Every barrel or standard cubic foot that enters a facility has to go somewhere - out a sales line, into a tank, up a flare, or into the fuel system. Production balance is the daily accounting exercise that checks whether all of those streams add up, and reconciliation is what an operator does when they do not. This guide explains how a volume or mass balance is closed across inlet, outlet, fuel, flare, and inventory change, what unaccounted-for volume means, and why a persistent imbalance triggers a measurement investigation.
Production Balance & Reconciliation in one line: Production balance is the practice of accounting for all of the fluid entering and leaving a facility over a period, so that inlet equals outlet plus fuel, flare, and the change in inventory. Reconciliation is the process of adjusting or investigating the numbers when they do not close - the leftover difference is the unaccounted-for volume or imbalance. A small, stable imbalance is normal and expected from measurement uncertainty, but a large or growing one points to a metering error, a leak, or an unmetered stream that needs to be found.
The core of a production balance is a conservation statement: over a defined period, usually a gas day or an operating day, what came in has to equal what went out plus whatever accumulated inside. For a tank battery or a gas plant that means adding up the metered inlet from wells or gathering lines and setting it against every outlet - sales at the custody meter, gas burned as fuel, gas or liquids sent to flare, and any recycle or transfer stream. Inventory change closes the loop: if the tanks held more oil at midnight than they did the day before, that increase is a legitimate destination for volume that was measured in but never shipped out.
Each of those streams has its own measurement, and each measurement has its own accuracy. A custody sales meter may be a proven turbine or Coriolis meter reading within tight tolerance, while fuel gas might be estimated from a small orifice plate and flare volume derived from a much less accurate ultrasonic or even a calculated figure. When the operator sums everything up, the two sides of the equation rarely match to the last unit, and the residual is the unaccounted-for volume. Whether that residual is acceptable depends on how well each stream is measured, which is why the balance is always read alongside the known uncertainty of its inputs rather than in isolation.
A production balance is most useful not as a single day's snapshot but as a trend. A facility that closes to within a fraction of a percent every day, with the sign of the imbalance flipping randomly around zero, is behaving normally - the leftover is just the noise of many imperfect measurements. The moment the imbalance starts to drift consistently in one direction, or jumps to a size the measurement uncertainty cannot explain, the balance stops being an accounting artifact and becomes a diagnostic. It is telling the operator that something physical has changed.
The list of causes an investigation works through is fairly consistent. A meter may have drifted out of calibration, so it is now reading high or low against the true flow. A stream that was always there may have gone unmetered - an open bypass, a recently added recycle line, or condensate dropping out somewhere it is not measured. There may be a genuine loss, such as a leak, a theft, or evaporation that the inventory figures do not capture. Because the balance cannot by itself say which of these is happening, its real value is as an early alarm: it flags that reconciled volumes no longer match reality and points the operator toward the streams worth checking first, usually the least accurate or least frequently proven ones.
Historically the daily balance was a spreadsheet compiled from meter tickets, tank gauges, and flow-computer printouts, assembled hours or a day after the fact. A cloud SCADA platform such as Merobix changes the timing by holding every input the balance needs in one place as it happens - the custody meter totals, fuel and flare volumes, and the tank-level readings that give inventory change - all timestamped and read from the field over Modbus, DNP3, OPC UA, and MQTT. The same conservation arithmetic can then be applied continuously instead of once a day, so the imbalance is visible while the operating day is still running.
That immediacy is what turns a balance from a monthly reconciliation chore into an operational tool. When the running imbalance for a battery starts to widen, an operator can see it the same shift, cross-check the contributing tags, and decide whether to prove a meter, walk down a line, or gauge a tank by hand - long before the discrepancy shows up in an end-of-month statement. Cloud SCADA also keeps the full history of both the raw stream measurements and the computed imbalance, so an investigation can look back and pinpoint the day the numbers began to diverge, which is often the fastest route to identifying the meter or stream at fault.
Shrinkage is a specific, expected loss of volume - such as gas coming out of solution or liquid volume shrinking as it is stabilized - that is quantified with a shrinkage factor. A production balance is the broader accounting check that all streams add up, and it uses corrected volumes that already account for known shrinkage. An imbalance in the balance is what is left over after all the expected effects, including shrinkage, have been applied.
There is no single universal number, because it depends on how accurately each stream in the balance is measured. A facility whose streams are mostly custody-quality meters can expect the balance to close very tightly, while one relying on estimated fuel and flare figures will see a larger residual. The practical test is whether the imbalance stays within the combined measurement uncertainty of its inputs and does not drift or grow over time.
Because every measurement feeding the balance carries some uncertainty, and those small errors accumulate when many streams are summed. Even well-calibrated meters read slightly high or low, tank gauges have a resolution limit, and some streams like flare are estimated rather than metered. A tiny non-zero residual that changes sign day to day is normal; it is a large or persistent one-directional residual that signals a real problem.
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.