Gas Turbine Meters: Compare K-Factor Curves at Multiple Pressures
Abstract
A turbine gas meter's K-factor is often summarized as a single number, but the underlying test is a curve across flow. Yang and colleagues reported curves for DN50 and DN80 meters using a negative-gauge-pressure setup and a high-pressure closed loop. This article compares point-by-point values with the source's cross-pressure statistics. Pressure and flow both matter to the comparison. The reported summary values were 0.502% for DN50 and 0.282% for DN80, while the curves give additional information near minimum flow. A printed pressure-unit discrepancy is documented so readers can trace the pressure labels to the methods and figure legend.
Background
A turbine meter translates the motion of an impeller into a count or pulse rate. The source describes the housing, inlet and outlet guides, rotor, cartridge and counting mechanism. How the measured output changes with flow matters to anyone who must translate pulses into transported gas volume. The paper's generic characteristic curve establishes the expected distinction between the low-flow end and the usable linear interval. The model and generic curve in its opening figures frame the later calibration results.
Figure 1. English technical redraw of the source meter section, source Figure 1, p. 46. Geometry is schematic.
Figure 2. English redraw of source Figure 2, p. 47. The response is conceptual and is separate from the measured curves.
Research question
When the calibration pressure changes, how much does a K-factor curve move across the tested flow points, and what can a single cross-pressure statistic show? The comparison uses the TBQM-G65-DN50 and TBQM-G100-DN80 specimens named in the source. It is useful only if the reader keeps track of which meter, pressure and flow point produced each value.
Experimental method
The study defines K = N/V, where N is the pulse count and V is the reference volume in cubic metres. Its Equation (2) relates observed pulse frequency f to volumetric flow as qv = f/K. The study first collected negative-gauge-pressure results on a Dresser gas metering rig. For tests at 0.1, 0.8 and 2.0 MPa it used a separate high-pressure closed-loop standard facility. Section 2.1 reports facility uncertainty figures of 0.25% and 0.33%, respectively. These are facility characteristics stated by the authors. They do not quantify all uncertainty for a transferred result.
Figure 3. English editorial reconstruction from the negative-gauge-pressure rig photograph, source Figure 3, p. 47. The illustration represents visible components.
Figure 4. English editorial reconstruction from the closed-loop rig photograph, source Figure 4, p. 48. The reported layout is represented schematically.
The two meters were tested at seven flow points each. The sample flows ranged from 6.5 to 100 m³/h for DN50 and 8 to 160 m³/h for DN80. The paper states three trials per flow point. To summarize a given pressure, its Equation (5) takes the midpoint between the highest and lowest point K-values over qt to qmax. That is a defined estimator over a subset of flow conditions. Its Equation (6) then expresses the spread of four such pressure-specific factors relative to their mean. Keeping these calculations separate helps interpret the reported result.
Results and discussion
The DN50 curve at the lowest tested flow shows a markedly different point K-factor from the point at maximum flow. The four pressure curves are closer over much of the operating range, although the relative position of traces changes with flow. The DN80 result also preserves a visible difference near low flow. These observations come directly from the curves based on Tables 1 and 2.
Figure 5. English redraw from Table 1 and source Figure 5, p. 49. Pressure series are negative gauge, 0.1 MPa, 0.8 MPa and 2.0 MPa.
Figure 6. English redraw from Table 2 and source Figure 6, p. 49. Pressure labels follow the figure legend and the stated test method. See the printed-unit note below.
Specimen | Test-flow range (m³/h) | Cross-pressure statistic, Eq. (6) | Largest difference reported for the stated linearity comparison |
TBQM-G65-DN50 | 6.5–100 | 0.502% | 0.558% |
TBQM-G100-DN80 | 8–160 | 0.282% | 0.904% |
The last two columns describe different calculations in the paper. The first statistic summarizes variability of the four factor estimates. The last column is the author's comparison of linearity over the specified range. Neither is a guarantee of error at a particular low-flow point. Where a project spends long periods near minimum flow, the point results are especially important. A technically meaningful comparison also needs to say whether pressure is reported as gauge or absolute and whether the volume basis matches the calibration report.
A pressure-series curve should remain linked to its individual flow values when exchanged as a data file. Rounding the factor before plotting can hide a small difference between two pressure runs. The table below carries the source point values in reciprocal cubic metres. It also helps a reviewer trace each plotted marker back to the correct meter and condition. The two sizes have different pulse factors and flow ranges, so compare relative variation within each specimen before comparing how similar their curves appear.
Specimen | Flow (m³/h) | Negative gauge | 0.1 MPa | 0.8 MPa | 2.0 MPa |
DN50 | 6.5 | 103719.18 | 103906.588 | 104110.544 | 105009.252 |
DN50 | 10 | 105157.767 | 105213.566 | 104715.614 | 105618.098 |
DN50 | 20 | 104962.476 | 104890.039 | 105721.677 | 105890.917 |
DN50 | 25 | 104842.901 | 104684.982 | 105976.907 | 105902.294 |
DN50 | 40 | 104724.477 | 104560.637 | 105904.585 | 105743.981 |
DN50 | 70 | 104764.732 | 104708.593 | 105690.789 | 105863.374 |
DN50 | 100 | 105208.094 | 104987.463 | 105386.821 | 105988.341 |
DN80 | 8 | 25939.893 | 25480.984 | 25470.243 | 25533.989 |
DN80 | 16 | 26087.176 | 25629.112 | 25559.371 | 25635.501 |
DN80 | 32 | 25935.767 | 25759.038 | 25647.180 | 25627.137 |
DN80 | 40 | 25885.873 | 25741.826 | 25653.668 | 25584.684 |
DN80 | 64 | 25778.352 | 25684.734 | 25627.529 | 25573.169 |
DN80 | 112 | 25698.554 | 25681.281 | 25643.844 | 25637.744 |
DN80 | 160 | 25682.651 | 25678.001 | 25637.166 | 25730.677 |
Conclusion
The reported aggregate values provide one summary of pressure sensitivity for two samples. Use the full K-factor curves for purchasing decisions involving different flows and pressures. In an inquiry, supply expected gas, pressure basis, operating flow range and volume convention. Request measured K-factors at corresponding conditions, the calibration method and its uncertainty.
Limitations and outlook
The reported study uses two specimens and specific equipment. It does not establish a universal behaviour for all sizes or other installations. Verification for another process should compare its actual operating conditions against the evidence supplied for the offered meter. The standards discussion in the paper is historical context for that study.
Notes
Source Table 2 headers print 0.8 kPa and 2.0 kPa, while Section 2.1, Table 4 and Figure 6 use MPa. This redraw uses MPa for the pressure series and keeps the point values exactly as Table 2 lists them. All K-factors are in reciprocal cubic metres. Reported percentages retain the paper's own definitions.
References and Image Sources