WIXLIB

Calibration of Bubble Flow MetersSpeaker/Author: M. W. DlaminiCo-authors: D. Jonker, E. P. TarnowNational Metrology Institute of South Africa, Private Bag X34 Lynnwood Ridge 0040,Pretoria, South AfricaE-mail (corresponding author): mdlamini@nmisa.orgThe Gas Flow laboratory of NMISA currently calibrates the majority of gas flowinstrumentation in the range 0,5 ml/min to 50 l/min for the South African industry. This hasled to unacceptably long lead times for calibration slots, which has had an adverse effect onthe industry.In an effort to encourage industry to develop gas flow calibration capacity to improve thissituation, the Gas Flow laboratory of NMISA investigated several calibration methods thatcan be employed to perform bubble flow meter calibrations. These methods include the useof laminar flow meters (LFM) and positive displacement piston provers as referencestandards. Needle valves and Mass Flow Controllers were used to control the gas flow rate. Astandard Bubble Flow Meter Cell (20 ml/min – 6 l/min) was used as the unit under test(UUT) during the measurements.This paper discusses the different calibration methods, compares the results obtained andtheir associated measurement uncertainties.Keywords: thermal mass flow controller, bubble flow meter, piston prover1. IntroductionThe need for stack emission testing and environmental monitoring laboratories to havetheir measuring equipment calibrated by a ISO/IEC 17025 accredited facility has resultedin the NMISA Gas Flow laboratory receiving a vast amount of calibration requests forbubble flow meters (accuracy of 1 % of reading).The NMISA Gas Flow laboratory uses secondary reference standards of accuracy0,15 % to 0,25 % of reading to calibrate gas flow measurement instruments. Thelaboratory is assessing the feasibility of transferring the calibration of bubble flow metersto accredited commercial calibration laboratories. The NMISA Gas Flow laboratory willprovide traceability to these calibration laboratories, if requested.In the following sections the calibration of a bubble flow meter using a commerciallyavailable laminar flow meter system, a positive displacement piston prover and massflow controllers are discussed. These three gas flow reference standards utilisecompletely different principles. The positive displacement piston provers obtainmeasurement traceability from the SI units of time and length.Test and Measurement Conference

2. Flow ControlTo perform gas flow rate measurements at different flow rates, a reliable and stable gasflow rate control device is required. This can be a manually adjustable flow control valveor automated mass flow controller.2.1 Needle ValvesNeedle valves are equipped with a finely threaded stem which fits precisely into a seatover the inlet of the valve[1]. This allows the operator to precisely adjust the flow byrotating the handle. The stem packing provides an air tight seal to atmosphere andsome needle valves are equipped with metal-to-metal seat seal. These devices are agood choice for high temperature applications with lighter and less viscous fluids[1].As the devices have no readout unit, a separate flow meter may be necessary tomonitor the flow rate.2.2 Thermal Mass Flow ControllersThermal mass flow controllers (MFC) are complex mechatronic devices designed tomeasure and control the flow. Thermal mass flow controllers warm up gas passingthrough a capillary stainless steel tube. The change in gas temperature along the gasflow path is then measured by thermal elements coiled around the capillary tube. Themass flow 𝑄𝑀 is given by the equation[2] 𝑇𝑄𝑀 𝐶𝑝𝐾(1)where T is the temperature difference, 𝐶𝑝 is the specific heat capacity of the gas andK is a calibration constant. Thermal mass flow controllers are gas specific since theyuse the specific heat capacity of the calibration gas for flow metering. Therefore,when the process gas differs from the calibration gas, a conversion factor is requiredand is defined as[2]𝑐𝑝1 𝑝1𝐶 𝑐𝑝2 𝑝2(2)where 𝑝𝑖 is the density at normal conditions and the subscripts 1 and 2 representcalibration and process gas, respectively. The electronics of the MFC then open orclose the solenoid valve to increase or decrease the flow.A mass flow controller is itself a mass flow meter as it has to first measure flow,compare it with the input set point and then adjust the solenoid valve.3. Flow Metering3.1 Volumetric Flow MetersVolumetric flow meters measure the time, t t1-t2, it takes for a gas to fill a volumeV. A volumetric flow meter consists of a flow cell and a base unit. The base unithouses the main electronics of the meter. The flow cell houses a flow tube throughwhich a piston rises as the gas fills the tube. Two motion sensors, one at the bottomTest and Measurement Conference

and one at the top of the tube, start and stop a time interval counter as the pistonpasses by. The volumetric flow rate is determined by the equation[3]𝑄𝑣 𝑡𝜋𝑟 2 ℎ𝑉2 𝑡1 𝑡2 𝑡1(3)where r is the flow tube radius and h is the distance between the two motion sensors.Some volumetric flow meters are equipped with barometric pressure transducers andthermometers for instant conversion of volumetric flow to standardised flow. Bubbleflow meters are volumetric meters using soap film instead of a piston and measurevolumetric flow rate only.3.2 Laminar Flow MetersA laminar flow meter is made of a circular cross sectional tube with flow restrictionsinside to create a pressure drop. The pressure drop inside the meter is defined to beproportional to the flow rate in accordance with the Hagen-Poiseuille law for laminarflow pressure drop[5]. The flow rate, Qs, is given by 𝑃𝑄𝑠 𝜋𝑑 4 128µ(4)where d is the inner diameter of the capillary tube, P is the pressure drop and µ is theviscosity of the fluid. The viscosity is a function of the temperature.Laminar flow meters require the Reynolds number (Re) of the flowing fluid to be lessthan 2000 indicating laminar flow conditions are maintained. To increase the range ofthese meters, manufacturers design their laminar flow meters with several capillarytubes to limit Re to less than 2000 in each capillary tube. The total flow rate throughthe meter is given by𝑄 𝑁𝑄𝑠(5)where N is the number of capillary tubes in the meter[5].These flow meters are good to measure low to extremely low flow rates and havelinear calibration and low noise[6]. To minimize errors, these meters must be operatedin a controlled temperature environment as the density of gases varies significantlywith changing temperature[6].4. Bubble Flow Meter Calibration Against a Piston Prover4.1 Calibration SetupPrior to calibration, the instruments were left in the laboratory to achieve thermalequilibrium. The calibration setup is shown in Figure 1 with the reference standardupstream of the bubble flow meter to avoid contamination. The system waspressurized to 200 kPa at a flow rate of 2500 sccm (standard cubic centimetre perminute) and leak checks were performed using a liquid leak detector. The bubble flowmeter was then calibrated against two flow cells in the flow ranges 0,5 sccm to50 sccm and 50 sccm to 5 000 sccm. The reference standard, which is a volumetricTest and Measurement Conference

flow meter, was set to read standardized (at 25 oC and 101,325 kPa) flow rate valuesand the bubble flow meter indicated volumetric flow rate.Figure 1. Schematic of calibration setup at NMISA, 1: gas cylinder; 2: pressuregauge and regulator; 3: MFC; 4: NMISA reference standard; 5: UUT (BFM); 6:MFC control box.The volumetric flow rate, Qv, measured by the bubble flow meter, was converted tostandardized flow rates, Qs, using the equation[3]𝑇𝑃𝑄𝑠 𝑄𝑣 (𝑇𝑠 𝑃𝑎)𝑎 𝑠(5)where Ts and Ps are the temperature and pressure values at standard conditions, Ta andPa are the actual temperature and pressure, respectively. These values were measuredby the reference standard.4.2 Measurement Uncertainty EvaluationThe uncertainty of measurement was calculated and expressed in accordance to theGuide to the Expression of Uncertainty in Measurement (GUM). The reporteduncertainty is stated as the standard uncertainty of measurement multiplied by acoverage factor of k 2, which for a normal distribution approximates to a level ofconfidence of 95,45 %.For measurement uncertainty estimation the following sources of uncertainty wereconsidered: Reference standard calibration: expanded uncertainty at 95 % level ofconfidence, therefore divided by k 2 and infinite degrees of freedom. Reference standard and UUT resolution: least significant digit as displayedtreated as a rectangular distribution, divided by 2 and the square root of threewith infinite degrees of freedom. Variability: For the reference standard, estimated standard deviation of themean (ESDM) was used and n-1 degrees of freedom. For the UUT, theestimated standard deviation was used and n-1 degrees of freedom.The piston prover reference standard has an expanded uncertainty of 0,25 % and 0,15 % of reading for the ranges 0,5 sccm to 50 sccm and 50 sccm to 5 000 sccm,respectively[4].Test and Measurement Conference

5. Bubble Flow Meter Calibration Against Laminar Flow Meters5.1 Calibration SetupThe calibration of the bubble flow meter against the laminar flow meters (LFM)followed the same procedure as explained in section 4.1. In this case, the LFMreference standard was connected upstream of the MFC and bubble flow meter. TheMFC and LFM were both connected to a Molbloc molbox1 to control and read theMFC and the LFM. System leak checks for vacuum coupling radiation fittings (VCR)were performed using the leak check function of the reference standard. For Swagelokconnections, the liquid leak detector was used to check for leaks.The bubble flow meter was calibrated against the laminar flow meters 5E1-VCR-V-Q(0 sccm to 50 sccm), 5E2-VCR-V-Q (50 sccm to 500 sccm) and 5E3-VCR-V-Q(500 sccm to 5 000 sccm).5.2 Measurement Uncertainty EvaluationThe measurement uncertainty evaluation for calibrating the bubble flow meter againstthe LFMs was performed in a similar way as explained in section 4.2. The laminarflow meters have an expanded uncertainty of 0,2 % of reading.6. Bubble Flow Meter Calibration Against a Mass Flow Controller6.1 Calibration SetupThe calibration of the bubble flow meter against the mass flow controllersfollowed the same procedure as explained in section 4.1. The bubble flow meterwas calibrated against two mass flow controllers with flow ranges 0 sccm to 500sccm and 500 sccm to 5 000 sccm. The calibration setup was as shown in Figure1without the piston prover (item 4). The mass flow controller control box (item 6)was used to set the flow rate and as a readout unit for the mass flow controller.6.2 Measurement Uncertainty EvaluationThe measurement uncertainty for calibrating the bubble flow meter against the massflow controllers was performed in a similar way as explained in section 4.2. Anadditional uncertainty contributor due to the interpolation of the mass flowcontroller’s calibration was added to the uncertainty budget. The standard error of theinterpolation was used as the uncertainty estimate with n-2 degrees of freedom.The 500 sccm mass flow controller has an expanded uncertainty of about 2,2 % at50 sccm and 1,1 % at 120 sccm. The 5000 sccm mass flow controller has an expandeduncertainty of 0,8 %.Test and Measurement Conference