The objective of this project is to characterize temperature effects on flow meters used in cryogenic liquids, e.g. turbine and Coriolis flow meters. With a validated physical model of the flow meter, room temperature water calibrations can be extrapolated to measure cryogenic flows such as liquified natural gas (LNG). Current work is focused on developing a proven physical model for Coriolis meters to extrapolate room temperature calibrations based on the resonance frequency of the flow tube(s) and literature values for the temperature dependence of the properties of the construction materials (Young’s modulus and Poisson’s ratio). Coriolis meters are often used in the custody transfer of valuable cryogens like liquid oxygen and LNG. NIST’s cryogenic liquid flow standard was decommissioned in 2020, leaving few options for meter users to calibrate their meters in an actual cryogenic flow.
The Fluid Metrology Group is using NIST’s 15 kg/s Liquid Flow Standard (LFS) [1] (Fig. 1) and data from the NIST cryogenic flow facility (decommissioned in 2020) [2] to better understand how coriolis meters are affected by the liquid and environment temperatures. The NIST water flow standard is a closed-loop, dynamic liquid flow calibration facility that is fully automated. Calibrations can be performed in the water flow facility at ambient temperatures or in a custom-built insulated chamber (Fig. 2) that allows the meter environmental temperature to be controlled between 10 °C and 50 °C. The major components of the LFS include: 1) the flow generation and control system consisting of a variable flow pump, reservoir tank, check standard flow meter, butterfly valve, and data acquisition system with digital proportional-integral-derivative (PID) controller; 2) a test section that accommodates a meter under test with pipe diameter ranging from 1.25 cm to 5 cm; and 3) the dynamic weighing system comprised of a collection tank and weigh scale.
NIST has developed a physical model based on Young’s modulus and Poisson’s ratio to hypothesize a correction factor for how Coriolis meters will behave at cryogenic temperatures (Fig. 3) [3]. As a first step, we tested our model over the narrow liquid temperature range of 285 K to 318 K. The temperature dependence predicted by the model agrees with experimental data within ± 0.08 %. The model uncertainty is 0.16 % (95 % confidence level) over this temperature range. We are currently analyzing data from several coriolis meters at liquid nitrogen (LN2) temperatures (≈ 77 K) to further validate our model. Figure 4 shows a coriolis meter installed in the cryogenic flow facility.
Presently, various liquid temperatures are achieved in the 15 kg/s LFS by initially cooling the water in the reservoir and then heating it over time by the mechanical action of the pump (no active control). A future goal is to add a heater / chiller to actively control the temperature of the water in the flow standard to temperatures between 10 °C and 40 °C. This will facilitate research on temperature effects on many meter types and reduce repeatability errors due to the water temperature varying during a calibration or test.