Professional Documents
Culture Documents
Wafer API 22 2 Paper For Mexico 2006 March 14 Rev Accepted
Wafer API 22 2 Paper For Mexico 2006 March 14 Rev Accepted
Abstract: The paper describes the testing of the Wafer V-Cone Meters in accordance with the new API 22.2
“Testing Protocol for Differential Pressure Flow Measurement Devices” in the Colorado Test Facility. The use
of this new API standard and some of the points which had to be addressed in order to implement it are
recorded. The results of testing 2” and 4” Wafer V-Cone meters in gas will be presented. The new non-
standard testing requirements in the API 22.2 will be discussed. The conclusions reached were: the results
verify the conditioning effect of the V-Cone as it meters the fluid; API 22.2 tests the claims of the
manufacturer of the meter in a more demanding manner than API 5.7. The results of the Wafer V-Cone
Testing Uncertainty are discussed.
Keywords: API 22.2 Test Protocol, Wafer V-Cone Meter.
1
6th ISFFM May 16-18, 2006
2
6th ISFFM May 16-18, 2006
3
6th ISFFM May 16-18, 2006
4
6th ISFFM May 16-18, 2006
5
4.2 Gas Flow Testing
Fig. 2 Air Test System
Gas testing was performed using compressed air in passing through the meter under test. The air was
a gas blow-down system, as shown in Fig 2 then released to the atmosphere.
Dry compressed air flows from the pressure vessel
through a flow control valve, which controlled the Pressure transducers, type T thermocouples, and a
mass flowrate of air through the test system. The air data acquisition system were used to measure and
then passed through the Wafer V-Cone being record system parameters. The mass flowrate of air
tested. A second control valve downstream of the was initially taken to the highest pipe Reynolds
Wafer V-Cone was used to maintain constant line number required for the tests. After stabilizing the
pressure at the meter. After the second control mass flowrate, the line pressure at the Wafer V-
valve, the air passed through the CEESI Master Cone was adjusted to the desired level. The data
Meter (a critical flow Venturi (CFV ))which was used acquisition system stored the average values at the
to accurately measure the mass flowrate of air end of the 30 system scans for each parameter.
5. Testing and Results and are shown on the plots as a dashed line. The
Mid Cd values produced under the above
Standard testing was performed to establish conditions were considered to be the baseline test
characteristic curves for each of the Wafer V-Cone results.
flowmeters. Calibrations maintained favorable ratios of
Wafer V-Cone meters are sold with a 10:1 turndown differential pressure to static line pressure (dP/P).
on flow, consequently the meters were tested over
that range. The 4” meters were tested over a pipe
Reynolds number range of 50,000 to 500,000 and Line Beta Pipe Reynolds
the 2” meter from 30,000 to 300,000. Size Number Range
The meter sizes and pipe Reynolds number ranges 0.45 50,000 to 500,000
are tabulated in Table 1.The tests were performed 4” 0.5 50,000 to 500,000
at a line pressure of approximately 87 psia. 0.65 50,000 to 500,000 *
2” 0.45 30,000 to 320,000
5.1 Analysis Method Applied to Wafer V-Cone Test
Results Table 1 Proposed Pipe Reynolds Number Ranges
for each Wafer V-Cone
Typically, to calculate flow for the Wafer V-Cone a
single discharge coefficient (Cd) value is used * On the baseline test for the 0.65 Beta insert in
through the entire flow range of the meter. The the 4” line size, the lowest Reynolds number which
single Cd value is the result of an analysis of could be tested was 93 962 instead of the planned
calibration data covering the desired flowrate range. 50 000. This is important in the determination of
The average Wafer V-Cone Cd is arrived at by Mid Cd, and is discussed further in the conclusions.
determining the appropriate upper and lower Cd
bounding values and then calculating the Mid Cd
value using the following simple equation.
C d max C d min
MidC d (1)
2
Mid Cd values were determined for each of the
Wafer V-Cone tests. The Mid Cd values are listed
for each test in Table 2.
The analysis of these test results has been based
on the difference in Mid Cd between any one test
and the baseline data for that Wafer V-Cone (Table
2 Column B). The percent difference uncertainty
(Table 2 Column C) is based on the uncertainty
analysis results. These values are explained
further in Section 6, Uncertainty Analysis.
The analysis included estimates of the standard 7. Acoustic Noise Testing (API Ch. 22.2 Section
uncertainties of: the measurement of mass flowrate; 4.8)
all secondary instrumentation; and the random
effects observed during testing. The uncertainty of This section is one of the revisions to the standard
the measurement of mass flowrate includes a and now reads “If excessive noise is encountered
detailed analysis of the primary and secondary during these tests, it shall be documented in the
standards used at CEESI. The actual uncertainties report”
for the secondary instrumentation were derived There was no noticeable noise from the V-Cone
from the manufacturer’s specifications and CEESI’s meters during these tests.
statistical process control program, utilizing historic
calibration data for the instrumentation. Random
effects were shown in the parent report (CEESI 8. Conclusions for the Testing of the Wafer V-
03RN-6506) to have minimal impact on the overall Cone Meters in Air
uncertainty of a data point.
8.1 Air tests were performed on 4 Wafer V-Cones -
The uncertainty associated with the measurement One 2 inch Wafer V-Cone with a beta of 0.45 and
of mass flowrate using a Critical Flow Venturi (CFV) three 4” Wafer V-Cones with betas of 0.45, 0.5, and
is approximately +/-0.35%. 0.65.
The performance of the Wafer V-Cone was 8.2 Testing was performed using compressed air on
introduced into the uncertainty analysis as the Cd all of the Wafer V-Cones at a line pressure of 87
Standard Uncertainty term. The Cd Standard psia to establish baseline performance. These tests
Uncertainty term was found by calculating the revealed that the characteristic curves of all of the
standard deviation of the individual data acquisition Wafer V-Cones were very similar. The similarity of
scans from the Wafer V-Cone characteristic curve. the characteristic curves indicates that the
The standard deviation was then multiplied by 2 to expansibility equation used with the Wafer V-Cone
produce the 95% confidence level. The 95% is correct.
confidence levels for each of the Wafer V-Cones
are shown in Table 2. 8.3 During the baseline test for the 4” line size with
the 0.65 Beta insert, the lowest Reynolds number
Table 2 contains three columns that are used in range tested was 93 962. At the time of the test, it
determining if an installation has an effect on the would have taken exceedingly long to cover the low
meter that is significant. If the Mid Cd value end of the range, so the decision was made to
changes due to an installation by more than the move on to other testing. Due to the shape of the
uncertainty associated with the calculation of the Cd Cd vs. Reynolds number curve, this has a
value, the installation had a statistically significant significant effect on the calculation of the Mid Cd.
effect on the meter. Column A represents the In Table 2, the results are first shown with the
installation effects test values over the entire 8.5 Noise measurements were made during all of
Reynolds number range from 50 000 to 500 000. the low pressure air testing performed on the four
This shows that there is a significant difference Wafer V-Cone meters. It was not possible to
between the Mid Cd values of each installation differentiate between the background noise in the
effects test from the baseline test. Due to the non- test area and the noise produced by the Wafer V-
linearity of the Cd vs. Reynolds number curve, the Cone meters.
proper way to compare Mid Cd values is to
compare the installation effects tests over the same 8.6 In conclusion the McCrometer Wafer V-Cone
range that the baseline test was run. These results, meter in these tests exhibited an exceptional ability
limiting the installation effects Reynolds number to operate effectively downstream of extreme flow
range from 90 000 to 500 000, are at shown as a disturbances. McCrometer will now advise their
separate section at the bottom of Table 2. customers of the Upstream and Downstream
straight pipe requirements between such
8.4 The conclusions for the half-moon disturbance disturbances and the Wafer V-Cone Meter to
tests are as follows: achieve the claimed uncertainty.
2” 0.45 Meter.
There is no statistical difference from the
baseline test when
Half Moon Plate 5D Upstream
Half Moon Plate 1D Downstream
Half Moon Plate 5D Upstream and 1D
Downstream
4” 0.45 Meter
There is no statistically difference from the
baseline test when
Half Moon Plate 5D Upstream
Half Moon Plate 2D Downstream
It is concluded that there would be no
statistical difference if there is
Half Moon Plate 5D Upstream and 1D
Downstream
4” 0.50 Meter
There is no statistically difference from the
baseline test when
Half Moon Plate 5D Upstream
Half Moon Plate 1D Downstream
Half Moon Plate 5D Upstream and 1D
Downstream
4” 0.65 Meter
Restricted Re range of 90,000 to 500,000
There is no statistically difference from the
baseline test when
Half Moon Plate 5D Upstream
Half Moon Plate 2D Downstream
It is concluded that there would be no
statistical difference if there is
Half Moon Plate 5D Upstream and 1D
Downstream
9. Discussion on the API Standard 22.2
“Differential Pressure Flow Measurement
Devices”
A. FIGURES
4. Baseline and Water Data for 2” 0.45 Beta Meter
5. Baseline and Water Data for 4” 0.45 Beta Meter
6. Baseline and Water Data for 4” 0.50 Beta Meter
7. Baseline and Water Data for 4” 0.65 Beta Meter
8. Installation Effect Data for 2” 0.45 Beta Meter
9. Installation Effect Data for 4” 0.45 Beta Meter
10. Installation Effect Data for 4” 0.50 Beta Meter
11. Installation Effect Data for 4” 0.65 Beta Meter
12. Uncertainty Data for 2” 0.45 Beta Meter
13. Uncertainty Data for 4” 0.45 Beta Meter
14. Uncertainty Data for 4” 0.50 Beta Meter
15. Uncertainty Data for 4” 0.65 Beta Meter
B. DETAILED DATA
1. Water Data for 2” 0.45 Beta Meter
2. Water Data for 4” 0.45 Beta Meter
3. Water Data for 4” 0.50 Beta Meter
4. Water Data for 4” 0.65 Beta Meter
0.900
0.895
0.890
D ischarge Coefficient
0.885
0.880
0.875
0.870
0.865
0 50000 100000 150000 200000 250000 300000
Reynolds Number
0.900
0.895
0.890
D ischarge Coefficient
0.885
0.880
0.875
0.870
0.865
0 100000 200000 300000 400000 500000 600000
Reynolds Number
0.895
0.890
D ischarge Coefficient
0.885
0.880
0.875
0.870
0.865
0 100000 200000 300000 400000 500000 600000
Reynolds Number
0.950
0.938
0.925
D ischarge Coefficient
0.913
0.900
0.888
0.875
0 100000 200000 300000 400000 500000 600000
Reynolds Number
0.895
0.890
Discharge Coefficient
0.885
0.880
0.875
0.870
0.865
0 50000 100000 150000 200000 250000 300000 350000
Reynolds Number
Baseline Data 3D Upstream 5D Upstream 5D and 1D 1D Downstream
1.00
0.98
0.96
D ischarge Coefficient
0.94
0.92
0.90
0.88
0.86
0 100000 200000 300000 400000 500000 600000
Reynolds Number
0.98
0.96
D ischarge Coefficient
0.94
0.92
0.90
0.88
0.86
0 100000 200000 300000 400000 500000 600000
Reynolds Number
1.00
0.98
0.96
D ischarge Coefficient
0.94
0.92
0.90
0.88
0.86
0 100000 200000 300000 400000 500000 600000
Reynolds Number
0.90
0.80
Uncertainty Values (95% C onfidence)
0.70
0.60
0.50
0.40
0.30
0.20
0.10
0.00
0 50000 100000 150000 200000 250000 300000 350000
Reynolds Number
1.00
0.90
0.80
Uncertainty Values (95% C onfidence)
0.70
0.60
0.50
0.40
0.30
0.20
0.10
0.00
0 100000 200000 300000 400000 500000 600000
Reynolds Number
0.90
0.80
Uncertainty Values (95% C onfidence)
0.70
0.60
0.50
0.40
0.30
0.20
0.10
0.00
0 100000 200000 300000 400000 500000 600000
Reynolds Number
0.60
0.50
Uncertainty Values (95% C onfidence)
0.40
0.30
0.20
0.10
0.00
0 100000 200000 300000 400000 500000 600000
Reynolds Number