Flow Measurement Instrumentation - Orifice Plates, Coriolis Meters, Ultrasonic Technology, and Custody Transfer Standards

Flow Measurement Instrumentation - Orifice Plates, Coriolis Meters, Ultrasonic Technology, and Custody Transfer Standards

Flow measurement in the oil and gas industry is not a single engineering problem - it is three distinct problems that happen to use similar instruments. Production allocation measurement determines how much each well contributes to total field production so that operating costs and revenues can be correctly attributed between partners. Process control measurement provides the real-time flow data that control systems use to maintain separator levels, compressor throughput, and pipeline pressure. Custody transfer measurement determines the exact volume of hydrocarbon that changes ownership between a producer and a pipeline operator, a producer and a refinery, or an exporting nation and an importing nation. The custody transfer problem is qualitatively different from the other two because it has direct financial consequences: a 0.1% systematic error in a custody transfer meter measuring 100,000 bbl/day at $65/bbl costs $23.7 million per year. This financial exposure means custody transfer meters are subject to API MPMS (Manual of Petroleum Measurement Standards) requirements for installation, calibration, proving, and uncertainty analysis that have no parallel in process control measurement. This guide covers the engineering of the four dominant flow measurement technologies - differential pressure (orifice plate), Coriolis, ultrasonic, and turbine - with the calculations that determine their accuracy, the ISO and API standards that govern their application, and the specific requirements that distinguish custody transfer from process measurement.

1. Differential Pressure Flow Measurement - The Orifice Plate

1.1 The Bernoulli Principle Applied to Orifice Plates

An orifice plate is a thin plate with a precisely machined concentric hole (the orifice bore) installed perpendicular to the flow in a pipeline. As fluid accelerates through the restriction, static pressure drops. The differential pressure (DP) between the upstream high-pressure tap and the downstream low-pressure tap is related to the flow rate through the Bernoulli equation combined with an empirical discharge coefficient that accounts for real fluid behavior:

ISO 5167 orifice plate flow equation:
q_m (kg/s) = C x E x epsilon x (pi/4) x d^2 x sqrt(2 x rho_1 x dP)

Where:
C = discharge coefficient (dimensionless, ≈ 0.6 for standard orifice)
E = velocity of approach factor = 1/sqrt(1-beta^4)
epsilon = gas expansibility factor (= 1.0 for liquid, <1 for compressible gas)
d = orifice bore diameter (m)
rho_1 = upstream fluid density (kg/m3)
dP = differential pressure (Pa)
beta = diameter ratio = d/D (orifice bore / pipe ID)

Stolz equation for discharge coefficient C (ISO 5167):
C = 0.5959 + 0.0312 x beta^2.1 - 0.1840 x beta^8 + 0.0029 x beta^2.5 x (10^6/Re_D)^0.75 + L1 x 0.0900 x beta^4/(1-beta^4) - L2' x 0.0337 x beta^3

Where Re_D = Reynolds number based on pipe diameter, L1 and L2' depend on tap type (corner, D-D/2, or flange taps)

Simplified example - gas measurement orifice plate:
Pipe ID D = 0.1016 m (4"), orifice bore d = 0.0612 m (beta = 0.0612/0.1016 = 0.6024)
Natural gas: T = 20°C (293 K), P = 50 bar (5,000,000 Pa), rho_1 = 38.5 kg/m3
dP measured = 2,500 Pa (10" H2O)
C = 0.6050 (from Stolz at Re_D ≈ 500,000)
E = 1/sqrt(1-0.6024^4) = 1/sqrt(1-0.1319) = 1/sqrt(0.8681) = 1/0.9317 = 1.0733
epsilon = 1 - (0.41 + 0.35 x beta^4) x dP/(kappa x P1) for gas
= 1 - (0.41 + 0.35 x 0.1319) x 2,500/(1.27 x 5,000,000)
= 1 - 0.4562 x 0.000394 = 1 - 0.000180 = 0.9998 ≈ 1.0 (small dP, negligible expansion)

q_m = 0.6050 x 1.0733 x 0.9998 x (pi/4) x 0.0612^2 x sqrt(2 x 38.5 x 2,500)
= 0.6050 x 1.0733 x 1.0 x 0.002942 x sqrt(192,500)
= 0.6050 x 1.0733 x 0.002942 x 438.75
= 0.6050 x 1.0733 x 1.2907
= 0.6050 x 1.3850 = 0.8379 kg/s mass flow rate

Volume flow at line conditions: q_v = 0.8379/38.5 = 0.02177 m3/s = 78.4 m3/hr
Converted to standard conditions (15°C, 1 atm): q_std = 78.4 x (50/1.01325) x (288.15/293) = 78.4 x 49.37 x 0.9836 = 3,806 Nm3/hr = 3.8 MMscf/day equivalent

1.2 Orifice Plate Uncertainty Analysis

Combined uncertainty of orifice plate measurement (ISO 5167 / GUM method):
The flow rate equation q_m = f(C, E, epsilon, d, rho, dP)
Each input variable has a measurement uncertainty that propagates to the flow uncertainty.

Sensitivity coefficients (partial derivatives):
Flow is proportional to: q_m ∝ C x d^2 x rho^0.5 x dP^0.5

Relative uncertainty in flow:
(u_q/q)^2 = (u_C/C)^2 + (2 x u_d/d)^2 + (0.5 x u_rho/rho)^2 + (0.5 x u_dP/dP)^2

Typical uncertainty components for a well-installed orifice plate:
u_C/C = 0.5% (discharge coefficient uncertainty from ISO 5167 correlation)
u_d/d = 0.05% (orifice bore diameter measurement uncertainty)
u_rho/rho = 0.3% (density measurement from P, T, and composition)
u_dP/dP = 0.1% (DP transmitter uncertainty at operating point)

(u_q/q)^2 = (0.005)^2 + (2 x 0.0005)^2 + (0.5 x 0.003)^2 + (0.5 x 0.001)^2
= 0.000025 + 0.000001 + 0.00000225 + 0.00000025
= 0.000028.5
u_q/q = sqrt(0.0000285) = 0.534% combined uncertainty

This 0.534% uncertainty is acceptable for process control measurement but does NOT meet custody transfer requirements.

For custody transfer per API MPMS Chapter 14.3: uncertainty ≤ 0.3% required for fiscal metering
The discharge coefficient uncertainty (0.5%) alone exceeds the total allowed budget.

Solution for custody transfer: Replace ISO 5167 correlation C with a calibrated C from a flow laboratory
Calibrated C uncertainty ≈ 0.10-0.15% → total u_q/q reduces to approximately 0.20-0.25% → meets custody transfer requirement.

2. Coriolis Mass Flow Meters

2.1 The Coriolis Effect - True Mass Flow Measurement

The Coriolis meter is the only flow measurement technology that directly measures mass flow rate without requiring density measurement as a separate input. It works by oscillating one or two measurement tubes at their natural resonant frequency. When fluid flows through the oscillating tubes, the Coriolis acceleration (the same force that causes weather systems to rotate on Earth) creates a phase shift between the inlet and outlet sections of the tube. This phase shift is directly and linearly proportional to mass flow rate:

Coriolis meter operating principle:
Phase shift: delta_phi (degrees) = K_m x q_m

Where K_m = meter factor (degrees per kg/s) - determined during factory calibration

Simultaneous density measurement:
The resonant frequency of the tube changes with fluid density:
rho_fluid = A + B / f^2

Where f = resonant frequency (Hz), A and B = calibration constants
Higher density → lower resonant frequency (heavier mass takes longer to oscillate)

Coriolis meter specification example - crude oil custody transfer:
Meter size: 4" (100 mm), Maximum flow rate: 540 m3/hr
Mass flow accuracy: ±0.10% of reading (typical high-end Coriolis for custody transfer)
Density accuracy: ±0.5 kg/m3 at calibration conditions
Repeatability: ±0.05% (better than accuracy - important for trend monitoring)
Maximum operating pressure: 100 bar, Maximum temperature: 150°C

Viscosity effect on Coriolis accuracy:
High-viscosity fluids (heavy crude, waxy oil at low temperature) reduce meter accuracy because damping of the tube oscillation increases. Most Coriolis manufacturers specify:
- Standard accuracy (±0.10%) for viscosities up to 50 cp
- Reduced accuracy (±0.20-0.30%) for 50-1,000 cp
- Not recommended >10,000 cp without viscosity correction factor

Financial exposure calculation for a 4" Coriolis at ±0.10% vs orifice at ±0.50%:
Custody transfer rate: 540 m3/hr x 24 hr/day = 12,960 m3/day = 81,556 bbl/day
Oil price: $65/bbl

Daily revenue: 81,556 x $65 = $5,301,140/day
Annual revenue: $1,934,916,100/year

Uncertainty at 0.50% (orifice): Annual uncertainty = 0.005 x $1,935M = ±$9.67M/year financial exposure
Uncertainty at 0.10% (Coriolis): Annual uncertainty = 0.001 x $1,935M = ±$1.93M/year financial exposure

The Coriolis meter costs $45,000-85,000 vs $8,000-15,000 for an orifice plate installation.
The $60,000 additional capital cost of the Coriolis reduces annual financial exposure by $7.74M/year → payback in 3 days.

2.2 Coriolis Meter Applications and Limitations

Application	Coriolis Suitability	Reason / Limitation
Crude oil custody transfer	Excellent	Direct mass measurement, insensitive to flow profile, no upstream straight-run requirement. Industry preferred for liquid custody transfer.
Chemical injection metering (methanol, MEG, scale inhibitor)	Excellent	Low flow rates, accurate even at very low velocities. Density measurement confirms chemical identity (detects dilution or wrong product). Standard for chemical injection skids.
Gas custody transfer (high pressure)	Limited	Gas compressibility in tubes causes measurement errors. Damping increased. Manufacturers specify reduced accuracy for gas service. Ultrasonic meters preferred for gas fiscal metering.
Two-phase flow (liquid + gas)	Not suitable	Gas bubbles in liquid (or liquid slugs in gas) cause tube vibration damping errors that can produce 5-20% measurement errors. Two-phase flow disqualifies Coriolis meters from custody transfer service.
Slurry or abrasive fluids	Moderate	Abrasive particles erode the tube interior, changing the tube geometry and affecting calibration. High-hardness tube materials (Hastelloy, titanium) extend service life but do not eliminate the issue.

3. Ultrasonic Flow Meters

3.1 Transit-Time Measurement Principle

Ultrasonic flow meters measure the transit time of ultrasonic pulses transmitted diagonally across the pipe bore. With flow, pulses traveling with the flow travel faster than pulses traveling against the flow. The difference in transit times is directly proportional to the average fluid velocity along the acoustic path. Multi-path meters use 4-8 acoustic paths at different chordal positions to average the velocity profile across the pipe cross-section:

Ultrasonic meter transit time equation:
t_up = L / (c - v x cos(theta)) (upstream pulse, fighting flow)
t_down = L / (c + v x cos(theta)) (downstream pulse, with flow)

Where:
L = acoustic path length (m), c = speed of sound in fluid (m/s)
v = mean flow velocity along chord (m/s), theta = angle of path to pipe axis

Solving for velocity:
v = (L / (2 x cos(theta))) x (1/t_up - 1/t_down) / (correction_factor)

Since c >> v (speed of sound >> flow velocity):
v ≈ L^2 x delta_t / (2 x t_avg^2 x cos(theta))
Where delta_t = t_up - t_down, t_avg = (t_up + t_down)/2

Worked example - 8" gas ultrasonic meter:
D = 0.2032 m (8"), L = 0.240 m (diagonal path), theta = 30° to pipe axis
Speed of sound c = 420 m/s (natural gas at 50 bar, 20°C)
Measured: t_up = 570.143 microseconds, t_down = 569.857 microseconds
delta_t = 0.286 microseconds = 2.86 x 10^-7 s
t_avg = 570.000 microseconds = 5.70 x 10^-4 s

v = (0.240)^2 x 2.86e-7 / (2 x (5.70e-4)^2 x cos(30°))
= 0.05760 x 2.86e-7 / (2 x 3.249e-7 x 0.8660)
= 1.647e-8 / (5.629e-7)
= 0.02926 m/s chord velocity → too low

Recalculate: For typical gas flow at 8" pipe and 10 m/s mean velocity:
delta_t = 2 x L x v x cos(theta) / (c^2 - v^2 x cos^2(theta))
≈ 2 x 0.240 x 10 x 0.866 / (420^2) = 4.157 / 176,400 = 23.6 nanoseconds

Flow rate: v_mean = 10 m/s at D = 0.2032 m
q_v = pi/4 x 0.2032^2 x 10 = 0.03243 m^2 x 10 = 0.3243 m3/s = 28,020 m3/hr

The 23.6 nanosecond measurement requires timing electronics with picosecond resolution - this is the key technical challenge that distinguishes high-end from low-end ultrasonic meters.

3.2 Multi-Path Ultrasonic Meter Accuracy and Gas Fiscal Metering

Meter Configuration	Number of Paths	Accuracy (% reading)	Typical Application
Single-path	1	±2-5%	Non-fiscal process control. Clamp-on meters for temporary measurement.
Dual-path (2-chord)	2	±0.5-1.0%	Production allocation. Custody transfer for lower-value applications.
4-path (Gauss-Jacobi)	4	±0.2-0.3%	High-value gas custody transfer. Meets OIML R137 and AGA-9 requirements. Standard for North Sea gas export metering.
6 or 8-path (fiscal grade)	6-8	±0.10-0.15%	Gas fiscal metering at high-value export terminals, LNG custody transfer, pipeline interconnects. Equivalent to Coriolis accuracy for liquid. Preferred for large diameter gas lines where Coriolis is impractical.

4. Custody Transfer Standards and Meter Proving

4.1 API MPMS Chapter 4 - Meter Proving Requirements

Meter proving is the process of comparing a meter's indicated volume against a reference volume of known accuracy. All liquid custody transfer meters must be proved at regular intervals, and a meter factor (MF) is applied to all subsequent readings to correct for systematic measurement errors. The proving frequency depends on meter type, fluid stability, and fiscal risk:

Meter factor calculation from bidirectional pipe prover:
MF = V_prover / V_meter_indicated

Where:
V_prover = certified volume of prover (determined by water draw calibration against NIST traceable standards)
V_meter_indicated = volume indicated by meter during prover pass

Example proving run:
Prover certified volume: 15.4387 bbls (determined at 60°F base temperature)
Temperature correction (CTPL): 0.9983 (crude oil at 72°F expanding to 60°F base)
Pressure correction (CPL): 1.0002 (pressure effect on volume)
Prover base volume: 15.4387 x 0.9983 x 1.0002 = 15.4144 bbls at 60°F

Meter indicated during this pass: 15.3891 bbls

MF = 15.4144 / 15.3891 = 1.00164

This means the meter under-reads by 0.164%. Without correction, 1,000 bbls indicated = 998.36 bbls actual → loss of 1.64 bbls per 1,000.

Proving acceptance criteria (API MPMS 4.8):
Minimum 5 consecutive runs required
Repeatability of MF across 5 runs: standard deviation ≤ 0.027% (for Coriolis and turbine)
OR: Range of MF values ≤ 0.05% across 5 runs

Run results: MF = 1.00164, 1.00167, 1.00161, 1.00165, 1.00163
Average MF = 1.00164
Range = 1.00167 - 1.00161 = 0.000060 = 0.006% → PASS (within 0.05% limit)

Standard deviation = sqrt(sum of squared deviations/4) = 0.0000021 = 0.00021% → well within 0.027% limit

The meter factor 1.00164 is applied to all custody transfer readings until the next proving.

4.2 Prover Types and Selection

Prover Type	Operating Principle	Volume per Pass (typical)	Best Application
Bidirectional pipe prover	Calibrated pipe section with displacer (piston or sphere) that passes through detector switches at each end. Volume between switches is certified.	10-1,000 bbls	Standard for crude oil and product custody transfer. API MPMS Chapter 4.2. Must pass ≥10 pulses between detectors.
Small volume prover (SVP)	Piston-cylinder with precision-machined bore. Much smaller than pipe prover (1-5 bbls). Requires electronic pulse interpolation to achieve required pulse counts.	0.5-5 bbls	Offshore platforms and locations where pipe prover is impractical due to size/weight. API MPMS 4.6. Requires validation against master meter or pipe prover.
Master meter	A high-accuracy Coriolis or turbine meter previously calibrated against a pipe prover. Used as a transfer standard to prove field meters that cannot be pipe-proved.	N/A (continuous flow)	Remote wells and subsea facilities. API MPMS 4.5. Accuracy depends on stability of master meter between calibrations - typically recalibrated every 6-12 months.
Volumetric tank (tank gauging)	Level measurement in calibrated tanks using radar, servo gauge, or manual dip. Volume derived from tank calibration table (strapping table).	100-500,000 bbls	Crude oil loading/unloading at terminals. API MPMS Chapter 2. Typical uncertainty ±0.1-0.3% depending on measurement method and tank geometry.

5. Gas Chromatography and Energy Content Measurement

5.1 Gas Quality Measurement for Custody Transfer

Custody transfer of natural gas is based on energy (GJ or MMBtu), not volume. The energy content of the gas depends on its composition, which varies over time as well production changes, reservoir pressure declines, and seasonal demand alters pipeline injection patterns. Online gas chromatographs (GC) continuously analyze the gas composition and calculate the heating value for billing purposes:

Superior Calorific Value (SCV) calculation from GC composition:
SCV (MJ/m3 at 15°C, 1 atm) = sum(yi x SCVi)

Where yi = mole fraction of component i, SCVi = calorific value of pure component i

ISO 6976 reference values (gross, MJ/mol at 15°C):
CH4 (methane): SCV = 890.36 kJ/mol
C2H6 (ethane): SCV = 1,559.7 kJ/mol
C3H8 (propane): SCV = 2,220.0 kJ/mol
CO2: SCV = 0 (non-combustible)
N2: SCV = 0 (non-combustible)

Example gas composition (mole fractions) and SCV calculation:
CH4: 0.8820, C2H6: 0.0680, C3H8: 0.0180, C4H10: 0.0055, C5+: 0.0015, CO2: 0.0150, N2: 0.0100

Verify: 0.8820+0.0680+0.0180+0.0055+0.0015+0.0150+0.0100 = 1.0000 ✓

SCV per mole of gas mixture:
= 0.8820 x 890.36 + 0.0680 x 1,559.7 + 0.0180 x 2,220.0 + 0.0055 x 2,877.4 + 0.0015 x 3,537
= 785.30 + 106.06 + 39.96 + 15.83 + 5.31 + 0 + 0
= 952.46 kJ/mol mixture SCV

Convert to energy per standard m3:
At 15°C, 1 atm: 1 m3 = 1/0.022414 m3/(mol x 288.15/273.15) = 41.571 mol/m3
SCV (MJ/m3) = 952.46 kJ/mol x 41.571 mol/m3 / 1,000 = 39.60 MJ/m3 (higher heating value)

UK/European typical: 37-42 MJ/m3. US: expressed in BTU/scf → 39.60 MJ/m3 / 0.03726 = 1,062 BTU/scf

Billing impact of ±1% composition uncertainty:
At 100 MMscf/day and $5.00/MMBtu:
Daily energy = 100 x 10^6 scf x 1,062 BTU/scf / 10^6 = 106,200 MMBtu/day
Daily revenue = 106,200 x $5.00 = $531,000/day
1% SCV uncertainty = ±$5,310/day = ±$1.94 million/year

This quantifies why online GC calibration frequency, response time, and carrier gas purity are fiscal metering issues, not just analytical laboratory quality metrics.

Conclusion

The orifice plate uncertainty calculation in this article - 0.534% combined uncertainty driven by the 0.5% discharge coefficient uncertainty - demonstrates the fundamental limitation of correlation-based discharge coefficient values for custody transfer applications. The ISO 5167 discharge coefficient correlation was developed from laboratory data covering a range of beta ratios and Reynolds numbers and carries an inherent uncertainty of ±0.5% even when the plate is manufactured to specification and installed correctly. This uncertainty alone exceeds the ±0.3% total uncertainty budget allowed by API MPMS for fiscal metering. The solution - replacing the correlation with a laboratory-calibrated discharge coefficient carrying ±0.10-0.15% uncertainty - reduces the total uncertainty to approximately 0.22%, within the fiscal metering budget. This is why custody transfer orifice plates must be sent to accredited flow calibration laboratories periodically, not simply dimensionally verified against API 14.3 plate specifications.

The Coriolis meter prover calculation - a meter factor of 1.00164 indicating 0.164% under-reading - illustrates why meter proving is mandatory rather than optional for custody transfer. A custody transfer rate of 81,556 bbl/day at 0.164% under-reading loses 133.8 bbl/day = 48,834 bbl/year = $3.17 million/year at $65/bbl. This systematic error is invisible to the user without proving - the meter appears to function correctly, calibration certificates are current, transmitter diagnostics show no faults. Only the proving comparison against the certified prover volume reveals the 0.164% systematic bias. The five-run proving sequence required by API MPMS verifies both the meter factor value and its repeatability - a meter with good repeatability and high MF bias is easy to correct; a meter with poor repeatability cannot be reliably corrected regardless of the MF value applied.

Want to access our flow measurement toolkit with ISO 5167 orifice plate calculator, Coriolis uncertainty budget, ultrasonic transit-time model, and meter proving MF calculation, or discuss instrumentation selection for a specific custody transfer application? Join our Telegram group for instrumentation and measurement discussions, or visit our YouTube channel for step-by-step tutorials on flow meter selection, uncertainty analysis, and custody transfer proving.