Hydraulic

Control valve sizing for liquids: Cv, Kv and cavitation

IEC 60534-2-1 gives the Cv/Kv required to pass the flow rate with the available pressure drop. For liquids the calculation only closes when you also check the cavitation index, incipient cavitation and the FR viscosity correction at both extreme operating cases.

When to use

Use this when specifying a liquid control valve — on a network branch or on a pump discharge — where the flow rate varies and the regulation must stay stable and quiet. Sizing delivers the required Cv (or Kv) at each point, the percent opening, the required rangeability and, above all, the cavitation check. For viscous fluids (oils, solutions, low flow) apply the FR correction from Annex C; for fluids near their vapor pressure (hot water, light hydrocarbons) check the cavitation index against the manufacturer's incipient sigma.

Why liquids demand more than the Cv equation

Sizing a control valve looks, at first glance, like a one-line calculation: given an available ΔP and a flow rate, what flow coefficient is required? The basic IEC 60534-2-1 equation delivers that immediately:

Kv = Q · √(SG / ΔP)

The catch is that, for liquids, this equation is only valid under two assumptions that rarely coexist in practice: fully turbulent flow and the absence of cavitation. When the fluid is viscous, the equation undersizes; when the local pressure drops below the vapor pressure, it oversizes (flow no longer grows with ΔP). An honest sizing treats all three pillars together: capacity (Cv/Kv), cavitation and viscosity correction (FR).

The capacity equation and the Cv ↔ Kv conversion

Kv is the flow of water, in m³/h, that passes through the fully open valve with 1 bar of drop. Cv is the imperial analog (US gpm, 1 psi). The relationship is Cv ≈ 1.156·Kv. The standard defines the required Kv from the effective drop — and here lies the first subtlety: if cavitation occurs, the ΔP that matters is not the actual one but the ΔP_max (allowable drop), because above it the flow saturates.

Cavitation: choked, incipient and the cavitation index

As the liquid accelerates through the restriction, the internal pressure plunges at the vena contracta and then partially recovers. If the minimum pressure falls below the vapor pressure Pv, bubbles form and collapse on recovery — cavitation. The IEC model defines the allowable drop:

ΔP_max = FL² · (P1 − FF·Pv), with FF = 0.96 − 0.28·√(Pv/Pc)

  • P1 and Pv are absolute (saturation is referenced to absolute zero).
  • FL is the pressure recovery factor: high in globes (~0.9, cavitates late), low in butterfly and ball valves (~0.6, cavitates early).
  • Once ΔP_max is reached, the flow is choked — it saturates and full cavitation is present.

But the damage (noise, vibration, erosion) starts well before the choked point. That is why serious design uses the cavitation index:

σ = (P1 − Pv) / ΔP

The higher σ, the farther from cavitation. You compare the operating σ with the incipient σ (σ_i) the manufacturer publishes for the trim and opening (the ISA-RP75.23 methodology). Acceptance is σ ≥ σ_i, not merely “not choked.”

The FR viscosity correction

The Kv equation assumes turbulence. In oils, glycols, solutions or at low flow rates, the flow can be laminar or transitional, and the valve passes less flow than the turbulent Kv predicts. The IEC introduces the valve Reynolds number:

Rev = 76000 · Fd · Q / (ν · √Kv0)

where Fd is the trim style modifier and ν the kinematic viscosity [mm²/s]. For Rev ≥ ~10000, the factor FR = 1 (nothing to correct). Below that, FR < 1 and the required Kv rises: Kv = Kv0 / FR. Forgetting this correction is a classic cause of a valve that “won’t pass flow” in the field with a viscous fluid.

How the method works, step by step

  1. Define the two extreme cases. Valve open (maximum flow, weaker system/curve) and closed (minimum flow, stronger system/curve). Each imposes a different ΔP.
  2. Compute the available ΔP across the valve in each case, in bar.
  3. Check cavitation: compute ΔP_max and σ. If actual ΔP ≥ ΔP_max, use ΔP_max in Kv; in every case compare σ against σ_i.
  4. Compute Kv0 with Kv0 = Q·√(SG/effective ΔP).
  5. Apply FR (viscosity) to get Kv = Kv0/FR; convert to Cv.
  6. Select by the larger Kv (the open case sets the size), check the opening (70–85% at maximum, >10–20% at minimum) and the required rangeability Q_max/Q_min.

Practical design considerations

  • Right geometry for the ΔP: high drop and proximity to Pv call for a globe or multi-stage anti-cavitation trim, not a butterfly.
  • Margin by sigma, not by choked: size against the incipient σ, leaving slack on noise and erosion.
  • Installed characteristic: in systems with high distributed loss, the equal-percentage characteristic linearizes the gain — a common rule on pump discharges.
  • Absolute pressures in cavitation; gauge pressures only in the head balance.
  • Material and trim consistent with the fluid and with residual cavitation (hardened stainless steel, ceramic trim, multi-stage).

In short, sizing a control valve for liquids means cross-checking capacity (Cv/Kv per IEC 60534-2-1), the cavitation index (with the incipient σ check) and the FR viscosity correction, always at the two extreme cases — only then does the selection have a useful opening, a guaranteed flow rate and quiet operation across the full range.

Formulas and fundamentals

Flow coefficient Kv (turbulent liquid, IEC 60534-2-1) Kv = Q · sqrt(SG / ΔP)

Required Kv [m³/h per 1 bar] to pass flow rate Q [m³/h] with pressure drop ΔP [bar] and specific gravity SG (water=1). It is the hydraulic size of the valve at that point. Imperial equivalent: Cv = Q[gpm]·sqrt(SG/ΔP[psi]); conversion Cv ≈ 1.156·Kv.

Allowable pressure drop (choked flow / cavitation limit) ΔP_max = FL² · (P1 − FF·Pv)

Pressure drop at which flow saturates (chokes) and full cavitation occurs. FL is the valve's pressure recovery factor, P1 the upstream absolute pressure, Pv the vapor pressure and FF the critical pressure ratio factor. If the actual ΔP ≥ ΔP_max, use ΔP_max in the Kv calculation and the valve operates cavitating.

Critical pressure ratio factor FF FF = 0.96 − 0.28 · sqrt(Pv / Pc)

Corrects the saturation reference at the vena contracta. Pv is the vapor pressure and Pc the thermodynamic critical pressure of the fluid (water ≈ 221 bar abs). For water at normal conditions FF ≈ 0.96. It feeds the allowable pressure drop ΔP_max.

Cavitation index σ σ = (P1 − Pv) / ΔP

Ratio of the upstream pressure margin above vapor pressure to the imposed pressure drop. The HIGHER the σ, the farther from cavitation. Compare the operating σ with the manufacturer's incipient σ_i: σ ≥ σ_i prevents the onset of cavitation; σ below σ_mv indicates advanced/erosive cavitation.

FR viscosity correction (valve Reynolds number) Rev = 76000 · Fd · Q / (ν · sqrt(Kv0)); Kv = Kv0 / FR

For Rev ≥ ~10000 the flow is turbulent and FR=1. For lower Rev (viscous fluid, low flow), FR < 1 corrects the turbulent Kv0 upward. Fd is the valve trim style modifier, ν the kinematic viscosity [mm²/s] and Q in m³/h. Annex C of IEC 60534-2-1.

Standards & methods

  • IEC 60534-2-1 (sizing equations — flow capacity, incompressible liquids)
  • IEC 60534-2-1 Annex C (FR correction for non-turbulent / viscous flow)
  • ISA-75.01.01 (Flow Equations for Sizing Control Valves — ANSI equivalent)
  • ISA-RP75.23 (Considerations for Evaluating Control Valve Cavitation — sigma index σ)
  • IEC 60534-8-3 / IEC 60534-8-4 (aerodynamic and hydrodynamic noise prediction)

Typical reference values

Quantity Typical range Note
Recovery factor FL — globe 0.85 to 0.95 Cage/plug globe recovers little pressure; cavitates late. Typical design value 0.90.
Recovery factor FL — butterfly/ball 0.55 to 0.70 Recover a lot of pressure at the vena contracta; cavitate at a much lower ΔP than globes.
FF factor (critical) — water ≈ 0.96 FF drops as Pv approaches Pc; for cold/warm water it stays around 0.96.
Valve Reynolds number Rev (turbulent limit) ≥ 10,000 Above this FR=1; below it, apply the viscosity correction (viscous liquids / low flow).
Incipient cavitation index σ_i 1.5 to 5 (trim dependent) Manufacturer value; multi-stage anti-cavitation trim tolerates lower σ than standard trim.
Design opening at maximum flow 70% to 85% of travel Leaves a reserve without saturating; at minimum flow keep above ~10–20%.

Worked example

Warm-oil control valve on a pump discharge

Inputs

Maximum flow (open case)
80 m³/h
Minimum flow (closed case)
35 m³/h
ΔP across valve — open / closed
2.0 / 3.6 bar
Specific gravity SG
0.92
Kinematic viscosity ν
40 mm²/s
P1 abs / Pv / FL (globe)
5.0–6.2 / 0.12 / 0.90 bar, —

Results

Required Kv — open (FR=0.98)
≈ 55.3
Required Cv — open
≈ 64
ΔP_max (choked) — closed
≈ 4.9 bar
Cavitation index σ — closed
≈ 1.7
Required rangeability
≈ 2.3:1

In the open case, Kv0 = 80·√(0.92/2.0) ≈ 54.3; since ν=40 mm²/s and the flow is moderate, the valve Reynolds number drops to ~8800 and FR≈0.98 raises Kv to 55.3 (Cv≈64) — and it is this larger value the valve is bought against. In the closed case the pump rides up its curve: ΔP=3.6 bar stays below ΔP_max=0.90²·(6.2 − 0.96·0.12)≈4.9 bar, so there is NO choked flow. However the cavitation index σ=(6.2−0.12)/3.6≈1.7 is already low; if the chosen trim's σ_i is greater than 1.7, incipient cavitation (noise/erosion) is present even without saturating the flow. Practical decision: standard globe if σ_i<1.7, otherwise multi-stage anti-cavitation trim. The required rangeability of only 2.3:1 leaves ample margin against the 50:1 of an equal-percentage globe.

Run your own calculation Create a free account and get this for your own inputs in seconds.

Common mistakes

  • Sizing Kv only at the design flow rate and ignoring the 'valve closed' case (minimum Q), where the pump rides up its curve, the ΔP grows and cavitation appears.
  • Using a generic FL (0.9) for a butterfly or ball valve: those geometries have FL ~0.6 and cavitate at a much lower drop; the computed ΔP_max ends up far too optimistic.
  • Confusing incipient cavitation with choked flow: flow saturates at ΔP_max, but noise, vibration and erosion start well before that, at σ above 1 — which is why you use σ_i, not just the choked point.
  • Forgetting the FR correction for viscous fluids or low flow: with Rev < 10000 the turbulent Kv undersizes the valve and it will not pass the flow rate.
  • Working with gauge pressures in the cavitation equation: P1 and Pv must be ABSOLUTE (saturation is referenced to absolute zero).
  • Oversizing 'to be safe': the valve then operates at 5–15% opening, in the non-linear, low effective rangeability zone, degrading control.

Frequently asked questions

What is the difference between Cv and Kv?

They are the same flow capacity coefficient in different unit systems. Cv uses US gallons per minute and psi; Kv uses m³/h and bar. The conversion is Cv ≈ 1.156·Kv. IEC 60534-2-1 standardizes Kv (and Cv in the imperial version). Pick one and keep units consistent throughout the pressure balance.

What is the cavitation index and how do I use it in practice?

It is σ = (P1 − Pv)/ΔP, with P1 and Pv absolute. The higher it is, the farther from cavitation. You compare the operating σ with the incipient σ_i the manufacturer publishes for that trim and opening: if σ ≥ σ_i there is no onset of cavitation; if σ falls below the damage index (σ_mv) there is erosion. It is more conservative and more useful than waiting for choked flow.

What is the difference between incipient cavitation and choked flow?

Choked flow is the limit where ΔP_max = FL²·(P1 − FF·Pv) is reached and flow saturates — increasing ΔP no longer increases flow. Incipient cavitation begins MUCH earlier than that point, when the first bubbles form at the vena contracta (high σ). Designing only against choked flow lets noise and erosion through; that is why ISA-RP75.23 works with the sigma indices.

When do I need to apply the FR viscosity correction?

When the flow through the valve is no longer fully turbulent — viscous fluids (oils, solutions, glycols) or low flow rates in small valves. Compute the valve Reynolds number Rev = 76000·Fd·Q/(ν·√Kv0); if Rev < ~10000, FR < 1 and the required Kv rises (Kv = Kv0/FR). Ignoring FR undersizes the valve and it will not deliver the flow rate.

Why does the FL factor change everything about cavitation risk?

FL measures how much pressure the valve recovers downstream of the vena contracta. Geometries that recover little (globe, FL≈0.9) only cavitate at high drops. Butterfly and ball valves recover a lot (FL≈0.6), so the internal minimum pressure plunges and they cavitate at a much lower ΔP. Using a generic FL of 0.9 for a butterfly produces an optimistic ΔP_max and an unsafe selection.

Why check two cases and not just the design flow rate?

With a fixed-speed pump or a variable-pressure network, the ΔP across the valve is not constant. At maximum flow the valve opens and sets the LARGER Cv (the one you buy against). At minimum flow it throttles, upstream the pump rides up its curve, the ΔP grows and the cavitation index drops — this is the critical case. Only by checking both extremes is the selection robust.

Glossary

Cv / Kv
Valve flow coefficient: the flow that passes with 1 psi (Cv) or 1 bar (Kv) of drop, at the fluid SG. It quantifies the hydraulic capacity at each opening.
Cavitation index σ
Dimensionless index (P1−Pv)/ΔP that measures the distance to cavitation. Compared against the manufacturer's incipient sigma (σ_i) to accept or reject the application.
Incipient cavitation
Initial stage: the first vapor bubbles appear at the vena contracta. It occurs at high σ, before choked flow; it marks the threshold of noise and erosion.
Choked flow
Condition where the drop reaches ΔP_max = FL²·(P1 − FF·Pv) and flow saturates. Beyond it, more ΔP yields no more flow; full cavitation is present.
FL factor
Liquid pressure recovery factor. High (globe) means little recovery and late cavitation; low (butterfly/ball) means heavy recovery and early cavitation.
FR correction
Factor that adjusts Kv when the flow is not turbulent (low valve Reynolds number), typical of viscous fluids. Defined in Annex C of IEC 60534-2-1.