July 2023, Vol. 250, No. 7
Features
Why Flowmeter Calibration Sometimes Lacks Accuracy
Chris Mills, Senior Consultant Engineer, TÜV SÜD National Engineering Laboratory
(P&GJ) — The oil and gas industry appears to be favoring a move toward using “newer” and more “advanced” flow measurement technologies, such as ultrasonic and Coriolis devices as an alternative to turbine and positive displacement meters.
Coriolis flow meters offer the distinct advantage of providing a direct measurement of mass flow rate and product density with stated uncertainties as low as 0.05% for mass and 0.05 kg/m3 for density for light hydrocarbons, meaning measurements of gas and oil flow are highly accurate.
The exact specification differs by meter manufacturer and model type. While the Coriolis forces for gas use are of a magnitude of three times smaller than in liquid use, a Coriolis flow meter can measure single-phase liquid or single-phase gas without any variation in model type.
Though the adoption of Coriolis flow meters is a logical move for oil and gas pipelines, their measurement uncertainty at elevated conditions is generally not well understood by end users. Several factors affect the performance of Coriolis devices, including temperature, pressure, fluid viscosity and the Reynolds number.
The Reynolds number helps to predict flow patterns in different fluid flow situations, such as laminar or turbulent flow. However, meter manufacturers have made corrections to compensate for these effects.
While these effects could potentially be ascertained by calibrating “in situ” at service conditions, industry appears to be moving away from proving meters onsite. Partly due to a lack of space, maintenance and cost, provers are becoming scarce in oil and gas pipeline terminals. The more favored approach appears to include Coriolis master and duty flow meters.
The duty meter remains in situ and the master meter is periodically sent to an accredited laboratory for a flow calibration to minimize oil and gas production downtime. The performance of the duty meter is then compared with the calibrated master meter.
However, the temperature, pressure and fluid properties of oil and gas flowing in a pipeline can differ considerably from conditions at the calibration laboratory. The standard practice for calibrating flow meters has been to match the fluid viscosity and, if possible, the fluid temperature and pressure. Unfortunately, matching all parameters is seldom possible due to the replication limitations of the calibration facilities.
Consequently, the parameter that is most matched is the fluid viscosity. This partly stems from the known effect of viscosity on conventional liquid flow meters, such as turbine and positive displacement devices. However, a limitation of the above approach is that temperature and pressure variations are also known to influence the overall measurement uncertainty of a meter.
To address this, TÜV SÜD National Engineering Laboratory commissioned a fully accredited elevated pressure and temperature (EPAT) oil flow facility to investigate the performance of flow meters under these conditions. It also enables liquid flow calibrations to be completed close to service conditions.
The ISO 10790:2015 standard provides guidance on the selection, installation and use of Coriolis flowmeters. However, it includes little practical guidance for their operation at elevated pressures, temperatures, and viscosities. There isn’t a complete lack of awareness in the oil and gas industry.
TÜV SÜD National Engineering Laboratory’s research in this critical area prompted the U.K. Oil and Gas Authority (OGA) to stipulate that temperature and pressure compensation applied to any flow meter between its calibration conditions and its pipeline conditions must be “agreed in advance” and must be “traceable and auditable.” Unfortunately, the methodology for calibrating and operating Coriolis meters at elevated conditions appears fragmented.
Our research has explored the performance of Coriolis flow meters that have been calibrated in the elevated pressure and temperature (EPAT) oil flow facility and the U.K. National Standards oil flow facility in Glasgow. We analyzed the calibration results in terms of fluid viscosity, Reynolds number, temperature, pressure and flow rate to identify trends and to ascertain whether manufacturers’ performance claims were valid.
The experimental results used were from a combination of DSIT-funded research, Joint Industry Projects, internal TÜV SÜD National Engineering Laboratory research and commercial calibrations. The Coriolis flow meter manufacturers were not active participants in the investigations, except for the Joint Industry Projects.
Overall, the research results reinforce the concept that Coriolis flow meters cannot simply be utilized at pipeline conditions without suitable consideration, characterisation and calibration. In summary, Coriolis water calibration does not replicate service conditions, and it is vital that end users remember that pressure corrections published by manufacturers are not fully traceable at present.
Pressure effects are more significant than temperature effects and viscosity/Reynolds number effects can be significant and should be considered. From our research, we would make the following recommendations for calibrating Coriolis flow meters when onsite proving is not available.
Temperature – The end user cannot easily modify the temperature compensation coefficient. Instead, a more practical approach would be to calibrate the device as close to the pipeline temperature as possible. This would allow the end user to ascertain whether temperature effects are significant, and a correction allowed for via an adjustment to the Coriolis mass factor.
The following calibration procedure is recommended for temperature effects:
- Zero device at operating temperature & pressure
- Calibrate device at operating temperature and pressure “as found”
- Perform calibration at ±10 °C to determine temperature offset
- If required, perform an “as left” calibration
Pressure - A traceable pressure correction should be used where possible. If the pipeline conditions are stable, then a static fixed pressure correction could potentially be applied. This involves the device being adjusted for the effects of pressure via an adjustment to the device mass factor or to the flow computer.
However, it should be noted that if the pressure effect is significant (e.g., –0.020 % per bar) then even a 5-bar variance could produce a meter offset of –0.10 % using a static fixed correction.
The preferred method would be to use a dynamic “live” correction via a pressure transmitter. This utilises a pressure measurement at the Coriolis flow meter that is either supplied to the Coriolis transmitter or to a flow computer to adjust the meter for the effects of pressure.
The meter compensation coefficient would still be a set value, but the amount of adjustment to the Coriolis flow meter will vary with the measured pressure.
The following calibration procedure is recommended for pressure effects:
- Zero device at operating temperature and pressure
- Calibrate device at operating temperature & pressure “as found”
- Additional pressure compensation calibration at ± 10 bar to derive (linear) pressure compensation coefficient
- If required, adjusting the Coriolis mass factor and then perform an “as left” calibration
Viscosity/Reynolds Number – If operating in high viscosity conditions, a Coriolis flow meter should be characterised against Reynolds number with a suitable fluid to ascertain the effects. However, correcting for the adverse effects of viscosity/Reynolds number can be challenging. Depending on the manufacturer, the device might apply a Reynolds number correction. It should also be noted that installation has a significant effect on the Reynolds number at which the laminar-turbulent transition occurs. Hence, the robustness of any Reynolds number correction might require further investigation at alternative entry lengths.
The following calibration procedure is recommended for viscosity/Reynolds number effects:
- Specify Reynolds number range of device from service condition density, viscosity, and flowrate
- Match Reynolds number range with high viscosity fluid at two or more temperatures at an accredited flow laboratory
- Where possible, recreate the installation entry lengths
- Zero device at close to operating conditions, if feasible
- Calibrate device at Reynolds number “as found”
- Decide if Reynolds number effect is significant
- If required, perform an “‘as left” calibration
Temperature is a significant effect for Coriolis flow meters. However, Coriolis flow meters have an onboard-resistance temperature detectors (RTD) and incorporate algorithms to correct for temperature effects on the flow tube material. This means that the temperature effect is automatically corrected, so the uncertainty measurement effect of temperature appears to be a magnitude less than pressure effects.
As the pressure effect has been shown to be linear, it can be corrected either via an adjustment to the meter mass factor, a static fixed pressure correction or a dynamic “live” correction via a pressure transmitter. This means that a traceable dynamic “live” pressure correction via a pressure transmitter should be used where possible.
It is also important for the end user to remember that the performance of Coriolis meters from one manufacturer are not necessarily like meters from other manufacturers as there are many variables such as meter design, flow tube dimensions, patented corrections and the quantity and quality of any internal R&D.
It is also hoped that ISO 10790 will be updated soon with the latest available traceable data on temperature, pressure, and viscosity/Reynolds number effects.
Author: Chris Mills is a senior consultant engineer at TÜV SÜD National Engineering Laboratory, a world-class provider of technical consultancy, research, testing and program management services.
Comments