TURBINE FLOW METERS

MODULE T10

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 1

 CONTENTS

1.0 DEVELOPMENT OF THE TURBINE METER ......................................................... 3

2.0 METER CONSTRUCTION.................................................................................... 3
2.1 METER ASSEMBLY ............................................................................... 3
2.2 METER HOUSING.................................................................................. 4
2.3 ROTOR ASSEMBLY............................................................................... 4
2.4 PICK-OFF COILS ................................................................................... 4
2.5 PRE-AMPLIFIERS.................................................................................. 5
3.0 THE METER TUBE............................................................................................... 5
3.1 CONFIGURATION.................................................................................. 5
3.2 FLOW CONDITIONING........................................................................... 5
3.3 OTHER CONSIDERATIONS ................................................................... 5
4.0 OPERATING A TURBINE METER......................................................................... 6
4.1 TURBINE METER PERFORMANCE CHARACTERISTICS ....................... 6
4.2 DETERMINATION OF THE METER K-FACTOR ...................................... 6
4.3 FLOW MEASUREMENT ......................................................................... 7
5.0 THEORY AND UNCERTAINTY ............................................................................. 8
5.1 FLOW EQUATIONS................................................................................ 8
5.2 MEASURED DATA ................................................................................. 9
5.3 UNCERTAINTY EQUATIONS................................................................ 10
5.4 TURBINE EQUILIBRIUM ...................................................................... 11
5.5 ALTERNATIVE ANALYSIS.................................................................... 12
5.6 TEMPERATURE CORRECTION (Ctlm)................................................... 13
5.7 PRESSURE CORRECTION (Cplm) ......................................................... 15
5.8 IMPLEMENTATION of Ctl & Cpl.............................................................. 18
5.9 EQUILIBRIUM VERSUS REFERENCE.................................................. 19
6.0 PULSE INTEGRITY ............................................................................................ 19
6.1 TURBINE OUTPUT SIGNAL ................................................................. 19
6.2 IP 252.................................................................................................. 20
6.3 SOURCES OF ERROR......................................................................... 21
6.4 BAD PULSE ......................................................................................... 21
7.0 PULSE INTERPOLATION................................................................................... 21
7.1 METER CALIBRATION......................................................................... 21
7.2 PULSE COUNT .................................................................................... 22
7.3 DOUBLE TIMING METHOD.................................................................. 22
7.4 THE FOUR TIMING METHOD............................................................... 22
7.5 PULSE MULTIPLICATION.................................................................... 23

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 2

1.0 DEVELOPMENT OF THE TURBINE METER

The turbine meter is not a new flow measurement device with records showing the
first patent for a turbine meter was issued in 1886. However it was not until the
1950’s that the turbine meter was developed into a precise flow measurement device
for use in the development of jet engines and liquid propellant rockets. The turbine
meter began to be used in the petroleum industry in the mid-1960’s and its use was
confirmed and expanded by the publication of API Standard 2534 “Measurement of
Liquid Hydrocarbon by Turbine Meter Systems” in March 1970.

Turbine meters infer flow rate from the measurement of rotational movement (angular
velocity) of a bladed rotor or impeller suspended in the flow stream. This requires the
adoption of two basic principles:

- volumetric flow rate is proportional to average stream velocity

- average stream velocity is proportional to rotor angular velocity

The degree to which these two basic assumptions are correct determines the
accuracy of the turbine meter which is therefore influenced by design, installation and
application variables.

2.0 METER CONSTRUCTION

2.1 METER ASSEMBLY

The basic construction of a modern turbine meter consists of a bladed rotor running
on bearings supported by a central shaft with a whole assembly contained within a
pressure housing, usually in the form of a flanged pipe spool. The design of the rotor
assembly varies considerably between manufacturers, especially in terms of the
shape and number of blades and the design of the upstream and downstream
hangers.

One or more pick-off coils, mounted in the housing wall, detect the rotation of the
rotor blades resulting in a generated pulse train where each pulse is equivalent to a
discrete volume of fluid. Totalisation of these pulses gives the volume passed and
the pulse frequency gives the flow rate.

In some applications this pulse output can be fed directly to a local totaliser, but for
fiscal metering applications the pulses are transmitted to a flow computer via local
pre-amplifiers.

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 3

2.2 METER HOUSING

The meter housing basically only consists of a flanged or threaded pipe spool, of a
material compatible with the process fluid and be non-magnetic. For hydrocarbon
applications this is normally carbon steel or stainless steel, and the meter calibration
applies to the complete unit, i.e. the internals together with the housing. Some
designs of turbine meter (e.g. Faure Herman ‘Heliflu’) incorporate a removable liner
within the housing, and in this instance it is the liner and internals that form the
calibrated unit and meter replacement is therefore cheaper.

2.3 ROTOR ASSEMBLY

The rotor assembly of turbine meters varies significantly between the different
manufacturers, who include Faure-Herman, Daniel Industries, ITT Barton, Kent
Instruments and Flow Automation (ex AOT-Hydril). In general the assembly consists
of a rotor located between upstream and downstream hangers, sometimes referred to
as stators.

The most common design of rotor is multi-bladed but the Heliflu meter incorporates a
helicoidal rotor. Either the blades, blade tips or positions within the hub of the rotor
are magnetised or constructed of a paramagnetic material (a material which is
attracted by a magnet).

The design of the upstream and downstream hangers also varies between
manufacturers with some stating that their construction assists in straightening the
fluid flow profile.

Typical materials of construction for the meter internals are stainless steel with
tungsten carbide bearings.

2.4 PICK-OFF COILS

Turbine meters can have up to three pick-off coils but for high accuracy metering of
hydrocarbons two pick-off coils are used to provide dual pulse signals which enable
signal comparison and checking. The pick-off coils are external to the fluid flow and
detects the movement of the meter rotor by registering the magnetic flux from the
magnetised portion of the rotor. This magnetic flux generates a voltage within the
pick-off coils at a frequency proportional to the angular velocity of the rotor and hence
the fluid flow rate. This system is equivalent to a standard electrical generator, but
the output produced is only of a low voltage (typically millivolts) with a high
impedance.

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 4

2.5 PRE-AMPLIFIERS

In high accuracy systems where the pulse frequency must be transmitted to a remote
flow computer or totaliser, it is necessary to incorporate pre-amplifiers into the
system. These units are separate from the turbine meters and are usually mounted
locally to the meters, connected by a short length of cable. There is one pre-amplifier
for each pick-off coil, although more than one pre-amplifier can be contained in a
single housing. The pre-amplifier increases the signal amplitude and reduces the
source impedance to create a relatively high voltage (typically 12 volts) and low
impedance signal which is more suitable for signal transmission with a reduced
susceptibility to external noise sources.

3.0 THE METER TUBE

3.1 CONFIGURATION

Turbine meters are normally installed in a horizontal plane, often in a number of

parallel streams, in a manner minimising stress, strain and vibration. As turbine
meters are susceptible to mechanical damage it is normal to install an upstream filter
or strainer to provide protection from any debris or sizeable particles within the flow.
Valving, secondary instrumentation and proving systems are usually provided as part
of an overall metering system, depending on the operational requirements.

3.2 FLOW CONDITIONING

The performance of turbine meters is sensitive to the velocity profile and flow
characteristics of the fluid. To try and eliminate the occurrence of a distorted velocity
profile or the inducement of fluid swirl straight lengths of pipe are incorporated
upstream and downstream of turbine meters, usually with the addition of flow
straighteners.

The installation requirements to provide suitable flow conditioning are detailed in the
API Manual of Petroleum Measurement Chapter 5, Section 3. A typical system is
shown to require an upstream length of ten pipe diameters, with a flow straightener,
and a downstream length of five diameters. The design of the flow straightening
elements varies between manufacturers, and is sometimes provided as a ‘matched
pair’ with a turbine meter.

3.3 OTHER CONSIDERATIONS

The installation of a turbine meter must also ensure that there is sufficient back
pressure in, and immediately downstream of, the meter. This is to ensure that
cavitation, or vaporisation of the liquid stream, does not occur. Cavitation can have a
marked impact on meter performance at high flow rates.

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 5

4.0 OPERATING A TURBINE METER

4.1 TURBINE METER PERFORMANCE CHARACTERISTICS

As previously described each pulse generated by a turbine meter is equivalent to a

discrete volume of fluid. Therefore for each meter a calibration characteristic called a
K-factor is applied. This K-factor is expressed in pulses per volume, normally cubic
meters or litres. For each make, model and size of turbine meter there will be a
stated nominal K-factor. However each individual meter within such a grouping will
have a slightly different K-factor, that will change with variations in the flow rate.

The variation of a turbine meters K-factor with reference to the flow rate is termed
linearity and is demonstrated by plotting the meter K-factor against flow rate. A
typical meter characteristic shows a reasonably flat central section with steeper
profiles at low and high flow rates. This central section covers the flow limits for the
meter and will normally allow a flowrate turndown ratio of 10:1. For a 4 inch meter
this would typically be 30 to 300 m3/h.

The specified linearity for a typical high accuracy turbine meter is ± 0.25%, although
±0.15% is sometimes stated, especially for a reduced turndown ratio.

The meter K-factor is not only affected by flow rate but also temperature, pressure
and especially viscosity. Therefore the characteristics curve needs to be produced at
constant operating conditions, and is only valid for the specified type of fluid.
Therefore an initial factory calibration is only really applicable where the operating
conditions remain reasonably close to those for the calibration. Otherwise in order to
provide the optimum operating accuracy turbine meters are proved in-situ against a
master measure to allow new K-factors to be calculated for changing conditions.

The performance of turbine meters can also include a reference to repeatability.

This is basically a measure of how close calculated K-factors are for identical process
conditions. The typical repeatability for high accuracy turbine meters is ± 0.02%, and
must be achieved to provide consistent results when the meters are proved.

4.2 DETERMINATION OF THE METER K-FACTOR

In order to determine the meter K-factor, which is the number of pulses generated by
the meter for each unit of volume (e.g. m3 or litre) throughput, the meter must be
calibrated by another flow measuring device.

The meter must be calibrated over its full operating range of flowrate because the
meter K-factor changes with flowrate.

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 6

 The meter K-factor also changes with fluid viscosity. In theory the meter should be
calibrated using its operating process fluid at the operating pressure and temperature.
It is for this reason that fiscal turbine flowmeters have an on-line calibration facility,
i.e. the meter prover.

The calibration is compromised if the K-factor is determined with a different fluid.

Manufacturers do not have access to hydrocarbon pipelines so a manufacturer’s
calibration must be carried out using available fluids. Manufacturers normally
calibrate using water.

4.3 FLOW MEASUREMENT

Once the meter K-factor is defined the flow through the meter can simply be
determined by dividing the number of pulses generated by the meter K-factor:

Pulses
Volume =
Meter K - factor

Similarity the flow rate can be calculated by dividing the pulse frequency by the meter
K-factor:

Pulse frequency
Flow rate =
Meter K - factor

The units of the flow rate will be consistent with those of the pulse frequency and
meter K-factor. For example, for a turbine meter with a K-factor of 2000 pulses per
cubic metre, registering a pulse frequency of 10,000 Hertz, the flow rate will be:

10,000 pulses per sec

Flow rate = 3
= 5 m3 /sec
2,000 pulses per m

This is a volume flow rate measured at the flowing temperature and pressure, i.e. the
observed volume or gross volume flow rate. To allow flows at different conditions
to be compared it is normal practice to adjust the observed volume to either a mass
or a standard volume.

A mass total or mass flow rate is obtained by dividing the gross volume figure by the
flowing density:

Gross volume
Mass =
Density at the same conditions

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 7

 The standard volume total or flow rate are corrected from the gross volume conditions
to the metric standard conditions (MSC) of 1.01325 bar absolute and 15 degrees
celsius. This is achieved by undertaking a volume correction to account for the
temperature and pressure effects from meter conditions to standard conditions. The
two correction factors are termed Ctlm (effect of temperature on the liquid at the
meter) and Cplm (effect of pressure on the liquid at the meter) and are fully defined
below. The product of these two terms is called the volume correction factor
(VCF):

VCF = C tlm × C plm

The standard volume can thus be calculated by applying the volume correction factor
to the gross volume indicated by the turbine meter:

Standard volume = Gross volume × VCF

5.0 THEORY AND UNCERTAINTY

5.1 FLOW EQUATIONS

The turbine meter is classified as an “inferential” meter as strictly it is sensitive to the

velocity of the flowing fluid. To determine (i.e. “infer”) the fluid flow the cross sectional
area of the pipe must be known.

The turbine meter measures the “gross observed volume” i.e. the volume at operating
pressure and temperature:

f
qv =
k

Where:

qv = gross observed volume flowrate m3/h

f = turbine meter output pulses (Hz or pulses/second)
k = the meter K-factor (pulses/m3)

The mass flowrate may be determined as:

qm = q v × ρ f

Where:

qm = mass flowrate kg/s

ρf = density at operating conditions kg/m3

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 8

 The volume at reference conditions may be determined as:

qvs = q v × VCF
or
ρ
qvs = q v × f
ρs
or
q
qvs = m
ρs

Where:

q vs = volumetric flowrate at reference (“standard”) conditions, usually 15°C and

1.01325 bara
ρs = density at reference (“standard”) conditions, usually 15°C and 1.01325 bara
VCF = volume correction factor

For liquid the volume correction factor may be determined by:

VCF = C tlm × C plm

For gas the volume correction factor may be determined as:

Pf Ts Z s
VCF = × ×
Ps Tf Z f

i.e. the gas law correction for Pressure (P), temperature (T) and compressibility (Z)
for “flowing” (f) to “standard” (s) conditions.

5.2 MEASURED DATA

If the input measurements detailed in section 5.1 are considered then:

“f” the turbine meter signal output can be measured to very high accuracy

The operating pressure (p) and the temperature (t) can be measured within the
accuracy limits required.

“ρf” the density at operating conditions can be measured using a vibrating element
transmitter (densitometer).

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 9

5.3 UNCERTAINTY EQUATIONS

The uncertainties in the flow measurements may be determined by:

E qv = E 2f + Ek2

Where:

Eqv = The uncertainty in the measurement of gross observed volume

Ek = The uncertainty in the determination of the meter K-factor
Ef = The uncertainty in measuring the flow meter output. This is to very high
integrity so is several orders of magnitude less than Ek.

Thus, E f = 0 in practical terms.

Therefore:

E qv = E k

The uncertainty in gross observed volume is the uncertainty in meter K-factor.

Similarly the uncertainty in mass flow (Egm) is:

E gm = E 2qv + E 2ρ

Where Eρ is the uncertainty in measuring operating density.

The uncertainty in volume flow at reference conditions (Eqvs ) is:

E qvs = E 2qv + E 2vcf

Where E vcf is the uncertainty in determining the volume correction factor.

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 10

5.4 TURBINE EQUILIBRIUM

The theoretical relationship between the fluid mean velocity (which is proportional to
flow rate) and the turbine meter rotor speed is very complex.

If a number of simplifying assumptions are made the equation of equilibrium for rotor
torque can be expressed as:

[ ]
τ f + τpu = n b × ∫ 1 / 2 × ρ × ν2r × (Cl × sin θ - CD × cos θ) × lb × r dr

Where:

τf = Torque due to bearing friction

τpu = Torque due to pick-up drag
nb = The number of blades on the rotor
ρ = Density
υr = The relative velocity between the fluid flow (υ) and the blade (υb) = (υ - υb)
Cl = The coefficient of lift
θ = The relative velocity angle (from the velocity vector diagram)
CD = The coefficient of drag
lb = The blade length
r = The radius

These terms will vary as conditions change by the following factors (not necessary
the exhaustive list)

τf = The bearing friction will vary with rotor speed (ϖ), temperature (t) and fluid
viscosity (µ)
τpu = The pick up torque will vary with rotor speed (ϖ)
ρ = The fluid density will vary with temperature (t), pressure (p) and composition
υr = The relative velocity will vary with flow rate (q, i.e. υ), rotor speed (ϖ, i.e. υr
and torque (τ)
Cl = The coefficient of lift will vary with viscosity (µ)
θ = The relative viscosity angle will vary as per υr
CD = The coefficient of drag will vary with flowrate (qv, i.e. Reynolds number) and
viscosity (µ)
lb = The blade length will vary with temperature

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 11

 This equation may be restated as:

Driving Torque = Retarding Torque

The ideal turbine meter would be designed to balance this equation of equilibrium
throughout the range to give a linear relationship between rotor speed (ϖ) and
flowrate (qv) and be insensitive to process parameter changes i.e. it would have a
constant K-factor.

It is physically impossible to design this turbine meter, the K-factor of practical turbine
meters varies with flowrate and fluid viscosity. Note that fluid viscosity, temperature,
density, composition and pressure are interactive. All of these parameters will affect
the K-factor.

5.5 ALTERNATIVE ANALYSIS

A considerable amount of theoretical work has been carried out on the forces acting
upon a turbine meter rotor without any real alternative to calibration being obtained.

One of the alternative analytical equations is:

a2 a3
K = a1 + -
q q2

Where:

K = Meter k-factor (pulses per unit volume)

q = Flow rate
a1 = Fixed leakage term
a2 = Variable leakage term
a3 = Bearing friction term

At low level flow rates the bearing friction term (a3) dominates thus giving a low
K-factor.

As the flowrate increases the variable leakage term (a2) increases which increases
the K-factor.

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 12

 At higher flowrates the bearing friction (a3) and variable leakage term (a2) tend to
become small in comparison to the fixed leakage term (a1). Consequently the
K-factor tends towards a1 at high flowrate i.e.:

K = a1

This approach can be used to illustrate the conventional turbine meter calibration
curve of K-factor against flowrate. This is a steep rise in K-factor at low flowrate, a
positive kink, and a relatively linear section over the normal operating range.

5.6 TEMPERATURE CORRECTION (Ctlm)

Hydrocarbon fluids exhibit a significant change in volume with changes in

temperature. The meter K-factor normally determines the volume at actual operating
(flowing/line) conditions. The volume at the reference temperature must be
calculated.

The correction applied to refer the volume to reference temperature is defined in the
API Manual of Petroleum Measurement Chapter 11.1 Volume X (alias API 2540, alias
ASTM D1250, alias IP 200) as:

C t l = exp [- α15 × ∆t (1 + 0.8 × α15 × ∆ t)]

Where:

α15 = Coefficient of thermal expansion at the reference temperature of 15°C

∆t = The temperature deviation from reference (15°C)
= (tf - 15)
tf = operating (flowing) temperature

The basic definition of thermal coefficient (α) is:

1 dV
α = ×
V dt

Where V is the volume at any temperature.

From the experimental work it has been determined that:

α = α15 + β ∆t

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 13

 Where β is function of α and is independent of temperature. It was found that:

β = constant × α 15
2

The constant has been evaluated as 1.6.

Therefore:

α = α 15 + 1.6 × α15
2
× ∆t
or
α = α15 (1 + 1.6 × α15 × ∆t )

Combining with expressions for α:

1 dV
× = α15 + 1.6 × α 215 × ∆ t
V dt

This can be re-arranged and integrated between the reference and flowing
temperature to become:

 V  1.6
ln  f  = α15 × ∆t + × α15
2
× ∆t 2
 V15  2

which becomes:

V15
Vf
[
= exp - α15 ∆ t + 0.8 × α15
2
× ∆t 2 ]
and
V
C tl = 15
Vf
or
V15 = Vf × C tl

Where:

V15 = Volume at reference temperature (15°C)

Vf = Volume at operating (flowing) temperature

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 14

 The thermal coefficient (α15) can be calculated by:

K0 K1
α15 = +
ρ15
2
ρ 15

Where:

ρ15 = The density at reference conditions. K0 and K1 are empirical constants

determined by experimental work with a selection of hydrocarbon fluids.

α15 may also be determined by the laboratory analysis of a sample of the actual
metered fluid.

This equation can also appear in the form:

K 0 K1
α15 = + + K2
ρ15
2
ρ15

Where K2 is also an empirical constant. In most cases K2 = 0 so is excluded from

the equation.

This correction may be applied at different locations on the metering station to

account for temperature changes between their locations.

The Ctl at the meter becomes Ctlm.

5.7 PRESSURE CORRECTION (Cplm)

5.7.1 Cpl

Hydrocarbon fluids also exhibit a significant change in volume due to compressibility.

The volume must be corrected to the equilibrium pressure, or to atmospheric
pressure (whichever is higher).

The compressibility correction applied to refer the volume to reference pressure is

defined in the API Manual of Petroleum Measurement Chapter 11.2.1.M (alias ISO
9770) and Chapter 11.2.2.M. These two standards contain tables of correction
factors.

Chapter 11.2.1.M applies to hydrocarbons in the density range 638 to 1074 kg/m3.

Chapter 11.2.2.M applies to hydrocarbons in the density range 350 to 637 kg/m3.

The suffix ‘M’ refers to the metric edition of the standard.

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 15

 The equation used is:

Ve 1
= Cpl =
Vf 1 - F ∗ (Pf - Pe )

Where:

Ve = The volume at equilibrium (or atmospheric) pressure

Vf = The volume at operating (flowing) pressure
Pf = The operating (flowing) pressure
Pe = The equilibrium pressure (see section 5.9 below)
F = The correction factor, known as the “fluid isothermal secant compressibility”,
more commonly known as “β” in European usage.

This equation also appears in the form:

1
Cpl =
1 - β × (Pf - Pe )

The pressures Pf and P e must be expressed in the same terms, i.e. both as absolute
or both as gauge. For Custody Transfer Metering reference pressure Pref will replace
Pe (see section 5.9 below).

If the pressure is expressed as ‘bar’ then F (or β) must be expressed as “/bar”.

Similarly if the pressure is expressed as “Pascal” then β must be “/Pascal”.

5.7.2 Crude Oil API 11.2.1 M

For hydrocarbons in the density range 638 to 1074 kg/m3 the fluid isothermal secant
compressibility (β) may be calculated as per API Chapter 11.2.1 (M) as follows:

 0.87096 t 
β = exp - 1.62080 + 0.00021592 × t + + 0.0042092 × 2  × 10 - 4 bar -1
 ρ 15
2
ρ15 

Where:

t = Operating temperature in °C
ρ15 = Density at reference conditions in kg/litre NOT kg/m3
kg/m3
kg/litre =
1000

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 16

 This equation may be expressed in terms of kg/m3 as follows:-

 870960 t 
β = exp - 1.62080 + 0.00021592 × t + + 4209.2 × 2  × 10 - 4 bar -1
 ρ 15
2
ρ15 

β may alternatively be determined by laboratory analysis.

Prior to the issue of API Chapter 11.2.1.M β (‘F’) was determined by the Downer
equation as given by:

  ρ15 
1.38315 + 0.00343804 × t - 3.02909 × ln  
 1000 
β = exp  bar -1
  ρ15  
- 0.0161654 × t × ln   
  1000  

 ρ 
ρ15 is in kg/m3 i.e.  15  represents density in kg/litre.
 1000 

The Downer equation is now superseded but may still apply to some existing
contracts.

5.7.3 Condensate API 11.2.2 M

For hydrocarbons in the density range 350 to 637 kg/m3 the fluid isothermal secant
compressibility (β) may be calculated as per API Chapter 11.2.2 (M) as follows:

1
F = β =
A + (Pf - Pe ) × B

Where:

Pf = The operating (flowing) pressure

Pe = The equilibrium pressure
F = The correction factor, known as the “fluid isothermal secant compressibility”,
more commonly known as “β” in European usage.

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 17

 The terms ‘A’ & ‘B’ are determined as follows:

A = 6.894757 × (-2.146589 1 * 10 -6 × TR 2 + 1.577439 * 10 -5 × TR 2 × G 2 ...

. .. - 1.0502139 * 10 - 5 × TR 2 × G 4 + 2.8324481 * 10 -7 × TR 3 × G6 - 0.9545939 ...
... + 7.2900662 * 10 - 8 × TR 3 × G 2 - 2.7769343 * 10 -7 × TR 3 × G 2 + 0.03645838 × TR × G
... - 0.05110158 × TR × G + 0.00795529 × TR + 9.1311491 × G) * 10 5

B = (-6.0357667 * 10 -10 × TR 2 + 2.2112678 * 10 -6 × TR × G 2 + 0.00088384 × G ...

... - 0.00204016 × G 2 ) * 10 5

Where:

TR = Temperature in °R
9
= (t f × + 32) + 459.67 Where tf is temperature in Celsius
5
G = Relative Density at 60°F / 60°F

5.8 IMPLEMENTATION of Ctl & Cpl

The standards detailed above which set out the correction factors are published as a
set of TABLES. The equations given above are supplementary to these tables. As
published, the standards dictate that if the equations are used then a specified routine
of rounding and truncating must be carried out so that the calculated value obtained
must agree with the published table.

It should be noted that the standards, as published, are intended to be used to correct
bulk cargo (stock tanks or ship’s cargo).

Developments in computing power and the abilities of modern flow computers have
changed the scenario. The flow computers use the volume correction factors to
correct small flow increments in a continual ‘real time’ cycle.

Implementation of the rounding and truncating routines can lead to significant errors
in a ‘real time’ situation and it is now considered inadvisable to apply rounding and
truncating to the Flow Computer algorithms.

This is however a controversial issue and the implementation of the rounding and
truncating routines should be considered in each situation.

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 18

5.9 EQUILIBRIUM VERSUS REFERENCE

The pressure correction factor (Cpl), as published in the standard, is intended to

correct a bulk cargo to the lowest pressure (at reference temperature) that the oil can
exist in the liquid state (vapour equilibrium pressure – pe).

It is conventional for a Custody Transfer Metering Station to ‘correct’ to reference

(base) conditions (normally 1.01325 bar.a).

This can result in a ‘pseudo’ volume where a ‘standard volume’ is quoted for a
product which cannot exist, as a liquid, at the pressure and temperature quoted. This
is particularly true for Condensates.

It should be remembered that the Custody Transfer Metering Station is calculating

financial accountancy values and that a ‘pseudo standard volume’ is appropriate for
this purpose.

Therefore Cpl should be calculated to the reference (base) pressure ( pref alias pb) and
NOT to equilibrium pressure (pe):

1
C pl =
1 - β × (Pf - Pref )

Note that the term (Pf – Pref) represents the operating GAUGE pressure (pf in bar.g or
Pa.g) so that:

Pref = 1 if pressure is expressed as ABSOLUTE

Pref = 0 if pressure is expressed as GAUGE

It should however be appreciated that the standard volume so calculated may not be
able to exist as a ‘real’ liquid.

6.0 PULSE INTEGRITY

6.1 TURBINE OUTPUT SIGNAL

As stated in the uncertainty analysis in section 5.3 above the error in counting the
pulses from the turbine meter is low. This is achieved by applying high integrity
design criteria to the flow computer pulse counting circuits.

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 19

6.2 IP 252

The IP Petroleum Measurement Manual, Part XIII, Section 1, (originally IP 252/76)

“Fidelity and Security of Measurement Data Transmission Systems” (and also section
3) specifies techniques for achieving integrity of pulse transmission.

Five levels of integrity are specified and have been designated as security protection
levels ‘E’ to ‘A’ as follows.

6.2.1 Level ‘E’ (Single Pick-Up)

Error reduction is achieved solely by correctly installed apparatus of good quality.

This is straightforward scaler totaliser system.

6.2.2 Level ‘D’ (Single Pick-Up)

Manual error monitoring at specified intervals by methods of comparison. This level

of security is intended to give protection against functional errors and failures and is a
method of verification by manual action. It has the means for checking the read-out
visually against an independent totalising system.

6.2.3 Level ‘C’ (Dual Pick-Up)

Automatic error monitoring and error indication at specified intervals by methods of

comparison. This level of security is intended to give protection against functional
errors and failures and this may be achieved by design methods acceptable to an
approving authority (if appropriate). The time intervals for error monitoring are
subject to revision in the light of experience gained.

6.2.4 Level ‘B’ (Dual Pick-Up)

Continuous monitoring, error indication and alarm signalling by methods of

comparison. This level of security is intended to give warning of transients and other
spurious influences, supply borne and radiated, in addition to functional errors and
failures.

6.2.5 Level ‘A’ (Dual Pick-Up)

Continuous verification and correction by methods of comparison. Errors must be

signalled even though they are corrected. This level of security is intended to given
protection against transients and other spurious influences, supply borne and
radiated, in addition to functional errors and failures.

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 20

6.3 SOURCES OF ERROR

Two significant sources of error must be considered.

6.3.1 Noise

Spurious pulses may be generated by electro magnetic interference. Unless the flow
computer can discriminate between genuine and noise pulses an over-reading would
occur.

6.3.2 Missed Pulses

Under certain circumstances, e.g. a broken blade on the turbine meter, pulses may
be lost and the meter will under-read.

6.4 BAD PULSE

For security levels ‘A’, ‘B’ and ‘C’ dual pick-ups are fitted to the turbine meter. The
two pulse trains can be compared and any difference registered as a “Bad Pulse” or
as a “Differential Pulse Alarm” (DPA) to use different company terminology.

Bad pulses will occur on meter start-up and on meter shut-down and under these
circumstances an alarm is not significant.

During normal operation an occasional bad pulse alarm may occur. Occasional
alarms are not significant. If frequent alarms occur the meter stream must be shut
down. By common practice the flow computer will annunciate after one hundred
errors have been detected.

The bad pulses must be totalised.

7.0 PULSE INTERPOLATION

7.1 METER CALIBRATION

The meter K-factor should be expressed to five digits (Reference ISO 4267-2 / BS
7286 Part 2, Table 1). The normal method of calibrating a turbine meter is to count
the number of pulses (f) transmitted by the meter as a known volume of fluid (V)
passes through the meter.

Therefore:

f Number of Pulses
K= =
V Volume

In order to express K to 5 digits a minimum number of pulses must be counted. The

DTI Guidelines specify this minimum to be 20,000 pulses.

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 21

7.2 PULSE COUNT

Electronic pulse counters count in integers, hence the requirement for the minimum of
20,000 discrete pulses.

If fractions of a pulse could be measured then less pulses (f) would be required OR a
smaller known volume (V) would be required to maintain the same resolution.

Pulse interpolation is the term applied to methods used to interpolate fractions of a

pulse.

The following methods of pulse interpolation are applied to pipe provers and compact
provers where “START” and “STOP” signals gate the meter pulse count (f) over the
calibrated volume (V).

7.3 DOUBLE TIMING METHOD

If the elapsed time (τ) is measured between the detector start and stop gate signals
(τsd) and the first pulse leading (or trailing) edge after each gate signal (τpc ) then this
ratio can be applied to the pulse count (f).

τpc
flp = f ×
τsd

Where:
f is an integer
flp is the interpolated pulse count and is a fraction

This method assumes steady state conditions throughout the meter calibration which
is not often achieved.

7.4 THE FOUR TIMING METHOD

The four timing method can accommodate flow rate changes between the start and
end of the meter calibration.

If the periodic time (the time per pulse i.e. the inverse of frequency) of the turbine
meter signal is measured at the start (τm1) and the end (τm2) of the calibration and the
time period between the detector gate signal and the first leading (or trailing) edge of
a pulse is measured at the start (τg1) and end (τg2) of the calibration then:

τg1 τg2
flp = f + -
τm1 τm2

This method does not account for instability during the meter calibration.

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 22

7.5 PULSE MULTIPLICATION

If each pulse is multiplied by a factor (Fpm) before the turbine meter signal input is
connected to the pulse counter the resolution can be improved.

i.e.

flp = f × Fpm

The multiplication MUST be exact for each pulse received. For example, if the pulse
multiplication (Fpm) is set to 10 then 10 pulses MUST be generated for EACH turbine
meter pulse received. Pulse multiplications of 9 (or less) or 11 (or more) cannot be
accepted.

The electronic technique employed is termed the “phase locked loop”. An electronic
control loop multiplies each pulse and monitors the pulse signals. If an error is
detected the “phase lock” is lost and the output fails totally. The pulse counter can
detect this total failure and will abandon (abort) the meter calibration (proof).

The Spectra-Tek PIM which is widely used in the North Sea applies this technique.

F:\Central\Tyro\ TYRO T10.doc Rev 1 October 1999 Page 23