Flow [m³/h] - NMi Euroloop Gas Calibration

Urban legends in legal
and industrial metrology
Jos van der Grinten
Chief Metrologist, NMi EuroLoop
1
Contents
• Urban legends
• Definitions and terminology
• Cases
• Calibration on air, application with natural gas
• Forward curve adjustment used for reverse application
• No recalibration required when using two meters in
series
• Calibration up to 70% extrapolation to 100%
• Z configurations
• 50 bar calibration for 60 bar operation
• How to deal with statements
• Conclusions and recommendations
2
Urban legends
Ethel Portnoy (1927–2004)
• 1978 - Broodje Aap.
De folklore van de postindustriële samenleving.
(Urban legends. The folklore of
the post industrial society.)
• Characteristics
•
•
•
•
•
•
Fictional, told as true
Sources are missing
No exact data, names, locations
Sensational or bizarre story
Latent fears or prejudices
Pseudoscientific nonsense
1979 Photo: Hans van Dijk - Anefo
3
Definitions and terminology – 1
• Step 1: Calibration  deviation ± uncertainty
reference value
GUM
deviation
Instrument value
Uncertainty (k=2)
(Guide to the expression of Uncertainty in Measurement, 1993)
Uncertainty
• the range of values that can reasonably be attributed to
the measurand
• a measure for the amount of missing information
4
Definitions and terminology – 2
• Calibration results on certificate
Flowrate [m³/h]
Deviation [%]
Uncertainty [%]
• Validity of certificate expires after
• After the end of the re-calibration period
• Accident with the meter
• Modifications made to the meter
(parameter list part of the certificate)
5
Definitions and terminology – 3
• Step 2: conformity assessment  compliance with
regulations, e.g. tolerances (≥ 95% confidence)

?
tol–

?

tol+
• Test / verification: conforming or not
• Inspection: non-conforming or not
6
Definitions and terminology – 4
Gas
meters
MID, annex IV Normative document
TSO
OIML R 137
contracts
(MI – 002)
Class
1.5
1.0
1.5
1.0
0.5
Qmin ≤
Q < Qt
3%
*
2%
*
3%
6%
2%
4%
1%
2%
–
–
Qt ≤ Q
≤ Qmax
1.5%
*
1%
*
1.5%
3%
1%
2%
0.5%
1%
0.3%
0.3%
WME
–
–
–
–
0.2%
–
0.3%
0.3%
0.6% 0.4%
–
–
• In-service requirements  national law
• TSO: high-volume  financial risk  high accuracy
7
Calibrate on air, use on natural gas?
year < 1991
0.8%
8
Calibrate on air, use on natural gas?
year < 1991
0.8%
9
Calibrate on air, use on natural gas?
Meter T
0.5
Lab A 41 barg As Found
Lab A 41 barg polyn. powers -2, -0.2, -0.33, 0
Lab B atm. air As Found
Lab B atm. air polyn. powers -2, -0.2, -0.33, 0
year 2016
error [%]
0.0
-0.5
0.9%
-1.0
-1.5
0
1000
2000
3000
4000
5000
6000
Flow [m³/h]
10
Calibrate forward – Adjust reverse
Claim
“The adjustment and verification of a meter carried out for
one direction is sufficient and valid for a usage in both flow
directions, if the meter run including input and output pipes
is carried out as described in the type examination
certificate for a bidirectional usage.
In this case the meter provides results which fulfill the MPE
independent from the flow direction.”
• Claim is for USM of a specific brand
• No restrictions to adjustment method
11
0.50
0.40
0.30
0.20
0.10
0.00
-0.10
-0.20
-0.30
-0.40
-0.50
Forward, adj.fact = 0.9982
Reverse, adj.fact. = 1.0062
Error [%]
Error [%]
Calibrate forward – Adjust reverse
year 2008
0
5000
10000
15000
0.50
0.40
0.30
0.20
0.10
0.00
-0.10
-0.20
-0.30
-0.40
-0.50
Forward, adj.fact = 0.9988
Reverse, adj.fact. = 0.9971
year 2013
0
20000
5000
20000
0.70
Forward, adj.fact = 1.0000
0.50
Error [%]
• 0.80% difference
• 0.17% difference
• < 0.05% difference !
15000
Flowrate [m³/h]
Flowrate [m³/h]
• USMs as found
• Adjust factors
10000
Reverse, adj.fact. = 1.0000
0.30
0.10
-0.10
year 2016
-0.30
0
2000
4000
6000
Flowrate [m³/h]
12
8000
10000
Calibrate forward – Adjust reverse
• Newer meters perform better
• Method only demonstrated for linear adjustments
• Reckon with at least 0.1% difference between
forward and reverse
• No check for polynomial adjustments
13
In series – no recalibration?
• In Situ comparison showed 0,4% difference
10D
5D
Slave
Manufacturer A
5D
10D
Master
Manufacturer B
• In reality average shift = 1% !
14
Extrapolation
• What to do if you cannot reach Qmax?
• Take volume 30 of PTB
Testing Instructions, p 46
• Calibrate up to 70% Qmax
• Low pressure gas or air curve
in full range must be available
• Shape of the curve between
0.2·Qmax and highest Q
• Extrapolate in Reynolds
domain until 100% Qmax
• Intended for turbine meters, also used for USMs
15
Extrapolation
• Result of a recalibration
Cal.
Lab
• Often
• No low-pressure or air curve available (e.g. recalibration)
• Base extrapolations on smooth curves
• Higher order polynomials will be unstable outside the
calibrated range
16
Extrapolation
• Turbinemeter, 42 bar, natural gas
0.00
As Found
As Left Adjusted
Fit
Error [%]
-0.50
-1.00
-1.50
0
1000
2000
3000
4000
5000
6000
7000
Flow [m3/h ]
17
Extrapolation
Meter curve
1.00
24” G16000 Turbine gasmeter
0.90
0.80
deviation e (%)
0.70
0.60
0.50
0.40
Calibration data
Straatsma
0.30
3rd order polynomial
4th order polynomial
0.20
WME
0.10
0.00
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
Non-dimensional flowrate Q/Qmax
18
0.90
1.00
Z configurations
• Parallel use
• Serial comparison
③
②
①
①
②
1
③
2
• Points of attention
• Flow disturbances and flow conditioning ①
• Noise ② (100 dB(A))
• Pulsations caused by closed side branches ③
19
Z configurations
• Separate calibration of two meters
1
• 0.5% difference with or without side branch
•  Recommendation: test the entire configuration
• Example
• Hilko den Hollander, EFMWS 2015, Noordwijk
• Calibration of a complete Z skid and its functionality
20
Pressure dependency – 1
Claim
• Calibrations @ 50 bar are cheaper than @ 60 bar
• So even when operating conditions are 60 bar we
order a 50 bar calibration
• There will be not that much difference ?
• Pressure dependency
21
Pressure dependency – 2
Meter U
1.0
Lab A 48 barg As Left
Lab A 60 barg As left
error [%]
0.5
0.0
+ 0.01% / bar
-0.5
0
5000
10000
15000
20000
25000
Flow [m³/h]
22
Pressure dependency – 3
0.50
USM (clean without corrections)
0.40
0.30
- 0.01% / bar
Deviation [%]
0.20
0.10
0.00
-0.10
-0.20
-0.30
-0.40
61 bara as found
41 bara as found
31 bara uncorrected
41 bara Chord B failure
77.2 bara initial 5 years ago
-0.50
0.0
5.0
10.0
15.0
20.0
Flow velocity [m/s]
25.0
30.0
23
35.0
How to deal with statements?
• No recalibration necessary?
• Only for meters on the scrap yard
• Change or damage to the meter  recalibrate
• End of recalibration period  recalibrate
• Innovative declarations issued by notified bodies
• Made on request of manufacturers
• So ask the manufacturer for
•
•
•
•
Magnitude of the change  additional uncertainty
Systematic proof, not just one case
Publications in the open literature
Results of studies and tests
preferably conducted by an independent third party
24
Conclusions and recommendations
• Metrological principle
• Change anything on the meter and the calibration
certificate will become void
• Legal metrology
• Some modifications are possible without jeopardizing
the performance with respect to legal tolerances
• Use acceptance criteria smaller than legal tolerances
• However legal metrology tolerances are too wide for
industrial applications (TSOs)
• Systematic testing and reporting required
25
Conclusions and recommendations
Cases
• Air calibration
Δe
0.9%
• High-pressure air, not documented yet
• How to scale the meter curve? Meter principle?
• Forward calibration – reverse adjustment
> 0.1%
• Newer meters perform better
• No recalibration with two meters in series
~ 1%
• Extrapolation
> 0.1%
• Z configuration
> 0.5%
• Pressure dependency
± 0.01% / bar
26
Conclusions and recommendations
Future
• Well documented test results  Transparency
• Transparency  Confidence
• Confidence  Trust
27
Questions?
Jos van der Grinten
NMi EuroLoop
[email protected]
28