’LNG measurement uncertainty, review and progress

LNG measurement uncertainty,

review and progressKianoosh Hadidi

LNG workshop, Aberdeen

25th October 2018

Workshop and Training 2018

The content of this presentation has been taken from the work carriedout by the partners in EMPR JRP, ENG03 LNG & ENG60 LNG II.

The related sources have been mentioned at the footnotes.

Introduction

Start ENG03 2009: Real estimation of measurement uncertainty of LNG energy,

1% Equivalent economic value of a reduction of 0.5 % in the

energy uncertainty was predeicted to be 150 M€/year

𝐸 = 𝑉𝐿𝑁𝐺 ∙ 𝐷𝐿𝑁𝐺 ∙ 𝐺𝐶𝑉𝐿𝑁𝐺 + 𝐸𝐺𝐴𝑆 𝐷𝐼𝑆𝑃𝐿𝐴𝐶𝐸𝐷 ± 𝐸𝐺𝐴𝑆 𝑡𝑜 𝐸𝑅

These values can be eitheragreed on a certain value or determind, negligablecontribution in many cases

The main contribution to the energy uncertainty

MassLNG

The energy content in the tranfered LNG

LNG performance: Methane Number, an indication of the knocking behavier of a LNG

These two quantities have direct economic impact on LNG trade

Introduction

Traceability

Development of calibration standards

Produce Reference data

Uncertainty evaluation

Measurement function

Clear estimation of the uncertainty sources

Uncertainty budget

Modification and developelment

New measurement methods/devices

Calculation methods

Uncertaintyreduction

Content

Mass flow measurement

Small scale mass flow standard

Volumetric flow measurement

Uncertainty evaluation of tank gauging systems

(LDV) based standard, and mid scale flowmeter calibration standard

Density measurement/calculations

State-of-the-art primary density standard

Development of a speed of sound (SOS) sensor

Development of a new fandamental equation of state

Gross calorific value, model calculation and uncertainty estimation

Energy uncertainty budget in large scales

The effect of small composition variation on the uncertainty of the LNG density and calorific value

The effect of temperature changes on the uncertainty of the LNG density

Methane number

Mass flow measurementSmall scale mass flow standard

A primary LNG mass flow standard at

Rotterdam for small scale test and

calibration facility based on weighing

method

Start point in development traceability

Measurement model of the reference mass

𝜑𝑀 𝑢𝑇 =𝑚𝑀𝑢𝑇 −𝑚𝑟𝑒𝑓

𝑚𝑟𝑒𝑓× 100

𝑚𝑟𝑒𝑓 = ∆𝑚𝑠𝑐𝑎𝑙𝑒 +𝑚𝑣𝑎𝑝𝑜𝑢𝑟 + 𝐶𝑖𝑛𝑐𝑙𝑖𝑛𝑎𝑡𝑖𝑜𝑛 + 𝐶𝑙𝑖𝑛𝑒𝑝𝑎𝑐𝑘

𝑚𝑣𝑎𝑝𝑜𝑢𝑟= total vapour mass flowing out of the weighing tank during calibration time window

∆𝑚𝑠𝑐𝑎𝑙𝑒 = accumulated cryogenic liquid in the weighing tank during the test time window

𝐶𝑙𝑖𝑛𝑒𝑝𝑎𝑐𝑘= correction due to change in trapped liquid mass between MuT and weighing tank,

𝐶𝑖𝑛𝑐𝑙𝑖𝑛𝑎𝑡𝑖𝑜𝑛= correction due to non-constant inclination of the calance

Results of the evaluation and preliminary validation of a primary LNG mass flow standard Metrologia 51 (2014) 539–551 doi:10.1088/0026-1394/51/5/539

Mass flow measurementSmall scale mass flow standard

The targeted uncertainty of the standard, 0.05%

The reached value of uncertainty 0.12% - 0.15%,

The main contributions to the irreproducibility are related to nonreversible parasitic forces in the weighing system

Potential improvement to reduce the CMC down to 0.1%

Results of the evaluation and preliminary validation of a primary LNG mass flow standard Metrologia51 (2014) 539–551 doi:10.1088/0026-1394/51/5/539

Volumetric flow measurement

Uncertainty evaluation for tank gauging systems

Ship tank volumetric measurement model.

Applicable for Membrane tank and a Moss tank

This work was fully in accordance with GUM and included real shipment data

Model measurement function

𝑉 = 𝑉𝑡𝑎𝑏𝑙𝑒 + 𝑐𝑉 𝐶𝑡𝑎𝑛𝑘,𝑡 𝑇 𝐶𝑡𝑎𝑛𝑘,𝑝 𝑝∆𝑉𝑇𝐴𝐵𝐿𝐸

∆ℎℎ − 𝑇𝑟𝑢𝑛𝑐(ℎ)

h = ℎ𝑖𝑛𝑑 𝐶𝑔𝑎𝑢𝑔𝑒,𝑇 𝑇𝑔𝑎𝑢𝑔𝑒 𝐶𝑔𝑎𝑢𝑔𝑒,𝑝 𝑝𝑔𝑎𝑢𝑔𝑒 + ∆ℎ𝑡𝑟𝑖𝑚 + ∆ℎ𝑙𝑖𝑠𝑡 + ∆ℎ𝜌 +

∆ℎ𝑐𝑜𝑚𝑝 + ∆ℎ𝑐𝑎𝑙 + ∆ℎ𝑑𝑟𝑖𝑓𝑡

Trim expressed in metres or fractions of a metre, according to the difference in bow and stern drafts

List represented by the angle α in degrees to port. In this illustrative case, the correction will be negative

Evaluation uncertainty in transferred LNG, https://lngmetrology.info/publications/project-reports/

Moss type Membrane type

Volumetric flow measurementUncertainty evaluation for tank gauging systems

An overview of relevant input quantities is given. Red color indicates that they may significantly influence the measurement of transferred volume.

Tank

Calibration

Drift/stabilit

Resolution

Temp. dim. structure

Hydrostaticpressure

Sagging/ Hogging

Temp. sensors

Inclinometer

Calibration

Drift/stabilit

Resolution

Disper. in readings

Sagging/ Hogging

Pressure gauge

CalibrationDrift

Float level gauge

Tape temp.

Liquid densityBoyancy

Calibration Location

Drift

Disper. in readings

Calibration

Drift/stabilit

Disper. in readings

Radar level gauge

Calibration

Mount. position

Temp.

Drift

Disper. readings

Surface detection

Measurand ValueUncertaint

y

Distributio

n

Standard

uncertainty

Rel.

Uncertaint

y

SensitivityContributio

n

hind,stop [m] 4,000 0,0075 normal 0,00 0,09 % 1 0,0038

Dhcal [m] 0,000 0,002 normal 0,00 NA 1 0,0010

Dhdrift [m] 0,000 0,01 normal 0,01 NA 1 0,0050

Dhtrim,stop[m] -0,009 0,000 rectangular 0,00 0,10 % 1 0,0000

Trim [m] 0,033 0,01 rectangular 0,01 15,38 % -0,038 -0,0002cTrim,loc,cal [m] 0,200 0,1 normal 0,05 25,00 % -0,038 -0,0019

Dhlist,stop[m] -0,013 0,002 rectangular 0,00 -7,50 % 1 0,0010

List [°] 0,019 0,01 rectangular 0,01 26,32 % -0,007 0,0000cList,loc,cal [m] 2,000 0,5 standard 0,50 25,00 % -0,007 -0,0033

Ttank,start [°C] -150,000 5 rectangular 2,50 -1,67 % 4,00E-06 0,0000

Tref,tank [°C] -160,000 2 standard 2,00 -1,25 % -4,00E-06 0,0000

Tgauge,start [°C] -130,000 5 standard 5,00 -3,85 % -4,00E-06 0,0000

Tref,gauge [°C] 20,000 0,5 standard 0,50 2,50 % 4,00E-06 0,0000

a 1,000E-

06standard

0,000,00 % 40 0,0000

b 1,000E-

06standard

0,000,00 % -600 0,0000

hstop 3,978 uh,empty 0,0075

Uh,empty 0,0149

Uh,empty* 0,38 %


Volumetric flow measurementUncertainty evaluation of tank gauging systems

Uncertainty budget for Moss type tank

Volumetric measurementUncertainty evaluation of tank gauging systems

• 𝐔𝑽=0.21%, GIIGL third edition 2010

• 𝐔𝑽=0.20 % to 0.55 % (k = 2) GIIGL fifth edition 2017

Measurand Value Uncertainty DistributionStandard

uncertainty

Rel. st.

uncertaintySensitivity

Contributi

on

Vtable (Trunc(hstart)) 34000 70,00 normal 35,000 0,10 % 1 35,00

hstart 22,84943 0,03774 standard 0,03774 0,17 % -1273,062204 -48,05

DVSaggingHogging,start 0,000 70,000 rectangular 35,000 NA 1 35,00

DVHydrostatic,start 0,000 70,000 rectangular 35,000 NA 1 35,00

DVTable,drift,start 0,000 0,000 rectangular 0,000 NA 1 0,00

rectangular 0,000 NA 0,00

Vtable (Trunc(hstop)) 1600 3,20 rectangular 1,600 0,10 % 1 1,60

hstop 0,15000 0,00786 standard 0,00786 5,24 % -19,05461974 -0,15

DVSaggingHogging,stop 0,000 3,500 rectangular 1,750 NA 1 1,75

DVHydrostatic,stop 0,000 3,500 rectangular 1,750 NA 1 1,75

DVTable,drift,stop 0,000 0,800 rectangular 0,400 NA 1 0,4

Ttank,start (°C) 20 20,0 rectangular 10,000 50,00 % 1,1E+00 11,22

Ttank.stop (°C) 20 20,0 rectangular 10,00 50,00 % 5,3E-02 0,53

Ttank,ref (°C) 20,00 2 rectangular 1,00 5,00 % -1,1E+00 -1,12

a 1,10E-05 1,1E-06 rectangular 0,00 5,00 % 0 0,0

Vtank unloaded 33242,13 uVloaded 87.97

UV,loaded 175.95

UV,loaded* 0.53%


Uncertainty budget for membrane type tank

• Larger than twice the uncertainty mentioned in third edition of GIIGNL 2010

• Better agreement in GIIGNL 2017

Volumetric flow measurementLDV based standard

Uncertainty budget flow rate measurements in LNG by Laser Doppler Velocimetry https://lngmetrology.info/publications/project-reports/

A new Laser Doppler Velocimetry (LDV) based standard

Mid-scale flow meter calibration has been built at Rotherdam

It is presently under testing

The master flow meter calibrated based on weighing method in the first LNG project

When operational this facility willenable treacible calibration with an established uncertainty

Developed in ENG03 LNG and ENG60 LNG II

Further modification in LNG III to be validated in croyogrnic canditions and defined as primary reerence standard for LNGstandard

LDV standard can be a great support to cross valide this novel and traceable calibration standards facility

Measurement function:

Uncertainty budget flow rate measurements in, LNG by Laser Doppler Velocimetry https://lngmetrology.info/publications/project-reports/

Volumetric flow measurementLDV based standard

𝑄𝑣 =𝑣𝑎𝑥𝑖𝑠 ∙ 𝜋 ∙ 𝑑

2

4(𝑎 + 𝑏 ∙ 𝑙𝑛(𝜌 ∙ 𝑣𝑎𝑥𝑖𝑠∙ 𝑑

𝜇 + 𝜖)

𝑄𝑣 Volumetric flowrate obtained from the LDV system

𝑣𝑎𝑥𝑖𝑠 Measured axial velocity using the LDV system

d Internal diameter of the LDV convergent throat

𝑎 Intercept of the model function

b Slope of the model function

𝜌 Density of LNG at local conditions of pressure and temperature

𝜇 Viscosity of LNG at local conditions of pressure and

𝜖 Model function error

A full uncertainty budget was defined for LNG measurements. Field data was not acceible to apply. Using random data from liquid nitrogen measurement at laboratory conditions shows a relative uncertainty of 0.63%.

A new Laser Doppler Velocimetry (LDV) based standard

Density measurementState-of-the-art primary density standard

Results of the LNG density measurements including uncertainty, https://lngmetrology.info/publications/project-reports/

Density measurement Single-Sinker Densimeter for cryogenic liquid mixtures

A special single-sinker densimeter

T-range: 90 K to 300 K

p-range: 0.05 MPa to 12 MPa

Density measurementState-of-the-art primary density standard

Source of uncertainty Expanded uncedrtainty Distribution Standard uncertainty

Density measurement 0.0080% Normal 0.0040% 𝑢 𝜌

Temprature measurement 0.0100%· p max Rectangular 0.0001% 𝑢 𝑇

Pressure measurement 15 mK Rectangular 0.0030% 𝑢 𝑝

Composition of gas mixture 0.0105% Rectangular 0.0060% 𝜌 𝑥

Reproducibility of the measurements

0.0200% Normal 0.0100% 𝑢 𝜌𝑟𝑒𝑝𝑟𝑜

Density correction (FTE) 0.0350% Normal 0.0175% 𝑢 𝜌𝑐𝑜𝑟𝑟

Relative combined expanded uncertainty in density (k = 2 ): 0.0440% FTE, the largest source of unceratinty. Reproducibility, Gas analysis

On three different LNG mixtures Along isotherms 115, 125 and 135 K Pressure range up to to 8.6 MPa

Uncertainty Improvement:Reporetd in 2014:0.06-0.08%Reported in 2015:0.044%Reported in 2017: 0.02%

Results of the LNG density measurements including uncertainty, https://lngmetrology.info/publications/project-reports/

Uncertainty budget

Speed of sound measurements in liquid Methane at cryogenic temperature and for pressure up to 10 MPa, https://lngmetrology.info/publications/presentations/

Measuring techniqueDouble Pulse–Ecko

𝜌 𝑝0, 𝑇 = 𝑑𝑒𝑛𝑠𝑖𝑡𝑦𝑐𝑝 = 𝑖𝑠𝑜𝑏𝑎𝑟𝑖𝑐 𝑠𝑝𝑒𝑠𝑖𝑓𝑖𝑐 ℎ𝑒𝑎𝑡 𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦

𝛼𝑝 = 𝑡ℎ𝑒𝑟𝑚𝑎𝑙 𝑒𝑥𝑝𝑎𝑛𝑠𝑖𝑜𝑛 𝑐𝑜𝑎𝑓.

SOS can be implemented to develop EOS

Density measurementDevelopment of a speed of sound (SOS) sensor

Speed Of Sound sensor (SOS)

Ultrasonic densimeter for simultaneous measurement of online speed-of-sound (SoS) and density in cryogenic flowing liquids

Suitable to be mounted on existing production facilities and monitor on-line density

In particular, the SoS device will be useful to calibrate the commercial on-line ultrasonic flowmeters.

Tested with liquified methane, T = 103-153 K and p = 10 bar

LNG measurement uncertaintyDensity measurement

Speed of sound measurements in liquid Methane at cryogenic temperature and for pressure up to 10 MPa, https://lngmetrology.info/publications/presentations/

Speed Of Sound sensor (SOS), uncertainty budget

Estimated Overall Uncertainty (k=2) ~0.216%

Density calculationsDevelopment of a new fandamental equation of state

Development of a new fundamental Equation of State to calculate saturation densities, based on experiments with a primary cryogenic densimeter

Enhanced revised Klosek and McKinley ERKM versus revised Klosek McKinley method (RKM)

This includes an additional pressure-dependent term

Thus the developed version, ERKM, can be used at temperatures higher than 115 K: 100𝐾 ≤𝑇≤135 𝐾

𝑝s ≤𝑝 ≤10 𝑀𝑃𝑎

The estimated uncertainty: 0.1 % for 100𝐾 ≤𝑇≤115 𝐾

0.15 % for 115𝐾 ≤𝑇≤135 𝐾,

Equation Of States, EOS

Report on the development of the selected EOS for calculation of the saturated liquid density of LNG, https://lngmetrology.info/publications/project-reports/

Gross calorific valueModel calculation and uncertainty estimationStoichiometric combustion equation of a hydrocarbon,

𝐶𝑛𝐻𝑚 + 𝑛 +𝑚

4𝑂2 → 𝑛𝐶𝑂2 +

𝑚

2𝐻2𝑂

Calculating effective stoichiometric coefficients to applyone reaction equation for LNGmix instead of all LNGmix

elements’equations, gives: 1.0000 ± 0.0008 ∙ 𝐿𝑁𝐺 + 2.1792 ± 0.0016 . 𝑂2 → 1.1215 ± 0.0008 ∙𝐶𝑂2 + (2.1155 ± 0.0016) ∙ 𝐻2𝑂

𝐻𝑠 = ∆𝐶𝑂2𝐻 + ∆𝐻2𝑂𝐻 − ∆𝐿𝑁𝐺𝐻 − ∆𝑂2𝐻 − ∆𝑁2𝐻

In ambient conditions ∆𝐿𝑁𝐺𝐻 = ∆𝐿𝑁𝐺𝐻0

∆𝑁2𝐻: correction term, non-combustible component but includes the evaporation heat of the dissolved nitrogen. Thus changes the total enthalpy of the LNG mixture.

Calculation of LNG enthalpies and calorific values, https://lngmetrology.info/publications/presentations

Difference in standard formation energies ∆𝑋𝐻0

In case of LNG𝐻𝑠 = ∆𝐶𝑂2𝐻 + ∆𝐻2𝑂𝐻 − ∆𝐿𝑁𝐺𝐻 − ∆𝑂2𝐻 − ∆𝑁2𝐻

∆𝑚𝑖𝑥𝐻 = 𝑎 ∙ (∆𝑚𝑖𝑥ℎ 𝑇, 𝑝 + ∆𝑚𝑖𝑥𝐻0)

∆𝑚𝑖𝑥𝐻0 : LNG standard enthalpy of formation

∆𝑚𝑖𝑥ℎ 𝑇, 𝑝 ∶ temperature and pressure dependent change of the LNG enthalpy; calculated with the GERG-2004 equations. It represents the additional internal energy gained during the NG phase transition from gas to the liquid phase, LNG.

Calculation of LNG enthalpies and calorific values, https://lngmetrology.info/publications/presentations/

Gross calorific valueModel calculation and uncertainty estimation


Calculation of LNG enthalpies and calorific, https://lngmetrology.info/publications/presentations/

The enthalpies and caloprific values of five chosen LNG examples were calculated at temperatures from -80°C down to the actual liquid state using the GERG2004

Expansion factor K=2


Calorific value, main (relative) uncertainty contributions in calculation (in %)



Uncertainties (k=2) in the liquid state:

Enthalpies: 1.0 – 1.3 %

Calorific values: 0.14 – 0.22 %

Main uncertainty contributions:

Enthalpy of formation of the pure components

Composition analysis of the LNG

Enthalpy difference approx. 900 J/g between standard conditions and liquid state

Neglecting the LNG cooling potential:

Effective loss of energy (by the regassification process)

Up to 500 k€ per tank ship


Energy uncertainty budget

𝐸 = 𝑉𝐿𝑁𝐺 ∙ 𝐷𝐿𝑁𝐺 ∙ 𝐺𝐶𝑉𝐿𝑁𝐺 + 𝐸𝐺𝐴𝑆 𝐷𝐼𝑆𝑃𝐿𝐴𝐶𝐸𝐷 ± 𝐸𝐺𝐴𝑆 𝑡𝑜 𝐸𝑅

The effect of composition variation on the uncertainty of the LNG density and calorific value

Coloured cells highlight the modified

composition. %component is the percentage

of the component in LNG

Ucomponent is the combined uncertainty on the

LNG component

SENSITIVITY STUDY OF LNG ENERGY TRANSFER UNCERTAINTY FROM COMPOSITION AND TEMPERATURE CHANGES, https://lngmetrology.info/publications /reports

The method used here is to vary the proportions of

individual components to determine the effects of each

on the overall mixture GCV and density of a specific

LNG example

Variation in composition has more influence on density


The uncertainty budget for the density using temperature sensitivity values determined by calculating the effect of a small change in temperature (e.g. 0.5oC) on the density (i.e., ∂ρ/∂T = ∆ρ/(T+-T-)).



Comparing the different methods gives an insignificant difference in the density uncertainty of ±0.014%.


Methane number

A new algorithm, based on existing algorithms such as the AVL and MWM, has been developed

The new algorithm pays special attention to the influence of heavier hydrocarbons as the concentration of these components in stored LNG increases over time.

None of the other algorithms provide information about the uncertainties in their results

Resulting uncertainty strongly depends on LNG composition

Uncertainty values vary between 0.2 and 0.7 MN or 0.3 and 0.8% (k=2)

Novel algorithm shows a good agreement with other popular methods for the set of exemplar LNG mixtures


Achivements towards uncertainty reduction

Development of a primary LNG mass flow standard (25 m3/h) with a CMC of 0.12 - 0.15%

Establishment of a real uncertainty budget for (state of the art) ship-based tank-gauging methods by performing a comprehensive metrological study

Standard based LDV device and associated uncertainty to support the validationof the mid-scal flow calibration facility

Development and validated an advanced primary LNG densimoter system (single-sinker) to produce new reference data with a target uncertainty of 0.02%. The achieved reported uncertainty in 2016, 0.02%.

Development and validation of new fundamental EOS (ERKM ) for LNG density prediction for a broader range of temperature (115𝐾 ≤ 𝑇 ≤ 135 𝐾) and the associated uncertainty. A prototype speed of sound sensor has been developed for on-line density and speed of sound measurements to transfer the metrological traceability to field operation.

A new algorithm for methane number, based on existing algorithms such as AVLand MWM, has been developed, associated uncertainty between 0.3 to 0.8% (K=2).

Thank you for your attention

Documents

’LNG measurement uncertainty, review and progress