Measuring the flow from every well with a dedicated multiphase meter is expensive, and on many wells there simply is no meter at all. A virtual flow meter closes that gap by calculating the flow rate from measurements a well already has - pressures, temperatures, and the choke position - fed through a model rather than a physical sensor in the stream. This guide explains how a VFM infers rate, where its accuracy comes from and where it falls short, and how it supports allocation and well surveillance without new hardware in the flow path.
Virtual Flow Meter in one line: A virtual flow meter, or VFM, is a software model that estimates flow rate from indirect measurements - typically upstream and downstream pressure, temperature, and choke opening - instead of a physical meter placed in the fluid stream. It uses the known physics of flow across a choke and up a wellbore, tuned against occasional real measurements such as a well test, to infer how much oil, gas, and water a well is producing. Because it relies on a model rather than direct measurement, it is less accurate than a good physical meter but provides a continuous rate on wells that would otherwise have none.
The physics a virtual flow meter exploits is that flow through a restriction is governed by the pressure drop across it and the size of the opening. Across a well's choke, the difference between upstream and downstream pressure, combined with how far the choke is open and the properties of the fluid, determines how much passes through. A VFM encodes that relationship in a model - sometimes a simple choke correlation, sometimes a full transient model of the wellbore and flowline - and solves it continuously using the live pressure, temperature, and choke-position readings the well already reports.
Because the raw physics needs to know the fluid's composition and behavior, a VFM is anchored to reality through calibration. A periodic well test, run through a test separator or a portable multiphase meter, gives a trustworthy measurement of the actual rates, and the model is tuned so its inferred output matches that test. Between tests, the VFM tracks changes in the measured pressures and temperatures and adjusts its estimate accordingly, effectively interpolating and extrapolating from the last known good measurement using physical relationships rather than guessing. This is why a VFM is often described as model-based or inferential measurement: the number is computed, not read off a sensor in the flow.
A virtual flow meter is a compromise, and understanding the tradeoff is what keeps it useful rather than misleading. Its accuracy depends on the quality of its inputs, how well the model represents the real well, and how recently and reliably it was calibrated. When the choke is barely open, when the flow regime shifts, or when the well's behavior drifts from the conditions of the last test, the inferred rate degrades. It will never match a well-maintained physical multiphase meter for a single well, and it should not be treated as if it does.
What it offers instead is coverage and continuity. A VFM can estimate rates on every well in a field, all the time, at effectively no incremental hardware cost, which is exactly where physical metering is impractical. That makes it well suited to well surveillance - spotting when a well's production has stepped down, when water is breaking through, or when a choke change did not have the expected effect - because those are relative changes the model captures well even if its absolute number carries uncertainty. It also feeds allocation, where a field's total is measured accurately at a custody point and the VFM estimates are used to split that verified total among wells in proportion to their inferred contribution, so small per-well errors wash out at the field level.
A virtual flow meter is only as good as the live data feeding it, which is where a SCADA platform becomes essential. The wellhead and choke pressures, flowing temperatures, and choke-position feedback that a VFM consumes are exactly the tags a cloud SCADA such as Merobix already gathers from the field over Modbus, DNP3, OPC UA, and MQTT. Delivering those measurements continuously and reliably to the model is what lets the VFM produce a rate that keeps up with the well in real time rather than lagging behind.
Because the platform also stores the history of every input and every well test, it provides both halves of what a VFM needs: the live stream to infer from and the archived tests to calibrate against. An operator can compare the inferred rate to the last physical test, watch the estimate track through a choke change, and flag when the model and a fresh test have diverged enough that a recalibration is due. Presented alongside the well's other tags on a browser or tablet, a virtual flow rate becomes just another live value the operator watches - one that extends production visibility to wells that would otherwise be metered only occasionally, or never at all.
No. A well-maintained physical multiphase meter measures the flow directly and will be more accurate for a single well than a model-based estimate. A virtual flow meter trades that per-well accuracy for coverage and cost, providing a continuous rate on every well at no extra hardware in the stream. It is most trustworthy for tracking relative changes and for allocation where errors average out across a field.
It is tuned against real measurements, most commonly a periodic well test through a test separator or a portable multiphase meter. The model's parameters are adjusted until its inferred rates match the tested rates, and it then tracks changes in the measured pressures and temperatures between tests. The estimate is most reliable soon after a test and degrades as conditions drift away from the calibration point, which is why regular tests still matter.
At minimum it needs the measurements that determine flow through the well's choke - typically upstream and downstream pressure, flowing temperature, and choke position - plus knowledge of the fluid properties. More sophisticated VFMs also use wellbore and flowline models and additional pressure and temperature points along the system. All of these are values a SCADA system commonly already collects, which is part of why a VFM needs no new sensor in the flow path.
This page references the protocol specifications 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.