Fluids and Solids Handling

Size Control Valves for

Lab-Scale Laminar Flow
Paul Martin, P.Eng. Existing equations used to size control valves are
Grant Girouard
Zeton Inc. inadequate for laboratory- and pilot-scale work.
Use the simple equation introduced in this article
to accurately size needle control valves
for laminar and transitional flow.

S
izing valves for laboratory- or pilot-scale use must Consider the flow of a fluid through a needle control
account for flow in the laminar and transitional regimes. valve sized for a lab or pilot plant. The valve’s trim, which
Conventional equations for calculating the flow coeffi- consists of the seat and the plug, defines the flow path.
cient (CV) — the essential parameter for sizing control valves (Figure 2 shows three types of trims commonly found in
— work only for turbulent flow (Re > 10,000). needle control valves.) The annulus formed between the
This article provides an overview of the available equa- seat hole (on the order of 0.060 in. dia.) and the needle or
tion to calculate CV and discusses its applicability to turbu- plug (on the order of perhaps 0.0595 in. dia.) is very small,
lent flow. It then offers a new way to size valves for flow
in laminar and transitional regimes. The discussion focuses Stem Packing
on the needle valve (Figure 1), which is the most accurate Packing Nut
among the various types of control valves and is most often
used for lab- and pilot-scale work.

Laminar and transitional flow

The flow coefficient (CV) is the volumetric flowrate
(gpm) of water that can pass through a valve with a 1-psi Stem
pressure drop across the valve at standard temperature. Valve
Body Plug or Needle
manufacturers provide flow capacities expressed in terms of
CV for their valves of varying sizes and types.
The conventional equation used to calculate CV is:

Flow
where Q is the volumetric flowrate (gpm), SG is the specific
gravity of the fluid relative to water (dimensionless), P1 is Seat
the inlet pressure (psi), and P2 is the outlet pressure (psi).
This equation is valid only for turbulent flow.
In the small valves used in the lab or pilot plant, how- p Figure 1. A typical needle control valve consists of a body, stem and
ever, the flow of liquids is often not turbulent, despite what plug, seat, and packing. The body encases the valve parts and includes the
fluid flow path and the inlet and outlet connections. An actuator moves the
many may believe.
stem up and down through the packing in the body. The plug (or needle) is
* This article is based on a paper presented at the AIChE Annual Meeting, a machined part on the end of the stem that fits into the seat of the valve.
Nov. 3–8, 2013, San Francisco, CA. The internal parts of the valve are collectively referred to as the valve’s trim.

48  www.aiche.org/cep  March 2014  CEP Copyright © 2014 American Institute of Chemical Engineers (AIChE)
 Open Open
Closed Closed Closed Open

where CVL is the flow coefficient for laminar or transitional

flow (gpm/psi0.5) and CVT is the flow coefficient for turbulent
Plug flow (gpm/psi0.5). FR approaches 1 as flow approaches fully
developed turbulent conditions, and becomes significantly
Seat less than 1 as flow becomes laminar.
For laminar conditions, ISA defines FR as:
Plug

Seat
where a and b are constants that depend on whether the goal
Seat-Guided Vee-Grooved Stem-Guided is to size a valve for a particular flowrate and pressure drop,
Needle Trim Needle Trim (Full-Size) Trim
to calculate the flowrate given the CVT and the pressure
p Figure 2. Three common types of needle valve trims are seat-guided, drop, or to predict the pressure drop required to obtain a
vee-grooved, and stem-guided trims. The trim type determines the fluid given flowrate through a valve of known CVT. For transi-
flow path through the valve. In a seat-guided valve (left), the flow path is
a circular annulus that varies in both length and inner diameter with the
tional flow, FR simply transitions smoothly between the two
travel of the valve, and the plug is never completely removed from the equations FR = aReb and FR = 1, and is generally given in
seat, even when the valve is 100% open. In a vee-grooved valve (center), tabular or graphical format.
vee-shaped grooves are milled into the plug and the fluid flows along these Unfortunately, determining the valve-trim Reynolds
grooves. In the stem-guided trim (right), which is also referred to as full-
size trim, opening the valve raises the plug and creates a circular hole that
number (Re in the annulus between the plug and the seat) in
the fluid passes through when the valve is 100% open. valve trims of the size used in lab and pilot plant equipment
(CV < 1) is not straightforward. The ISA method does not
so the velocity of the fluid moving through this valve clearly specify how to determine or select the appropriate
will be very high. But high fluid velocity alone does not variables (velocity, diameter, etc.) to use in the equations.
mean that the flow is turbulent (Re > 10,000), because the And, valve manufacturers consider the geometric dimen-
Reynolds number depends on more than just the velocity. sions of their trims a proprietary part of their design and
In these small needle valves, the surface area (or wetted seldom publish that information.
perimeter) of the annulus is large enough to stabilize lami- Apply Page’s equation. George W. Page described a
nar flow, even at very high velocities. Thus, Eq. 1 is not simpler and more elegant method of sizing valves in the
applicable to pilot-scale needle valves. laminar and transitional flow regimes (2). This method
Applying Eq. 1 to a system operating at laminar or tran- models a control valve in the transitional flow regime as
sitional flow conditions will result in valves that are signifi- two valves in series: one valve in purely laminar flow
cantly undersized, sometimes by many trim sizes. (representing flow through the body piping), and one in
turbulent flow (representing flow through the throttling ori-
Methods to correct for laminar flow fice). Applying the simple equations in Crane (3) for each
Rely on vendor expertise. Some manufacturers of lab- of the two conditions, Page arrives at a new equation for the
and pilot-scale control valves are aware of the consequences laminar/transitional flow sizing correction factor, FR:
of laminar flow on sizing and have developed their own
methods to correct for this phenomenon. However, we have
found that vendors’ sizing predictions for viscous liquids do VT
not match measurements; frequently, the predictions and the
measurements can differ by 200–500%. Therefore, we do
not recommend relying solely on vendor sizing for laminar where v is the fluid kinematic viscosity (cSt), L is the length
flow conditions. of pipe between the upstream and downstream pressure-
Use the ISA method. The International Society of Auto- measurement taps (in.), and DO is the diameter of a circle
mation (ISA) has developed a rather complex and cumber- that has the same area as the valve opening (in.):
some method to size valves in the laminar and transitional
flow regimes (1). This iterative method is often simplified to
nomographs that can be used to determine an effective valve
Reynolds number and the associated valve sizing correction where A is the area of the annulus between the plug and the
factor (FR). FR is a “fudge factor” and is expressed as: seat (in.2).

Copyright © 2014 American Institute of Chemical Engineers (AIChE) CEP  March 2014  www.aiche.org/cep  49
Fluids and Solids Handling

Page demonstrated good agreement between the results hypothesized that families of trims of similar geometry
of Eq. 4 and experimental data. However, this equation would need only a single parameter to describe their laminar
is not easy to use because DO is not always known. DO is flow performance relative to that observed in turbulent flow,
calculated from the valve geometry, and is equal to the valve since the same underlying geometric factors define pressure
size only for full-sized globe-type trims. drop for a particular trim type under any flow regime.
For a lab- or pilot-scale valve, frictional loss in the inlet The ISA (1, 5) provides the relationship FR = aReb,
and outlet tubing is not the controlling resistance to flow with the exponent b set to 0.5, 0.67, or 1, depending on the
and seldom contributes detectably to the overall pressure intended use. We tested the relationship for b = 0.5.
drop, so L is meaningless in the laminar flow situation. Consider a valve of a particular geometry with a particu-
Thus, Eq. 4 is not valid for nonturbulent conditions in lar fluid flowing through it at a constant valve opening (%).
reduced-port valves. The Reynolds number, Re = DVρ/µ (where D is the diameter
Use Zeton’s K-factor method. Rearranging Page’s equa- of the opening in meters and V is the velocity in m/sec) varies
tion (Eq. 4) reveals that L/DO2 is an adjustable parameter directly with the volumetric flowrate. Taking the definitions
with little practical physical meaning. Indeed, it is essen- of K (Eq. 7) and FR (Eq. 3), and then grouping all of the pro-
tially a parameter of fit for the Fanning/Darcy equation (4). portionality constants together as av, we obtain the following
When flow is purely laminar, Page’s equation (Eq. 4) (2) relationship:
reduces to:

where av and K are both lumped parameters of best fit (with-

VT out a clear physical interpretation). Substituting the laminar
flow geometric parameter av into Page’s equation (Eq. 8)
Upon rearrangement of this equation, it appears that a gives the following simple relationship:
new parameter, K, can be constructed to account for the con-
tributions of the lumped parameter L/DO2, the fluid density
(ρ, kg/m3), and a unit conversion factor. K can be obtained
from a plot of flow versus ΔP/µ, where ΔP is the pressure
difference across the valve (psi) and µ is dynamic viscosity Next, the equations that define CV (Eq. 1) and FR (Eq. 2)
(cP). This plot is a straight line through the origin for a valve can be combined with Eq. 10 to derive a new sizing equation:
in fully laminar flow (Figure 3). Accordingly, the value of K
for any particular valve can be determined through a single
experiment with a Newtonian fluid viscous enough and at
a flowrate low enough to ensure fully laminar flow through
the valve trim. (The flow must be fully laminar such that the Solving for Q yields a somewhat more complex equation:
flowrate varies linearly with ΔP/µ.)
Thus, for fully laminar flow:

Equation 12 reduces to Eq. 1, which is the definition of

Page’s equation then reduces to: CV for turbulent flow, when av = 0 or when the product µav2
is sufficiently small. A high av value for a particular trim
geometry indicates that the trim has a tendency to stabilize
laminar flow to a greater extent than trims with a lower av
value, perhaps due to the higher ratio of wetted perimeter to
Separate experiments for each trim size are required cross-sectional area in the flow path.
to determine the value of K, which varies with the CVT This model uses a single parameter to correct for the
of the valve. laminar flow behavior of a series of geometrically similar
trims, across a wide range of CVT values. For two popular
Zeton’s new laminar flow sizing parameter trim brands, we found that av values of 0.135–0.14 provide
An equation relating K and CVT could reduce the number an excellent fit.
of experiments needed to determine K for a series of trims This simple, one-parameter model given in Eqs. 11
that have similar geometry but different CVT values. We and 12 may be cautiously extended to use with other seat-

50  www.aiche.org/cep  March 2014  CEP Copyright © 2014 American Institute of Chemical Engineers (AIChE)
guided annular needle-type valves. In this case, the model and used with the liquid density to calculate volumetric
equations should not be trusted completely and should be flowrate.
verified with experiments. However, it will surely give bet- Data were collected over a range of applied pressures
ter results than either the turbulent sizing equation, or the (5–500 psig) for several trim sizes. Brand A’s manufacturer
ISA sizing equations without values for Fd and D validated sells valves with trim sizes designated by letters ranging from
by real testing on viscous liquids. A (largest) to P and micro-flow trims designated as P1 (larg-
Equations 11 and 12 with a conservative value of av est of the micro trims) to P14. The verification tests evaluated
(e.g., 0.18) can be used for the approximate sizing of annular trim sizes M, P1, P3, and P5.
needle-type trims in the absence of other data. Equation 11 Gas flow measurements were taken to verify the actual
with av = 0.18 can also be used as a test of whether laminar- CVT of the valves. A rotameter and thermal mass flowmeter
flow correction is necessary. If the calculated FR is signifi- were used to measure the flowrate of bottled nitrogen, with
cantly less than 1, it would be prudent to either perform measurements made in choked flow (i.e., P1 > 2P2).
validation experiments with your particular trim and fluid, or Figure 3 presents the results of the first set of tests, which
to have your valve supplier do so for you. used two fluids — Liquid 1 (a kerosene-like solvent with
µ = 0.8 cP and ρ = 788 kg/m3 at 20°C) and Liquid 2 (a paraf-
Model verification for seat-guided needle trims fin white mineral oil with µ ~ 29 cP and ρ = 846 kg/m3 at
The model (Eqs. 11 and 12) was verified for a pneumati- 20°C) — flowing through a 0.25-in. Brand A control valve
cally actuated needle control valve (Brand A). The valve’s with an M trim size. The vendor specifies this valve’s CVT to
trim, which is a type widely used in lab- and pilot-scale be 0.01 nominal; the measured CVT based on nitrogen flow
work, has a proprietary design, the dimensions of which measurements is 0.012.
are not published. In general, however, the trim involves As shown in the Q vs. ΔP/µ plot in Figure 3, the model
a small-bore circular seat with a tapered needle plug that of Eq. 8 is in excellent agreement with the experimental
is never completely removed from the seat even when the data for Liquid 2 in purely laminar flow in this trim, with FR
valve is fully open (i.e., the plug is seat guided). under these conditions ranging between 0.25 and 0.6. The
The testing apparatus consisted of a reservoir of liquid more-general model of Eq. 10 does a good job of predicting
that could be pressurized with air or nitrogen, a filter to the experimental data for the much-less-viscous Liquid 1 (FR
remove fine particulate matter, an accurate differential- ranging from 0.93 to about 1) in transitional to turbulent flow.
pressure transmitter with tubing tees mounted immediately The next set of tests used more-viscous fluids, Liquid 3
upstream and downstream of the valve body to connect the (µ ~ 90 cP) and Liquid 4 (µ ~ 296 cP), and Brand A valves
pressure-measurement impulse lines, a bucket or vial, and with the smaller trims of P1 (CVT ~ 0.002), P3 (CVT ~ 0.001),
an accurate scale. Data on weight and time were collected and P5 (CVT ~ 0.0004). The unique K values obtained for each
trim size were in excellent agreement with Eqs. 8 and 10.
0.08 Figure 4 is a plot of K–0.5 vs. measured CVT for all of the
0.07
valve trims of Brand A that were tested, including the M
trim at varying stem positions. The slope of this plot is the
0.06
laminar flow sizing coefficient, av, which for this brand and
Flowrate, gpm

0.05

0.04 0.10
0.09
0.03
0.08
0.02
0.07
0.01 0.06
K

0.00 0.05
1

0 10 20 30 40 0.04
ΔP/μ, psi/cP
White Mineral Oil Solvent 0.03
Experimental Experimental 0.02
Model (Eq. 8) Model (Eq. 10) 0.01
0.00
p Figure 3. The volumetric flowrate as a function of ΔP/µ is plotted for
0 0.002 0.004 0.006 0.008 0.010 0.012 0.014
two liquids. The model of Eq. 8 fits well for Liquid 2 (mineral oil) in purely
laminar flow (FR under these conditions ranges between 0.25 and 0.6), C VT
while the more-general model of Eq. 10 does a good job of predicting the
experimental data for transitional to turbulent flow for the much-less- p Figure 4. K–0.5 vs. measured CVT for all of the Brand A valve trims. The
viscous Liquid 1 (solvent) (FR ranges from 0.93 to about 1). slope of this plot is the laminar-flow sizing coefficient, av.

Copyright © 2014 American Institute of Chemical Engineers (AIChE) CEP March 2014 www.aiche.org/cep 51
Fluids and Solids Handling

these trims had a value of about 0.135. The graph demon- Using the same setup as for the seat-guided valves, av
strates the excellent fit of the new model (Eqs. 11 and 12) values were measured for the vee-grooved needle valve
to the entire set of measured data for the entire family of in the range of 0.05–0.065. The av values decreased as the
trims. valve’s open area decreased, falling to 0.02 in some of the
Subsequent testing used a different brand of control trims. Even at av values in the range of 0.02–0.06, the FR
valves with similar seat-guided needle-type trims in a similar values for these valves were significantly less than one with
trim size range (CVT = 0.01–0.0001). In these tests, av had liquids of even modest viscosity. Assuming that flow in these
similar values, in the range of 0.12–0.14. The value of av valves is turbulent, therefore, will lead to the selection of
did not change much with trim travel, similar to what was valves far too small for the duty.
observed for Brand A.
Low-flow control valves manufactured by other vendors Stem-guided trims
with similar seat-guided needle-type valve trims were found to Based on our observations, we would not expect av to
have av values of similar magnitude. An experiment on a single remain constant with varying stem position for valves with
valve from a third manufacturer also demonstrated that the av full-size trims. These valves are fully stem guided rather
value was constant as the valve’s open area changed. than seat guided, and the trim becomes an unrestricted
circular orifice when the valve is fully open. Such valves
Vee-grooved needle trims include some of the fine-metering manual needle valves
The new equations were also demonstrated for a vee- offered by several tube-fitting manufacturers. Individual
grooved needle valve. This valve has a trim consisting sizing experiments using the K factor previously discussed
of a cylindrical plug with one or more triangular grooves would be required for accurate sizing of this type of valve
of varying depth machined into the plug. The triangular in laminar flow. CEP

groove trim design has a smaller ratio of wetted perimeter

to cross-sectional area than a trim with a purely annular
cross section, which translates into a higher Reynolds num-
Acknowledgments
The authors wish to thank Kevin Nold, Min Kwan Kim, Johnathan Caguilat,
ber at a given flowrate and viscosity than a circular annular Kenneth Edwards, and Sun Young Gong for their painstaking work in collect-
trim of the same CVT. Thus, this type of valve would be ing the data on which this article is based.
expected to have a smaller av value.

Literature Cited
Nomenclature 1. International Society of Automation, “Flow Equations for
A = area of the annulus between the plug and the seat Sizing Control Valves,” ISA/ANSI 75.01, Research Triangle
of the valve, in.2 Park, NC (1975).
av = trim family valve-sizing factor for laminar flow, 2. Page, G., “Simplified Valve Sizing for Laminar Flows,” Chem.
(gpm/cP)0.5 Eng., 105 (11), pp. 195–198 (Oct. 1998).
CVL = flow coefficient for laminar and transitional flow, 3. Crane Co., “Flow of Fluids Through Valves, Fittings, and Pipes,”
gpm/psi0.5 Technical Paper No. 410, Crane, New York, NY (1982).
CVT = flow coefficient for turbulent flow, gpm/psi0.5 4. Sakiadis, B. C., “Fluid and Particle Mechanics,” in Green, D. W.,
DO = diameter of a circle with the same area as the and R. H. Perry, eds., “Perry’s Chemical Engineers’ Handbook,”
valve opening, in. 6th ed., Section 5, p. 5-24, McGraw-Hill, New York, NY (1984).
FR = laminar flow adjustment ratio (dimensionless) 5. Baumann, H. D., “A Unifying Method for Sizing Throttling
K = single trim valve-sizing factor for laminar flow, Valves under Laminar or Transitional Flow Conditions,” Journal
psi/gpm-cP of Fluids Engineering, 115 (1), pp. 166–169 (Mar. 1993).
L = length of pipe between upstream and down-
stream pressure measurement taps, in.
P1 = valve inlet pressure, psi PauL Martin, P.Eng., is a senior project manager and engineer at Zeton
P2 = valve outlet pressure, psi Inc. in Burlington, Ontario, Canada (Phone: (905) 632-3123; Email:
[email protected]). He has 17 years of experience designing pilot-
ΔP = pressure difference across valve, psi and demonstration-scale plants for the chemical process industries. He
Q = volumetric flowrate, gal/min received his BASc and MASc from the Univ. of Waterloo, Ontario, both in
Re = Reynolds number (dimensionless) chemical engineering. He is a member of the AIChE Process Develop-
ment Div. and a licensed P.Eng. in Ontario.
Greek Letters
ρ = density, kg/m3 Grant Girouard is a project engineer at Zeton Inc. in Burlington, Ontario,
Canada (Phone: (905) 632-3123; Email: [email protected]). He
µ = dynamic viscosity, cP received a BASc from the Univ. of Waterloo, Ontario, in chemical
υ = fluid kinematic viscosity, cSt engineering.

Copyright © 2014 American Institute of Chemical Engineers (AIChE)

52 www.aiche.org/cep March 2014 CEP