In Stage 1, the NIST 26 m3 PVTt flow standard [vvi - vii] was used to successively calibrate four
CFVs in air over a pressure range extending from 350 kPa to 700 kPa. For these CFVs this
pressure range corresponded to a Reynolds numbers range from 1.1 × 106 to 2.4 × 106. The four
CFVs are referred to throughout this document as the low pressure (LP) CFVs. The expanded
uncertainty of each of the LP CFVs was 0.10 % as verified in Section 4.1.
11
In Stage 2, the four LP CFVs were combined in parallel and positioned downstream of a single
CFV. Under choked flow conditions, the stagnation pressure (P0) of the upstream CFV equaled
four times that of the four downstream LP CFVs. In this way the upstream CFV is calibrated at
pressures and flows four times greater than any one of the downstream LP CFVs. This
calibration process was done in dry air. We repeated this procedure four times so that a total of
four CFVs were calibrated in Stage 2, which are herein referred to as the medium pressure (MP)
CFVs. The Reynolds number range of each MP CFV extended from 4.7 × 106 to 8.4 × 106, and
the expanded uncertainty of each MP CFVs was 0.13 %. The calibration results and uncertainty
analysis are documented in Section 4.2.
Stage 3 is analogous to Stage 2. The four MP CFVs were combined in parallel and used to
calibrate a single upstream CFV in dry air at four times the pressure. A total of eight CFVs were
calibrated in Stage 3, which are herein referred to as high pressure (HP) CFVs. The bootstrapping
approach implemented in Stages 2 and 3 resulted in eight HP CFVs that were traceable
to the 26 m3 PVTt primary standard, but with a flow capacity nearly sixteen times the Stage 1
calibration. The calibrated Reynolds number range of each HP CFV was 20 × 106 to 27.5 × 106
and the expanded uncertainty of each of the HP CFVs was 0.17 %. The calibration results and
uncertainty analysis are documented in Section 4.3.
In Stage 4 the eight HP CFVs were combined in parallel and used to calibrate a total of nine
TMWS, one at a time, in natural gas, at the nominal pressures for which they are used, and
mounted in the location of pipeline where they are used. Reynolds number matching was used to
apply the air-based calibration of the HP CFVs to natural gas. At these high Reynolds numbers
(i.e., above 16 × 106), theoretical predictions indicated a difference in the CFV discharge
coefficient between dry air and natural gas of less than 0.01 % [viii]. Each of the TMWS was
calibrated over a nominal volumetric flow range from 0.25 m3/s (3.2 × 104 acfh) to 1 m3/s
(1.3 × 105 acfh) at nominal pipeline pressures of 7500 kPa and ambient temperatures. The
volumetric flow was changed by varying the number of HP CFVs. The expanded uncertainty of
each of the TMWS ranged from 0.24 % to 0.25 % depending on flow. The calibration results and
uncertainty analysis are documented in Section 4.4.
In Stage 5, the nine TMWS are combined in a parallel array, and used to calibrate customer
flowmeters (MUT) in natural gas at ambient temperatures and at nominal pipeline pressures of
7500 kPa. The flow range of the facility extends from a minimum of 0.25 m3/s (3.2 × 104 acfh)
when only a single TMWS is used to a maximum of 9 m3/s (1.1 × 106 acfh) when all nine
TMWS are used together at their full capacity. The expanded uncertainty of a MUT (not
including the repeatability of the MUT) is 0.25 % at the highest flow and increases to 0.27 % at
the lowest flow. The uncertainty analysis is documented in Section 4.5.