Fiscal Metering Handbook
Fiscal Metering Handbook
Fiscal Metering Handbook
February 2018
BERGEN - 18.01.2018
Ref. No.: CMR-17-A211576-RA-1
Rev.: 01
REPORT
Handbook of uncertainty
calculations for ultrasonic,
turbine and Coriolis oil flow
metering stations
Documentation of uncertainty
models and internet tool
Authors
Kjell-Eivind Frøysa, Per Lunde, Gaute Øverås Lied, Camilla Sætre, Ranveig N.
Bjørk
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 3 of 79
Table of Contents
Disclaimer ............................................................................................................................................... 6
1 Introduction ................................................................................................................................ 7
2 Description of metering stations .............................................................................................. 9
3 Oil measurement uncertainties .............................................................................................. 11
3.1 Temperature measurement ....................................................................................................... 11
3.2 Pressure measurement .............................................................................................................. 12
3.3 Density measurement ................................................................................................................ 13
4 Functional relationships.......................................................................................................... 15
4.1 Overall functional relationships .................................................................................................. 15
4.1.1 Volumetric flow rate at standard conditions ............................................................... 15
4.1.2 Volumetric flow rate at line conditions ....................................................................... 17
4.1.3 Mass flow rate – from measured volume flow rate .................................................... 18
4.1.4 Mass flow rate – when mass flow rate is primary measurement............................... 19
4.1.5 Volumetric flow rate – from measured mass flow rate .............................................. 20
4.2 Oil volume expansion coefficients .............................................................................................. 21
4.2.1 Overall expression ..................................................................................................... 21
4.2.2 Temperature volume expansion coefficient ............................................................... 21
4.2.3 Pressure volume expansion coefficient ..................................................................... 22
4.3 Steel volume expansion coefficients .......................................................................................... 22
4.3.1 Temperature volume expansion coefficient ............................................................... 23
4.3.2 Pressure volume expansion coefficients ................................................................... 23
5 Uncertainty models .................................................................................................................. 26
5.1 Volumetric flow rate at standard conditions ............................................................................... 26
5.1.1 Uncertainty model when volume flow rate is the primary measurement ................... 26
5.1.2 Uncertainty model standard volume flow rate estimated from measured mass flow 26
5.2 Volumetric flow rate at line conditions ........................................................................................ 27
5.2.1 Uncertainty model when volume flow rate is the primary measurement ................... 27
5.2.2 Uncertainty model when volume flow rate is estimated from measured mass flow .. 28
5.3 Mass flow rate ............................................................................................................................ 28
5.3.1 Uncertainty model when mass flow rate is estimated from measured volume flow .. 28
5.3.2 Uncertainty model when mass flow rate is the primary measurement ...................... 29
5.4 Calibration uncertainties............................................................................................................. 29
5.5 Proving uncertainties.................................................................................................................. 30
5.6 Metering uncertainties ................................................................................................................ 31
5.7 Oil and steel expansion factor uncertainties .............................................................................. 33
5.7.1 Volumetric flow rate at standard conditions, from measured volume flow ................ 33
5.7.2 Measured volumetric flow rate at line conditions ....................................................... 35
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 4 of 79
5.7.3 Mass flow rate – from measured volume flow rate .................................................... 36
5.7.4 Volume flow rate at standard conditions, from measured mass flow rate ................. 38
5.7.5 Volume flow rate at flow meter conditions, from measured mass flow rate .............. 39
5.8 Expansion factor model uncertainties ........................................................................................ 41
5.8.1 Oil expansion factor - introduction ............................................................................. 41
∆m ,∆ p
5.8.2 Oil expansion factor Aliq .................................................................................... 42
m ,∆p
5.8.3 Oil expansion factor Aliq ....................................................................................... 43
∆p
5.8.4 Oil expansion factor Aliq .......................................................................................... 44
6 Program .................................................................................................................................... 49
6.1 Software platform ....................................................................................................................... 49
6.2 Installation and use .................................................................................................................... 49
6.3 Program overview ...................................................................................................................... 49
6.3.1 Specify metering station template ............................................................................. 50
6.3.2 Oil Properties Page ................................................................................................... 52
6.3.3 Equipment Page ........................................................................................................ 53
6.3.4 Calibration Page ........................................................................................................ 55
6.3.5 Proving Page ............................................................................................................. 57
6.3.5.1 Proving with Ultrasonic or Turbine Master Meter ...................................................... 57
6.3.5.2 Proving with displacement prover.............................................................................. 59
6.3.6 Metering Page ........................................................................................................... 60
6.3.7 Results Page ............................................................................................................. 63
6.3.8 Uncertainty Budget Charts Page ............................................................................... 66
6.3.9 Uncertainty Range Plots Page .................................................................................. 67
6.3.10 Uncertainty Report Page ........................................................................................... 67
6.3.11 Note about “Save” and “Open” functionality .............................................................. 69
6.3.12 Note about “opening” an uncertainty analysis file ..................................................... 69
7 Summary ................................................................................................................................... 70
8 References ................................................................................................................................ 71
Appendix A Detailed formulas for the linearity contribution to the proving uncertainty ......... 73
A1 Functional relationship ............................................................................................................... 73
A2 Uncertainty model ...................................................................................................................... 74
Appendix B List of symbols ............................................................................................................ 77
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 5 of 79
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 6 of 79
Disclaimer
CMR is not liable in any form or manner for the actual use of the documents, software or other results
made available for or resulting from a project and does not warrant or assume any liability or
responsibility for the completeness or usefulness of any information unless specifically agreed otherwise
in the tender and resulting contract document.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 7 of 79
1 Introduction
Documentation of uncertainty of flow rates measured by fiscal flow metering stations is essential as part
of the evaluation of the condition of such metering stations. Authorities have requirements with respect
to maximum uncertainty in order to secure the national interests. The partners selling the oil have
interests in the uncertainty to secure their incomes. Finally, buyers of oil have interest in ensuring that
they are not getting a lower amount of oil than what they pay for.
In order to get all parties to accept an uncertainty analysis, it is important to obtain standardized ways
of carrying out such analyses. The ISO Guide to the expression of uncertainty in measurement, [ISO
GUM] provides general methodology for carrying out uncertainty analyses. This methodology can also
be applied in uncertainty analysis of fiscal oil metering stations. However, the ISO GUM does not give
detailed methods for the specific uncertainty analyses for such metering stations (or other applications).
Therefore models have to be developed based on the ISO GUM methodology. Similarly, the ISO 5168
Measurement of fluid flow -- Procedures for the evaluation of uncertainties [ISO 5168], provides general
procedures for evaluation of uncertainty for the measurement of fluid flow. Also, the procedures in this
standard have to be developed further in order to approach the uncertainty evaluation of a specific
metering station.
The Norwegian Society for Oil and Gas Measurement (NFOGM) in cooperation with the Norwegian
Petroleum Directorate (NPD) and The Norwegian Society of Graduate Technical and Scientific
Professionals (Tekna) have previously issued “Handbook for uncertainty calculations for gas metering
stations” [Frøysa et al, 2014], a handbook in agreement with the ISO GUM methodology. Calculation of
uncertainties according to this handbook can be done free of charge using a Microsoft Silverlight based
calculation program available at www.nfogm.no.
Prior to the handbook [Frøysa et al, 2014], some more technology-specific uncertainty handbooks for a
fiscal ultrasonic gas metering station [Lunde et al, 2002] and for a fiscal orifice gas metering station and
a turbine oil metering station [Dahl et al, 2003] have been issued. These handbooks are also in
agreement with the ISO GUM methodology, and were based on a previous version of the ISO GUM
from 1995. Calculation of uncertainties in these handbooks are based upon an Excel spread sheet
available for download from www.nfogm.no free of charge. In addition, uncertainty models for fiscal
turbine oil metering stations [Dahl et al, 2003] and fiscal ultrasonic oil metering stations [Lunde et al,
2010] have been established.
The present work is a similar Handbook as [Frøysa et al, 2014], but valid for fiscal oil metering stations
with ultrasonic, turbine or Coriolis flow meters used as duty meters. The intention of this work is to
establish an uncertainty analysis model covering common fiscal oil metering station configurations
applied on the Norwegian Sector. The intention is also to make a tool in which a complete uncertainty
analysis for an oil metering station can be performed within one tool in a minimum of time. This is
achieved as the tool calculates all necessary parameters from a minimum of inputs, based upon
reasonable default values and default input values for uncertainty in accordance with requirements in
the Norwegian measurement regulations and NORSOK. Furthermore, a main focus is to make it easy
to define the most common metering station configurations in the tool.
The uncertainty model is flexible, allowing (i) ultrasonic master meter prover (ii) turbine master meter
prover (iii) volume displacement prover or (iv) Coriolis master meter prover. The uncertainty model is
implemented on a web-based Microsoft Silverlight technology. This can be accessed for free from
www.nfogm.no.
This Handbook is a documentation of the uncertainty models developed and the web-based calculation
tool. It should be noted that the example input values in the calculation tool are pure example values,
and should not be regarded as recommended values by NFOGM, CMR, NPD or any other party.
Chapter 2 describes on an overview level the metering stations covered in the Handbook. In Chapter 3,
uncertainties related to secondary instrumentation temperature, pressure and density are covered.
Chapter 4 presents the functional relationships defining the metering stations, Chapter 5 documents the
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 8 of 79
uncertainty models for the metering stations and Chapter 6 documents the web-based uncertainty
calculation program. Chapter 7 includes a brief summary of the Handbook.
Appendix A contains some details with respect to the uncertainty model related to adjustments of a flow
meter after flow calibration. Appendix B contains a list of symbols.
The uncertainty models presented in this Handbook are based upon the ISO GUM uncertainty
methodology. The measurement regulations by the Norwegian Petroleum Directorate and the NORSOK
standard I-106 on fiscal measurement systems for hydrocarbon liquid and gas [NORSOK I-106] have
been important references with respect to layout of the metering stations and requirements to the
uncertainty of individual instruments and the operation of the metering station as a whole. A series of
ISO, API MPMS and other international standards and reports have also been essential in this work.
The details are covered in the relevant sections of the Handbook. It is also referred to the reference list
in Chapter 8.
The present work has been carried out for Norwegian Society for Oil and Gas Measurement (NFOGM)
with financial support also from Norwegian Petroleum Directorate and Tekna. A reference group
consisting of the following members has reviewed the work:
Dag Flølo has been especially involved with regular project meetings and discussions throughout the
project.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 9 of 79
The different primary flow meters can be proved according to the following setup:
7. Configuration 7: Coriolis (mass flow) prover [API MPMS 4.5 and 5.6]
Note that the USM and turbine flow meters are proved with volume flow meters, while the Coriolis flow
meter is proved with mass flow meters (Coriolis).
Figure 2-1: Overview of the seven different configurations of primary flow meter versus proving device
It is assumed that the proving of the primary flow meter is carried out at a single flow rate. The calibration
of the proving device will in case of master meter provers be carried out at a series of flow rates. In case
of a displacement prover, the prover is calibrated at a single flow rate only.
The metering station is also equipped either with densitometer giving the density at the densitometer
pressure and temperature conditions, or provided e.g. with sampling and laboratory analysis, which
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 10 of 79
gives the standard density with a given uncertainty. If the density is from a Coriolis meter, the uncertainty
tool described in this handbook, will consider this as a densitometer. The uncertainty input should in that
case be given at an overall level. Note that the Coriolis meter providing the density measurement, is not
the same Coriolis meter providing the mass flow rate.
Water in oil is not covered. Therefore, uncertainty related to sampling and analysis is not explicitly
covered.
The densitometer, the flow meter and the proving device are all equipped with pressure and temperature
measurements.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 11 of 79
• Overall level.
• Detailed level.
In case of the overall level, the absolute uncertainty in the measured temperature is specified directly
by the user of the uncertainty calculation program.
where
u (T elem , transm ): standard uncertainty of the temperature element and temperature transmitter, calibrated
as a unit. Typically found either in product specifications or in calibration certificates.
u (T stab , transm ): standard uncertainty related to the stability of the temperature transmitter, with respect
to drift in readings over time. Typically found in product specifications.
u(TRFI) : standard uncertainty due to radio-frequency interference (RFI) effects on the
temperature transmitter.
u (T temp ): standard uncertainty of the effect of temperature on the temperature transmitter, for
change of oil temperature relative to the temperature at calibration. Typically found in
product specifications.
u (T stab , elem ): standard uncertainty related to the stability of the temperature element. Instability may
relate e.g. to drift during operation, as well as instability and hysteresis effects due to
oxidation and moisture inside the encapsulation, and mechanical stress during
operation. Typically found in product specifications.
u(Tmisc) : standard uncertainty of other (miscellaneous) effects on the temperature transmitter.
This uncertainty model is quite generic, and can be used on a series of industrial products. In cases
where this model does not fit with the product specifications, the miscellaneous uncertainty contributions
can be used for specification of other uncertainty contributions. Alternatively, the uncertainty of the
temperature measurements can be calculated manually, and the result can be given to the program
using the overall input level.
When the average of two temperature measurements is used, it is assumed that the two temperature
measurements are uncorrelated. The reason for this assumption is that often the two probes are not
calibrated at the same time. This means that even if they are calibrated using the same procedure, the
time difference generates an uncorrelated drifting term, both in the reference and in the temperature
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 12 of 79
measurement itself. This means that the uncertainty in the average of two temperature measurements
is assumed to be equal to the uncertainty for one measurement, divided by the square root of two.
• Overall level.
• Detailed level.
In case of the overall level, the relative uncertainty in the measured pressure is specified directly by the
user of the uncertainty calculation program.
where
This uncertainty model is quite generic, and can be used on a series of industrial products. In cases
where this model does not fit with the product specifications, the miscellaneous uncertainty contributions
can be used for specification of other uncertainty contributions. Alternatively, the uncertainty of the
pressure measurements can be calculated manually, and the result can be given to the program using
the overall input level.
When the average of two pressure measurements is used, it is assumed that the two pressure
measurements are uncorrelated. The reason for this assumption is that often the two probes are not
calibrated at the same time. This means that even if they are calibrated using the same procedure, the
time difference generates an uncorrelated drifting term, both in the reference and in the pressure
measurement itself. This means that the uncertainty in the average of two pressure measurements is
assumed to be equal to the uncertainty for one measurement, divided by the square root of two.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 13 of 79
• Overall level.
• Detailed level.
In case of the overall level, the relative uncertainty in the measured density is specified directly by the
user of the uncertainty calculation program.
In case of the detailed level, the uncertainty model is more complicated than for the temperature and
pressure measurements above. The density measurement consists of several steps:
This will in total form the functional relationship for the density measurement as follows:
ρ ={ρu [1 + K18 (Td − Tc )] + K19 (Td − Tc )}⋅ (1 + [K20A + K20B ⋅ (Pd − Pc )]⋅ (Pd − Pc ))
+ [K 21A + K21B ⋅ (Pd − Pc )]⋅ (Pd − Pc ) .
(3-3)
In this equation subscript “d” means densitometer conditions and subscript “c” means calibration
conditions. The following variables are used in this equation:
By using the general uncertainty model approach in ISO GUM [ISO GUM, 2008], the uncertainty model
will be
u c2 ( ρ ) = s ρ2 u u 2 ( ρ u ) + u 2 ( ρ stab ) + u 2 ( ρ rept ) + s ρ2 ,Td u 2 (T d ) + s ρ2 , Pd u 2 ( Pd )
(3-4)
+ u 2 ( ρ temp ) + u 2 ( ρ pres ) + u 2 ( ρ misc ),
where
ρu .
u ( ρ rept ) : standard uncertainty of the repeatability of the indicated (uncorrected) density, ρu .
u(Td ) : standard uncertainty of the oil temperature in the densitometer, Td .
u(Pd ) : standard uncertainty of the oil pressure in the densitometer, Td .
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 14 of 79
u ( ρ temp ) : standard uncertainty of the temperature correction factor for the density, ρ representing
the model uncertainty of the temperature correction model used,
{ρu [1+ K18(Td −Tc )] + K19(Td −Tc )} and the pressure correction model used,
ρ ⋅ (1 + [K 20 A + K 20 B ⋅ (Pd − Pc )]⋅ (Pd − Pc )) + [K 21 A + K 21 B ⋅ (Pd − Pc )]⋅ (Pd − Pc ) . This
includes also the uncertainty of the various K-coefficients, and the measurement of the
pressure and temperature during calibration.
u(ρmisc) : standard uncertainty of the density, accounting for miscellaneous uncertainty contributions,
such as due to:
• reading error during measurement (for digital display instruments),
• possible deposits on the vibrating element,
• possible corrosion of the vibrating element,
• mechanical (structural) vibrations on the oil line,
• variations in power supply,
• self-induced heat,
• flow in the bypass density line,
• possible liquid viscosity effects,
• effect of a by-pass installation of the densitometer,
• other possible effects.
The sensitivity coefficients in Eq. (3-4) can be calculated from the functional relationship Eq. (3-3) by
use of the ISO GUM methodology:
sρu = {[1+ K18(Td −Tc )] + K19(Td −Tc )}(1+[K20A + K20B ⋅ (Pd − Pc )]⋅ (Pd − Pc )) , (3-5)
sρ,Td = {ρu K18 + K19}(1+ [K20A + K20B ⋅ (Pd − Pc )]⋅ (Pd − Pc )) , (3-6)
These expressions are in practice obtained by partially derivation of the density (Eq. (3-3)) with respect
to ρu , Td and Pd , respectively.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 15 of 79
4 Functional relationships
In this Chapter, overall functional relationships for the volumetric flow rates at standard and line
conditions, and for mass flow rate are presented in Section 4.1. The functional relationships for the oil
expansion coefficients for the expansion of oil due to pressure and temperature are covered in Section
0 The functional relationships for the steel expansion coefficients for the expansion of steel due to
pressure and temperature are covered in Section 4.3.
• The standard volume of oil measured by the primary flow meter is compared to the standard
volume of oil measured by a proving device using a single flow rate. This is denoted “proving”.
• The standard volume of oil measured by the proving device is compared to a reference standard
volume of oil using a single flow rate if the proving device is a displacement prover, and by
multiple flow rates if the proving device is a master meter. This is denoted “calibration”.
• The reference standard volume is provided by an external party. The traceability of this device
is outside the scope of this Handbook.
V0calibratio n
V0 proving metering
V0 , meas = , ref
proving
, prover
V0 , flowmeter . (4-1)
V calibratio n V
0 , prover 0 , flowmeter
Here,
V0 , meas : the standard volume of oil measured by the primary flow meter, after corrections from
the proving and calibration.
calibration
V 0, ref : the standard volume of oil measured by the reference instrumentation during calibration
of the proving device.
calibration
V 0, prover : the standard volume of oil measured by the proving device during calibration of the
proving device.
proving
V 0, prover : the standard volume of oil measured by the proving device during proving of the primary
flow meter.
proving
V 0, flowmeter: the standard volume of oil measured by the primary flow meter during proving of the
primary flow meter.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 16 of 79
V0metering
, flowmeter: the standard volume of oil measured by the primary flow meter at metering, without the
corrections from the proving and calibration.
The second parenthesis is the correction from the proving (using a single flow rate) and the first
parenthesis is the correction from the calibration, for the flow rate used at proving.
The standard volume of oil through the primary flow meter and the proving device is typically found from
the actual volume of oil (at a measured pressure and temperature), through volume correction factors.
However, at calibration, standard volumes are compared. After a calibration, the calibration certificate
including uncertainty is usually given for this comparison of standard volumes. Therefore, the above
equation is modified as follows:
V calibratio n
C tlpprov C plp
prov proving
V prover met met metering
V0 , meas = 0calibratio
, ref
C tlm C plmV flowmeter . (4-2)
V n C prov C prov V proving
0 , prover tlm plm flowmeter
Volumes without subscript “0” are here actual volumes at the relevant pressure and temperature.
Furthermore Ctlx is the temperature volume expansion coefficient for oil and C plx is the pressure volume
expansion coefficient for oil, from actual temperature and pressure to standard temperature and
pressure. “x” is replaced by “m” when the actual oil temperature and pressure are the ones at the primary
flow meter. “x” is replaced by “p” when the actual oil temperature and pressure are the ones at the
proving device. The superscript “prov” or “met” indicates whether the actual temperature and pressure
during proving or during normal measurement shall be used.
The volumes defined by the flow meter and the proving device also have to be corrected for steel
expansion due to pressure and temperature, relative to a reference temperature and pressure, at which
a nominal volume is given:
V0calibration
V0, meas = cal cal calibration ×
, ref
C C V
tsp psp 0, nom, prover
(4-3)
CtlpprovC plp Ctsp C psp Vnom, prover
prov prov prov proving
prov prov prov × Ctlm
met met met met metering
C plmCtsm C psmVnom, flowmeter .
C C C C provV proving
tlm plm tsm psm nom, flowmeter
Here Ctsx is the temperature volume expansion coefficient for steel and C psx is the pressure volume
expansion coefficient for steel, from a base temperature and pressure to actual temperature and
pressure. “x” is replaced by “m” when the actual steel temperature and pressure are the ones at the
primary flow meter. “x” is replaced by “p” when the actual steel temperature and pressure are the ones
at the proving device. The superscript “cal”, “prov” or “met” indicates whether the actual temperature
and pressure during calibration, during proving or during normal measurement shall be used.
Furthermore,
V0calibratio
n
, nom, prover : the standard volume of oil that would have been measured by the proving device during
calibration of the proving device, if temperature and pressure expansions in steel not
had been taken into account.
proving
Vnom , prover: the actual volume of oil (line conditions) that would have been measured by the proving
device during proving of the primary flow meter, if temperature and pressure expansions
in steel not had been taken into account.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 17 of 79
proving
Vnom , flowmeter : the actual volume of oil (line conditions) that would have been measured by the
primary flow meter during proving of the primary flow meter, if temperature and pressure
expansions in steel not had been taken into account.
metering
Vnom , flowmeter : the actual volume of oil (line conditions) that would have been measured by the
primary flow meter volume at metering, without the corrections from the proving and
calibration, if temperature and pressure expansions in steel not had been taken into
account.
C prov C plp
prov met
C tlm met
C plm Ctspprov C psp
prov met
C tsm met
C psm metering proving
Vnom V0calibratio n
V0 , meas = tlp prov ⋅ cal cal prov Vnom , flowmeter proving, prover , ref
. (4-4)
Ctlm C plmprov C C C C prov Vnom , flowmeter V0calibratio n
tsp psp tsm psm , nom , prover
C prov C plp
prov met
C tlm met
C plm
m , ∆p
Aliq = tlp prov , (4-5)
prov
C tlm C plm
and
The three superscripts m, p and c in these expressions refer to metering, proving and calibration,
respectively. The ∆ in the superscripts is given when there are two temperature and pressure corrections
of relevance for the given process (metering, proving or calibration).
proving
m , ∆p m , ∆ p , c metering Vnom V0calibratio n
The volumetric flow rate at standard conditions, qv0, meas , can now be written as
Vnom
proving
V0calibratio n
Asteel qnom , flowmeter proving
m , ∆ p m , ∆p , c metering
qv 0 , meas = Aliq , prover
, ref , (4-8)
V V calibratio n
nom , flowmeter 0 , nom , prover
metering
where qnom , flowmeter is the volumetric flow rate of oil at line conditions that would have been measured
by the primary flow meter during metering, without the corrections from the proving and calibration, and
if temperature and pressure expansions in steel not had been taken into account.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 18 of 79
qv 0,meas
qv ,meas = met met . (4-9)
Ctlm C plm
By use of Eqs. (4-5), (4-7) and (4-8) above, the volumetric flow rate at line conditions, Eq. (4-9), can be
written as
Vnom
proving
V0calibratio n
qv , meas = Aliq Asteel q nom , flowmeter proving
∆ p m , ∆ p , c metering , prover
calibratio
, ref
, (4-10)
V V n
nom , flowmeter 0 , nom , prover
where
∆p C tlpprov C plp
prov
Vnom
proving
V0calibratio n
q m , meas = ρ 0 Aliqm , ∆p Asteel q nom , flowmeter proving
m , ∆ p , c metering , prover
calibratio
, ref
. (4-12)
V V n
nom , flowmeter 0 , nom , prover
When the density is found from laboratory analysis, this is the relevant functional relationship. When the
density is measured by a densitometer, oil volume correction factors must be applied to get the standard
density. In that case, Eq. (4-12) has to be elaborated on in the following manner:
Vnom
proving
V0calibratio n
qm , meas = ρ 0 Aliq Asteel qnom , flowmeter proving
m , ∆p m , ∆p , c metering , prover
calibratio
, ref
V V n
nom , flowmeter 0, nom , prover
ρ dens Vnom
proving
V0calibratio n
= A m , ∆p m , ∆p , c metering
A q proving, prover
calibratio
, ref
(4-13)
C Cmet met liq steel nom , flowmeter V V n
tld pld nom , flowmeter 0 , nom , prover
met met
Ctlm C plm Ctlpprov C plp
prov
Vnom
proving
V0calibratio n
= ρ dens A m , ∆p , c metering
q proving, prover
calibratio
, ref
.
met met
Ctld prov
C pld Ctlm prov
C plm
steel nom , flowmeter V V n
nom , flowmeter 0, nom , prover
The condition “d” in the oil volume correction factors means that the actual temperature and pressure
are the ones at the densitometer (i.e. densitometer conditions). ρdens is the density at the densitometer
conditions. This means that the functional relationship for the mass flow rate in the case where the
density is measured by a densitometer can be written as
Vnom
proving
V0calibratio n
q m , meas = ρ dens Aliq Asteel q nom , flowmeter proving
∆ m , ∆ p m , ∆ p , c metering , prover
, ref , (4-14)
V V calibratio n
nom , flowmeter 0 , nom , prover
where
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 19 of 79
met met
∆m , ∆p C tlm C plm C tlpprov C plp
prov
4.1.4 Mass flow rate – when mass flow rate is primary measurement
This section covers the analysis for mass flow meters (Coriolis flow meter). The mass of oil measured
by the primary flow meter is traceable through the following chain:
• The mass of oil measured by the primary flow meter is compared to the mass flow rate of oil
measured by a proving device using a single flow rate. This is denoted “proving”.
• The mass of oil measured by the proving device is compared to a reference mass flow rate of
oil over multiple flow rates. This is denoted “calibration”.
• The reference mass is provided by an external party. The traceability of this device is outside
the scope of this Handbook.
M ref
calibration
M prover
proving
metering
M meas = calibration proving M flowmeter . (4-16)
M
prover M flowmeter
Here,
M meas : the mass of oil measured by the primary flow meter, after corrections from the proving
and calibration.
calibration
M ref : the mass of oil measured by the reference instrumentation during calibration of the
proving device.
calibration
M prover : the mass of oil measured by the proving device during calibration of the proving device.
proving
M prover : the mass of oil measured by the proving device during proving of the primary flow meter.
proving
M flowmeter
: the mass of oil measured by the primary flow meter during proving of the primary flow
meter.
metering
M flowmeter: the mass of oil measured by the primary flow meter at metering, without the corrections
from the proving and calibration.
The second parenthesis is the correction from the proving (using a single flow rate) and the first
parenthesis is the correction from the calibration, for the flow rate used at proving.
M prover
proving
M ref
calibration
qm , meas = q metering
m , flowmeter proving calibration , (4-17)
M flowmeter M prover
where qmmetering
, flowmeter is the mass flow rate of oil that would have been measured by the primary flow meter
during metering, without the corrections from the proving and calibration.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 20 of 79
When the density is found from laboratory analysis, this is the relevant functional relationship. The
functional relationship for the volume flow rate at flow meter conditions, in the case where the density is
found from laboratory analysis can be written as:
=
,
, (4-19)
= =
, ,
, (4-20)
met
C tld met
C pld M prover
proving
M ref
calibration
qv 0, meas = q metering
proving calibration . (4-21)
ρ dens m , flowmeter
M flowmeter M prover
The condition “d” in the oil volume correction factors means that the actual temperature and pressure
are the ones at the densitometer (i.e. densitometer conditions). The superscript “met” indicates the
actual temperature and pressure during normal measurement is used in the calculation. ρdens is the
density at the densitometer conditions. This means that the functional relationship for the volume flow
rate at flow meter conditions, in the case where the density is measured by a densitometer can be written
as
qmmetering M prover
proving
M ref
calibration
= ∆m
, flowmeter
qv ,meas proving calibration , (4-22)
Aliq ρ dens M flowmeter M prover
where
met met
∆m
C tlm C plm
Aliq = met met
. (4-23)
C tld C pld
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 21 of 79
where Ctlx is the temperature expansion coefficient and C plx is the pressure expansion coefficient. Here
“x” denotes the condition (either “p” for prover, “m” for flow meter or “d” for densitometer).
where
K0 K1
α= + + K2 , (4-26)
ρ 2
0 ρ0
and
This formula corrects the volume from the temperature Tx to the standard reference temperature T0. ρ0
is the density at standard reference pressure and temperature. K0, K1 and K2 are coefficients that depend
on the type of oil.
The formula above includes two minor approximations compared to the API MPMS 11.1 standard.
The first approximation is that no calculation between the old ITS-68 and the new ITS-90 temperature
scales have been included. For the purpose of this uncertainty model, this is acceptable as the difference
between the two scales is minimal, much less than the typical uncertainty of a temperature
measurement.
The second approximation is that these formulas here are used with another standard reference
temperature than 60 °F. To be strict, a calculation of the volume change from the temperature Tx to the
standard reference temperature of 15 °C shall according to API MPMS 11.1 be carried out by first
calculating the volume change in the temperature change from Tx to 60 °F, with 60 °F as standard
reference temperature, and standard reference density at 60 °F. Thereafter the volume change in a
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 22 of 79
API MPMS 11.1 gives several sets of values for the coefficients K0 and K1, depending on the type of oil.
Among these are (i) crude oil, (ii) fuel oil, (iii) jet group and (iv) gasoline. Crude oil is there classified with
API gravity at 60 °F between 100 and -10, corresponding to reference density at 15 °C and 1 atm
between 611.16 and 1163.79 kg/m³.
In API MPMS 11.1.6.1, the coefficients K0 and K1 are given for the four liquid hydrocarbon types
mentioned above. These coefficients are multiplied by 1.8 to convert from Fahrenheit to Celsius
temperature scale. The data set then obtained is shown in Table 4.1.
Table 4.1 Temperature expansion coefficient for selected types of oil, for use with temperature on
Celsius scale.
Crude Oil Fuel Oil Jet Group Gasoline
K0 613.97226 186.9696 594.5418 346.42277
K1 0 0.48618 0 0.43883
K2 0 0 0 0
1
C plx =
1−100(Px − Pe )F
, (4-28)
where
• A = -1.6208,
• B = 0.00021592,
• C = 0.87096,
• D = 0.0042092.
In Eqs. (4-28) and (4-29) it is important to use the unit bar for the pressure and °C for the temperature.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 23 of 79
For displacement provers, API MPMS 12.2 gives the following expression for single-walled provers:
where Gc is the mean coefficient of cubic expansion per degree temperature of the material of which the
container is made, between the temperatures Tb and T.
For ultrasonic flow meters, ISO 12242 gives the following expression in Annex A:
where Eh and Er are the linear temperature expansion coefficients for the meter housing and the meter
rotor, respectively. The notation from the NORSOK-standard is used here. In newer versions of
NORSOK I-105 and the successor NORSOK I-106, this formula is not present. Furthermore, it has not
been possible for the authors to identify another general formula for the temperature expansion for a
turbine flow meter in international standards.
In this work, the following expression will be used for displacement provers, ultrasonic, and turbine flow
meters:
This is a valid approximation of Eqs. (4-30), (4-31) and (4-32) as long as the correction factor is not far
from 1. That means that extreme temperature differences are not taken into account. It also assumes
that in the case of a turbine meter, the rotor is of the same material (metal type) as the meter housing.
For displacement provers, API MPMS 12.2 gives the following expression for single-walled provers:
C psx = 1 +
(Px − Pb )ID ,
(4-34)
E ⋅ WT
where
ID: Inner diameter of pipe.
E: Young’s modulus of the pipe metal.
WT: Pipe wall thickness.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 24 of 79
For ultrasonic flow meters [Lunde et al, 2007] has demonstrated that the expansion depends on a series
of issues. These include
In ISO 12242, Appendix A, and in ISO 17089-1, Appendix E, this is addressed. A worst case expression
is given as follows:
R2 + r2 (P − Pb )
C psx = 1 + 4 2 + µ x , (4-35)
R −r
2
E
where
r: Inner diameter of pipe.
R: Outer diameter of pipe.
E: Young’s modulus of the pipe metal.
µ: Poisson’s ratio of the pipe metal.
The notation from the ISO-standards is used here. This expression will be used here, because the topic
is an uncertainty model. Indicative numbers of the size of the coefficient and on dependencies on the
pressure are therefore sufficient for the purpose in focus here.
For turbine flow meters, the NORSOK I-105, Rev 2 from 1998 contains the following expression:
C psx = 1 + (P − Pb )
(2 − e)2 R
,
(
E 1 − AT /(πR ) 2t
2
) (4-36)
where
R: Inner diameter of pipe.
t: Pipe wall thickness of pipe.
E: Young’s modulus of the pipe metal.
e: Poisson’s ratio of the pipe metal.
AT: The area in the pipe cross section that is occupied by the rotor blades. (In the program
to be given in percent of the total cross sectional area.)
The notation from the NORSOK-standard is used here. In newer versions of NORSOK I-105 and the
successor NORSOK I-106, this formula is not present. Furthermore, it has not been possible for the
authors to identify another general formula for the pressure expansion for a turbine flow meter in
international standards. Therefore, this formula will be used here.
This means that generally, the pressure expansion coefficient can by written as
C psx = 1 + β (P x − Pb ) . (4-37)
Displacement prover:
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 25 of 79
R
β= . (4-38)
E ⋅ dw
4 (R + d w ) + R 2
2
β = + µ . (4-39)
E (R + d w ) − R
2 2
β=
(2 − µ )2R
(
E 1 − AT /(πR2 ) 2d w ) . (4-40)
Here the notation of the steel expansion factors is uniformed between the different technologies, as
follows:
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 26 of 79
5 Uncertainty models
In this Chapter, the uncertainty models for volumetric flow rate at standard conditions, volumetric flow
rate at line conditions and mass flow rate are given in Sections 5.1, 5.2 and 5.3, respectively. These
uncertainty models are quite general, and the various components of them are detailed in the next
sections. In Sections 5.4, 5.5, and 5.6, respectively, the uncertainty contributions related to the
calibration process, the proving process and the flow metering are addressed. In Section 5.7, the
uncertainty contribution related to the oil and steel expansion coefficient is addressed. In Section 5.8
the model uncertainty of these oil and steel expansion coefficients is addressed more in detail.
5.1.1 Uncertainty model when volume flow rate is the primary measurement
The relative standard uncertainty of the volumetric flow rate at standard conditions can be deduced from
Eq. (4-8), and is here written as follows:
q prov q met
0 0
(5-1)
q ,∆p ,c .
v 0,meas Aliq Asteel qv 0 v0 v0
The interpretation of this equation is that the uncertainty consists of contributions related to expansion
of oil and steel (first term), contributions related to the calibration process (second term), contributions
related to the proving process (third term) and contributions related to the flow metering (last term).
The first term on the right hand side of Eq. (5-1), related to the expansion of oil and steel, is discussed
further in Section 5.7.
The second term on the right hand side of Eq. (5-1), related to the calibration process, is discussed
further in Section 5.4.
The third term on the right hand side of Eq. (5-1), related to the proving process, is discussed further in
Section 5.5.
The fourth term on the right hand side of Eq. (5-1), related to the flow metering, is discussed further in
Section 5.6.
u ( q v 0, meas ) u ( ρ 0 ) u ( qm ) u ( qm ) u ( qm )
2 2 2 2 2
cal prov met
= + + + . (5-2)
q v 0, meas ρ q cal
q mprov q mmet
0 m
The interpretation of this equation is that the uncertainty consists of contributions related to density
measurements (first term), contributions related to the calibration process (second term), contributions
related to the proving process (third term), and contributions related to the flow metering (last term).
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 27 of 79
When the density is measured by a densitometer, the relative standard uncertainty of the volumetric
flow rate at standard conditions can be deduced from Eq. (4-21), and is here written as follows:
The interpretation of this equation is that the uncertainty consists of contributions related to density
measurements and expansion of oil (first term), contributions related to the calibration process (second
term), contributions related to the proving process (third term), and contributions related to the flow
metering (last term).
The first term on the right hand side of Eq. (5-3), related to the measured density and expansion of oil,
is discussed further in Section 5.7.
The second term on the right hand side of Eq. (5-3), related to the calibration process, is discussed
further in Section 5.4.
The third term on the right hand side of Eq. (5-3), related to the proving process, is discussed further in
Section 5.5.
The fourth term on the right hand side of Eq. (5-3), related to the flow metering, is discussed further in
Section 5.6.
5.2.1 Uncertainty model when volume flow rate is the primary measurement
The relative standard uncertainty of the volumetric flow rate at standard conditions can be deduced from
Eq. (4-10), and is here written as follows:
(
∆p m ,∆p ,c
Asteel ) 2
( )
u qvcal0
+ cal
2
+
(
u qvprov )
+
2
u qvmet ( ) 2
q prov q met
0 0
(5-4)
q A ∆p A m ,∆p ,c .
v ,meas liq steel qv 0 v0 v0
The interpretation of this equation is that the uncertainty consists of contributions related to expansion
of oil and steel (first term), contributions related to the calibration process (second term), contributions
related to the proving process (third term) and contributions related to the flow metering (last term).
The first term on the right hand side of Eq. (5-4), related to the expansion of oil and steel, is discussed
further in Section 5.7.
The second term on the right hand side of Eq. (5-4), related to the calibration process, is discussed
further in Section 5.4.
The third term on the right hand side of Eq. (5-4), related to the proving process, is discussed further in
Section 5.5.
The fourth term on the right hand side of Eq. (5-4), related to the flow metering, is discussed further in
Section 5.6.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 28 of 79
5.2.2 Uncertainty model when volume flow rate is estimated from measured
mass flow
For the case where the density is found from laboratory analysis, the relative standard uncertainty of the
volumetric flow rate at line conditions from measured mass can be deduced from Eq. (4-20):
" # " #
= + +
,
(5-5)
,
The relative standard uncertainty of the volumetric flow rate at line conditions from measured mass can
be deduced from Eq. (4-22), and is here written as follows, for the case when a densitometer is used:
= + + + . (5-6)
q v , meas Aliq ∆m
ρ q mcal q mprov q mmet
dens
The interpretation of these equations is that the uncertainty consists of contributions related to density
measurements and expansion of oil (first term), contributions related to the calibration process (second
term), contributions related to the proving process (third term), and contributions related to the flow
metering (last term).
The first term on the right hand side of Eqs. (5-5) and (5-6), related to the measured density and
expansion of oil, is discussed further in Section 5.7.
The second term on the right hand side of Eq. (5-6), related to the calibration process, is discussed
further in Section 5.4.
The third term on the right hand side of Eq. (5-6), related to the proving process, is discussed further in
Section 5.5.
The fourth term on the right hand side of Eq. (5-6), related to the flow metering, is discussed further in
Section 5.6.
5.3.1 Uncertainty model when mass flow rate is estimated from measured
volume flow
The relative standard uncertainty of the mass flow rate when it is estimated from measured volume flow
rate can in the case when the density is determined from laboratory analysis be deduces from Eq. (4-12),
and is here written as follows:
u (qm,meas ) u ρ 0 Aliq
=
2
(
m ,∆p m ,∆p ,c
Asteel ) 2
( )
u qvcal0
+ cal
2
+
(
u qvprov )
+
2
u qvmet ( ) 2
q prov q met
0 0
(5-7)
q ρ Am ,∆p Am ,∆p ,c .
m ,meas 0 liq steel qv 0 v0 v0
When the density is measured by a densitometer, the relative standard uncertainty of the mass flow rate
can be deduced from Eq. (4-14), and is here written as follows:
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 29 of 79
=
(
u (qm,meas ) u ρ dens Aliq
2 ∆m ,∆p m ,∆p ,c
Asteel ) 2
( )
u qvcal0
+ cal +
2
(
u qvprov ) +
2
( )
u qvmet
2
q prov q met
0 0
(5-8)
q .
m ,meas ρ dens Aliq Asteel
∆m , ∆p m ,∆p ,c
qv 0 v0 v0
The interpretation of these two equations is that the uncertainty consists of contributions related to
density measurement and expansion of oil and steel (first term), contributions related to the calibration
process (second term), contributions related to the proving process (third term) and contributions related
to the flow metering (last term).
The first term on the right hand side of Eqs. (5-7) and (5-8), related to the density measurement and
expansion of oil and steel, is discussed further in Section 5.7.
The second term on the right hand side of Eqs. (5-7) and (5-8), related to the calibration process, is
discussed further in Section 5.4.
The third term on the right hand side of Eqs. (5-7) and (5-8), related to the proving process, is discussed
further in Section 5.5.
The fourth term on the right hand side of Eqs. (5-7) and (5-8), related to the flow metering, is discussed
further in Section 5.6.
5.3.2 Uncertainty model when mass flow rate is the primary measurement
The relative standard uncertainty of the mass flow rate when mass flow is the primary measurement
can be deduced from Eq. (4-17), and is here written as follows:
u (qm,meas ) u qmcal ( ) ( ) ( )
2 2 2 2
u qmprov u qmmet
= cal + +
q q q prov q met . (5-9)
m , meas m m m
The interpretation of this equation is that the uncertainty consists of contributions related to the
calibration process (first term), contributions related to the proving process (second term) and
contributions related to the flow metering (last term).
The first term on the right hand side of Eq. (5-9), related to the calibration process, is discussed further
in Section 5.4.
The second term on the right hand side of Eq. (5-9), related to the proving process, is discussed further
in Section 5.5.
The third term on the right hand side of Eq. (5-9), related to the flow metering, is discussed further in
Section 5.6.
For volume flow measurements, the calibration uncertainty refers to the second term on the right hand
of Eqs. (5-1), (5-4), (5-7) and (5-8), and can be written as follows:
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 30 of 79
( )
u qvcal0
cal =
2
(
u qvcal0 ,ref
) 2
(
u qvcal0 ,rept − prover
+
) 2
. (5-10)
q cal qvcal0
qv 0 v0
Similar for mass flow rate measurement, referring to the second terms on the right hand side of Eqs.
(5-2), (5-3), (5-6) and first term of Eq. (5-9):
= + . (5-11)
q mcal q mcal q mcal
For both Eqs. (5-10) and (5-11) the first terms on the right hand side are the relative standard
uncertainties of the calibration references:
• If the proving device is a displacement prover (relevant for volume flow), this uncertainty term
refers to the reference system used at the on-site calibration of the prover.
• If the proving device is a master meter (flow meter), there are two options:
o If the master meter is calibrated off-site, at a flow laboratory, the uncertainty term refers
to the uncertainty of the flow reference at the flow laboratory.
o If the master meter is calibrated on-site, typically by use of a compact portable prover
with transfer meter, the uncertainty term refers to the uncertainty in that prover/transfer
meter system.
The second terms on the right hand side of both Eqs. (5-10) and (5-11) are the relative standard
uncertainties due to the repeatability obtained during calibration of the proving device. This is found from
the repeatability checks carried out under such calibrations. Note that the term relates to the uncertainty
and not the repeatability itself.
For volume flow measurements, the proving uncertainty refers to the third term on the right hand of Eqs.
(5-1), (5-4), (5-7) and (5-8) and can be written as follows:
(
u q vprov
prov 0 ) =
2
(
u q vprov
0 , rept − prover ) 2
u q vprov
+
(
0 , rept − flowmeter ) 2
(
u q vprov
+
0 ,linearity ) 2
(
u q vprov
+
0 , profile ) 2
. (5-12)
q q vprov q vprov q prov q prov
v0 0 0 v0 v0
Similar for mass flow rate measurement, referring to the third terms on the right hand of Eqs. (5-2), (5-3),
(5-6) and the second term of Eq. (5-9):
= +
, % , %
' , () ' , )
+& * +& * (5-13)
+ +
, + , +
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 31 of 79
The first two terms on the right hand sides of both Eqs. (5-12) and (5-13) correspond to the repeatability
of the prover and the flow meter. Normally they can be merged to a single term, which is found from the
repeatability check carried out under proving. Note that the term relates to the uncertainty and not the
repeatability itself.
The third terms on the right hand sides of both Eqs. (5-12) and (5-13) account for the effect that the
proving is not carried out at the same flow rate as used in the flow calibration. It is only relevant when
the proving device is a master meter. This is dealt with in the same way as in [Frøysa et al, 2014]
(Uncertainty of the correction factor estimate). The uncertainty contribution is described in Appendix A.
It is calculated from the deviation between the master meter flow rate and the flow rate measured by the
reference meter at flow calibration, at a series of flow rates. The adjustment of the master meter is
assumed to be carried out by linear interpolation. The actual expression for any uncorrected percentage
deviation, δp , and the related uncertainty of the master meter after adjustment of the master meter is
given in Appendix A.
As described in Appendix A, the relative standard uncertainty of the correction factor estimate can be
written as
(
u q vprov
0 ,linearity ) = u (K ) = 1 ∂p
u( p) =
δp 3 .
(5-14)
qv 0 K K ∂K 100 + p
Similar for mass flow rate (swap unit of volume flow rate with unit of mass flow rate in Appendix A):
u ( qmprov
,linearity )
u ( K ) 1 ∂p δp 3
= = u ( p) = . (5-15)
qm K K ∂K 100 + p
The last term on the right hand side of Eq. (5-12) and the fourth term of Eq. (5-13), account for the effect
on the master meter by changes in flow profile from the flow calibration to the proving. Typically, this
effect is larger when the master meter has been calibrated at an off-site calibration facility than when
the master meter has been calibrated on-site. The size of this term depends on the care taken for having
upstream pipe work at proving as close as possible to the pipe work at flow calibration. Flow meter
specifications and type tests can give indications of the size of this term in case of an off-site calibration.
It is, however, difficult to give specific and general numbers for this term. In the case of on-site
calibration, the term usually is expected to be smaller, because the master meter is not physically moved
between flow calibration and proving.
The two last terms of Eq. (5-13) are related to the influence of pressure and medium temperature,
respectively, if the calibration pressure differs from the process pressure during proving, or if the zero
point adjustment temperature at calibration differs from the process temperature during proving.
For volume flow measurements, the metering uncertainty refers to the last term on the right hand side
of Eqs. (5-1), (5-4), (5-7) and (5-8), and can be written as follows:
u (qvmet
0 )
u (qv 0,rept − flowmeter ) u (qv 0,linearity ) u (qv 0, profile )
2 met met2 met 2 2
met = .
+ + q met
(5-16)
q q met
q met
v0 v0 v0 v0
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 32 of 79
Similar for mass flow measurements, the last term on the right hand side of Eq. (5-9);
" #
= + +
, % , ( ,
(5-17)
+ +
, + , +
The first terms on the right hand sides of both Eqs. (5-16) and (5-17) correspond to the repeatability of
the flow meter. This is usually not measured directly (as in the flow calibration and proving operations)
but can be based on vendor specifications, proving repeatability tests or found in other ways. Note that
the term relates to the uncertainty and not the repeatability itself.
For a Coriolis flowmeter, the repeatability is assumed to be specified as follows, where zero point stability
in unit kg/h is input for the uncertainty estimation:
The second term on the right hand sides of both Eqs. (5-16) and (5-17) account for the effect that the
flow meter is not measuring on the same flow rate as the flow rate where proving was carried out. It is
calculated based on specification of the total linearity of the meter over the calibrated range of volumetric
flow rate at standard conditions. This linearity, given in percent, is denoted L. It is assumed that as a
maximum there is a linear drift of L % over the calibrated flow rate range of the flow meter. This means
that when the flow meter is proved at a volumetric flow rate at standard conditions q vprov 0
,and used at a
volumetric flow rate at standard conditions q vmet0 , the linearity uncertainty contribution can be written as
u (qvmet
0,linearity )
2
qvmet0 − qv 0
prov
= L
. (5-19)
q met 3 q cal
− q cal
v 0 v 0 , max v 0 , min
It is here assumed a rectangular probability function for the uncertainty, which means that the relative
standard uncertainty is found by dividing the maximum drift by the square root of 3.
u ( q mmet,linearity )
2
L qm − qm
met prov
= . (5-20)
q mmet 3 q mcal,max − q mcal,min
The last term on the right hand side of Eq. (5-16) and third term of Eq. (5-17) account for the effect on
the duty meter by changes in flow profile from the proving to normal operation. Typically, this term is
expected to be small, because the duty meter is not physically moved between flow calibration and
proving. In addition a new proving is carried out at each ship loading. In case of continuous operation a
new proving is carried out with some days time interval.
The two last terms of Eq. (5-17) are related to the influence of pressure and medium temperature,
respectively, if the proving pressure differs from the process pressure during metering, or if the zero
point adjustment temperature during proving differs from the process temperature during metering.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 33 of 79
In Section 5.7.1 the model for the uncertainty of the expansion factor found in Eq. (5-1) and related to
the volumetric flow rate at standard conditions, is presented.
In Section 5.7.2 the model for the uncertainty of the expansion factor found in Eq. (5-4) and related to
the volumetric flow rate at line conditions, is presented.
In Section 5.7.3 the models for the uncertainty of the expansion factors found in Eqs. (5-7) and (5-8)
and related to the mass flow rate, are presented.
Section 5.7.4 presents the models for the uncertainty found in Eq. (5-3) related to the volumetric flow
rate at standard conditions from measured mass flow rate.
Section 5.7.5 presents the models for the uncertainty found in Eq. (5-6) related to the volumetric flow
rate at line conditions.
C tlpprov C plp
prov met
C tlm met
C plm prov
C tsp prov
C psp met
C tsm met
C psm
Asteel =
m , ∆p m , ∆p ,c
Aliq . (5-21)
prov
C tlm prov
C plm cal
C tsp cal
C psp prov
C tsm prov
C psm
When the expressions for all expansion coefficients (see Sections 4.1.4 and 4.3) are inserted, Eq. (5-21)
can formally be written as
m,∆p m,∆p,c
Aliq (
Asteel = f Tpcal , Ppcal , Tpprov, Ppprov, Tmprov, Pmprov, Tmmet , Pmmet , ρ0 ). (5-22)
Each of these input parameters have uncertainty. In addition, there are material constants for the oil and
the steel. These also have uncertainty. In this work they are considered as part of the model uncertainty
m,∆p m,∆p,c
for each type of expansion factors that are included in Aliq Asteel . That means that uncertainty in
temperature expansion coefficient and pressure expansion coefficient for oil, and the same for steel are
considered to be part of the model uncertainty.
It should be noted here that the temperature Tpcal (the temperature at the proving device during
calibration) and the temperature T pprov (the temperature at the proving device during proving) are
measured by the same temperature measurement device. Similarly pressure Ppcal (the pressure at the
proving device during calibration) and the pressure Ppprov (the pressure at the proving device during
proving) are measured by the same pressure measurement device. However, there can be several
months or more between the measurements during calibration and the measurements during proving.
The temperature measurement device and the pressure measurement device can drift in-between these
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 34 of 79
measurements, and there may also be re-calibrations of these equipment. Therefore, in the uncertainty
model, Tpcal and Ppcal will be assumed to be uncorrelated with T pprov and Ppprov .
In the same way, the temperature T mprov (the temperature at the flow meter during proving) and the
temperature T mmet (the temperature at the flow meter during ordinary flow metering) are measured by
the same temperature measurement device. Similarly pressure P mprov (the pressure at the flow meter
during proving) and the pressure Pmmet (the pressure at the flow meter during ordinary flow metering)
are measured by the same pressure measurement device. Opposite to the case for the temperature
and pressure measurements at the proving the device, the temperature and pressure measurements at
the flow meter (during proving and at ordinary flow metering) are carried out within few days or less.
Therefore, in the uncertainty model, T mprov is considered to be totally correlated with T mmet , and P mprov
is considered to be totally correlated with Pmmet .
m,∆p m,∆p,c
The uncertainty model for Aliq Asteel can then be written as
( m , ∆p m , ∆p , c
u Aliq Asteel )
2
1
= m, ∆p m, ∆p , c
2
∂f
( )
cal
2
∂f
+ u Ppcal ( )
2
(
u Aliqm,,∆mod
p
)
+
2
m, ∆p , c
u Asteel ( ) 2
+
, mod
Am, ∆p , c ,
∂ρ0
m , ∆p
Aliq steel
where the two last terms represent model uncertainties. It will further be assumed that when a
temperature or pressure device is used two times (in both calibration and proving operation or both in
proving and normal flow metering operation), the absolute standard uncertainty of this temperature or
pressure device is the same in the two cases. This is consistent with the specifications of most
temperature and pressure devices and in agreement with the uncertainty models for temperature and
pressure given in Sections 3.1 and 3.2. This means that the notation will be simplified as follows:
u( Pm ) = u (Pmprov ) = u (Pmmet ).
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 35 of 79
u (Aliq ) = 1
2 2 2 2
m , ∆p m , ∆p , c
A ∂f ∂f u (T )2
m , ∆p msteel cal + prov
A A , ∆p , c Am , ∆p Am , ∆p , c ∂T p
liq steel liq steel ∂Tp p
∂f 2
∂f
2
2
+ cal + u (P )2 + ∂f + ∂f u (T )2
∂P ∂P prov p ∂T prov ∂T met m
p p m m
(5-25)
∂f ∂f 2 ∂f
2
2
+ prov + met u (Pm ) + u (ρ 0 )
∂Pm ∂Pm ∂ρ 0
u (Aliq , mod )
u (Asteel , mod )
2 2
m , ∆p m , ∆p , c
+ + .
A m , ∆ p A ,c
m , ∆ p
liq steel
( m ,∆p
u Aliq ,mod )
The model uncertainty term for oil expansion factors, m ,∆p , is addressed in Section 5.8.3. The model
Aliq
m , ∆p , c
, is addressed in Section 5.8.7.
Asteel
C tlpprov C plp
prov prov
C tsp prov
C psp met
C tsm met
C psm
A A ∆p m , ∆ p ,c
= . (5-26)
liq steel C prov C prov C cal C cal C prov C prov
tlm plm tsp psp tsm psm
When the expressions for all expansion coefficients (see Sections 4.1.4 and 4.3) are inserted, Eq. (5-26)
can formally be written as
∆p m,∆p,c
Aliq (
Asteel = f 2 Tpcal , Ppcal , Tpprov , Ppprov , Tmprov , Pmprov , Tmmet , Pmmet , ρ0 ). (5-27)
∆p m,∆p,c
The methodology and assumptions for finding the uncertainty Aliq Asteel is similar to the methodology
m,∆p m,∆p,c
and assumptions for finding the uncertainty Aliq Asteel in Section 5.7.1. The relative standard
∆p m,∆p,c
uncertainty of A A
liq steel then becomes
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 36 of 79
u (Aliq ) = 1
2 2 2 2
∆p m , ∆p , c
∂f 2 ∂f 2
cal + prov u (Tp )
A
∆p msteel 2
A A , ∆ p , c A∆ p Am , ∆ p , c
liq steel liq steel ∂Tp ∂Tp
∂f 2
∂f 2
2
2
+ cal2 + u (P )2 + ∂f 2 + ∂f 2 u (T )2
∂P ∂P prov p ∂T prov ∂T met m
p p m m
(5-28)
∂f 2 ∂f 2
2
∂f
2
+ prov + met u (Pm )2 + 2 u (ρ 0 )
∂Pm ∂Pm ∂ρ 0
u (Aliq , mod )
u (Asteel , mod )
2 2
∆p m , ∆p , c
+ + .
A ∆ p A ,c
m , ∆ p
liq steel
(
∆p
u Aliq ,mod )
The model uncertainty term for oil expansion factors, ∆p
Aliq
, is addressed in Section 5.8.4. The model
m , ∆p , c
, is addressed in Section 5.8.7.
Asteel
C tlpprov C plp
prov met
C tlm met
C plm prov
C tsp prov
C psp met
C tsm met
C psm
ρ 0 Aliqm ,∆p Asteel
m , ∆ p ,c
= ρ0 . (5-29)
prov prov cal cal prov
C tlm C plm C tsp C psp C tsm C psm prov
When the expressions for all expansion coefficients (see Sections 4.1.4 and 4.3) are inserted, Eq. (5-29)
can formally be written as
ρ0 Aliqm,∆p Asteel
m,∆p,c
= f 3 (Tpcal , Ppcal , Tpprov, Ppprov, Tmprov, Pmprov, Tmmet , Pmmet , ρ0 ) . (5-30)
Each of these input parameters have uncertainty. In addition, there are material constants for the oil and
the steel. These have also uncertainty. In this work they are considered as part of the model uncertainty
m,∆p m,∆p,c
for each type of expansion factors that are included in Aliq Asteel . That means temperature expansion
coefficient and pressure expansion coefficient for oil, and the same for steel.
The methodology and assumptions for finding the uncertainty ρ0 Aliq
m, ∆p m,∆p ,c
Asteel is similar to the
m,∆p m,∆p,c
methodology and assumptions for finding the uncertainty Aliq Asteel in Section 5.7.1. The relative
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 37 of 79
u (ρ 0 Aliq Asteel )
2 2 2 2
m , ∆p m , ∆p , c
∂f 3 ∂f 3
cal + prov u (Tp )
1
=
2
ρ Am , ∆p Am , ∆p ,c ρ Am , ∆p Am , ∆p ,c
0 liq steel 0 liq steel ∂Tp ∂Tp
∂f 2 ∂f 2 ∂f 3 ∂
2
u (Aliq , mod )
u (Asteel , mod )
2 2
m , ∆p m , ∆p , c
+ + .
A m , ∆ p A ,c
m , ∆ p
liq steel
( m ,∆p
u Aliq ,mod )
The model uncertainty term for oil expansion factors, m ,∆p , is addressed in Section 5.8.3. The model
Aliq
m , ∆p , c
, is addressed in Section 5.8.7.
Asteel
When the expressions for all expansion coefficients (see Sections 4.1.4 and 4.3) are inserted, Eq. (5-29)
can formally be written as
ρdensAliq Asteel = f4 (Tpcal , Ppcal , Tpprov, Ppprov, Tmprov, Pmprov, Tmmet , Pmmet , Tdmet , Pdmet , ρ0 , ρdens ).
∆m, ∆p m, ∆p, c
(5-33)
It is here assumed that the standard density, ρ0 , is calculated from the measured density at the
densitometer, and used in the volume correction coefficients. Each of these input parameters have
uncertainty. In addition, there are material constants for the oil and the steel. These also have
uncertainty. In this work they are considered as part of the model uncertainty for each type of expansion
∆m,∆p m,∆p ,c
factors that are included in Aliq Asteel . That means temperature expansion coefficient and pressure
expansion coefficient for oil, and the same for steel.
The methodology and assumptions for finding the uncertainty ρdens Aliq∆m,∆p Asteel
m,∆p ,c
is similar to the
m,∆p m,∆p,c
methodology and assumptions for finding the uncertainty Aliq Asteel in Section 5.7.1. The relative
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 38 of 79
(
u ρ dens Aliq∆m, ∆p Asteel
m , ∆p , c
)
2
2
2 2
∂f 4 ∂f 4
cal + prov u (Tp )
1
= 2
ρ A∆m , ∆p Am, ∆p , c ρ A∆m, ∆p Am, ∆p , c
dens liq steel dens liq steel ∂Tp ∂Tp
∂f 2 ∂f 2 ∂f 4 ∂
2
(
u A∆m , ∆p
+ ∆liqm,,mod
)
+
2
(
m , ∆p , c
u Asteel , mod ) 2
A ∆p m , ∆p , c .
liq Asteel
It should be noted here that a small approximation has been carried out. It is assumed that the
uncertainty of the relative density used in the volume correction coefficients is uncorrelated with the
density ρdensthat is explicitly written in Eq. (5-32). As the sensitivity for the relative density in the volume
correction coefficients is quite small, this is a reasonable approximation.
( ∆m ,∆p
u Aliq ,mod )
The model uncertainty term for oil expansion factors, ∆m ,∆p
Aliq
, is addressed in Section 5.8.2. The model
m , ∆p , c
, is addressed in Section 5.8.7.
Asteel
5.7.4 Volume flow rate at standard conditions, from measured mass flow
rate
The functional relationship and the uncertainty model for the volume flow rate depend on whether the
density is determined from laboratory analysis or measured by an online densitometer. These two cases
must therefore be addressed individually.
The uncertainty model for the volume flow rate at standard conditions from measured mass flow rate,
when the density is determined by densitometer, is given in Eq. (5-3). The relative standard uncertainty
of Ctldmet Cpld
met
ρdens is part of that equation. When the expressions for all expansion coefficients (see
Sections 4.2.2 and 4.2.3) are inserted, this ratio can formally be written as
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 39 of 79
It is here assumed that the standard density, ρ0 , is calculated from the measured density at the
densitometer, and used in the volume correction coefficients. Each of these input parameters have
uncertainty. In addition, there are material constants for the oil. These also have uncertainty. In this work
they are considered as part of the model uncertainty for each type of expansion factors that are included
in Ctldmet Cpld
met
. That means temperature expansion coefficient and pressure expansion coefficient for oil.
The methodology and assumptions for finding the uncertainty Ctldmet Cpld
met
ρdens is similar to the
m,∆p m,∆p,c
methodology and assumptions for finding the uncertainty A liq A
steel in Section 5.7.1. The relative
It should be noted here that a small approximation has been carried out. It is assumed that the
uncertainty of the relative density used in the volume correction coefficients is uncorrelated with the
density ρdensof the ratio Ctldmet Cpld
met
ρdens . As the sensitivity for the relative density in the volume
correction coefficients is quite small, this is a reasonable approximation.
u ( Ctldmet,modCpld
met
,mod )
The model uncertainty term for oil expansion factors, , is addressed in Section 5.8.5.
Ctldmet,modCpld
met
,mod
5.7.5 Volume flow rate at flow meter conditions, from measured mass flow
rate
The functional relationship and the uncertainty model for the volume flow rate depend on whether the
density is determined from laboratory analysis or measured by an online densitometer. These two cases
must therefore be addressed individually.
The uncertainty model for the volume flow rate at line conditions from measured mass flow rate, when
the density is determined by densitometer, is given in Eq.(5-6). The relative standard uncertainty of
1 ( Aliq
∆m
ρ dens ) is part of that equation. When the expressions for all expansion coefficients (see Sections
4.2.2 and 4.2.3) are inserted, this ratio can formally be written as
met met
It is here assumed that the standard density, ρ0 , is calculated from the measured density at the
densitometer, and used in the volume correction coefficients. Each of these input parameters have
uncertainty. In addition, there are material constants for the oil. These also have uncertainty. In this work
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 40 of 79
they are considered as part of the model uncertainty for each type of expansion factors that are included
in Aliq∆m . That means temperature expansion coefficient and pressure expansion coefficient for oil.
1⁄B∆ C E
1⁄B∆ C E
1 HIJ HIJ HIJ
= G "KE # + "LE # + "K #
1⁄ B ∆
E HKE HLE HK
(5-38)
C
It should be noted here that a small approximation has been carried out. It is assumed that the
uncertainty of the relative density used in the volume correction coefficients is uncorrelated with the
density ρdensthat is explicitly written in Eq.(5-37). As the sensitivity for the relative density in the volume
correction coefficients is quite small, this is a reasonable approximation.
u ( Aliq
∆m
,mod )
The model uncertainty term for oil expansion factors, ∆m
, is addressed in Section 5.8.6
Aliq
DENSITY FROM LABORATORY ANALYSIS
The uncertainty model for the volume flow rate at flow meter conditions from measured mass flow rate,
when the density is determined by laboratory analysis, is given in Eq. (5-5). The relative standard
uncertainty of is part of that equation. When the expressions for the expansion coefficients
(see Sections 4.2.2 and 4.2.3) are inserted, this ratio can formally be written as
1
= IN "K ,L , # (5-39)
The standard density, ρ0 , is derived from the laboratory analysis, and are assumed used in the volume
correction coefficients. Each of these input parameters have uncertainty. In addition, there are material
constants for the oil. These also have uncertainty. In this work, they are considered as part of the model
O
uncertainty for each of the expansion factors. The methodology and assumptions for finding the
uncertainty of STQ STQ is similar to the methodology and assumptions for finding the uncertainty
PQRS PURS
m,∆p m,∆p,c O
STQ STQ
PQRS PURS
A liq A
steel in Section 5.7.1. The relative standard uncertainty of then becomes
1⁄ V28 V28
8<V 5<V 0
1⁄ V28 V28
8<V 5<V 0
HL H H E
E V28
8<V
V28
5<V
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 41 of 79
+ V28
8<V,V4W
V28
5<V,V4W
The model uncertainty term for oil expansion factors, V28 V28 , is addressed in Section 5.8.5.
8<V 5<V
∂Ctlx
u(Ctlx )mod = u(BT ) . (5-41)
∂BT
Similarly, it is assumed that the pressure correction factor depends on a parameter. The model
uncertainty of a pressure expansion coefficient can then be written as
∂C plx
u (C plx )mod = u (BP ) . (5-42)
∂BP
This will be the basis for the derivations in the next three Sections.
It should also be mentioned that in API MPMS 11.1, the model uncertainty of the volume correction
coefficients (C , Cplx) are specified as a fixed percentage (fixed relative uncertainty) over wide ranges
tlx
of pressure and temperature. This is implicitly used below in the derivation of the uncertainty models.
An alternative, and maybe better approach would have been a model for the model uncertainty similar
to the approach for steel expansion factors, see Section 5.8.5. However, in such a case, the model
uncertainty of the oil volume correction factors would have to be defined as a linear function of the
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 42 of 79
difference between line and standard temperature and between line and standard pressure. Because
this is not the case in API MPMS 11.1, such an approach has not been selected here.
∆m ,∆ p
5.8.2 Oil expansion factor Aliq
∆m ,∆ p
The oil expansion factor Aliq is given in Eq. (4-15) and repeated here for convenience:
met met
∆m , ∆p C tlm C plm C tlpprov C plp
prov
∆m ,∆p ∆m , ∆p 2 2
∂Aliq ∂Aliq
(u(A ∆m ,∆p
liq ) )
mod
2
=
∂B
u (BT ) +
∂B
u ( BP )
T P
( (C )C (C ))C ( (C )C (C )) 2
= C plpprovC plm u (BT )
∂ prov ∂ prov met ∂ prov ∂
tlm + Ctlp ∂B tlm Ctld −Ctlp Ctlm ∂BT tld + Ctlm ∂B
prov met met prov met prov met met
C provC met
met ⋅
∂BT tlp T
tlm tlm T
tld (5-44)
plm pld ( )
prov met 2
Ctlm Ctld
C prov C met ( (C plp )C plm +C plp ∂BP (C plm ))C plm C pld −C plp C plm (∂BP (C plm )C pld +C plm ∂BT (C pld ))
2
+ CtlpprovCtlmmet ⋅ ∂BP u (BP ) .
∂ prov met prov ∂ met prov met prov met ∂ prov met prov ∂ met
It is now assumed that all the derivatives related to temperature are equal and that all derivatives related
prov
to pressure are equal. This is similar to assuming that the temperatures ( Tp , Tmprov, Tmmet , Tdmet ) in
question here are not too far away from each other, and similarly that the pressures
(P p
prov
, Pmprov, Pmmet, Pdmet ) in question here are not too far away from each other. This means that the
notation will be simplified as follows:
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 43 of 79
(u(A ∆m , ∆p
liq ) )
mod
2
( ) ( )
2
C plp
prov met met
C plm Ctlm + Ctlpprov Ctlm Ctld − CtlpprovCtlm
prov met met prov
Ctlm + Ctld
met
∂
= prov met ⋅ (Ctl )u (BT )
C plm C pld
prov met
Ctlm Ctld
2
( ∂BT )
( ) ( )
2
C provC met C met + C prov C provC met − C provC met C prov + C met ∂
+ prov met ⋅
tlp tlm plm plp plm pld plp plm plm pld
(C )u ( B )
Ctlm Ctld
prov met 2
C plm C pld ( ∂B P )pl P
(5-46)
( ) ( )
2
C plp
prov met met
C plm Ctlm + Ctlpprov Ctlm Ctld − CtlpprovCtlm
prov met met prov
Ctlm + Ctld
met
= ⋅ u (C )
C plm
prov met
C pld C prov met 2
tlm C tld ( tl
)mod
( ) ( )
2
C provC met C met + C prov C provC met − C provC met C prov + C met
+ prov met ⋅
tlp tlm plm plp plm pld plp plm plm pld
u (C ) .
Ctlm Ctld
prov met 2
C plm C pld ( )
pl mod
( ∆m , ∆p
u Aliq )
mod
2
met met met
= 1 + Ctlm − Ctlm − Ctlm
2
u (Ctl )mod
2
m ,∆p
5.8.3 Oil expansion factor Aliq
m ,∆p
The oil expansion factor Aliq is given in Eq. (4-5) and repeated here for convenience:
C tlpprov C plp
prov met
C tlm met
C plm
A m , ∆p
= ⋅ (5-48)
liq prov
C tlm prov
C plm
m ,∆p 2
m , ∆p 2
∂Aliq ∂Aliq
(u (A ) )m , ∆p
liq mod
2
=
∂B
u (BT ) +
∂B
u ( BP )
T P
C prov C met
= plp provplm ⋅
( (C )C
∂
∂BT
prov
tlp
met
tlm + Ctlpprov ∂
∂BT (C met
tlm ))C prov
tlm − Ctlpprov Ctlm
met ∂
∂BT (C ) u (B )
prov
tlm
2
C plm
(C prov 2
tlm ) T
(5-49)
( (C )C (C ))C (C ) u (B )
2
C prov C met ∂ prov met
+ C plp
prov ∂ met prov
− C plp
prov met
C plm ∂ prov
+ tlp provtlm ⋅
∂BP plp plm ∂BP plm plm ∂BP plm
.
Ctlm
(C prov 2
plm ) P
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 44 of 79
As in Section 5.8.2, it is assumed that all the derivatives related to temperature are equal and that all
derivatives related to pressure are equal. Eq. (5-49) then simplifies to
( m , ∆p
u Aliq )
mod
2
met met
= 1 + Ctlm − Ctlm
2
u (Ctl )mod
2
met
C plm
+ 1 + prov
met
C plm
− prov
2
u (C pl )mod
2
. (5-51)
Aliq m , ∆p Ctlpprov Ctlmprov
Ctlm
met
C plp C plm C met
plm
∆p
5.8.4 Oil expansion factor Aliq
∆p
The oil expansion factor Aliq is given in Eq. (4-11) and repeated here for convenience:
∆p C tlpprov C plp
prov
∆p 2
∆p 2
∂Aliq ∂Aliq
(u (A ) )∆p
liq mod
2
=
∂B
u (BT ) +
∂B
u ( BP )
T P
=
C plp
prov
⋅
∂
∂BT (C )C
prov
tlp
prov
tlm − Ctlpprov ∂
∂BT (C ) u (B )
prov
tlm
2
C plm
prov
(C ) prov 2
tlm
T
(5-53)
(C )C − C (C ) u (B )
2
C prov C met ∂ prov prov prov ∂ prov
+ tlp provtlm ⋅
∂BP plp plm plp ∂BP plm
.
Ctlm
(C ) prov 2
plm
P
As in Sections 5.8.2 and 5.8.3, it is assumed that all the derivatives related to temperature are equal
and that all derivatives related to pressure are equal. Eq. (5-53) then simplifies to
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 45 of 79
2
C plp − Ctlpprov ∂
(u(A ) )
prov prov
Ctlm
∆p 2
= ⋅ (C )u ( B )
liq mod
C plm
prov prov 2
Ctlm ( ∂B T
tl
) T
2
C prov C prov − C prov ∂
+ prov ⋅
tlp plm plp
(C )u ( B )
Ctlm
prov 2
C plm (
∂BP
pl P
)
2
(5-54)
C prov
C prov
−C prov
= ⋅ u (Ctl )mod
plp tlm tlp
C
prov
plm (C ) prov 2
tlm
2
C prov C prov − C prov
+ prov ⋅
tlp plm plp
u (C ) .
Ctlm
prov 2
C plm ( pl mod
)
Expressed with relative standard uncertainties, Eq. (5-54) becomes
( )
∆p
u Aliq mod
2
C prov
= 1 − tlp
2
u (C tl )mod
2
C plp
prov
+ 1 − prov
2
u (C pl )mod
2
. (5-55)
Aliqm , ∆p C prov prov C prov
tlm C tlm C plm plm
H" # H" #
' ) = "^_ # + "^` #
] ] ] ]
], E ], E
H^_ H^`
H H
(5-56)
= " # "^_ # + "^` #
]
H^_ ] ]
H^` ]
As in Sections 5.8.2, 5.8.3 and 5.8.4 it is assumed that all the derivatives related to temperature are
equal and that all derivatives related to pressure are equal. Eq. (5-56) then simplifies to
H H
' ) = " # "^_ # + "^` #
], E ], E ]
H^_ ]
H^` (5-57)
= ] " # E +' ] )
E
" #
= +
], E ], E ] ] E ] ] E
] ] ] ] ] ] ] ] (5-58)
+ PUR
=. @ + . P STQS[\ @
+"PQR #S[\
PQRa
STQ
URa
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 46 of 79
∆m
5.8.6 Oil expansion factor Aliq
The oil expansion factor Aliq∆m is given in Eq. (4-23) and repeated here for convenience:
met met
∆m
C tlm C plm
Aliq = met met
. (5-59)
C tld C pld
HB∆ C HB∆ C
' B∆ C ) =& "^_ #* + & "^` #*
E H^_ H^`
H H
d ' " # − " # ) h
H^_ E H^_ E
=c "^_ #g
E " E # (5-60)
b f
H H
d ' − ) h
H^` E H^` E
+c "^` #g
E E
b f
As in Sections 5.8.2, 5.8.3 and 5.8.4 it is assumed that all the derivatives related to temperature are
equal and that all derivatives related to pressure are equal. Eq. (5-60) then simplifies to
' B∆ C )
E
− H − H
=& " # "^_ #* + & "^` #*
E E
− −
= " # +
E E
E " E # E
E
E
E
+ X∆S STQ
PURS + PUR
= .1 − @ . @ + .1 − P STQ @ . @
RYZ S[\ STQ
PQRS +"PQR #S[\ S[\
S,∆U
XRYZ P STQ PQRS
STQ
PURS
STQ
(5-62)
QR\ UR\
m ,∆p ,c
5.8.7 Steel expansion factor Asteel
m ,∆p ,c
The steel expansion factor Asteel is given in Eq. (4-6) and repeated here for convenience:
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 47 of 79
C tspprov C psp
prov met
C tsm met
C psm
m , ∆ p ,c
Asteel = cal cal prov prov . (5-63)
C C C C
tsp psp tsm psm
As discussed in Sections 4.3.1 and 4.3.2, the temperature and pressure volume correction factors for
steel can be written as
see Eqs. (4-33) and (4-37). The parameters α and β depends on the type of flow meter / prover and on
m ,∆p ,c
the type of steel quality. The model uncertainty of Asteel will be calculated from the uncertainty in the
α and β parameters.
• Same type of duty flow meter as master meter (ultrasonic and ultrasonic or turbine and turbine)
• Different type of duty flow meter than master meter
The first case, when an ultrasonic flow meter or a turbine meter is used as master meter for a
duty meter of the same type. Here, the α and β coefficients will be the same for the duty meter and
m ,∆p ,c
the master meter. In this case Asteel can be written as
m ,∆ p ,c
=
(1 + 3α (T m p
prov
− Tb ))(1 + β m (Ppprov − Pb ))(1 + 3α m (Tmmet − Tb ))(1 + β m (Pmmet − Pb ))
(1 + 3α (T − Tb ))(1 + β m (Ppcal − Pb ))(1 + 3α m (Tmprov − Tb ))(1 + β m (Pmprov − Pb ))
Asteel cal
. (5-66)
m p
m ,∆p ,c
The model uncertainty of Asteel can be written as
2 2
∂A m , ∆p , c ∂Am , ∆p , c
(u (A m , ∆p , c
))2
= steel u (α m ) + steel u (β m ) . (5-67)
∂α m ∂β m
steel , mod
By inserting Eq. (5-66) into Eq. (5-67) and carrying out the differentiation, the relative standard model
uncertainty can be written as
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 48 of 79
(m , ∆p , c
u Asteel , mod )
=
2
3α m Tpprov − Tb
+
( 3α m Tmmet − Tb ) ( )
Am , ∆p , c
steel
1 + 3α T prov − T
m p b (
1 + 3α m Tpprov − Tb ) ( )
−
(
3α m Tpcal − Tb
−
)
3α m Tmprov − Tb ( )
2
u (α m )
2
( )
1 + 3α m Tpcal − Tb 1 + 3α m Tmprov − Tb ( )
αm
( −P)
(5-68)
+
β m Ppprov b
+
β m Pmmet − Pb ( )
m p (
1 + β P prov − P 1 + β P prov − P
b m p b ) ( )
β m (Ppcal − Pb ) β m (Pmprov − Pb ) u (β m )
2 2
− − .
1 + β m (Ppcal − Pb ) 1 + β m (Pmprov − Pb ) βm
Next, the case of where a different type of meter than the duty meter is used as master meter is
addressed. In that case there will be different α and β coefficients for the flow meter and the proving
device, and they have to be treated as four uncorrelated variables (α ,α , β , β ) with respect to the
m p m p
m ,∆p ,c
uncertainty. In this case A steel can be written as
m , ∆ p ,c
=
(1 + 3α (Tp p
prov
))( ( ))( (
− Tb 1 + β p Ppprov − Pb 1 + 3α m Tmmet − Tb 1 + β m Pmmet − Pb ))( ( )) .
A steel
(1 + 3α (T
p p
cal
− Tb ))(1 + β (P p
cal
p − P ))(1 + 3α (T
b m m
prov
− Tb ))(1 + β (P m m
prov
− P ))
b
(5-69)
m ,∆p ,c
The model uncertainty of Asteel can be written as
2 2
∂Am,∆p ,c ∂Am,∆p ,c
(u(A m,∆p ,c
steel,mod )) 2
= steel u(α p ) + steel u(β p )
∂α ∂β
p p (5-70)
2 2
m ,∆p ,c m ,∆p ,c
∂A ∂A
+ u(α m ) +
steel
u(β m ) . steel
∂α m ∂β m
By inserting Eq. (5-69) into Eq. (5-70) and carrying out the differentiation, the relative standard model
uncertainty can be written as
( m ,∆p ,c
u Asteel
,mod )
=
2
3α p Tpprov − Tpcal ( )
2
u (α p )
2
Am,∆p ,c
steel p p ( b (
1 + 3α T prov − T 1 + 3α T cal − T
p p b ))( ( ))
α
p
+
β p Ppprov − Ppcal ( )
2
u (β p )
2
( p p (
1 + β P prov − P 1 + β P cal − P
b p p b ))( ( ))
β
p (5-71)
+
3α m T − T ( met prov
)
2
u(α m )
2
( ( ))( ( ))
m m
1 + 3α m Tm − Tb 1 + 3α m Tm − Tb αm
met prov
+
β m Pmmet − Pmprov ( )
2
u(β m )
.
2
( (
1 + β m Pm − Pb 1 + β m Pm − Pb
met prov
))( ( )) βm
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 49 of 79
6 Program
This Chapter documents the web-based computer program for carrying out uncertainty analyses based
on the uncertainty models described in this Handbook. It should here be emphasised that the example
input values in that calculation tool are just examples, and should not be regarded as recommended
values by NFOGM, CMR, NPD or any other party.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 50 of 79
The start page of the application (Figure 6-1) is also the page where the user specifies the metering
station template, meaning the general type of instruments and the layout of these. There are several
aspects of the metering station that is modeled:
• Flow Meter: what type of meter is used and what configuration of sensors is used to measure the
line temperature and line pressure. Available flow meters: ultrasonic, turbine and Coriolis.
• Stationary prover / master meter: what type of stationary prover / master meter is used
(Ultrasonic, Turbine or Displacement Prover, Coriolis if Coriolis flow meter) and what
configuration of sensors is used to measure the temperature and pressure of the prover / master
meter (single, dual or average).
• Density measurement: what type of density measurement is used, laboratory measurement or
installed densitometer. If an installed densitometer is used, what configuration of sensors is used
to measure the densitometer temperature and pressure (single, dual or average).
By specifying choices for each of these aspects, the user is in effect selecting a metering station
template. When the user then presses the “Accept and Continue”-button a copy of the selected template
is created and the application moves to the first of several input pages, “Oil Properties” (Figure 6-2). A
page navigation menu below the application header is also displayed, where the user now can move
freely between different pages (Figure 6-3), some related to input and others related to computed results
and visualizations. The pages typically organize content in several sections, and the user can select a
section with some form of navigation control.
The selected metering station template is set up with some example values, so the user can explore the
application functionality without first finishing all the data input.
The following pages are available after the metering station template has been selected:
• Metering Station: start page where the selected template is displayed. The user can also create
a new or open an existing from a file.
• Oil: input regarding oil product type and also other properties like base pressure and
temperature.
• Equipment: input regarding properties and uncertainty in the metering station equipment, for
example flow meter, master meter, and the temperature and pressure sensors used.
• Calibration: input regarding calibration conditions and uncertainty in the calibration procedure.
This input page is not applicable when the selected stationary prover is of type “Displacement
Prover”.
• Proving: input regarding proving conditions and uncertainty in the proving procedure.
• Metering: input regarding metering conditions and uncertainty in the metering procedure.
• Results: computed uncertainty of the main flow measurement variables (absolute volume flow,
standard volume flow and mass flow), volume correction factors and density measurement, all
displayed as uncertainty budgets tables.
• Charts: computed uncertainty of the main flow measurement variables (absolute volume flow,
standard volume flow and mass flow), volume correction factors and density measurement, all
displayed as uncertainty budgets tables.
• Plots: computed uncertainty of the main flow measurement variables (absolute volume flow,
standard volume flow and mass flow) as function of a selected flow rate range, displayed as
plots.
• Report: summary of the uncertainty analysis formatted as an on-screen report. This can be
printed, exported to Excel (XLSX) and PDF format, and it is possible to save the analysis in a file
for later use, sharing and reference.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 51 of 79
The user can move between the input pages in any order, but due to computational dependencies the
following work flow is recommended when input data: “Oil”->”Equipment”->”Calibration”->”Proving”-
>”Metering”. In addition, the logical flow between sections in each page is typically from left to right.
The following discusses each of the pages.
Figure 6-1 Oilmetering application start page, where the user specifies the metering station template. It
is also possible to open a previously saved file.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 52 of 79
Figure 6-2 “Oil”-page with “Product Type”-section selected, where the user specifies the oil product type
from a set of API standards, or model another product type by entering values for a set of API
constants.
Figure 6-3 Page navigation menu where the user can move freely between different pages, some related
to input and others related to computed results and visualizations.
There are two sections on the oil properties page, one concerning the reference density and API model
of the oil, and the other concerning the operating conditions:
• Product Type: specification of oil density at reference conditions and API standard oil product
type, defined by a set of API constants (Figure 1.2).
• Conditions: specification of base pressure and temperature and equilibrium vapor pressure
(Figure 6-4).
The available oil product types is as defined in API. An oil product not in the API standard can be
specified by choosing the “Other”-checkbox. The table listing the API standard constants (Figure 6-2)
will then be enabled (it is read-only otherwise) and relevant model parameters can be input and will be
used in the calculation. The model uncertainties for “Ctl” and “Cpl” is also given as either API standards,
or user defined. When specified as “API” the actual values (temperature and pressure dependent) is
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 53 of 79
displayed for reference in corresponding read-only text fields. When specified as “User Defined” these
text fields become enabled and the user can enter values directly.
In the “Conditions” section, the “normal process conditions” is defined. This is the typical system wide
pressure and temperature values, meaning that these values will be used as default values for different
pressure and temperature values until specified otherwise. For example when later specifying “Proving
Conditions” (Figure 6-9), the “flow meter conditions” and “master meter conditions” will both be equal to
the “normal process conditions”. The user can then change this as necessary.
Figure 6-4 “Oil”-page with “Conditions”-section selected, where the user specifies the oil operating
conditions.
The “Equipment”-page contains input regarding properties and uncertainty in the metering station
equipment (Figure 6-5). The content of this page depends on the selected metering station template,
but it can include
Some of the equipment uncertainty models can be specified either as an overall measurement
uncertainty or as a detailed model of a typical sensor. This choice is controlled by checkbox as shown
in (Figure 6-6) and (Figure 6-7) for a temperature sensor.
Note that in some views there is functionality for storing frequently used specifications in files for later
retrieval (as shown in Figure 6-7 where the detailed input for a temperature transmitter). The “Save”-
button can be used to save the complete specification to a file, and the “Load”-button can then later be
used for quickly loading the saved specification for a temperature transmitter.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 54 of 79
Figure 6-5 “Equipment”-page with “Flow Meter”-section selected. This page concerns uncertainty in the
metering station equipment.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 55 of 79
Figure 6-7 Detailed input for a temperature transmitter. The “Save”-button can be used to save the
complete specification to a file, and the “Load”-button can then later be used for quickly
loading the saved specification.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 56 of 79
Figure 6-8 “Calibration”-page with “Master Meter Calibration”-section selected. This page concerns
calibration conditions and uncertainty in the calibration procedure.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 57 of 79
The “Proving”-page contains input regarding proving conditions and uncertainty in the proving
procedure. There is significant differences between the scenario where proving is performed with an
ultrasonic or turbine master meter, and the scenario where a displacement prover is used. Therefore,
these are now described separately.
Figure 6-9 “Proving”-page with “conditions”-section for scenario with duty meter proving by ultrasonic
or turbine master meter.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 58 of 79
Figure 6-10 For measured volume flow rate: Uncertainty in proving of duty meter against master meter at
a given proving rate. Note the visualization of the uncertainty contribution due to difference
between proving flow rate and nearest calibration flow rate for the master meter.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 59 of 79
Figure 6-11: For Coriolis flow meter (measured mass flow rate): Uncertainty in proving of duty meter
against master meter at a given proving rate. Note the visualization of the uncertainty
contribution due to difference between proving flow rate and nearest calibration flow rate for
the master meter. Uncertainty contributions related to changes in pressure and temperature
from proving to operation are specified separately.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 60 of 79
Figure 6-12 “Proving”-page with “conditions”-section for scenario with duty meter proving by
displacement prover.
The “Metering”-page contains input regarding metering conditions and uncertainty in the metering
procedure. It contains the following sections:
• Metering Conditions: pressure and temperature conditions for flow meter at metering (Figure
6-14). These are used to calculate corresponding liquid and steel volume correction factors.
• Metering Uncertainty: specifies the metering flow rate, the operating range and the linearity of the
flow meter in the operating range, and the uncertainty in the flow measurement procedure at this
flow rate (Figure 6-15). In the uncertainty model, repeatability for flow meter at the metering rate
can be specified, and in addition an uncertainty due to flow profile and fluid effects on the flow
meter. There is also an uncertainty contribution due to difference between metering flow rate and
proving flow rate, and this is automatically computed using the linearity of the flow meter and
included in the model. This contribution is also visualized as shown in Figure 6-15. For the case of
Coriolis meter, the uncertainty contribution for the flow meter repeatability at metering is calculated
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 61 of 79
from the flow meter repeatability uncertainty and the specified zero point stability. The uncertainty
model for the Coriolis meter is shown in Figure 6-16.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 62 of 79
Figure 6-15 Measured volume flow rate: “Metering”-page with “Uncertainty”-section for uncertainty in
flow metering at a given flow rate. Note the visualization of the uncertainty contribution due
to difference between metering flow rate and proving flow rate.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 63 of 79
Figure 6-16: Coriolis meter (measured mass flow rate): “Metering”-page with “Uncertainty”-section for
uncertainty in flow metering at a given flow rate. Note the visualization of the uncertainty
contribution due to difference between metering flow rate and proving flow rate. Uncertainty
contribution for the flow meter repeatability at metering is calculated from the flow meter
repeatability uncertainty and the specified zero point stability.
This page is the first of several pages that displays the result of the uncertainty calculation based on the
input data (Figure 6-17). There is one section for each of the main flow measurement variables, standard
volume flow, absolute volume flow and mass flow. Then there is one section for each of the relevant
volume correction factor, where the number of factors depends on the chosen template. In addition,
depending on whether oil densitometer is used, there will be a section for the uncertainty in the
computation of reference density.
The uncertainty is displayed as uncertainty budgets tables, and the functional relationship is displayed
for reference. Depending on the model displayed, there can also be a list of “computed values”. These
are values computed from the input data for use in the uncertainty calculation and listed here for
convenience. An example of this is in Figure 6-18 displaying the uncertainty model for the volume
correction factor AliqΔp and where the different computed values for the related factors is listed for
reference (Ctlm, Cplm, Ctlp, Cplp, etc.).
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 64 of 79
Figure 6-17 Computed uncertainty of the main flow measurement variables (standard volume flow,
absolute volume flow and mass flow) and some other essential values, displayed as
uncertainty budgets tables.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 65 of 79
Figure 6-18 Uncertainty model for the volume correction factor AliqΔmΔp with the different computed
values for the related factors listed for reference.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 66 of 79
This page displays the same data as the “Result”-page, but as uncertainty budget charts (Figure 6-19).
That is, there is one chart for each of the main flow measurement variables, standard volume flow,
absolute volume flow and mass flow. Then there is one chart for each of the relevant volume correction
factors, where the number of factors depends on the chosen template. In addition, depending on
whether oil densitometer is used there will be a chart for the uncertainty in the computation of reference
density.
Note that the “Export Image”-button let the user save an image of the chart to a file.
Figure 6-19 Computed uncertainty of the main flow measurement variables (standard volume flow,
absolute volume flow and mass flow) and some other essential values, displayed as
uncertainty budgets charts. The “Export Image”-button let the user save an image of the chart
to a file.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 67 of 79
The “Plots”-page contains computed uncertainty of the main flow measurement variables (standard
volume flow, absolute volume flow and mass flow) as function of a selected flow rate range, displayed
as plots (Figure 6-20). It is possible to select the flow rate range and the flow rate unit (Sm3/h, kg/h,
m3/h).
Note that the “Export Image”-button let the user save an image of the plot to a file.
Figure 6-20 Computed uncertainty of the main flow measurement variables plotted over a selectable flow
range.
The “Report”-page contains a summary of the uncertainty analysis formatted as an on-screen report
(Figure 6-21). There are 4 essential features represented by the 4 buttons on the top of the page:
• Save uncertainty analysis: it is possible to save the analysis in a file for later use, sharing and
reference. The file is encrypted so that it cannot be tampered with (note that this does not imply
that it is password protected and it can still be shared with others).
• Print report: the on-screen report containing a summary of the analysis can be sent to a printer if
available.
• Save report to XLSX: a more detailed report containing the input data and analysis results can be
saved in a XLSX file. This is the standard file format for Excel, but several other spreadsheet
application is compatible with this format today.
• Save report to PDF: the same report as for XLSX can be saved in PDF format.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 68 of 79
Figure 6-21 “Report”-page contains a summary of the uncertainty analysis formatted as an on-screen
report. This can be printed and it is possible to save the analysis in a file for later use and
reference.
The more detailed XLSX and PDF report includes in addition the following:
• Detailed data regarding oil product type and operating conditions like base pressure and
temperature, as given by the user on the “Oil”-page.
• Uncertainty in the calibration procedure as given by the user on the “Calibration”-page.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 69 of 79
When the user saves an uncertainty analysis to file, it will always be a new file, named from a standard
“Save File As..”-dialog. It is not possible to save “changes” to an existing file. In practice, this is not a
limitation. If the user opens an uncertainty analysis file and want to “save changes”, it is always possible
to just use the same file name and thereby overwrite the file.
While this mechanism seems like an unnecessary limitation, it is in fact an important security feature of
Silverlight. A Silverlight application cannot generally access the file system on a computer. The only
exception to this is if the user is shown a file select dialog (controlled by the system, not the application)
and then selects a specific file to open and read (read-only) or a name for a file to create (write-only).
Through the system controlled file dialog the user has full control over what files the application can
read, and over what file areas and file name the application can write to.
When the application start page is first shown the two buttons at the bottom right “Accept and Continue”
and “Open From File” is both enabled. If the user chooses either of these the application moves to the
“Oil” page. If the user now goes back to the start page the “Open From File” button is no longer enabled
and the “Accept and Continue” button have changed name to “Create New”. It is therefore not possible
to open an uncertainty analysis file from this state. To either create a new uncertainty analysis or open
an existing from file the user must first press the “Create New” button. This returns the application to the
initial state where both the “Accept and Continue” and “Open From File” is enabled.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 70 of 79
7 Summary
This Handbook documents uncertainty models for fiscal oil metering stations using ultrasonic flow meter,
turbine flow meter or Coriolis flow meter. Proving device is either a displacement prover, an ultrasonic
flow master meter or a turbine flow master meter in case of volume flow meter. If the flow meter is a
Coriolis meter, the proving device will also be a Coriolis flow meter. The uncertainty models cover
volumetric flow rate at standard conditions, volumetric flow rate at line conditions and mass flow rate.
The density is either measured by an online densitometer, or obtained through sampling and laboratory
analysis. The uncertainty models are implemented on a web-based Microsoft Silverlight technology.
This can be accessed free of charge from www.nfogm.no.
The present work is related to a similar work on fiscal gas metering stations, see [Frøysa et al, 2013]. It
is also based on [Dahl et al, 2003], [Lunde et al, 2002] and [Lunde et al, 2010].
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 71 of 79
8 References
API MPMS 4.2, “Manual of Petroleum Measurement Standards – Chapter 4: Proving Systems, Section
2 – Displacement Provers,” American Petroleum Institute, Washington DC, 2003.
API MPMS 4.5, “Manual of Petroleum Measurement Standards – Chapter 4: Proving Systems, Section
5 – Master-Meter Provers,” American Petroleum Institute, Washington DC, 2011.
API MPMS 5.3, “Manual of Petroleum Measurement Standards – Chapter 5: Metering, Section 3 –
Measurement of Liquid Hydrocarbons by Turbine Meters,” American Petroleum Institute, Washington
DC, 2005.
API MPMS 5.8, “Manual of Petroleum Measurement Standards – Chapter 5: Metering, Section 8 –
Measurement of Liquid Hydrocarbons using Ultrasonic Flow Meters,” American Petroleum Institute,
Washington DC, 2011.
API MPMS 11.1, “Manual of Petroleum Measurement Standards – Chapter 11: Physical Properties
Data, Section 1 – Temperature and Pressure Volume Correction Factors (VCF) Software for
Generalized Crude Oils, Refined Products, and Lubricating Oils,” American Petroleum Institute,
Washington DC, 2004.
API MPMS 12.2, "Manual of Petroleum Measurement Standards – Chapter 12: Calculation of Petroleum
Quantities, Section 2 - Calculation of Petroleum Quantities Using Dynamic Measurement Methods and
Volumetric Correction Factors," Part 1-5, American Petroleum Institute, Washington DC, 1995 (Part 1),
2003 (Part 2), 1998 (Part 3), 1997 (Part 4), 2001 (Part 5).
Dahl, E., Frøysa, K.-E. and Lunde, P.: “Handbook of Uncertainty Calculations - Fiscal Orifice Gas and
Turbine Oil Metering Stations”, Issued by Norwegian Society for Oil and Gas Measurement (NFOGM)
and Norwegian Petroleum Directorate, 2003.
Frøysa, K.-E. and Lied, G. Ø.: “Handbook for uncertainty calculations for gas flow metering stations.
Documentation of uncertainty models and internet tool”, CMR report no. CMR-14-A10099-RA-1-Rev01,
issued by Norwegian Society for Oil and Gas Measurement (NFOGM) and Norwegian Petroleum
Directorate, 2014.
Lunde, P. and Frøysa, K.-E.: “Handbook of Uncertainty Calculations, ultrasonic fiscal gas metering
stations,” Issued by Norwegian Society for Oil and Gas Measurement (NFOGM) and Norwegian
Petroleum Directorate, 2002.
Lunde, P. and Frøysa, K.-E.: ”Ormen Lange ultrasonic gas flow meters. A study for establishment of
corrections for pressure and temperature effects.” CMR report no. CMR-06-A10048-RA-01, March
2007.
Lunde, P., Frøysa, K.-E., Lied, G., Skatvedt, R. and Vervik, S.: "Handbook of uncertainty calculations.
Fiscal oil metering stations based on ultrasonic duty and master meters," in proceedings from the 28th
International North Sea Flow Measurement Workshop, St. Andrews, Scotland, October 2010.
ISO 5168, “Measurement of fluid flow – Procedures for the evaluation of uncertainties,” ISO. Geneva,
2005.
ISO 12242, “Measurement of fluid flow in closed conduits – Ultrasonic transit-time meters for liquid,”
ISO, Geneva, 2012.
ISO 17089-1, “Measurement of fluid flow in closed conduits – Ultrasonic meters for gas – Part 1: Meters
for custody transfer and allocation measurement,” ISO, Geneva, 2010.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 72 of 79
NORSOK I-105: “Fiscal measurement system for hydrocarbon liquid,” NORSOK standard I-105, Rev.
2, 1998.
NORSOK I-106: “Fiscal measurement system for hydrocarbon liquid and gas,” NORSOK standard I-
106, Ed. 2, 2014.
NPD, “Regulations relating to measurement of petroleum for fiscal purposes and for calculation of CO2-
tax (The measurement regulations),” The Norwegian Petroleum Directorate (NPD), 2001, last translated
2012.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 73 of 79
A1 Functional relationship
After flow calibration, an adjustment of the flow meter shall be performed. The flow calibration is carried
out by comparing the output flow rate from the flow meter with the similar reading from a reference
measurement. This is carried out at a set of N different flow rates where the reference meter measured
the flow rate q v , ref , i and the flow meter measured the flow rate q v , Meter , i , i = 1, …, N. A full correction of
the flow meter at each of these flow rates can therefore be written as
where
qv ,ref ,i
Ki = . (A.2)
qv ,Meter ,i
The relative difference in per cent between the flow rate as measured by the flow meter and the
reference meter can be written as
qv ,Meter,i − qv ,ref ,i
pi = 100 . (A.3)
qv ,ref ,i
1 − Ki 100
pi = 100 ; Ki = . (A.4)
Ki 100 + pi
From these correction factors a general correction factor valid for all flow rates (and not only at the
specific flow rates where the flow calibration is carried out) is established. This can formally be written
as
qv = Kqv,Meter, (A.5)
where
K = f (K1, K2 ,Κ , KN , qv ) . (A.6)
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 74 of 79
1− K
p = 100 ; K = 100 (100+ p) . (A.7)
K
In practice, such a correction could have been carried out by different methods, including
i. no correction,
ii. a constant percentage correction,
iii. linear interpolation, and
iv. other methods (splines and other curve fittings).
If one of the two first methods is used, there will be known systematic errors left which are not corrected
for. This is not in accordance with the Norwegian Petroleum Directorate Measurement Regulations
[NPD, 2012], where one requirement in Section 8 is that “The measurement system shall be designed
so that systematic measurement errors are avoided or compensated for”. They will therefore not be
covered here.
In the third method, the adjustment will be based on a linear interpolation between the adjustment factors
established for the flow rates used in the flow calibration. Such an interpolation can be carried out either
on K, or on the percentage deviation p. Here, a linear interpolation in p is described. Both for the
correction and for the uncertainty analysis, the results will almost be the same whether the interpolation
is carried out on p or on K. The linear interpolation can be written as
pi +1 − pi
p = pi + (qv ,Meter − qv ,Meter ,i );
qv , Meter ,i +1 − qv ,Meter ,i
(A.8)
K can then be found from Eq. (A.7). It should be commented that this third method provides a correction
such that the flow meter’s flow rate will be corrected to the reference meter flow rate, when the flow rate
is equal to any of the flow rates used in the flow calibration. This case is therefore in agreement with the
Norwegian Petroleum Directorate Measurement Regulations.
The fourth method is a generalization of the third method, where the linear interpolation is replaced with
a non-linear interpolation (e.g. based on splines) or a partially linear interpolation where more
interpolation points than the ones used in the flow calibration (ref. method (iii)) are used. In such cases,
it is recommended that for the uncertainty analysis, it is treated as method (iii).
A2 Uncertainty model
The uncertainty of the flow rate due to the above mentioned adjustment of a flow meter after flow
calibration will now be described. This is the linearity contribution to the proving uncertainty, as
appearing in the third term on the right hand side of Eq. (5-12). It can be written as
u (qvprov
0 ,linearity ) u (K )
2
2
= . (A.9)
qvprov K
0
with a reference to Eq. (A.5). The term is related to the percentage difference, p, between flow rate from
the flow meter and the reference measurement, because of Eq. (A.4). The actual expression depends
on the adjustment method for the flow meter, and of any uncorrected percentage deviations, δp ,
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 75 of 79
between the flow meter and the reference meter. As discussed above, only a linear interpolation
correction method will be addressed here.
More specific, the correction is carried out using a linear interpolation in the percentage deviation
between the flow meter and the reference meter. The linear interpolation provides an approximate value
for the deviation from reference for flow rates between the ones used in the flow calibration. This is
illustrated in an example shown in Figure A.8-1, where a flow meter is flow calibrated at volume flow
rates at standard conditions of 500 m³/h and 2000 m³/h. The deviation from reference at 500 m³/h is in
this example 0.3 %. At 2000 m³/h it is 0.1 %. The blue curve represents the interpolated for volume flow
rates at standard conditions between 500 m³/h and 2000 m³/h. The correction of the meter is based on
this curve. However, such a linear interpolation is an approximation, and the exact shape of the deviation
curve is not known. In this work it is assumed that the true curve is somewhere inside the red
parallelogram. It is further assumed that the probability is the same for the curve to be anywhere inside
the parallelogram. The maximum (and unknown) uncorrected percentage deviation after correction is
therefore not larger than:
For flow rates outside the calibrated range, extrapolation is carried out for getting an estimate for the
uncorrected percentage deviation. In this case, the uncorrected percentage deviation increases as the
flow rate leaves the calibrated range, and is calculated as
qv 0,Meter − q v 0,Meter ,1
δp = − p 2 − p1 ;
qv 0,Meter , 2 − qv 0,Meter ,1
(A.12)
and
q v 0 ,Meter − q v 0 ,Meter ,n
δp = p n − p n −1 ;
q v 0 ,Meter ,n − q v 0 ,Meter ,n −1
(A.13)
The expression for δp is considered to be expanded uncertainty of p with 100 % confidence level and
rectangular distribution function. The standard uncertainty of p is then found by dividing δp with the
square root of 3. The relative standard uncertainty of the correction factor estimate can now be written
as
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 76 of 79
(
u q vprov
0 ,linearity ) = u (K ) = 1 ∂p
u( p) =
δp 3 .
(A.14)
q prov
v0 K K ∂K 100 + p
Figure A.8-1 Example of deviation from reference at flow calibration at a standard volume flow rate of 500
m³/h (here deviation of 0.3 %) and 2000 m³/h (here deviation of 0.1%). For the correction of the
flow meter, the deviation at flow rates between 500 m³/h and 2000 m³/h are found by linear
interpolation (blue curve). It is assumed that the “true” deviation curve is somewhere within
the red parallelogram.
Figure A.8-2 Relative standard uncertainty related to the correction factor for the example shown in Figure
A.8-1.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 77 of 79
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 78 of 79
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.
Handbook of uncertainty calculations for ultrasonic, turbine Ref. no.: CMR-17-A211576-RA-1
and Coriolis oil flow metering stations Rev.: 01 Date: 18.01.2018
Report Page 79 of 79
Conditions substituting “x”: Index “d” means densitometer conditions, index “m” means line (flow meter)
conditions, index “p” means proving device condition and index “c” means flow calibration conditions.
Superscript “met” means during normal metering, “prov” means during proving and “cal” means during
flow calibration.
© Christian Michelsen Research AS. This document is an original Christian Michelsen Research (CMR) document.