Rotary Gas Meters: How Much Weight to Give a Simulated Pressure Drop
Abstract
A pressure-loss calculation can help compare rotary meter concepts, but it needs a measured check before informing selection. Zhang, Su and Lin simulated internal air flow in a 50 mm rotary meter and compared the calculated pressure drop with four measured points. The simulation predicted less pressure drop at 60 m³/h and greater pressure drop at 80, 90 and 100 m³/h. This article explains how to use the pressure and velocity fields alongside this comparison. The four-point check is evidence for the studied setup and supports a request for additional measurements at the intended gas-service conditions.
Background
Pressure drop is one consideration when choosing a rotary gas meter for a station whose downstream equipment needs a defined inlet pressure. A positive-displacement meter has moving rotors and clearances around them. Those features make simulation useful for studying where pressure and velocity change during rotation. A numerical result for the complete meter also depends on the modeled fluid, geometry and boundary conditions. The reported study illustrates those differences using an identified meter and an air-flow comparison.
Research question
How closely did the model track measured pressure loss at the four flow rates reported in the paper? The answer requires checking both the size and direction of the measured-to-simulated difference. The figure showing internal flow at one operating rate answers a separate question about how the model represents conditions inside the chamber.
Method
The authors worked with a 50 mm CNIM-RM series Roots-type meter. The study identifies its casing and flanges and distinguishes the rotary lobes from the gearing and counting display. These details define which assembly was studied. The drawings below are reconstructions of the original photograph and the labeled component view.
Figure 1. Functional line drawing from source Figure 1, p. 462. It depicts the study meter and does not reproduce its dimensions.
Figure 2. English reconstruction of source Figure 2, p. 462. Component labels follow the paper's description.
The source created the fluid domain in SolidWorks and carried out transient calculations in PumpLinx. The study used air at 300 K and attached straight regions to the meter domain. The domain connected an inlet, an outlet and the moving rotor region, with a reported mesh count of approximately 820,000. The outlet static pressure was set to 101,325 Pa. Pressure was monitored at the inlet and outlet of the meter. The measured comparison used a 2,000 L bell prover for flow, reported as class 0.2, with a stated maximum applicable flow of 120 m³/h. This rating describes the prover in the paper.
Figure 3. English reconstruction of source Figure 3, p. 462. The hatch indicates where mesh regions are located and is illustrative.
Results and discussion
Figures 4 and 5 show simulated fields at 100 m³/h at eight successive rotor positions. They let a reader compare regions within the moving-domain model as the rotor phase changes. The comparison does not measure pressure loss within an installed natural gas meter. The printed field pictures were sampled to preserve the visible distribution and rotor phases. Some pressure scale digits are truncated in the publication's rendered image, limiting quantitative reading of that figure.
Figure 4. Grayscale reconstruction of source Figure 4, p. 463. Grey levels reflect the printed field, with pressure-scale digits clipped in that source.
Figure 5. Grayscale reconstruction of source Figure 5, p. 464. The source legend shows a velocity magnitude scale of 0 to 20 m/s.
For a check on the calculated overall loss, Table 1 of the paper lists the measured and simulated pressure drop. In the following table the signed difference is calculated as simulation minus measurement. The original percentage results use a separate reporting convention and are not combined with the difference column.
Flow (m³/h) | Measured drop (Pa) | Simulated drop (Pa) | Signed difference (Pa) |
60 | 138.41 | 101.99 | −36.42 |
80 | 265.35 | 337.96 | +72.61 |
90 | 312.25 | 432.52 | +120.27 |
100 | 408.15 | 552.14 | +143.99 |
Figure 6. Replotted from exact source Table 1 values and corresponding to source Figure 6, p. 464. The lines join reported points and do not estimate additional measurements.
The sign reverses across the four listed points. At 60 m³/h the model is below the measured curve and at 100 m³/h it is almost 144 Pa above. A uniform correction inferred from just one flow point would therefore misstate other points. The paper discusses possible influences of mechanical and pipeline losses at lower flow. Its comparison does not separately measure the effect of each loss. The total drop also needs to be compared to the permissible pressure-loss budget for the complete metering run, keeping the meter and installation contributions clear.
The simulations covered four air-flow values, with the published fields taken from the 100 m³/h run. When reviewing the model output, ask whether the quoted loss is taken at the same positions used for the measurement. A change in pressure tap location can include a different length of pipe and a different static-pressure recovery region. This is a comparison question for any new report. The published sketches show why identifying the modeled domain matters to the interpretation of the plotted totals.
A further distinction is the difference between loss and absolute pressure. The imposed outlet boundary is about atmospheric pressure, whereas the reported meter loss is the difference across the device. Compare like quantities in the technical submittal, especially when one dataset gives the loss of the meter assembly and another gives the pressure budget for the complete station. If test conditions differ by gas or by volumetric flow basis, list those differences beside the two curves before deciding whether they can be compared.
Conclusion
The model explains the changing pattern of the internal field, and the measured series establishes how much confidence to place in the simulated overall loss for this experiment. For procurement, specify the actual gas, flow range, pressure, temperature and acceptable pressure drop. Ask for measured values and corresponding test conditions for the offered meter and installation.
Limitations and outlook
Four air-flow points do not cover the full operating range at other pressures, temperatures or gas compositions. A follow-up study would benefit from further measurements near the lower end and repeated trials where the difference changes sign. The source is a method comparison on a specified test meter.
Notes
The pressure difference is reported in pascals. Volumetric flow rates describe the conditions of the source study. Figure 6 uses reported tabular values. Figures 4 and 5 are graphical reconstructions and are not numerical CFD data exports.
References and Image Sources