The AGA 3 overview describes the orifice method and the standard behind it, but it stops short of the arithmetic that actually turns a differential pressure into a flow rate. That arithmetic is worth understanding, because the equation has a twist: one of its key terms depends on the very flow rate you are trying to find, which forces the calculation to be solved by iteration rather than in a single pass. This guide walks through the AGA 3, also known as API 14.3, mass flow equation, shows how the discharge coefficient, expansion factor, and beta ratio fit together, and explains the iterative loop that resolves the circular dependence between the discharge coefficient, the Reynolds number, and the flow.
Orifice Flow Rate Calculation in one line: The AGA 3 orifice flow rate is calculated from a mass flow equation that multiplies a discharge coefficient, an expansion factor, the bore area, and the square root of the product of fluid density and the differential pressure across the plate, scaled by a term involving the beta ratio. The equation cannot be solved in one step because the discharge coefficient depends on the Reynolds number, which depends on the flow rate being solved for, so it is solved iteratively: assume a flow, compute the coefficient, recompute the flow, and repeat until it converges.
At its heart the AGA 3 equation says the mass flow through the orifice is proportional to the square root of the differential pressure across the plate multiplied by the density of the fluid. That square-root relationship is the fundamental physics of a differential-producer meter: constrict the flow with a plate, and the pressure drop it creates rises with the square of the flow, so the flow is recovered by taking the square root of the measured differential. Everything else in the equation is a set of coefficients that turn that proportionality into an accurate quantity for a real installation.
The bore area and the beta ratio set the geometry. The beta ratio is the orifice bore diameter divided by the pipe inside diameter, and it appears in the equation through a velocity-of-approach term that accounts for the fact that the fluid already has velocity in the pipe before it reaches the constriction. A larger beta, a bigger bore relative to the pipe, changes how the constriction behaves, and the equation captures that through this beta-dependent factor alongside the bore area itself. These geometric terms are fixed once the plate and pipe are known, at a given temperature, since thermal expansion slightly changes the dimensions.
Two more coefficients correct for real fluid behaviour. The discharge coefficient accounts for the difference between the ideal, frictionless flow the simple physics assumes and the actual flow, which contracts to a vena contracta and loses a little to friction and turbulence, so the coefficient is a shade below one. The expansion factor corrects for the fact that a compressible fluid such as gas expands as it drops through the pressure differential, reducing its density at the plate compared with upstream; for a liquid, which is essentially incompressible, this factor is effectively one. Multiply the square-root core by the discharge coefficient, the expansion factor, the geometric terms, and the bore area, and you have the AGA 3 mass flow.
The complication that makes this a computation rather than a plug-in is the discharge coefficient. It is not a constant; the standard gives it as a function of the beta ratio and the Reynolds number of the flow. The Reynolds number expresses the ratio of inertial to viscous forces and characterises the flow regime, and crucially it depends on the flow rate: faster flow means a higher Reynolds number. So the discharge coefficient depends on the Reynolds number, which depends on the flow rate, which is exactly the unknown the whole equation is trying to produce. That circularity is why a single-pass calculation is impossible.
The resolution is iteration. The computer makes an initial estimate of the flow, often by assuming a starting discharge coefficient at a high Reynolds number, and uses it to compute a first flow rate. From that flow it calculates the Reynolds number, from the Reynolds number and beta it recalculates the discharge coefficient, and with the updated coefficient it recomputes the flow. This new flow feeds a new Reynolds number, a new coefficient, and a new flow, and the loop repeats. Each pass brings the flow and the coefficient into closer agreement with each other.
Because the discharge coefficient changes only gently with Reynolds number over normal operating ranges, this loop converges quickly, usually in a small number of passes, and the computer stops when successive flow estimates differ by less than a tight tolerance. At that point the flow rate and the discharge coefficient are mutually consistent: the coefficient corresponds to the Reynolds number of the very flow it produced. This iterative solution runs every measurement cycle in the flow computer, which is why an orifice flow computer is doing far more than a square root behind the scenes, even though the result appears instantly.
In practice the iterative AGA 3 calculation is performed by the flow computer or the multivariable transmitter at the meter run, every measurement cycle, from live inputs. The differential pressure, static pressure, and temperature come from the transmitters; the density or the properties needed to compute it come from the gas composition and the pressure and temperature; and the plate bore, pipe diameter, and reference conditions come from the configuration. The computer feeds these into the equation, runs the iteration to convergence, and produces the flow rate, which it then integrates into volume and energy over the period.
This is where the earlier terms connect to the records that make a measurement defensible. Every input and constant the calculation uses, the beta ratio implied by the plate and pipe dimensions, the composition behind the density, the reference conditions, the differentials and pressures, is exactly what the configuration and quantity records preserve, precisely so the iterated result can be reproduced and audited later. The iterative computation is deterministic: given the same inputs and configuration, it converges to the same flow, which is what lets an auditor recompute an hour and confirm the reported quantity.
A cloud SCADA such as Merobix does not repeat the AGA 3 iteration itself, but it makes the inputs and the result visible and checkable across many meters. By trending the differential, pressure, and temperature that drive the calculation alongside the resulting flow, it lets an operator see when an input is behaving oddly, such as a differential drifting in a way that suggests a plugging impulse line, before it quietly biases the computed flow. Surfacing the live inputs and the computed rate together turns the flow computer's internal iteration into something an operator can trust and, together with the retained records, an auditor can later reproduce.
Because the discharge coefficient in the equation depends on the Reynolds number, and the Reynolds number depends on the flow rate, which is the unknown the equation is solving for. This circular dependence makes a single-pass calculation impossible. The computer instead assumes a flow, computes the coefficient, recomputes the flow, and repeats until successive estimates agree within a tight tolerance, at which point the flow and the discharge coefficient are mutually consistent.
The discharge coefficient corrects for the difference between ideal frictionless flow and real flow, which contracts to a vena contracta and loses a little to friction and turbulence, so it sits just below one and varies with beta and Reynolds number. The expansion factor corrects for a compressible fluid such as gas expanding as it drops through the pressure differential, which lowers its density at the plate. For an essentially incompressible liquid the expansion factor is effectively one, whereas the discharge coefficient always applies.
It is calculated in the flow computer or multivariable transmitter at the meter run, every measurement cycle, from live differential pressure, static pressure, and temperature together with the configured plate and pipe dimensions, gas composition, and reference conditions. The device runs the iterative solution to convergence and integrates the resulting rate into volume and energy. Because the calculation is deterministic, the same inputs and configuration reproduce the same result, which is what makes an audited hour reproducible.
Primary references from the standards bodies and regulators that define this topic:
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.