Why does the orifice equation use √ΔP and not ΔP directly? What does this mean for meter turndown?

Because the Bernoulli equation relates pressure difference to the square of velocity: ΔP = ½ρv². Solving for v gives v ∝ √ΔP, and since Q = Av, we have Q ∝ √ΔP. This is a direct consequence of kinetic energy being proportional to v². The practical consequence is poor turndown. If maximum flow gives ΔP_max = 100 kPa: At 50% flow: ΔP = 25 kPa (25% of max — not 50%). At 30% flow: ΔP = 9 kPa (9% of max). At 10% flow: ΔP = 1 kPa (1% of max — below most transmitter accuracy). This limits orifice meters to a practical turndown of 3:1 to 5:1, compared to 100:1 for Coriolis meters. For applications requiring wider flow range, consider using two transmitters with different ranges on the same taps, or specifying a senior orifice fitting with interchangeable bore plates.

How do I correct an orifice flow reading for a change in operating conditions?

The orifice equation shows Q ∝ √(1/ρ). If density changes at the same ΔP, actual flow changes. In automated flow computers (RTUs, DCS), pressure and temperature are read continuously and used to compute flowing density in real time — this is called flow compensation and is mandatory for custody transfer metering. For manual correction: Q_actual = Q_indicated × √(ρ_design / ρ_actual). Example: Gas line pressure drops from 10 barg to 8 barg (isothermal): ρ_actual/ρ_design = (8+1.013)/(10+1.013) = 0.819. Actual flow = indicated × √(1/0.819) = indicated × 1.104. The meter under-reads by 10.4%. For liquid temperature rise from 20°C to 60°C: water density drops from 998 to 983 kg/m³, giving only 0.8% error — usually negligible for liquids.

What is the difference between ISO 5167 and AGA-3? Which should I use?

ISO 5167-2 is the international standard covering orifice plates for liquids, gases, and steam. It allows corner taps, flange taps, and D&D/2 taps and uses the Reader-Harris/Gallagher (RG) Cd equation. AGA Report No. 3 (ANSI/API 14.3) is the North American standard specifically for natural gas custody transfer. It mandates flange taps, uses the same RG equation (harmonised with ISO 5167 in 1992), and requires gas compressibility per AGA-8. For most calculations, the two give essentially identical results (less than 0.1% difference). Use AGA-3 for natural gas billing/custody transfer in North America. Use ISO 5167 outside North America, for liquids or steam, or when local regulations specify it.

When should I not use an orifice plate?

Orifice plates are a poor choice for: High viscosity fluids (Re < 5000) — Cd becomes unstable; use a positive displacement meter or Coriolis instead. Two-phase flow (liquid with gas, or wet steam) — the equation is not valid; use a multiphase meter or separate phases upstream. Slurries and fluids with suspended solids — solids erode the bore edge and settle in impulse lines; use a magnetic flowmeter or Coriolis. Very low flow or wide turndown (greater than 5:1) — poor signal at low ΔP; use Coriolis, ultrasonic, or variable area meter. Very dirty gases — bore fouls rapidly; consider an annubar or averaging pitot that can be retracted for cleaning.

How accurate is the orifice plate and what contributes to measurement uncertainty?

For a standard orifice plate conforming to ISO 5167, the combined uncertainty is approximately ±0.5% to ±2% of reading. Main contributors: Discharge coefficient Cd (ISO 5167 claims ±0.5–0.75% in the preferred β range — the dominant term); bore diameter d (±0.1% error in d causes ±0.2% error in flow); differential pressure measurement (DP transmitter accuracy typically ±0.1–0.5% of span); fluid density (±0.5% error in ρ causes ±0.25% error in Q); installation effects such as insufficient straight run or gasket intrusion (can add 0.5–5% systematic bias). For custody transfer, an uncertainty budget per ISO 5167 Annex B or AGA-3 is required.

Orifice Plate Flow Calculator — ISO 5167, Reader-Harris/Gallagher, Beta Ratio | Free

Technical Reference

Orifice Flow —
Theory, Practice & Standards

A complete engineering reference covering ISO 5167 methodology, derivation from first principles, design procedures, and field troubleshooting. For engineering students and practising process engineers.

What is an Orifice Plate and Why Does It Work?

An orifice plate is a thin, flat disc with a precisely machined circular hole (the bore) installed perpendicular to a pipe's axis. When fluid flows through the pipe, the orifice forces the entire stream through a reduced area — velocity rises sharply and static pressure drops. Measuring that pressure differential tells you the flow rate.

This works because energy is conserved along a streamline. Bernoulli's principle states that for steady, incompressible, inviscid flow, the sum of pressure energy, kinetic energy, and potential energy is constant. When cross-sectional area decreases, continuity (mass conservation) demands velocity increase, and Bernoulli requires that velocity increase to come at the expense of pressure. The result is a measurable pressure differential directly related to flow rate.

Orifice plates are standardised under ISO 5167-2 and AGA Report No. 3 (for natural gas). When installed correctly, they achieve measurement uncertainties of ±0.5% to ±2% without individual calibration — purely by conforming to dimensional tolerances and installation requirements. This remarkable accuracy from a passive device with no moving parts explains why orifice plates remain dominant in industrial metering after more than a century of use.

ℹ️

Why orifice plates over other meters? Compared to venturi tubes and flow nozzles, orifice plates have higher permanent pressure loss — but they dominate because they are inexpensive to fabricate, can be inspected and replaced in-line without process shutdown (using a senior orifice fitting), tolerate high temperature and pressure, handle dirty fluids with periodic cleaning, and their uncertainty is fully characterised by four decades of international collaborative research.

The Governing Equation (ISO 5167)

The volumetric flow rate through an orifice plate per ISO 5167-2 is:

Q = Cd · (π/4 · d²) · (1/√(1−β⁴)) · √(2ΔP/ρ)

ISO 5167-2 Orifice Equation — Incompressible Flow

Cd = discharge coefficient (dimensionless) — from Reader-Harris/Gallagher equation β = d/D = beta ratio (dimensionless) d = orifice bore diameter at flowing temperature (m) D = pipe internal diameter at flowing temperature (m) ΔP = differential pressure P₁ − P₂ (Pa) ρ = fluid density at upstream flowing conditions (kg/m³)

The term 1/√(1−β⁴) is the velocity of approach factor E. It corrects for the fact that fluid already has measurable velocity in the pipe before reaching the orifice. For β = 0.3, E = 1.004 (negligible). For β = 0.7, E = 1.228 — a 23% correction that cannot be ignored.

For mass flow rate, rearrange using ṁ = ρQ. Note that ρ moves inside the square root:

ṁ = Cd · E · (π/4 · d²) · √(2 · ΔP · ρ)

Mass flow — preferred for gas and steam where volumetric flow depends on conditions

For compressible gases, apply the isentropic expansion factor Y (ε in some standards):

Q_gas = Cd · E · Y · (π/4 · d²) · √(2ΔP/ρ₁)

Y corrects for density change as gas expands. Y = 1.0 for liquids.

Y = 1 − (0.351 + 0.256β⁴ + 0.93β⁸) · [1 − (P₂/P₁)^(1/κ)] κ = isentropic exponent: 1.4 for air, 1.3 for steam, 1.28 for natural gas

⚠️

Most common calculation mistake: Using the simplified Q = Cd · A · √(2ΔP/ρ) and omitting the velocity of approach factor E. For β > 0.5, this underestimates flow by up to 15%. Always use the full ISO equation, or verify that the Cd value you are using already incorporates E (some older sources tabulate Cd·E together as C).

The Discharge Coefficient Cd — Reader-Harris/Gallagher Equation

Cd is the most critical parameter. It accounts for real-fluid effects — viscous losses, turbulence, and the jet contraction at the vena contracta — that ideal Bernoulli theory ignores. ISO 5167 uses the Reader-Harris/Gallagher (RG) equation, derived by regression of ~16,000 calibration data points from 9 independent laboratories:

Cd = 0.5961 + 0.0261β² − 0.216β⁸ + 0.000521(10⁶β/Re_D)^0.7
+ (0.0188 + 0.0063A)β³·⁵(10⁶/Re_D)^0.3
+ (0.043 + 0.080e^(−10L₁) − 0.123e^(−7L₁))(1−0.11A)(β⁴/(1−β⁴))
− 0.031(M₂' − 0.8M₂'^1.1)β^1.3

RG equation — valid for 0.1 ≤ β ≤ 0.75, Re_D ≥ 5000, per ISO 5167-2:2022 §5.3

A = (19000β/Re_D)^0.8 L₁ = tap distance upstream / D (0 for corner taps, 25.4/D for flange taps, 1.0 for D taps) M₂' = 2L₂'/(1−β) where L₂' = tap distance downstream / D

Because Cd depends on Re_D, and Re_D depends on velocity (and hence Q and Cd), the solution is iterative — start with Cd = 0.61, compute Q and Re_D, update Cd, repeat until convergence (typically 3–5 iterations, |ΔCd| < 0.0001). This calculator performs this iteration automatically.

Cd by Orifice Type (High Re Asymptote)

Sharp-edged, standard: 0.60–0.63
Quadrant-edge (low Re service): 0.77–0.82
Conical entrance: 0.73–0.80
Rounded entrance (nozzle-like): 0.95–0.99
Eccentric / segmental bore: 0.60–0.63

Effect of Edge Wear on Cd

The upstream edge must be sharp. A worn edge (radius e) increases Cd and causes over-reading. ISO 5167 specifies maximum edge radius: e/d ≤ 0.0004. For a 100 mm bore, that is 0.04 mm — less than a human hair. Even modest rounding (e/d = 0.01) increases Cd by ~0.7%, causing ~0.7% positive flow error. After years in erosive service, edge wear can cause 3–5% systematic over-reading.

Beta Ratio β — Design Trade-offs

The beta ratio β = d/D is the central design variable. It determines the differential pressure signal at a given flow rate, the permanent pressure loss, the required straight run, and the measurement uncertainty.

β Range	ΔP Signal	Permanent Loss	Uncertainty	Best For
0.20–0.30	Very high — may need small-range DP transmitter	~90% of ΔP	Higher — RG eq. less accurate at extremes	Rarely used; very low flow, small pipes
0.30–0.50	High	~75–85% of ΔP	Good (±0.6%)	General process metering
0.50–0.65	Moderate	~55–70% of ΔP	Best (±0.5%)	Preferred range — lowest uncertainty per ISO 5167
0.65–0.75	Lower — needs wider-range DP transmitter	~40–55% of ΔP	Good (±0.6%)	High flow, energy recovery important, large pipes

✅

Design rule: Size the orifice so that at maximum flow, ΔP equals the upper range of your DP transmitter (e.g., 25 kPa or 100 inH₂O). At minimum expected flow (say 30% of max), ΔP will be 9% of maximum (because Q ∝ √ΔP). Check that 9% of your transmitter URV is above its minimum reliable output — usually 0.5–1% of URV. If minimum flow gives ΔP below transmitter minimum, you need a smaller β or a wider-range transmitter.

Tap Configurations and Straight Run Requirements

Tap Type	Upstream Tap Location	Downstream Tap Location	Region / Standard
Corner Taps	Flush with plate face upstream	Flush with plate face downstream	Europe; ISO 5167; D < 50 mm
Flange Taps	25.4 mm (1 inch) upstream of plate	25.4 mm (1 inch) downstream	North America; AGA-3; oil & gas
D and D/2 Taps	1D upstream of plate	0.5D downstream	Large pipes; near vena contracta position

Minimum upstream straight pipe (from nearest fitting to orifice plate) per ISO 5167, for β = 0.6:

Upstream Disturbance	Required Straight Run (× D)
Single 90° elbow	18D
Two 90° elbows, same plane	28D
Two 90° elbows, perpendicular planes	54D
Gate valve, fully open	18D
Control valve, any position	50D — avoid if possible
Downstream (any fitting)	5D minimum

🚨

Reversed plate — the most common installation mistake: Orifice plates have a sharp upstream edge and a chamfered downstream bore. Installing backwards (sharp edge facing downstream) changes Cd by +10 to +15% and causes significant over-reading. Always verify the flow direction arrow engraved on the plate handle matches the pipe flow direction. Check during every maintenance inspection.

Step 1 — Continuity Equation (Conservation of Mass)

For steady, incompressible flow, mass flow rate is the same at every cross section. Since density ρ is constant, volumetric flow rate Q is also constant:

A₁ · v₁ = A₂ · v₂ = Q

A₁ = pipe area = π/4 · D² (section 1, upstream) A₂ = orifice area = π/4 · d² (section 2, at bore) v₁, v₂ = mean velocities at each section

From this: v₁ = v₂ · (A₂/A₁) = v₂ · (d/D)² = v₂ · β². So pipe velocity is smaller than orifice velocity by a factor of β². For β = 0.5, the pipe velocity is one-quarter of the orifice velocity.

Step 2 — Bernoulli Equation (Conservation of Energy)

Apply Bernoulli between section 1 (upstream tap) and section 2 (orifice throat), assuming steady, inviscid, incompressible, horizontal flow:

P₁ + ½ρv₁² = P₂ + ½ρv₂²

Each term has units of Pa (J/m³) — energy per unit volume

Rearranging:

P₁ − P₂ = ½ρ(v₂² − v₁²)

Substituting v₁ = β²v₂ from the continuity equation:

ΔP = ½ρv₂²(1 − β⁴)

β⁴ appears because v₁² = (β²v₂)² = β⁴v₂²

Solving for the orifice velocity v₂:

v₂ = √(2ΔP / ρ(1−β⁴)) = E · √(2ΔP/ρ)

where E = 1/√(1−β⁴) is the velocity of approach factor

Step 3 — Ideal Volumetric Flow Rate

Multiply orifice velocity by orifice area to get the theoretical (ideal) volumetric flow rate:

Q_ideal = A₂ · v₂ = (π/4 · d²) · E · √(2ΔP/ρ)

Assumes: no viscosity, no turbulence, no jet contraction beyond the bore

This is what you would measure if the fluid were ideal (inviscid) and the jet filled the bore exactly. In reality, two physical effects cause actual flow to differ.

Step 4 — Physical Reality: Vena Contracta and Viscous Losses

Effect 1 — Vena Contracta (jet contraction): Due to fluid inertia, streamlines converge beyond the sharp bore edge. The actual minimum jet cross-section (the vena contracta) is smaller than the bore area by the contraction coefficient Cc. For a sharp-edged orifice, Cc ≈ 0.61, meaning the jet uses only 61% of the bore area. This is the primary reason Cd ≈ 0.61.

Effect 2 — Viscous and turbulent losses: Real fluid has viscosity; energy is dissipated in the shear layer and turbulent mixing zone. A velocity coefficient Cv ≈ 0.98 accounts for this loss.

Combined into one coefficient:

Cd = Cc × Cv ≈ 0.61 × 0.98 ≈ 0.60

Physical interpretation of the discharge coefficient

Therefore the actual flow rate is:

Q = Cd · E · (π/4 · d²) · √(2ΔP/ρ)

The ISO 5167 orifice equation — full form for incompressible flow

💡

Key insight: Cd is not an empirical fudge factor — it has physical meaning. Cd ≈ 0.61 means that only 61% of the bore area is effectively carrying flow, because the jet contracts to 61% of the bore area at the vena contracta. A 100 mm orifice in a 200 mm pipe does not behave as if it were 100 mm of open pipe — the effective flow area is only about 61 mm equivalent diameter.

Step 5 — Iterative Solution for Cd

The challenge: Cd (via the RG equation) depends on Re_D, and Re_D depends on Q (which depends on Cd). This circular dependency is resolved by iteration:

Initial estimate

Start with Cd⁰ = 0.61 (or the high-Re asymptote of the RG equation for the given β).

Calculate flow and Reynolds number

Compute Q = Cd · E · A₂ · √(2ΔP/ρ), then Re_D = 4ṁ / (π · D · μ) = 4ρQ / (π · D · μ).

iii

Update Cd

Substitute Re_D into the RG equation to get Cd_new.

Check convergence

If |Cd_new − Cd_old| < 0.0001, stop. Otherwise set Cd_old = Cd_new and return to step ii. Typically converges in 3–5 iterations.

⚠️

At low Reynolds numbers (Re < 10,000): Cd drops noticeably below 0.61 and the RG equation uncertainty increases sharply. ISO 5167 requires Re_D ≥ 5,000 for standard orifice plates. For viscous fluids or very low flow, consider a quadrant-edge or conical-entrance orifice, which maintain stable Cd down to Re ≈ 500.

Step 6 — Compressible Gas Flow Correction

For gases, density decreases as pressure drops across the orifice. The fluid that exits at P₂ is less dense than at entry P₁, so actual volumetric flow is higher than the incompressible calculation predicts. The isentropic expansion factor Y corrects for this:

Y = 1 − (0.351 + 0.256β⁴ + 0.93β⁸) · [1 − (P₂/P₁)^(1/κ)]

P₁, P₂ = absolute pressures at upstream and downstream taps (Pa absolute) κ = ratio of specific heats (Cp/Cv): air 1.40, steam 1.30, natural gas ~1.28 Y = 1.0 for liquids; Y < 1.0 for gases (typically 0.96–0.99 for ΔP/P₁ < 10%)

When ΔP/P₁ < 2% (a common situation in process plants at operating pressure), Y > 0.997 and the correction is negligible (<0.3%). When ΔP/P₁ > 5%, Y must be applied. For ΔP/P₁ > 25%, the isentropic assumption breaks down and the full energy equation with real gas properties is needed.

How to Design an Orifice Plate from Scratch

Define Process Conditions

Collect from process datasheet and P&ID:

Fluid: name, phase, composition (if mixture or gas)
Flow range: Q_min, Q_normal, Q_max (in m³/hr or kg/hr)
Operating pressure P₁: upstream absolute pressure (kPa abs or psia)
Operating temperature T₁: upstream (°C or °F)
Pipe: nominal size, schedule, material — needed for actual ID
Metering purpose: indication only, control, or custody transfer (drives accuracy requirements)

Determine Fluid Properties at Operating Conditions

You need density ρ and dynamic viscosity μ at P₁ and T₁ — not at standard conditions. This is the most common source of large errors.

Liquids: Use Perry's, NIST WebBook, or process simulation. Water at 80°C: ρ = 971.8 kg/m³, μ = 355 μPa·s — very different from 20°C values (998.2 kg/m³, 1002 μPa·s).
Ideal gases: ρ = PM/(RT). For real gases: ρ = PM/(ZRT), where Z is the compressibility factor from AGA-8 or NIST equations.
Steam: Use IAPWS-IF97 steam tables. At 10 barg, 200°C: ρ = 5.74 kg/m³. At 10 barg, 250°C: ρ = 5.14 kg/m³ — a 10% difference that directly impacts flow accuracy.

Get the Actual Pipe Internal Diameter

Never use nominal pipe size. Look up actual ID from ASME B36.10 (carbon steel) or B36.19 (stainless) pipe tables for your schedule. Example: 4" pipe, Schedule 40: ID = 102.26 mm. Schedule 80: ID = 97.18 mm. Using the wrong schedule introduces a β error that propagates as a β⁴ error in the flow equation — easily 5–10% flow error.

For existing plant piping, verify with ultrasonic thickness measurement — corrosion, erosion, or scale may have changed the ID significantly from as-built.

Select DP Transmitter Range and Calculate Required β

Choose a target ΔP at Q_max — this is your DP transmitter upper range value (URV). Common choices: 25 kPa (100 inH₂O), 50 kPa, 125 kPa. Rearrange the orifice equation to solve for d:

d = D · √β where β is found by solving:
Q_max = Cd · E · (π/4 · β²D²) · √(2ΔP_max/ρ)

Since both Cd and E depend on β, iterate: start with β = 0.6, compute d, calculate E and Cd at this β and Re_D, recompute d, repeat until β converges (usually 3 iterations). Confirm β is in range 0.20–0.75. If outside, adjust DP URV or pipe size.

Check Reynolds Number at Minimum Flow

Minimum flow is the worst case for Reynolds number. Calculate:

Re_D = 4 · ρ · Q_min / (π · D · μ)

If Re_D < 10,000, measurement uncertainty increases significantly. ISO 5167 minimum is 5,000. Below this, consider a quadrant-edge orifice, or a different meter type (Coriolis or electromagnetic) if low-flow accuracy is essential.

Specify Plate Material, Thickness, and Edge Tolerance

Material: 316 SS for most services. Hastelloy C-276 for chlorides. Carbon steel for clean, non-corrosive hydrocarbons. Duplex SS for sour gas (H₂S-containing).
Bore tolerance: Per ISO 5167, bore diameter must be measured to ±0.1% of d. For a 50 mm bore, that is ±0.05 mm — requires precision bore gauging, not a standard ruler.
Edge sharpness: Upstream edge radius ≤ 0.0004 × d. For d = 100 mm, max radius = 0.04 mm. Inspect with 10× magnifier on receipt and after each removal from service. Any visible flat = replace.
Plate thickness: Per ISO 5167, minimum 3 mm; maximum depends on β. The bore bevel angle on the downstream face must be 30–45° per standard.

Verify Straight Run and Specify Installation

Map all fittings within 50D upstream and 10D downstream of the proposed orifice location. Look up required straight run from ISO 5167 Table A.1 (or AGA-3 Table 2) for your β and specific fitting combination. If straight run is inadequate, specify a CPA 50E or equivalent flow conditioner. Note on the datasheet: tap orientation (horizontal for liquids, top for gases, condensate-leg for steam), gasket material and thickness (affects effective bore diameter), and whether a senior orifice fitting is required for in-line plate changes.

Field Troubleshooting — Symptom / Cause / Fix

Symptom	Likely Cause	Diagnostic / Corrective Action
Consistently high reading, 5–15%	Worn/rounded upstream bore edge; reversed plate; low-pressure impulse line partially blocked	Pull plate, measure bore, inspect edge under magnifier; verify plate direction arrow; blow down LP impulse line
Consistently low reading, 5–15%	Partial bore blockage (debris, scale, wax); condensate in gas LP impulse line; blocked HP tap	Pull and clean plate; drain gas impulse lines; clear HP tap with probe rod or high-pressure N₂ purge
Erratic, noisy DP signal	Two-phase flow (flashing or wet gas); pulsation from reciprocating machines; partial tap blockage causing slug flow in impulse line	Check process conditions for flash point; install pulsation dampener or snubber; blow down both impulse lines simultaneously
Zero reading at known flow	Manifold in equalise position; both impulse lines blocked or cross-connected; DP transmitter failed low	Check 5-valve manifold configuration; trace both impulse lines physically; verify transmitter with hand pump
Negative DP reading	HP and LP connections swapped at transmitter or at taps	Swap HP/LP at transmitter first (safe); if still negative, re-identify which tap is actually upstream on pipe
Reading drifts slowly over months	Gradual bore erosion changing d and β; DP transmitter drift; slow impulse line fouling	Annual bore measurement in dirty service; recalibrate transmitter every 2 years; fit isolation valves for blowdown without process shutdown
Step change in reading after maintenance	Plate installed backwards; wrong plate (different β) reinstalled; gasket protruding into bore; impulse lines reconnected backwards	Pull plate and verify β marking matches datasheet; check that gasket ID ≥ pipe ID; verify HP/LP connections per P&ID

How to Use This Calculator Correctly

Pipe Diameter D: Pipe internal diameter from pipe schedule tables — not nominal size. At high temperature, apply thermal expansion: D_hot = D_cold × [1 + α(T−20°C)] where α ≈ 17×10⁻⁶/°C for 316 SS.
Orifice Diameter d: The certified bore diameter at 20°C from the plate manufacturer's certificate. Apply thermal expansion for hot service just as for pipe ID.
Fluid Density ρ: Density at upstream operating P and T. For water: use tables. For hydrocarbons: use process simulator or API correlations. For gases: use ρ = PM/(ZRT). Never use standard-condition density for gas — this is the single most common large error in orifice calculations.
Dynamic Viscosity μ: At operating T. For water, viscosity halves between 20°C and 60°C; a factor of 2 error in μ changes Re_D by factor 2 and Cd by ~0.5–1%.
Differential Pressure ΔP: The measured or design value. Confirm units: 1 kPa = 4.015 inH₂O = 0.145 psi = 10 mbar. Unit errors of 25× (kPa vs inH₂O) are more common than you'd expect.
Discharge Coefficient: Leave at the calculated ISO value unless you have a calibrated plate with a traceable Cd from a flow laboratory. A calibrated Cd is valid only for the specific plate, at the calibrated β and Re_D range.

⚠️

Gas metering reminder: Always enter density at flowing conditions (actual P, T). If you enter density at standard conditions (e.g., 1.2 kg/m³ for air at 15°C, 1 atm) but your line is at 5 barg, the actual density is ~6× higher and your calculated flow will be ~2.45× too high — a catastrophic error in sizing or accounting.

Frequently Asked Questions

Because the Bernoulli equation relates pressure difference to the square of velocity: ΔP = ½ρv². Solving for v gives v ∝ √ΔP, and since Q = Av, we have Q ∝ √ΔP. This is a direct consequence of kinetic energy being proportional to v².

The practical consequence is poor turndown. If maximum flow gives ΔP_max = 100 kPa:

At 50% flow: ΔP = 25 kPa (25% of max — not 50%)
At 30% flow: DP = 9 kPa (9% of max)
At 10% flow: ΔP = 1 kPa (1% of max — below most transmitter accuracy)

This limits orifice meters to a practical turndown of 3:1 to 5:1, compared to 100:1 for Coriolis meters. For applications requiring wider flow range (e.g., variable-flow processes), consider using two transmitters with different ranges on the same taps, or specifying a senior orifice fitting with interchangeable bore plates.

When fluid approaches a sharp-edged orifice, the streamlines must turn sharply inward to pass through the bore. Due to fluid inertia (momentum), the streamlines cannot make a 90° turn at the bore edge — they overshoot and continue converging beyond the bore face. The jet cross-section continues to decrease for a short distance downstream until reaching a minimum area at the vena contracta ("contracted vein" in Latin). Beyond that point, the jet expands and turbulent mixing with the surrounding slow fluid begins.

The ratio of vena contracta area to bore area is the contraction coefficient Cc. For a sharp-edged orifice, the streamlines' inability to negotiate the abrupt 90° corner results in Cc ≈ 0.611 — a result that can be derived analytically for a 2D slot using complex potential flow theory (the Kirchhoff solution) and is well confirmed experimentally for 3D orifices.

Since Cd = Cc × Cv and Cv ≈ 0.98 (small viscous loss correction), Cd ≈ 0.61 × 0.98 ≈ 0.60. The 0.61 figure is not arbitrary — it is fundamentally the geometric consequence of fluid inertia at a sharp-edged opening.

For a rounded-entry orifice, streamlines follow the curved surface smoothly with no separation, Cc → 1.0, and Cd → 0.98. This is why flow nozzles (rounded entry) have Cd ≈ 0.96–0.99.

The orifice equation shows Q ∝ √(1/ρ). If density changes at the same ΔP, actual flow changes. In automated flow computers (RTUs, DCS), pressure and temperature are read continuously and used to compute flowing density in real time — this is called flow compensation and is mandatory for custody transfer metering.

For manual correction when conditions deviate from design:

Q_actual = Q_indicated × √(ρ_design / ρ_actual)

Examples:

Gas line pressure drops from 10 barg to 8 barg (isothermal): ρ_actual/ρ_design = (8+1.013)/(10+1.013) = 0.819. Actual flow = indicated × √(1/0.819) = indicated × 1.104. The meter under-reads by 10.4% — a significant error in gas accounting.
Liquid temperature rises from 20°C to 60°C: water density drops from 998 to 983 kg/m³. Actual flow = indicated × √(998/983) = indicated × 1.0076. Only 0.8% error — usually negligible for liquids but matters in precise chemical dosing.
Gas composition changes (e.g., natural gas heating value changes): molar mass M changes, density changes proportionally, flow correction = √(M_design/M_actual).

The permanent pressure loss ΔP_perm is the irrecoverable portion of ΔP — the energy permanently converted to heat by turbulent mixing downstream of the vena contracta. It is approximately:

ΔP_perm ≈ ΔP × (1 − β²) / (1 + β²) [sharp-edged orifice, approximate]

For β = 0.5: ΔP_perm ≈ ΔP × 0.60 (60% lost). For β = 0.7: ΔP_perm ≈ ΔP × 0.34 (34% lost). Compare to a venturi tube where permanent loss ≈ 5–10% of ΔP.

Annual pumping cost of permanent pressure loss:

Cost ($/yr) = (ΔP_perm × Q × 8760 × C_elec) / (η_pump × 3,600,000)

ΔP_perm in Pa, Q in m³/s, C_elec in $/kWh, η_pump = pump efficiency (decimal)
Example: 150 mm pipe, β = 0.6, ΔP = 25 kPa, so ΔP_perm ≈ 13.5 kPa, Q = 50 m³/hr = 0.0139 m³/s, 8760 hr/yr, $0.10/kWh, η = 0.70: Cost ≈ $760/yr
For a large 600 mm natural gas header at high flow, the same calculation can yield $50,000–$200,000/yr — enough to justify replacing the orifice with a venturi tube in 1–2 years.

ISO 5167-2 is the international standard covering orifice plates for liquids, gases, and steam. It allows corner taps, flange taps, and D&D/2 taps. Uses the RG Cd equation. Published by ISO, adopted in Europe, Asia, and international projects. Applies to all fluids.

AGA Report No. 3 (ANSI/API 14.3) is the North American standard specifically for natural gas custody transfer. It mandates flange taps, uses the RG Cd equation (harmonised with ISO 5167 in 1992), requires gas compressibility calculated per AGA-8, and specifies additional quality requirements for custody transfer including flow computer standards, chart integration, and uncertainty budgets.

For most engineering calculations, the two give essentially identical results (<0.1% difference) when using the same tap configuration and RG equation. Use AGA-3 if: you are metering natural gas for billing/custody transfer in North America, or the contract specifies it. Use ISO 5167 if: you are outside North America, metering liquids or steam, or the contract/local regulations specify it. For process indication (non-custody), either standard is acceptable — just be consistent.

Orifice plates are a poor choice in the following situations:

High viscosity fluids (Re < 5000): Cd becomes unstable and inaccurate. Use a positive displacement meter, Coriolis, or laminar flow element.
Two-phase flow (liquid with gas, or wet steam): The equation is not valid and readings are unreliable. Use a dedicated multiphase meter or separate the phases upstream.
Slurries and fluids with suspended solids: Solids settle in impulse lines and erode the bore edge. Use a magnetic flowmeter (conductive liquids) or Coriolis.
Very low flow or wide turndown (>5:1): Poor signal at low ΔP. Use a Coriolis, ultrasonic, or variable area meter.
Very dirty gases (catalyst fines, heavy tars): Bore fouls rapidly. Consider an annubar or averaging pitot which can be retracted for cleaning without shutdown.
Cryogenic or very high temperature service where straight run is unavailable: The cost of piping modifications may exceed the cost of a more compact meter (Coriolis, ultrasonic) that requires less straight run.

For a standard orifice plate conforming to ISO 5167, the combined uncertainty is approximately ±0.5% to ±2% of reading. The main contributors are:

Discharge coefficient Cd: ISO 5167 claims ±0.5–0.75% for Cd in the preferred β range. This is the dominant term.
Bore diameter d: A ±0.1% error in d causes ±0.2% error in d² (and hence flow). Requires precision bore gauging — not a standard workshop tool.
Differential pressure measurement: DP transmitter accuracy typically ±0.1–0.5% of span. At low flow (small ΔP), this becomes the dominant uncertainty.
Fluid density: ±0.5% error in ρ causes ±0.25% error in Q (because Q ∝ √(1/ρ)). For gases, this depends on accuracy of P, T measurement and compressibility correlation.
Installation effects: Insufficient straight run, non-circular pipe cross-section, gasket intrusion — each can add 0.5–5% systematic bias.

For custody transfer applications, an uncertainty budget per ISO 5167 Annex B or AGA-3 is required, documenting each contributor with its Type A or Type B evaluation.

Engineering Glossary — Orifice Flow Measurement

Discharge Coefficient Cd

Ratio of actual volumetric flow to ideal (Bernoulli) flow through the orifice bore. Cd = Cc × Cv ≈ 0.61 for a sharp-edged orifice at turbulent flow. Calculated by the Reader-Harris/Gallagher equation as a function of β and Re_D. Decreases at low Re, increases with edge wear.

ISO 5167-2 §5.3 / AGA-3

Beta Ratio β = d/D

Ratio of orifice bore diameter to pipe internal diameter. Central design variable — determines ΔP signal, permanent loss, straight run requirement, and uncertainty. Valid range per ISO 5167: 0.1–0.75. Optimal for lowest uncertainty: 0.5–0.65. β appears as β⁴ in the velocity of approach factor — small errors in β have amplified effect on calculated flow.

ISO 5167-2 §5.1

Velocity of Approach Factor E = 1/√(1−β⁴)

Corrects the Bernoulli equation for the fact that fluid has pre-existing velocity in the pipe upstream of the orifice. E = 1.004 at β = 0.3 (negligible). E = 1.083 at β = 0.5. E = 1.228 at β = 0.7. Some older references absorb E into Cd and tabulate C = Cd·E — always confirm which convention is in use.

Contraction Coefficient Cc

Ratio of vena contracta jet area to orifice bore area. Cc ≈ 0.611 for a sharp-edged orifice — derived analytically from potential flow theory for a 2D slot (Kirchhoff solution). The primary physical reason Cd ≈ 0.61. Increases toward 1.0 as the entry is rounded (Cc ≈ 0.98 for a well-rounded flow nozzle).

Vena Contracta

The cross-section of minimum jet area and maximum velocity, located 0.3–0.8D downstream of the orifice plate face. Static pressure is minimum here. The vena contracta is not fixed — its position moves with β and Re_D. D&D/2 taps were historically called "vena contracta taps" because 0.5D downstream approximates the vena contracta location for typical β values.

Pipe Reynolds Number Re_D

Re_D = ρvD/μ = 4ṁ/(πDμ). Governs flow regime (laminar <2300, transitional 2300–4000, turbulent >4000) and determines Cd via the RG equation. ISO 5167 minimum: Re_D ≥ 5000. Best accuracy: Re_D ≥ 10,000. Note: orifice Re is defined on pipe diameter D, not orifice diameter d.

Isentropic Expansion Factor Y (or ε)

Correction factor for compressible gas flow accounting for gas density decrease across the orifice. Y = 1.0 for incompressible liquids. Y < 1 for gases. For air at ΔP/P₁ = 0.10: Y ≈ 0.963. Significant (>0.3%) when ΔP/P₁ > 2%. For ΔP/P₁ > 25%, the equation's accuracy degrades and full isentropic nozzle equations with real gas properties are needed.

ISO 5167-2 §5.3.3

Permanent Pressure Loss ΔP_perm

The irrecoverable fraction of ΔP converted to heat by turbulent mixing. For a sharp-edged orifice: ΔP_perm ≈ ΔP·(1−β²)/(1+β²). At β = 0.5: ~60% of ΔP is lost. At β = 0.7: ~34% lost. Much higher than venturi tubes (5–10%) — a significant operating cost for high-flow applications, justifying venturi tubes on large pipelines despite higher capital cost.

Reader-Harris/Gallagher Equation

The empirical equation for Cd adopted by ISO 5167 (1998 revision) and AGA-3. Derived by M.J. Reader-Harris and J.A. Gallagher from regression of ~16,000 calibration data points across 9 laboratories worldwide. Replaced the Stolz equation, which underestimated the Re_D dependence of Cd at low Re. Applies to corner, flange, and D&D/2 taps with appropriate L₁, M₂' coefficients.

ISO 5167-2:2022 §5.3.2

Flow Conditioner

Device installed upstream of an orifice to straighten the velocity profile and destroy swirl. Allows use of shorter upstream straight runs than tabulated in ISO 5167. Common types: Gallagher (19-tube bundle), CPA 50E (perforated plate), Laws (tabs). A conditioner at 10D upstream can replace the 40D that would otherwise be required after a double-elbow out-of-plane. Must be selected for the specific pipe Re and disturbance type.

Senior Orifice Fitting

A specialised pipe fitting that allows the orifice plate to be removed and replaced while the line is under pressure and flowing, without process shutdown. Uses two isolation valves and a sliding carrier mechanism. Standard for custody transfer metering where downtime costs are high. Available in sizes 2"–36", rated to ASME 2500 class.

Custody Transfer Metering

Flow measurement used to determine the quantity of fluid transferred between two parties for commercial billing. Requires the highest accuracy, documented uncertainty budgets, traceable calibration, tamper-evident sealing, and compliance with the applicable standard (AGA-3 for gas, API MPMS for liquids). Error of 0.5% in a large gas pipeline can represent millions of dollars per year.

Turndown Ratio

Ratio of maximum to minimum measurable flow within specification. For orifice meters: typically 3:1 to 5:1, limited by the √ΔP relationship (at 20% flow, DP is only 4% of max — below most transmitter accuracy). Extended to ~10:1 using dual-range DP transmitters in parallel. Compare: Coriolis 100:1, ultrasonic 50:1, magnetic 30:1.

Quadrant-Edge Orifice

An orifice with a rounded (quarter-circle) upstream bore entry instead of a sharp edge. Maintains stable Cd ≈ 0.77 down to Re ≈ 500, making it suitable for viscous fluids (heavy oil, polymers, glycols) where standard sharp-edged orifices would give unstable and inaccurate Cd. Standardised in ISO 5167-2 Annex B. Requires different Cd equations — not interchangeable with sharp-edge equations.

ISO 5167-2 Annex B

Flange Taps

Pressure measurement taps located exactly 1 inch (25.4 mm) from the upstream and downstream faces of the orifice plate. The North American industry standard, mandatory for AGA-3 custody transfer. The 25.4 mm dimension is fixed regardless of pipe size — unlike D&D/2 taps which scale with D. This means for large pipes (D > 250 mm), flange taps are close to the corner tap location; for small pipes they approach the D&D/2 location.

AGA-3 / ISO 5167-2

Compressibility Factor Z

Correction to the ideal gas law for real gas behaviour: ρ = PM/(ZRT). Z = 1.0 for ideal gas. For natural gas at typical pipeline conditions (50 barg, 20°C), Z ≈ 0.88 — using Z = 1.0 overestimates density by 14% and under-reads flow by 7%. Calculated per AGA-8 (custody transfer), NIST equations, or process simulators.

▸ FLOW — MULTIPLE UNITS
Mass Flow	—
Mass Flow (kg/s)	—
Mass Flow (t/hr)	—
Actual Vol. Flow (m³/hr)	—
Actual Vol. Flow (m³/s)	—
Nm³/hr (ref 0°C, 1.01325 bar)	—
Nm³/day (ref 0°C, 1.01325 bar)	—
Sm³/hr (ref 15°C, 1.01325 bar)	—
▸ HYDRAULICS & METER
Vel. in Pipe	—
Reynolds No. (pipe)	—
Expansibility factor Y	—
Velocity of Approach E	—
Cd — ISO RHG f(Re,β,tap) [NOT constant]	—
▸ PRESSURE LOSS
Perm. Pressure Loss (% of ΔP)	—
Perm. Loss (mmH₂O)	—
Perm. Loss (mbar)	—
Perm. Loss (bar)	—
Flow Uncertainty ±U (k=2)	—
ΔP/P₁ Ratio (ISO limit <0.25 for Y validity)	—