Every proving produces a slightly different factor, so the real question is never whether a factor changed but whether it changed by more than normal scatter would explain. A meter factor control chart answers that by applying statistical process control to the sequence of proved factors: a center line for the expected value and control limits around it that mark the boundary of ordinary variation. A point outside those limits is a signal, not noise, and it triggers an investigation.
Meter Factor Control Chart in one line: A meter factor control chart plots successive meter factors over time with a mean center line and upper and lower control limits derived from the factors' own scatter. Points inside the limits represent normal proving variation; a point outside signals a real shift in the meter and prompts investigation. It brings statistical process control to custody meter proving.
Statistical process control was built to answer one recurring question: is this variation normal, or is something actually different? A control chart does it by establishing what normal looks like from the data itself. For a custody meter you take a run of accepted factors from a period when the meter was known to be behaving, compute their average as the center line, and compute control limits set out from that center by a multiple of the factors' own variability. Those limits define the band within which the factor should stay if nothing has changed.
The power of the chart is that it separates two things people otherwise confuse. Common-cause variation is the ordinary, expected scatter that every proving has - small differences from run to run driven by minor changes in flow, temperature, and the measurement itself. Special-cause variation is a genuine shift in the meter or the process, something that has actually changed. Points bouncing around inside the control limits are common cause and require no action; a point that breaks through a limit is special cause and demands an explanation. The limits are the objective line between the two.
This matters because without a chart, people react to the wrong signals. A factor that happens to land a bit higher than the last one looks like a change and tempts someone to adjust or investigate, when it may be pure scatter. Conversely, a slow real shift can hide because no single value looks dramatic. The control chart replaces that eyeballing with a rule grounded in the meter's own historical variability, so the reaction threshold is set by statistics rather than by gut feel or by whichever number happened to look surprising.
The most basic out-of-control signal is a single point outside the upper or lower control limit. Because the limits are set so that ordinary scatter almost never crosses them, such a point is strong evidence that the meter has genuinely shifted rather than just varied, and it should stop the routine acceptance and start an investigation. That single rule catches abrupt changes cleanly, which is exactly the kind of change that a step in the meter's condition produces.
More subtle patterns catch gradual change that no single point flags. A long run of consecutive points all on the same side of the center line, or a steady march of points trending toward one limit, signals a shift or a drift even while every individual value is still technically inside the band. These pattern rules are what let a control chart detect slow degradation early, before it ever produces an out-of-limit point, which is precisely the drift case where waiting for a single dramatic value would mean catching the problem far too late.
There is an important discipline in how the limits are used: they are calculated from the meter's actual scatter, not set to the contractual or regulatory tolerance. Control limits tell you when the meter's behavior has changed; the tolerance tells you when the factor is unacceptable for custody. A meter can cross its control limits while still inside tolerance - that is the early warning - and it can also sit well inside its control limits and be perfectly fine. Keeping the two ideas separate is what makes the chart a leading indicator rather than a redundant restatement of the pass or fail check.
A control chart is only useful if it exists and is watched, and doing it by hand for every custody meter is exactly the tedious, easily-skipped task that gets dropped when people are busy. A cloud SCADA platform automates it: it collects each accepted factor as proves happen, maintains the center line and control limits, and updates the chart continuously so the current picture is always available. Merobix records each prove's factor with its measurement data and can present the resulting chart from any browser, so the statistical view lives right next to the meter it describes.
Automation also handles the flagging that turns a chart into an alarm. Rather than requiring someone to notice an out-of-limit point or a run of points on one side, the system can apply the control-chart rules automatically and raise a notification when a factor breaks a limit or when a pattern indicates a shift or drift. That routes the signal to a person while there is still time to act, and it does so consistently for every meter rather than only for the ones someone remembered to plot.
As always, the platform is the analysis and alerting layer, not the measurement itself. The prover and flow computer produce the factors through the certified proving process; the cloud SCADA takes those factors and applies SPC to them. This complements the plain trend view - a simple plot of factor versus date shows movement, while the control chart adds the statistical limits that say whether the movement is real - and it complements the pass or fail tolerance check by giving an earlier, statistically grounded warning that the meter is starting to change before it ever reaches an unacceptable value.
Control limits are calculated from the meter's own proving scatter and mark when its behavior has statistically changed, whether or not the factor is still acceptable. The tolerance is the contractual or regulatory boundary beyond which the factor is unacceptable for custody. A meter can cross its control limits while still inside tolerance, which is exactly the early warning the chart is meant to give, so the two should never be set equal to each other.
The clearest signal is a single factor falling outside the upper or lower control limit, which strongly indicates a real shift rather than normal scatter. Pattern rules add more sensitivity, such as a long run of consecutive points all on one side of the center line or a steady trend toward a limit, which catch gradual drift before any single point crosses. Any of these should stop routine acceptance and start an investigation.
Because the tolerance check only catches a problem once the factor is already unacceptable, whereas a control chart catches a meter that is starting to change while it is still within tolerance. The chart separates ordinary run-to-run scatter from a genuine shift using the meter's own variability, giving an earlier, statistically grounded warning. Used together, the control chart is the leading indicator and the tolerance check is the hard limit.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.