(4A / π) = 16.3 mm.
Off a hose chart, –8 (½ in, ~12.7 mm) is too small: velocity would sit near 6.6 m/s. –10 (~16 mm) is tight. –12 (¾ in, ~19 mm) brings v down to about 2.9 m/s, kinder if the hose is long or full of bends. In steel tube, DN 15 is frequently too narrow inside. DN 20, around a 21–22 mm bore, is comfortable.
Same pump, suction at 1 m/s: D = 32.6 mm. That is a –20 hose (1¼ in) or a DN 32 tube.
Matching the suction line to the pressure line out of habit is a mismatch. Return at 2.5 m/s lands near 20.6 mm — the same dash as discharge, or one size up if the return filter is a long way off. With the three diameters on the sketch, head loss is run on the real lengths. Pump displacement comes in if the motor is already chosen.
The three pages belong on the same job.
Glossary
- Flow rate
- Volume of fluid that crosses a section in a unit of time. Symbol Q. Hydraulic circuits state it in L/min; the same relations use m³/s, and 50 L/min is 8.33 × 10−4 m³/s.
- Mean velocity
- The velocity in Q = v × A, obtained by dividing flow rate by the internal area of the pipe. The suction, discharge and return bands on this page are values of that velocity.
- Internal diameter
- The bore, D in Q = v × π × D2 / 4. The hole the oil actually uses. Not the outside diameter, and not the DN label.
- Reynolds number
- Dimensionless number Re = v × D / ν. Below about 2,300 the pipe flow is laminar; above about 4,000 it is turbulent. The band between the two is not a design target.
- Nominal diameter
- A size label for tube and fittings, written DN under ISO 6708. NPS is the counterpart in inch-based catalogues. DN 20 is not a 20.0 mm bore. DN matches fittings; the flow calculation uses the inside.
- Cavitation
- What an inlet that is too fast is called. Local pressure falls below vapour pressure, or below the oil's air-release point. Bubbles form, then collapse in the first pump stage: pitting, a gravelly rattle, and volumetric efficiency gone.
- Kinematic viscosity
- ν in Re = v × D / ν. For an ISO VG 46 hydraulic oil it is about 46 mm2/s at 40 °C. Warmer oil has a lower ν, and Re rises in the same pipe at the same speed.
- Dash size
- A hose size stepped in sixteenths of an inch of inside diameter. –8 is ½ in, about 12.7 mm. –12 is ¾ in, about 19 mm inside, not 20. –10 is about 16 mm. –20 is 1¼ in.