Small hydrocarbon flow calibration facilities at NMIJ

APMP/TCFF Workshop Small hydrocarbon flow calibration facilities at NMIJ Yoshiya Terao

Original presentations Expansion of Calibration Flow Range of Small Liquid Hydrocarbon Flow Facility at NMIJ by Kar-Hooi Cheong, Ryouji Doihara Takashi Shimada and Yoshiya TERAO Development of prototype low-flow rate test rig applicable for highly volatile liquids by Ryouji Doihara, Takashi Shimada, Kar-Hooi Cheong and Yoshiya Terao at 9 th International Symposium on Fluid Flow Measurement (ISFFM2015), April 2015

LPG Gasoline Kerosene Light oil Heavy oil Flowrate (m 3 /h) 1.E+04 1.E+03 1.E+02 1.E+01 1.E+00 1.E-01 NATIONAL METROLOGY INSTITUTE OF JAPAN Primary standards for hydrocarbon at NMIJ Petroleum terminal Petroleum refinery Tanker Chemical plant Hydraulic machinery 100 L/h Primary standard for hydrocarbon 1.E-02 1.E-03 1.E-04 Fuel efficiency measurement Fuel mixer 1 L/h Small Hydrocarbon Flow Facility 1.E-05 0.02 L/h 1.E-06 0.1 1 10 100 Viscosity (mpa s)

JC08 mode fuel efficiency (km/l) Background and Motivation NATIONAL METROLOGY INSTITUTE OF JAPAN Top 10 car models with best fuel efficiency in Japan 40.0 37.0 36.4 35.0 33.0 33.0 32.6 30.8 30.4 30.2 30.0 30.0 30.0 25.0 20.0 15.0 10.0 5.0 0.0 hybrid vehicle mini / light vehicle Car models Source: http://e-nenpi.com

Automotive industry Main targets NATIONAL METROLOGY INSTITUTE OF JAPAN Fuel flowmeter Engine test bench Semiconductor industry Liquid mass flowmeter Vaporizer for manufacturing equipment

Automotive industry Fuel flowmeter Measurement conditions: Real gasoline (high volatility), Light oil, 200 L/h 0.02 L/h, Wide flow range (over 1000:1), Wide temperature range, (High pressure over 200MPa). Semiconductor industry Water, Alcohols, Solvents, Toxic liquids, 10 L/h 0.001 L/h, Wide temperature range (10 85 ), Liquid mass flowmeter

Measurement principles Gravimetric calibration: Advantage to accuracy and to simple traceability Disadvantage to evaporation error at the low flow rate range with volatile liquid Water is mostly used as test liquid. Volumetric calibration: SVPs have been adopted in many calibration facilities for oil (including real gasoline), because it can make a closed pipe line system. Some of them have a linear encoder to output many pulses. Syringe pump: Syringe pumps are widely used in medical and analysis fields for ultra-low flow rate. Fluctuation has been observed as a result of the stepper motor driving mechanism.

Cheong Kar-Hooi s gravimetric system

Calibration Facility Weighing section (liquid collection) Test section (DUT mounting) Flow generation section (tank, pump)

Calibration Facility Weighing section (liquid collection) Test section (DUT mounting) Flow generation section (tank, pump) thermostatic chamber (for better stability of liquid temperature controllable at any value between 15 ~35 ) Calibration method Gravimetric (static weighing) Standing-start-and-finish 2kg weighing system (liquid collection for 1~100 L/h) 100g weighing system (liquid collection for 0.02~1 L/h) Measurement point of temperature, T tm and pressure, p tm Test DUT meter Pulse signal Weighing vessel Weighing scale, m s Bypass to storage tank Pulse counter, Number of pulses, I p

Uncertainty Budget NATIONAL METROLOGY INSTITUTE OF JAPAN Uncertainty budget for 1 L/h ~ 100 L/h using 2 kg weighing system Uncertainty sources Light oil Kerosene Pulse count 8.2 10-5 8.2 10-5 Estimation of liquid density 2.6 10-4 3.0 10-4 Measurement of liquid mass 3.3 10-5 3.3 10-5 Dead volume effect 6.1 10-6 6.4 10-6 Relative combined standard uncertainty 2.8 10-4 3.2 10-4 Uncertainty budget for 0.02 L/h ~ 1 L/h using 100 g weighing system Uncertainty sources Light oil Kerosene Pulse count 8.2 10-5 8.2 10-5 Estimation of liquid density 2.6 10-4 3.0 10-4 Measurement of liquid mass 1.8 10-4 1.8 10-4 Dead volume effect 1.2 10-4 1.3 10-4 Relative combined standard uncertainty 3.5 10-4 3.8 10-4

Uncertainty Budget (Kerosene) NATIONAL METROLOGY INSTITUTE OF JAPAN estimation of liquid density 0.030% pulse count 0.008% dead volume effect 0.001% measurement of liquid mass 0.003% estimation of liquid density 0.030% pulse count 0.008% measurement of liquid mass 0.018% dead volume effect 0.013% 2 kg weighing system (1 L/h~100 L/h) 100 g weighing system (0.02 L/h~1 L/h)

Uncertainty budget for 0.02 L/h ~ 1 L/h using 100 g weighing system estimation of liquid density 0.030% pulse count 0.008% measurement of liquid mass 0.018% dead volume effect 0.013% Dead volume effect The value is evaluated in relative to the smallest amount of liquid collection which is 10 g. 10 g of liquid collection is performed at 0.02 L/h and it takes about 40 minutes for one collection. Some measures to cut down the dead volume effect: Dead volume is made as small as possible. Temperature variation of liquid in the dead volume is controlled and stabilized (±0.025 o C in one hour duration).

Contributing factors to uncertainty of liquid mass measurement NATIONAL METROLOGY INSTITUTE OF JAPAN estimation of liquid density 0.030% pulse count 0.008% measurement of liquid mass 0.018% dead volume effect 0.013% leakage possibility 0.009% correction factor of weighing scale 0.003% reading of weighing scale 0.001% liquid evaporation 0.015% buoyancy effect correction 0.001%

Evaluation of liquid evaporation NATIONAL METROLOGY INSTITUTE OF JAPAN Conditions for liquid collection Liquid temperature:15 ~35. Smallest liquid collection: 10 g (about 40 minutes at 0.02 L/h ) Method of evaluation An evaluation method thought to give the closest value to the actual condition was performed. Collection 100 g discharge Measurement of evaporation for 60 minutes Results At 35, the evaporation rate is 0.0015 g per hour for kerosene and 0.0007 g per hour for light oil. In relative to 10 g of liquid collection, this produces 1.5 10-4 (0.015 %) in terms of relative uncertainty.

Calibration and Measurement Capability Ultimate uncertainty of calibration (Calibration and measurement capability in italic, coverage factor: k=2) Flow range, Q / Weighing system Volumetric flowrate Mass flowrate 1 L/h Q 100 L/h (2 kg weighing system) 0.032 % (0.064 %) 0.010 % (0.020 %) 0.02 L/h Q < 1 L/h (100 g weighing system) 0.039 % (0.078 %) 0.025 % (0.050 %) Note: (ultimate standard uncertainty) 2 = (uncertainty due to facility) 2 + (uncertainty due to flowmeter) 2 Uncertainties due to flowmeter are as follows: Response of flowmeter to standing-start and finish Repeatability

Doihara s weighing and syringe system

Schematic diagram of the prototype test rig

Schematic diagram of the prototype test rig

Calibration procedure NATIONAL METROLOGY INSTITUTE OF JAPAN Calibration The syringe pump is calibrated using the gravimetric weighing tank.

Calibration procedure Then a DUT flowmeter is calibrated using the syringe pump. At a DUT calibration stage, the flow rig is able to make a closed pipe line system. NATIONAL METROLOGY INSTITUTE OF JAPAN Calibration

Standing method and flying method Standing method The syringe pump is calibrated with a standing start and stop method. What the weighing tank needs to calibrate is only the syringe pump (over 10 000 pulses). Characteristic of the syringe pump is essentially linear to flow rate. Thus the standing method is not significant problem in this syringe pump calibration. Calibration

DUT flowmeter is calibrated at a constant flow rate with a flying start and finish method. NATIONAL METROLOGY INSTITUTE OF JAPAN If the number of pulses of the flowmeter would be small, the pulse counting error can be reduced using a double chronometry pulse interpolation technique in the flying method. Calibration Flying method

Relative deviation from nominal value [ % ] 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0.00-0.01 Calibration results of the syringe They all agree within ±0.02%. Influence of the standing method was enough small. 0.01 0.10 1.00 10.00 100.00 Flow rate [ L/h ] NATIONAL METROLOGY INSTITUTE OF JAPAN Light oil: Jan. 2015 K SY P V SY SY ±0.02%. [Pulse/mL]

Evaporation effect and correction The weighing tank system has some evaporation controls to reduce evaporation. The weighing tank was surrounded by small liquid pools. The weighing tank has a lid with small holes to insert tubes. The changing of weight from an initial weight was measured in several conditions for one hour (Static). Condition Tank lid with Evaporation Evaporation rate small holes trap pool [ mg/h ] Static A 0.6 Static B 3 Static C 4 Static D 6

Relative deviation from nominal value [ % ] NATIONAL METROLOGY INSTITUTE OF JAPAN The syringe pulse factor was increased at low flow rate. Results with evaporation correction (Static A) were similar. 0.09 0.08 0.07 Without evaporation correction With evaporation correction (Static A) 0.06 0.05 0.04 0.03 0.02 0.01 0.00-0.01 0.01 0.10 1.00 10.00 100.00 Flow rate [ L/h ]

Evaporation effect and correction The atmosphere inside the wind shield was ventilated when the working liquid in the weighing tank was discharged. The actual evaporation rate was measured in the calibration sequence. The measured evaporation rate (actual condition E) was between static A and static C. Condition Tank lid with Evaporation Evaporation rate small holes trap pool [ mg/h ] Static A 0.6 Static B 3 Static C 4 Static D 6 Actual E 2 4 Since actual evaporation rates were not constant, the evaporation effect was corrected at each calibration.

Relative deviation from nominal value [ % ] NATIONAL METROLOGY INSTITUTE OF JAPAN Corrected results (Actual E) showed almost constant value over full flow rate range. 0.09 0.08 0.07 0.06 Without evaporation correction With evaporation correction (Actual E) 0.05 0.04 0.03 0.02 0.01 0.00-0.01 0.01 0.10 1.00 10.00 100.00 Flow rate [ L/h ]

Relative deviation from nominal value [ % ] NATIONAL METROLOGY INSTITUTE OF JAPAN Position dependency of syringe pulse factor The standing stop position was varied from the cylinder end position to 35 ml, keeping the calibration collection volume of 5 ml. 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0.00-0.01 0 5 10 15 20 25 30 35 40 Standing stop position from cylinder end [ ml ] 0 5mL 5mL

Relative deviation from nominal value [ % ] NATIONAL METROLOGY INSTITUTE OF JAPAN Position dependency of syringe pulse factor Volume domain of the cylinder has been fixed. Standing stop position: 5m L, Collection volume: 10 ml 0.09 0.08 0.07 0.06 5mL collection (light oil) 10mL collection (lightl oil) 0 5mL 10mL 0.05 0.04 0.03 0.02 0.01 0.00-0.01 0 5 10 15 20 25 30 35 40 Standing stop position from cylinder end [ ml ]

Relative deviation from nominal value [ % ] NATIONAL METROLOGY INSTITUTE OF JAPAN Calibration results of the syringe Difference depending on kinds of liquids has not been observed. 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0.00-0.01 Light oil: Jan. 2015 0.01 0.10 1.00 10.00 100.00 Flow rate [ L/h ] Industrial gasoline: Feb. 2015 Industrial gasoline: Mar. 2015

Preliminary uncertainty budget Uncertainty source Standard uncertainty Combined standard uncertainty Expanded uncertainty (k=2) Syringe pulse count 0.002% Mass measurement 0.008% Density at syringe 0.010% Change of mass in dead volume 0.007% Random effect of syringe calibration 0.008% Calibration for syringe pulse factor 0.017% 0.034% Linearity and long term reproducibility of reference syringe pulse factor 0.017% Density at syringe 0.010% Change of mass in dead volume 0.015% Random effect of DUT calibration 0.005% Calibration for mass flow rate 0.030% 0.060% Density at DUT 0.016% Calibration for volume flow rate 0.034% 0.068%

Flow rate: Specification 60 L/h 0.02 L/h (1000 ml/min 0.3 ml/min). Working liquid (as the first step): Light oil (diesel oil), Industrial gasoline (Flash point 40 ) Calibration principle: Gravimetric weighing tank, Syringe pump. Temperature and pressure (as the first step): 20 (Stability ±0.08 ) 0 0.3 MPa (at test section) Preliminary expanded uncertainty (k=2): 0.068 % (for volumetric flow), 0.060 % (for mass flow).

Summary Cheong s gravimetric system 0.02 L/h to 100 L/h Kerosene and light oil Liquid temperature: 15 to 35 Open for public calibration service Will be peer-reviewed in 2017 Doihara s gravimetric and syringe system 0.02 L/h to 60 L/h Applicable to volatile liquids Liquid temperature: 20 Flexible structure Used for research purpose

Thank you for your attention. NATIONAL METROLOGY INSTITUTE OF JAPAN