AGA-3 Natural Gas Orifice Meter Run Custody Transfer Flowmeter ANSI 300 600 Flanged

When two parties invoice each other against the same pipeline, a 1% disagreement is a commercial dispute, not a rounding error. This meter run is built as a single accountable assembly: honed meter tube, orifice fitting, concentric plate, DP and static pressure transmitters, RTD and a flow computer applying AGA Report No.3 with AGA-8 compressibility, delivered pressure tested and wet calibrated as one set.

The orifice meter run survives because it is the only gas measuring device whose uncertainty is published, traceable and accepted by both sides of a contract. Turbine, ultrasonic and Coriolis meters are all better on turndown; none of them carries a century of published discharge coefficient data, nor can a plate be inspected, measured and replaced by the buyer's own technician in an afternoon. That is the commercial reason the meter run is still specified at city gates, compressor stations and plant fuel gas headers.

The measurement chain is short but every link is defined. Gas accelerates through the bore, and the corner-tap differential pressure is related to mass flow through the same square-law equation used for liquids, extended by two gas-specific terms: the expansibility factor epsilon accounting for compressibility across the plate, and the density term evaluated at upstream conditions. Volume at base conditions then follows from the AGA-8 compressibility factor and the gas composition.

What separates a meter run from a plate in a pipe is the geometry around it. The discharge coefficient in the equation is only valid while the velocity profile arriving at the plate matches the one used to derive it, which is why the upstream and downstream lengths, the internal bore tolerance and the tap hole squareness are all part of the instrument rather than installation details. Change any of them and the equation is no longer the equation you were invoiced against.

Because DP follows the square of flow, the run is sized for the contracted flow band. Below about 30% of the DP span the square root amplifies transmitter noise and the reading becomes unreliable, and above the span the plate edge and tap holes take a disproportionate beating.

The same plate geometry appears in three different levels of assembly on this site, and the difference is what the buyer is actually paying for: a standard stainless orifice plate in a flanged carrier gives indication duty on plant air and nitrogen, an ASME orifice gas flowmeter covers general gas service to ASME MFC-3M, and the meter run adds the documented geometry and correction method that a custody contract needs. Where the line is large and permanent pressure loss dominates the energy bill, a pitot tube flowmeter installs through a single tapping at a fraction of the cost.

The comparison below is about how much of a contract each option can support, not only about accuracy on a datasheet.

Related pages: ASME orifice gas flowmeter · standard stainless steel orifice plate flowmeter · pitot tube flowmeter for large gas ducts · Venturi flowmeter for low permanent pressure loss · all orifice plate flowmeters · temperature compensated steam orifice plate flowmeter · sanitary tri clamp orifice plate flowmeter

A plate and a flow computer are two components; a meter run is a geometry. It includes the honed upstream and downstream tube sections cut to the length the beta ratio and upstream fitting require, the orifice fitting that holds the plate concentric and square to the bore, corner taps drilled to position, and a dimensional report recording the actual internal diameter the flow computer is configured with. Buy the parts separately and the equation's assumptions become an installation contractor's judgement. Buy the run and the manufacturer signs for the discharge coefficient that the contract was written around, with the paperwork to support it.

For biomethane the answer is generally yes, provided the composition is known well enough to evaluate density and the compressibility factor correctly, since the physical measurement itself is unchanged. For hydrogen the answer is no above a few percent by volume. The published discharge coefficient correlations were derived on natural gas, and the expansibility term and Reynolds number behaviour depart once the hydrogen fraction rises. Hydrogen blending beyond roughly 10-20% by volume needs a method written for it, such as AGA Report No.11 or ISO 20765 based calculation, plus a measurement principle that does not rely on the natural gas coefficient tables. Specify the composition at the enquiry stage rather than after the run is built.

API MPMS 14.3 treats annual inspection as the baseline, and more often where the gas carries liquids or solids. The checks that matter are the bore diameter against its stamped value, the sharpness of the upstream edge, flatness, and any deposit or erosion on the upstream face. An edge radius that has grown beyond about 0.05 mm, a bore that is out of round, or a plate that has bowed is a replacement, not a repair. Keep a spare plate and gasket set on site: a plate costs a fraction of the gas lost while a disputed meter is investigated, and swapping one is a ten minute job with a dual chamber fitting.

Because differential pressure follows the square of flow, resolution collapses as flow falls. Below roughly 30% of the transmitter span a small zero shift or a millivolt of noise translates into a large flow error, and below 10% the reading is indicative at best. Add to that a plate sized for peak winter demand running at summer flows and the effect compounds. The remedies are structural rather than electronic: size the bore for the normal flow band instead of the maximum, install two parallel runs with changeover logic for strongly seasonal loads, or lower the DP span and accept a narrower upper range. Re-ranging a transmitter does not restore turndown that the geometry never had.