A custody meter's factor should stay put if the meter is healthy, but it rarely stays perfectly still forever. Meter factor drift is the slow, directional creep of successive factors as the meter physically changes with age. Because it is gradual rather than sudden, it hides in the run-to-run scatter until you plot it - and catching that slope early is what lets an operator act before the meter slips out of tolerance.
Meter Factor Drift in one line: Meter factor drift is a gradual, one-directional change in a custody meter's factor over successive provings, caused by slow physical changes such as bearing wear, deposits, or erosion. It differs from an abrupt step change and only becomes visible when factors are plotted against date. Watching the trend lets operators intervene before the meter goes out of tolerance.
The most useful distinction when looking at a meter factor history is between drift and a step. Drift is a slow, roughly continuous trend: each new factor sits a little higher, or a little lower, than the last, and over many proves the values march steadily in one direction. A step is different in character - the factor holds steady, then suddenly jumps to a new level and stays there. The two point at different physical stories, so telling them apart tells you what to look for.
Gradual drift is the signature of slow wear and accumulation. In a turbine meter, bearing wear slowly changes the rotor's response; in any meter, gradual deposition of wax, scale, or solids narrows or roughens the flow path; erosion slowly changes internal dimensions. None of these happen overnight, so their effect on the factor accumulates prove by prove, producing a gentle slope rather than a cliff. Because the change per prove is small, often smaller than the normal scatter between runs, a single new factor rarely looks alarming on its own.
An abrupt step, by contrast, usually means a discrete event: a piece of debris lodged in the meter, a damaged rotor blade, a sudden change in the fluid, or a maintenance action such as a repair or a cleaning that reset the meter's condition. Steps demand a different response - find and fix the event - whereas drift calls for trending and eventual intervention on a schedule. Confusing the two leads to the wrong action, so distinguishing them from the shape of the factor history is the first analytical step.
Drift is essentially invisible from one prove to the next, because the per-prove change is buried in the ordinary scatter of proving runs. It only reveals itself when you plot the accepted factor from each prove against the date it was taken and look at the whole sequence. Then the eye picks out what individual numbers hide: a consistent upward or downward slope that no single point could show. This is why factor history, not the latest factor alone, is the object worth watching.
Reading such a plot is about separating signal from noise. Every prove has some scatter, so the points will not sit on a perfect line even when the meter is stable; the question is whether there is a persistent slope underneath the scatter. A cloud of points with no trend means a stable meter; a cloud that is clearly marching in one direction means drift. The steeper and more consistent the slope, the faster the meter is changing and the sooner it will reach the edge of its acceptable band.
The payoff of seeing the slope is foresight. Once you can measure the rate of drift from the trend, you can project when the factor will reach its tolerance limit and plan an intervention - cleaning, inspection, bearing replacement, or meter swap - before that happens, during scheduled downtime rather than as an emergency. Drift caught early is a maintenance-planning problem; drift caught late, after the meter has already gone out of tolerance, is a measurement-integrity and possibly a retroactive-correction problem. The plot is what turns the former into the outcome instead of the latter.
This is a topic where automation directly changes the outcome, because the whole point of drift is that people miss it by looking at one prove at a time. A cloud SCADA platform keeps the full history of accepted factors for each custody meter and plots them against date automatically, so the trend is always in front of the operator rather than something they would have to assemble by hand from paper reports. Merobix records each prove's factor alongside the meter's live measurement data and makes the history viewable from any browser, which is exactly the view that makes drift visible.
Beyond just showing the trend, the platform can watch it. By defining how far the factor is allowed to move and how close to the tolerance limit is too close, the system can raise an alarm when the trend shows sustained movement in one direction or when the projected path is heading toward the limit. That turns drift from something noticed by chance into something surfaced deliberately, giving maintenance the lead time to act during planned work. It also helps distinguish drift from a step: a run of small same-direction changes reads as drift, while a single large jump reads as an event, and the trend view makes that difference plain.
It is worth being precise about the division of labor. The prover and flow computer perform the actual proving and produce each factor; the cloud SCADA does not compute factors or replace the certified measurement chain. What it adds is the longitudinal view and the alerting on top of it - keeping every factor, plotting the trend, projecting toward the limit, and flagging when intervention is due - so that a slowly wearing meter is caught while it is still in tolerance rather than after it has quietly drifted out.
Drift is a slow, one-directional trend in which each successive factor moves a little further in the same direction, usually from gradual wear, deposits, or erosion. A step change is a sudden jump to a new level that then holds, usually from a discrete event such as debris, damage, a fluid change, or maintenance. They point at different causes and call for different responses, so distinguishing them from the shape of the factor history is the first step.
Because the change per prove caused by gradual wear is often smaller than the normal scatter between proving runs, a single new factor rarely looks unusual on its own. Only when many factors are plotted against date does the consistent slope beneath the scatter become visible. That is why the meter's factor history, not just its latest factor, is what you have to watch to catch drift.
Because a meter that is drifting will eventually reach the edge of its acceptable tolerance, and catching the trend early lets you measure the drift rate and project when that will happen. With that lead time you can plan cleaning, inspection, or a meter swap during scheduled downtime instead of reacting to an out-of-tolerance meter as an emergency. Drift caught late can also force a retroactive volume correction for the period the meter was quietly out of tolerance.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.