Professional Documents
Culture Documents
Aga PDF
Aga PDF
Brent Berry
ABB Automation, Totalflow
Bartlesville, OK
Page 3
conditions) and various geometries (diameters). edge sharpness, wall roughness, eccentricity and flow
Once all the data is taken then an empirical equation conditioning, etc. are so important. Ideally your flow
is derived which allows us to compute Cd over many measurement system would be exactly the same as
combinations of conditions. was used in the laboratory.
(1) It is assumed there is no energy loss between the Density at Base Conditions
taps.
Pb Mrair Gi
(2) The velocity profile (Reynolds number) Ub eq. 4
influences are not fully treated by the equation.
Zb R ( Tb N5 )
It is assumed that some installation effects and
causes of flow perturbations (changes) are Notice Pf , Tf , Z f , Gi , Pb , Tb , and Zb in eq 3
insignificant. and eq 4. These represent temperatures, pressures,
specific gravities and compressibilities. It is these
(3) Different tap locations affect the flow rate. Tap variables that eventually make there way into the old
location is assumed for a given Cd. factors Ftb, Fpb, Ftf, Fgr, and Fpv (see Section 3 of
this document for more information). Leaving
Through rigorous testing, you could develop a densities in the fundamental equation, rather than
unique Cd for each of your orifice meters. This hiding them in a plethora (abundance) of factors,
technique, referred to as in-situ calibration, is seems less confusing and more instructive.
something like proving a linear meter. However it is
somewhat bothersome since you need a unique Cd
for each expected flow rate. Economics usually Real VS Ideal Gas Specific Gravity
make in-situ procedures unfeasible. One other item of note regarding the density
equations is that they are based on Gi, ideal gas
Therefore, the goal is to develop a universal Cd that specific gravity. Most systems have historically
everyone can use. To accomplish this, one must provided Gr , real gas specific gravity, which is
control their orifice meter installation well enough so different. Additionally AGA-8 requires Gr as an
that it replicates the same orifice meters used in the input, not Gi. Strictly speaking, Gi, is related to Gr
laboratory from which the universal Cd equation was with the following equation.
derived. This is referred to as the law of similarity.
If your orifice meter system is acceptably similar to
the laboratory's then your Cd will be acceptably
similar to the laboratory derived Cd. That is why
Page 4
Gi Gr «
air
ª Pm Tm N 5 Z m º
gas gas
»
Section 2 - SUMMARY OF CHANGES TO THE
STANDARD
¬ gas
« Pm Tm N 5 Z m »
air air
¼
Change 1, Document Organization
Page 5
Change 2, The nomenclature of the fundamental
equation was modified slightly. Once the data was gathered and accepted, talented
people, using computers, derived a Cd equation that
It is important to note that the fundamental equation has a high degree of correlation with all the new data.
did not actually change. Since it is based on the This means that, within the stated uncertainties of
actual physics of an orifice meter you would not Part I of the new standard, you can feel confident that
expect it to change. However, nomenclature was when applied as specified by the standard, the new
slightly modified. Cd equation will produce dependable answers.
eq. 1, AGA-3-1985 Fundamental Equation The tricky part about Cd is that one needs to know
the flow rate to compute it. But one also needs the
Cd S Cd to compute the flow rate. This is a sort of catch
Y d2 2 gc U f 'P 22. In a situation like this we say the equation is not
1- E 4 4
Qv = in closed form. This is why you will hear people say
Ub an iterative solution is required to compute the new
Cd.
Page 6
diameters based on the difference between the Change 6, Fpv, supercompressibility is computed
reference temperature and the actual fluid using AGA-8
temperature.
Many people have been using NX-19 to compute Fpv
This means that virtually none of the equation can be for natural gas. The new standard specifies AGA-8.
pre-computed and re-used. Even though the new
equation does not have Fb, there are portions of the A new AGA-8 standard was published in late 1992.
equation that depend only on the diameters. In the That standard documents two possible ways to
past, we would compute those portions of the compute Fpv. One method is referred to as gross
equation only when the diameters manually changed. method, the other is referred to as detailed method.
Now, since the diameters are a function of The gross method is supposed to be simpler to
temperature they, and everything based on them, implement and require less computing power than the
must be computed on a continual basis. detailed method. Having worked with both, I can tell
you that compared to either of these methods NX-19
Assume the measurement system is supplied dr, processing requirements are relatively minuscule
orifice diameter at reference temperature and Dr, pipe (small).
diameter and reference temperature. Before these
diameters can be used anywhere in the flow rate As a user, there are two major distinctions between
calculation they must be corrected for thermal effects the gross and detailed methods you should consider.
with the following equations.
1. The gross method accepts the same composition
Corrected orifice diameter data you are used to supplying for NX-19
d >
d r 1 D 1 T f Tr @ (specific gravity, percent CO2 and N2). The
detailed method requires a total analysis. What
constitutes a total analysis depends on each
Corrected pipe diameter measurement site. Generally, composition
D >
Dr 1 D 2 T f Tr @ through C6s is considered a total analysis.
Sometimes C7s or C8s or C9s might need to be
broken out. The detailed method of the equation
Change 5, Downstream expansion factor, requires will support this if needed.
additional compressibility
2. The gross method is applicable over a narrower
The equations for upstream expansion factor have range of operating conditions than the detailed
not changed. However to compute the downstream method. The gross method was designed to be
expansion factor, real gas effects must now be applicable for pipeline quality natural gas at
accounted for. This means an additional Z, normal pipeline pressures and temperatures. For
compressibility calculation is required when example, the gross method supports up to 0.02%
computing the downstream expansion factor. Hydrogen Sulfide, while the detailed method
supports up to 100% Hydrogen Sulfide.
If your system measures static pressure downstream,
but you do not want to incur the additional The following table summarizes the range of
processing to compute another Z for the expansion applicability for the two methods. The Normal
factor there is something you can do. Range column applies to the gross method. The
Expanded Range column applies to the detailed
You can compute the upstream pressure as follows method.
and use it to compute the upstream expansion factor.
'P
Pf 1 Pf 2
N
where N is a conversion constant from differential
pressure to static pressure units.
If you employ this technique, you must be careful to
use Pf1 for all occurrences of static pressure in the
flow rate equation. You cannot use upstream
pressure in some places and downstream pressure in
others.
Page 7
Table 2-2 - AGA-8 Ranges of Applicability
Page 8
x Meter tube diameter tolerances were
Regarding this issue, a summary of statements taken changed
from Part 4 of the new standard follows:
For Any Diameter
The orifice equation in use through AGA-3-
1985 was based on data collected in 1932/33 old range of 0.1 to 0.75 %
under the direction of Professor S.R. Beitler at depending on Beta
Ohio State University (OSU). The results of
these experiments were used by Dr. Edgar new 0.25% regardless of Beta
Buckingham and Mr. Howard Bean to develop
the coefficient of discharge equation. For Max-Min Diameter
In the 1970's, researchers reevaluated the OSU old range of 0.1 to 0.75 % depending on
data and found a number of reasons to question Beta
some of the data points. This analysis identified
303 technically defensible data points from the new 0.5% regardless of Beta
OSU experiments. Unfortunately it is not
known which points were used by x Eccentricity requirement was changed
Buckingham/Bean to generate the discharge
coefficient equation. old H d 0. 03 Dm
Statistical analysis of the Regression Data Set 0. 0025 Dm
(the new data set) showed that in several new H d
regions, the Buckingham/Bean equations did 0.1 2. 3 E m4
not accurately represent that data.4
x Perpendicularity requirement added
This means that the uncertainty statement in the AGA-3- New statement that orifice plate plane must be
1985 standard cannot be substantiated in all cases. kept at an angle of 90 degrees to the meter tube
axis.
Changes 13-18, Changes to Part 2 - Specification and
Installation Guidelines This concludes the overview of changes in the new
Since this paper mostly deals with the equation, details orifice metering standard.
about changes to the installation requirements are only
mentioned in brief here.
Page 7
Section 3 - MORE ON FACTORS
In this section a table is developed to more clearly show the relationships between the fundamental and factored
equation forms (both AGA-3-1985 and AGA-3-1992). A complete derivation of factors will not be shown here.
Both the 1985 and 1992 standards already document those derivations. To be more instructive, the density terms of
equations 1 and 2 are shown being calculated using density equations 3 and 4. Additionally, within the density
equations, Gi is computed based on equation 5. These equation numbers refer to equations in this document.
1985-AGA-3 fundamental equation shown with AGA-3-1992 fundamental equation shown with
density equations included density equations included
Sustituting density equations (eq. 3 and eq. 4) and Substituting density equations (eq. 3 and eq. 4) and
ideal gas gravity equation (eq. 5) into the AGA-3- ideal gas gravity equation (eq. 5) into the AGA-3-
1985 fundamental equation (eq. 1) results in equation 1992 fundamental equation (eq. 7) results in equation
A-1 as shown below. A-2 as shown below.
Cd S S
Y d2 2 g c U f 'P N c Cd Ev Y d 2 2 U f 'P
4 Qv = 4
4
1- E
Qv =
Ub
Ub
Pb Mrair Gi
Pb Mrair Gi Ub
Ub Zb R ( Tb N5 )
Zb R (Tb N5 )
§ § ª Z º· ·
§ § ªZ º· · ¨ Pf Mr ¨Gr « bgas »¸ ¸
¨ Pf Mr ¨Gr « bgas »¸ ¸ ¨ air ¨ ¸
¨ air ¨ ¸ S © «¬ Zbair »¼¹ ¸
Cd S © «¬ Zbair »¼¹ ¸ Nc Cd Ev Y d 2 2gc ¨ ¸ 'P
Y d2 2gc ¨ ¸ 'P 4
¨ Zfgas R Tf N5 ¸
1E4 4
¨ Zfgas R Tf N5 ¸ ¨ ¸
¨ ¸ © ¹
© ¹ Qv
Qv § § ª Z º· ·
§ § ªZ º· · ¨ Pb Mr ¨Gr « bgas »¸ ¸
¨ Pb Mr ¨Gr « bgas »¸ ¸ ¨ air ¨ ¸
¨ air ¨ ¸
© «¬Zbair »¼¹ ¸ © «¬ Zbair »¼¹ ¸
¨ ¸
¨ ¸ ¨ Zbgas R Tb N5 ¸
¨ Zbgas R Tb N5 ¸ ¨ ¸
¨ ¸ © ¹
© ¹
Page 8
Table 3-1 Fundamental Equation Terms included in Factored Equation Terms
Y Y Y Y
Zbgas Zbgas
Z fgas Fpv Z fgas Fpv
Gr Gr
Gr Fgr Gr Fgr
1 1
( Tf N 5 ) ( Tf N 5 )
Ftf Ftf
Pb Fpb Pb Fpb
Tb Ftb Tb Ftb
Pf ' P Pf ' P
Extension Extension
Not Applicable Not Applicable Cd Fc Fsl
Not Applicable Not Applicable Ev Ev
S § M rair ·
Nc 2¨ ¸
4 © R ¹ Fn
§ § ª 1 º ··
Not Applicable Not Applicable ¨ Mr ¨ « » ¸¸ ¸
¨ ¨
«¬ Zbair »¼ ¹ ¸
air
©
¨ ¸
¨ R ¸
¨ ¸
© ¹
Page 9
Section 4 - PART 4's NEW EQUATION Calculate velocity of approach term
PROCEDURES
1
To date, most publications in the public domain, have Ev
included summaries of the new equations as 1 E4
presented in Parts 1 and 3 of the new standard. Since
Part 4 of the new standard is intended as a guide for Note: In the following equations A0 through A6 and
those taking on the task of implementing the S1 through S8 are references to constants that are
equation, it seems appropriate to include a summary documented in the standard.
of it here. This does not provide all the necessary
information to completely implement the new Calculate orifice coefficient of discharge constants
equation, but it should give you a feel for the scope
of work involved. N4
L1 L2
D
General Outline of Solution Procedures 2 L2
For this example it is assumed the fluid being M2
measured is natural gas. The general outline of the 1E
solution procedures for flange-tapped orifice meters E4
is as follows: Tu S2 S3e 8. 5 L1 S4 e 6 . 0 L1
1 E4
1. At Tf, calculate terms that depend only upon TD S6 M2 S7 M21. 3 E1.1
orifice geometry: d, D, E , Ev and Cd correlation
terms. If D ! A4 N 4
2. Calculate the upstream flowing pressure, Pf from Then Ts 0.0
either Pf2 or Pf1 and Dp.
§ D·
Else Ts A3 1 E ¨ A4 ¸
3. Calculate required fluid properties (Gi, Rhof, and © N4 ¹
Rhob) at Tf, Pf and other specified fluid
conditions. Additional Tap Term for small diameter pipe
4. Calculate the upstream expansion factor.
Cd 0 A0 A1E 2 A2 E 8 TU TD TS
5. Determine the converged value of Cd. 0.7
Cd1 A5 E 0.7 250
6. Calculate the final value of Qb. 0.35
Cd 2 A6 E 4 250
Detailed Outline of Solution Procedures 0.8
Cd 3 S1E 4 E 0.8 4.75 250 0.35
1. At Tf, calculate terms that depend only upon
orifice geometry: d, D, b, Ev and orifice
Cd 4 S 5TU S 8TD E 0.8 4.75 0.8
coefficient correlation terms.
D >
Dr 1 D 2 T f Tr @ 3. Calculate required fluid properties (Gi, Rhof,
and Rhob) at Tf, Pf and other specified fluid
Calculate Beta conditions.
'P Fc 0
Cd Cd X035
.
Cd Cd X08. X035
1
.
2
Cd X08. 3
4
x
N3 Pf Dc 07. C X
d1
035
.
035
. Cd 115 . Cd X08.
. Cd X X 08
2 3
08
.
035
.
4
Else
Compute expansion factor pressure constant Yp § B·
4
Fc 0 1
Cd Cd X07. Cd Cd X08. ¨A ¸ Cd X08.
© X¹ 2 3
4
0. 41 0. 35 E
Yp B § B·
k Dc 07 1
. Cd X07. Cd Cd X08.
2 3
X
. Cd ¨A ¸ X08. 08
08
© X¹
. Cd X08.
3 4
S
If FI 1000 FI N c Cd Ev Y d 2 2 U f 'P
Qv = 4
c p
Then FI
FI c Ub
FI p
Else FI 1000
Page 11
Section 5 - PART 3's NEW EQUATION Calculate corrected orifice diameter
FACTORS
For this example it is assumed the fluid being 2. Calculate the upstream flowing pressure, Pf
measured is natural gas and that the inch-pound units from either Pf2 or Pf1 and Dp
of measure are used. The general outline of the
solution procedures for flange-tapped orifice meters 'P
is as follows: Pf Pf1 Pf2
N3
1. At Tf, calculate terms that depend only upon
orifice geometry: d, D, E , Ev and Fn. 3. Compute Factors associated with densities
(Rhob and Rhof)
2. Calculate the upstream flowing pressure, Pf from
either Pf2 or Pf1 and Dp.
14. 73
Fpb
Pb
3. Calculate factors associated with densities at Tf,
Pf and other specified fluid conditions. These
factors include Fpb, Ftb, Ftf, Fgr, and Fpv. Tb 459. 67
Ftb
4. Calculate the upstream expansion factor. 519. 67
Page 12
4. Calculate the upstream expansion factor.
0. 7
§ 1,000,000E ·
Compute orifice differential to flowing pressure ratio, Fsl 0.000511 ¨¨ ¸¸
x © ReD
¹
0 .8 0.35
ª § 19,000 E · º 4 § 1,000,000 E ·
'P «
0.0210 0.0049 ¨¨ ¸¸ » E ¨¨ ¸¸
x « © Re ¹ »¼ © Re ¹
N3 Pf ¬ D D
Y 1 Yp x
Qv Fn Fc Fsl Y Fpb Ftb Ftf Fgr Fpv Pf1 hw
Qv Pb Gr
ReD 0. 0114541
P D Tb Zbair
Page 13
Section 6 - TOTALFLOW’s left to each system designer. Much of this section is
IMPLEMENTATION OF NEW ORIFICE devoted to describing techniques for integrating the
EQUATION FOR GAS fundamental flow rate equation to produce volume.
This section describes Totalflow’s realtime As illustrated below, portions of the equation are
implementation of the new orifice metering computed at different times. The possible times are:
equations. As previously shown in Section 2,
equation 7 of this document, the fundamental
equation for volumetric flow rate is stated as follows. Table 6-1 Names of Calculation Time Periods
Page 14
Substituting equation 3's density solution into periods. To illustrate those portions of Fip, the
equation 9, results in the following equation for Fip. following equations are provided.
2 Mrair
ª Pf M r Gi º Fipconst
Fip 2« air
» 'P R
«¬ Z f R Tf N 5 »¼ eq. 12 Constants within Fip equation
ª Zb º
Gi Gr « gas » Pf 'P
Ext pt
«¬ Z bair »¼ ( Tf N5 )
eq. 5 (restated) Gi computed from Gr eq. 14 Extension within Fip equation
Substituting equation 5's ideal gravity solution into Restating the Fip equation in terms of the variables
equation 10, results in the following equation for Fip. solved for in equations 12, 13 and 14 results in an
Fip equation of the following nomenclature.
ª ª Z bgas ºº
« Pf M rair Gr « »» Fipconst Gr
« «¬ Zbair »¼ » Fip Ext pt Fpv
Fip 2« » 'P Zbair
« Z f R Tf N 5 »
« » eq. 15 Fip with time dependent factors shown
¬ ¼
With this final representation of Fip, we can now
eq. 11 Fip with Gr used instead of Gi construct a table showing each portion of the
flowrate equation (equation 8) and their respective
Equation 11 shows the form of Fip used in this computation time periods. See Table 6-2 on
implementation to compute gas volumes. However, following page.
portions of Fip are computed on different time
Page 15
Table 6-2 Summary of Calculation Time Periods
Page 16
Static Pressure and Expansion Factor
As mentioned in Section 2 of this document, if downstream expansion factor is used then an additional Z
(compressibility) calculation must be performed. To avert the need for this additional processing, this
implementation always uses the upstream static pressure thereby allowing computation of the upstream expansion
factor.
The user is allowed to specify either up or down stream for location of the static pressure sensing element. If the
upstream location is specified, that pressure measurement is used without modification. However, if the
downstream location is specified then the upstream pressure is computed as:
'P
Pf 1 Pf 2
N
This logic and math execute each second thereby always providing the upstream static pressure for use throughout
the whole equation.
Averaging Techniques
Type 1 Averages
Averages constructed from one second samples taken only during times of flow are maintained for the real time
measured variables of differential pressure, static pressure, and flowing temperature.
Type 2 Averages
Averages constructed from all one second samples (regardless of flow) are also maintained for the same variables.
Type 1 averages are stored in the historical record for periods in which some quantity (volume or mass) accrued.
Type 2 averages are stored for periods in which zero quantity accrued. This technique provides adequate volume
adjustment averages for downstream processing but also supports site operations with averages for pressure and
temperature even when there is no flowrate.
In older Totalflow devices Type 1and 2 averages were always based on linear values. In newer Totalflow devices
either linear or square root averages can be specified.
x Different Z (compressibility) calculation methods are available. These include the latest AGA-8 methods and
NX-19. Additionally Fpv can be turned off if desired.
x VOLP, Volume calculation period defaults to one hour, but is user selectable. Selections offered are 1, 2, 5, 10,
30, and 60 minutes.
x Higher static pressure transducers are supported. Up to 3500 psi is currently in use.
CONST PERIOD
2 Mrair
Fipconst
R
IF (FpvMethod = AGA-8gross)
Compute AGA-8 gross method precalcs (e.g. AGA-8 terms that are function of composition)
Using AGA-8 gross method Compute Zbgas
ENDIF
Pb Mrair Gr
Ub
Zb air R ( Tb N5 )
SEC PERIOD
IF (Pressure Tap Downstream) Calculated Upstream Static Pressure, Pf
Dpsec
Pfsec Pf 1
N5
ELSE
Pfsec Pf 2
ENDIF
Secs Secs 1
Pfacc Pfacc Pfsec
Tfacc Tfacc Tfsec
IF (DP > DP_ZERO_CUTOFF) (If Flow Exists)
Pfsec Dpsec
Ext pt
(Tfsec N5 )
Extacc flow Extacc flow Ext pt
Dpacc flow Dpacc flow Dpsec
Pfacc flow Pfacc flow Pfsec
Tfacc flow Tfacc flow Tfsec
Secs flow Secs flow 1
ENDIF
Extacc flow
Extvolp
Secs
Dpacc flow
Dpvolp
Secs flow
Pfacc flow
Pfvolp
Secs flow
Tfacc flow
Tfvolp
Secs flow
At Tf, calculate terms that depend only upon orifice geometry: d, D, b, Ev and orifice coefficient correlation
terms.
d >
d r 1 D 1 T f Tr @
D D >1 D T
r 2 f T @
r
d
E
D
Calculate velocity of approach term
1
Ev
1 E4
If D ! A4 N 4 Then Ts 0.0
§ D·
Else Ts A3 1 E ¨ A4 ¸
© N4 ¹
IF (FpvMethod = OFF)
Fpv = 1.0
Zbgas
Fpv
Z fgas
Zbgas
Fpv
Z fgas
Dpvolp
x
N3 Pfvolp
Compute expansion factor pressure constant Yp
0. 41 0. 35 E 4
Yp
k
Compute expansion factor
Y 1 Yp x
Calculate Fip
Fipconst Gr
Fip Extvolp Fpv
Zbair
4000 Nic D P
Fic
Ev Y d 2
Cd Cd0
FI
X
Cd
Cd_step.3 Compute the correlation value Fc and it's derivative Dc, of Cd at the assumed flow, X
If ( X X c )
Fc 0
Cd Cd X 0.35 Cd Cd X 0.8 X 0.35 Cd X 0.8
1 2 3
4
Else
§ B·
Fc Cd Cd X 0.7
0 1
C d2 3
© X¹
Cd X 0.8 ¨ A ¸ Cd X 0.8 4
B § B·
Dc 0.7 Cd X 0.7 Cd Cd X 0.8
1
2 3
X
0.8Cd ¨ A ¸ X 0.8
© X¹
3
0.8
0.8Cd X 4
Cd_step.4 Calculate the amount of change to guess for Cd
Cd Fc
GCd
D
1 c
Cd
Cd Cd GCd
Cd_step.6 Repeat steps 2,3,4 and 5 until the absolute value of G C d is less than 0.000005.
Calculate the final value of qm, the mass flow rate at line conditions.
S
qm = N c Cd Ev Y d 2 Fip
4
Calculate the final value of Qv, the volumetric flow rate at base conditions.
qm
Qv
Ub
Calculate the final value of Volb, the volume at base conditions for the Volume Period
Secs
Vol b Qv *
N vtime
Section 7 - NOMENCLATURE
1. American Petroleum Institute Measurement on Petroleum Measurement Standards (API MPMS) Chapter 14.3,
Part 1; Also recognized as AGA Report No. 3 Part 1; Also recognized as GPA 8185-92, Part 3; Also
recognized as ANSI/API 2530-1991, Part 1
2. American Petroleum Institute Measurement on Petroleum Measurement Standards (API MPMS) Chapter 14.3,
Part 2; Also recognized as AGA Report No. 3 Part 2; Also recognized as GPA 8185-92, Part 2; Also
recognized as ANSI/API 2530-1991, Part 2
3. American Petroleum Institute Measurement on Petroleum Measurement Standards (API MPMS) Chapter 14.3,
Part 3; Also recognized as AGA Report No. 3 Part 3; Also recognized as GPA 8185-92, Part 3; Also
recognized as ANSI/API 2530-1991, Part 3
4. American Petroleum Institute Measurement on Petroleum Measurement Standards (API MPMS) Chapter 14.3,
Part 4; Also recognized as AGA Report No. 3 Part 4; Also recognized as GPA 8185-92, Part 4; Also
recognized as ANSI/API 2530-1991, Part 4
5. American Gas Association (AGA) Transmission Measurement Committee Report No. 8; Also recognized as
API MPMS Chapter 14.2.
6. Teyssandier, Raymond G.; Beaty, Ronald: New orifice meter standards improve gas calculations, Oil & Gas
Journal, Jan. 11, 1993
7. ANSI/API 2530: Second Edition, 1985, Orifice Metering Of Natural Gas and Other Related Hydrocarbon
Fluids; Also recognized as AGA Report No. 3; Also recognized as GPA 8185-85; Also recognized as API
MPMS Chapter 14.3, API 2530.