A calibration teaches a gas chromatograph what to report; a validation checks whether it is still telling the truth without changing anything. In a custody validation run, the analyzer is fed a gas of known composition, its answer is compared against the certificate, and the result is logged as a pass or a fail. Crucially, the operating calibration is left untouched. This non-adjusting check is what lets an operator prove the analyzer was accurate at a given moment, which is exactly what a contract or a regulator wants to see for audit.
GC Analysis Validation in one line: A GC validation run is a non-adjusting check in which a gas of certified known composition is injected, the analyzer's result is compared to the certificate, and the outcome is recorded as a pass or fail, without altering the operating calibration. It differs from a calibration, which actually re-derives the response factors and changes how the analyzer reports. Regulators and contracts require periodic validations because they document that the running calibration was still accurate at a known point in time.
Calibration and validation both involve running a gas of known composition through the analyzer, which is why they are so easily confused, but their purposes are opposite. Calibration is an adjusting operation: the analyzer measures the known gas and re-derives its response factors so that its output matches the certificate going forward. After a calibration, the analyzer behaves differently than it did before, because the numbers it uses to convert peak areas into mole percents have been rewritten.
Validation is a non-adjusting operation. The analyzer runs the known gas and reports what it would report for any sample, using the response factors it already has, and that reported composition is simply compared against the known values. Nothing is rewritten. If the comparison is within the agreed tolerance, the run passes and the operating calibration is confirmed to still be good; if it falls outside tolerance, the run fails and flags that a calibration or investigation is needed. The analyzer leaves a validation exactly as it entered it.
This distinction is not academic, because it changes what the record means. A calibration record says the analyzer was made correct at that moment. A validation record says the analyzer was found to still be correct at that moment, without anyone touching it. For custody purposes the second statement is the more powerful one, because it demonstrates that the readings taken during ordinary operation, on the very calibration that was running, were trustworthy. That is why a validation is often the evidence a dispute or an audit actually asks for.
A custody analyzer runs unattended for long stretches between calibrations, and money moves the whole time. A calibration proves accuracy on the day it is performed, but it says nothing about the days that follow, during which the detector can age and the response can drift. Periodic validation fills that gap by re-proving accuracy at intervals without disturbing the calibration, so both parties to the measurement have documented evidence that the analyzer stayed within tolerance across the period, not just on calibration day.
Because a validation is non-adjusting, it is also honest in a way a calibration cannot be. A calibration always leaves the analyzer reporting the certificate values, so a fresh calibration by itself cannot reveal how far the analyzer had drifted before it was corrected. A validation performed just before that calibration captures the drift: it records what the analyzer was actually reporting after weeks of running, which is the number that governed the gas that already flowed. That is precisely the information a contract or a regulator needs to know the measurement can be trusted retroactively.
The comparison also exercises repeatability, not just accuracy. Running the same validation gas periodically and comparing the results over time shows whether the analyzer returns a consistent answer, which is a separate quality from being close to the certificate on any single run. An analyzer that passes every validation but with results scattering more widely than before is telling you its performance is degrading even while it remains in tolerance, and that early signal is worth having before a run finally fails.
A validation is only useful if its result survives as a durable, timestamped record, and this is where a SCADA historian earns its place in the measurement chain. Each validation run produces a set of measured values against known values, a pass or fail verdict against the tolerance, and the time it occurred, and all of that belongs in a historian that keeps it unaltered for the life of the contract. When a dispute or an audit arrives months later, the historian answers the question of whether the analyzer was proven accurate during the period in question.
The historian also has to keep validation and calibration events distinct, because conflating them destroys the value of the record. A validation record must show that the operating calibration was not changed, so an auditor can rely on it as evidence about the readings that were actually in use. A calibration record, by contrast, marks a point where the analyzer's behavior changed. Keeping the two clearly labeled and separately trended is what lets an operator reconstruct the exact state of the analyzer at any past moment.
Beyond the audit trail, storing validations in a historian turns them into a health trend for the analyzer. Plotting the validation deviations over successive runs shows drift developing between calibrations, so an operator can schedule a calibration before a validation fails outright rather than reacting after a failed run has cast doubt on a period of measurement. The non-adjusting check that protects the audit record doubles as an early-warning signal, all without ever disturbing the calibration the site is billing on.
A calibration adjusts the analyzer: it re-derives the response factors so future readings match the known gas. A validation does not adjust anything: it runs a known gas, compares the analyzer's existing output to the certificate, and logs a pass or fail. A validation proves the running calibration was still accurate, while a calibration makes the analyzer accurate and resets its behavior.
Because a calibration erases the evidence of drift. Once you calibrate, the analyzer reports the certificate values again, so you can no longer see how far it had wandered. Running a validation first captures what the analyzer was actually reporting after weeks of operation, which is the number that governed the gas already measured, and that is exactly what an audit or dispute needs to know.
Because its value is proving accuracy at a specific past moment, which requires a durable, unaltered, timestamped record. A historian keeps each validation's measured-versus-known values and its pass or fail verdict for the life of the contract, clearly separated from calibration events. Trending those results over time also reveals drift developing between calibrations, so a calibration can be scheduled before a run fails.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.