The clean physics of an orifice plate predicts a certain flow for a given pressure drop, but real fluid never quite obeys the clean physics - friction, the way the jet contracts just past the bore, and the tap placement all shave the real flow below the ideal. The discharge coefficient, Cd, is the empirical number that bridges the gap between ideal and actual. This guide defines Cd, explains how the Reader-Harris/Gallagher correlation computes it from the beta ratio and the Reynolds number, and describes why the flow computer has to solve for it iteratively rather than plug in a fixed value.
Orifice Discharge Coefficient (Cd) in one line: The discharge coefficient, Cd, is the empirical factor in the orifice flow equation that relates the actual flow through the plate to the theoretical flow predicted by ideal physics. It accounts for real effects the ideal equation ignores, such as friction and the contraction of the jet downstream of the bore, and it is typically a bit less than 1. In AGA 3 practice Cd is not a constant but is computed from the beta ratio and the pipe Reynolds number using the Reader-Harris/Gallagher correlation, which the flow computer solves iteratively.
The idealized orifice equation assumes a frictionless fluid whose stream neatly fills the bore, and from the measured pressure drop it predicts a theoretical flow. Reality falls short of that ideal in ways the clean equation cannot capture. The fluid has viscosity, so it loses some energy to friction. More importantly, as the stream squeezes through the bore it does not stop contracting at the plate - it keeps narrowing for a short distance downstream to a point called the vena contracta, where the effective flow area is smaller than the bore itself. Both effects mean the real flow is less than the theoretical flow the ideal equation would compute.
The discharge coefficient is simply the ratio that reconciles the two: actual flow divided by theoretical flow. Because the real flow is less than ideal, Cd comes out a little below 1. Multiplying the theoretical flow by Cd pulls the calculation down to match what genuinely passes through the plate, so Cd is the term that makes the orifice equation agree with measured reality. It is empirical by nature - it was established from extensive laboratory flow tests rather than derived from first principles, because the vena contracta and friction effects resist clean theoretical prediction.
Crucially, Cd is not one universal number. How much the jet contracts and how friction bites depend on the geometry of the specific installation and on the flow regime, so the same plate can have a different Cd at different flow rates, and two plates of different bore-to-pipe ratio have different coefficients. That dependence is what turns Cd from a single tabulated value into a computed quantity - to get it right for a given meter at a given moment, you have to account for the geometry and the flow conditions it is operating under.
The relationship used in modern orifice metering to compute Cd is the Reader-Harris/Gallagher equation, adopted in the AGA 3 and related international standards. It is an empirical correlation - fitted to a large body of flow-test data - that expresses the discharge coefficient as a function of two main quantities: the beta ratio, which is the orifice bore diameter divided by the pipe diameter, and the pipe Reynolds number, which characterizes the flow regime by combining velocity, pipe size, density, and viscosity. The tap configuration also enters, because where the pressure is sensed relative to the plate changes the coefficient.
The beta ratio enters because the geometry of the constriction governs how the jet contracts: a bore that is a small fraction of the pipe produces a different contraction pattern than one that nearly fills the pipe, so Cd shifts with beta. The Reynolds number enters because the balance of inertial and viscous forces in the flow affects both friction and the shape of the contracting jet. At low Reynolds numbers, where viscous effects are more pronounced, Cd is more sensitive; as flow increases and the Reynolds number rises, Cd changes more gently. The correlation captures how the coefficient responds across the operating range.
The result is that Cd for a real meter is a specific number produced by feeding this meter's beta ratio, its tap type, and the current Reynolds number into the correlation - not a value looked up once and forgotten. Two meter runs with the same plate but different pipe sizes, or the same meter operating at different flows, will legitimately use different Cd values. This is why getting the physical configuration right in the flow computer matters: the beta ratio comes from the measured bore and pipe diameters, and the tap type must match the physical taps, or the correlation computes a coefficient for a meter that is not the one actually installed.
A subtle complication makes Cd more than a straightforward lookup: it depends on the Reynolds number, but the Reynolds number depends on the flow rate, and the flow rate is exactly what the equation is trying to compute - and computing it requires Cd. This is a circular dependency. You cannot get the flow without Cd, you cannot get Cd without the Reynolds number, and you cannot get the Reynolds number without the flow. The variables are entangled, so there is no way to solve for the flow in a single clean step.
The flow computer breaks the circle by iterating. It starts with an estimate - an assumed Cd or an assumed flow - computes the Reynolds number from it, uses that to evaluate Cd from the Reader-Harris/Gallagher correlation, and recalculates the flow. That new flow gives a better Reynolds number, which gives a better Cd, which gives a better flow, and the loop repeats. With each pass the numbers change less, converging quickly on a consistent set where the Cd, the Reynolds number, and the flow all agree with one another. Only then does the computer report the flow for that instant. It does this continuously, every calculation cycle, as conditions change.
This iterative solve is one reason orifice metering is done by a dedicated flow computer rather than a simple multiplier, and it is why the meter's configuration must be exact. Every cycle the computer re-solves the loop using the live differential pressure, static pressure, temperature, and the fixed geometry, so an error in the beta ratio, the tap type, or the fluid properties propagates through the iteration into the reported flow. In a cloud SCADA such as Merobix, the flow computer's inputs and the resulting flow are visible and logged, so an operator can confirm the live pressures and configuration that drive the Cd iteration are sensible, and catch a mis-entered bore diameter or tap setting before it quietly biases custody totals.
It represents the ratio of the actual flow through an orifice plate to the theoretical flow the ideal, frictionless equation would predict. Real flow falls short of the ideal because of friction and because the stream keeps contracting to a vena contracta just downstream of the bore, so the discharge coefficient is a little less than 1. Multiplying the theoretical flow by the discharge coefficient brings the calculation into line with what genuinely passes through the plate.
Modern orifice metering computes Cd from the Reader-Harris/Gallagher correlation, an empirical equation fitted to a large body of flow-test data. It expresses the coefficient as a function of the beta ratio - the orifice bore divided by the pipe diameter - and the pipe Reynolds number, with the tap configuration also entering. Because these depend on the specific installation and the current flow, Cd is a computed quantity for each meter and operating point rather than a fixed tabulated number.
Cd depends on the Reynolds number, the Reynolds number depends on the flow rate, and the flow rate is what the equation is trying to compute using Cd - a circular dependency that cannot be solved in one step. The flow computer breaks the loop by starting with an estimate, computing the Reynolds number, evaluating Cd, and recalculating the flow, then repeating until the values converge. It does this every calculation cycle using the live pressures and fixed geometry.
This page references the standards, specifications, and official documentation published by the organizations below. Editions, product capabilities, and documentation change over time - confirm current requirements and specifications directly with the source.
Last reviewed: July 27, 2026. Merobix is not affiliated with, endorsed by, or sponsored by these organizations; their names are used only to identify the standards and products discussed.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.