A custody gas chromatograph has to report a full composition quickly, several times an hour, so it cannot afford to wait for the heaviest hydrocarbons to crawl through the column one by one. Column backflush is the trick that solves this: after the light components have passed, the analyzer reverses the carrier flow through a short precolumn to sweep the heavy ends back out as a single group. This shortens the analysis and lumps the heavies into one C6-plus number, but it depends on valve timing that has to be exactly right.
Column Backflush in one line: Column backflush is a technique in which a custody GC reverses carrier gas flow through a precolumn at a set moment so the heavy hydrocarbons that entered the precolumn are swept back out together rather than eluting individually. This lets the analyzer report the heavies as a single C6-plus group and finish its analysis cycle much faster. The valve timing that controls when backflush begins is critical, because a mistimed switch corrupts the C6-plus result and the heating value.
In a chromatograph, lighter components move through the column quickly and heavier ones move slowly, so if you waited for every heavy hydrocarbon to elute in order, a single analysis would take far too long to keep up with a live custody stream. The heaviest components in natural gas are also present in small amounts and are of limited individual interest for custody purposes, where what matters most is their combined effect on energy content, not the exact split between one heavy isomer and the next.
Backflush exploits a two-column arrangement. The sample first passes through a short precolumn that separates the lighter components, which move on to the main column and detector to be measured individually, while the heavier components are still held back in the precolumn. At a chosen moment the analyzer reverses the carrier flow through the precolumn, driving those trapped heavies back out the way they came as a single combined slug rather than letting them creep forward one at a time.
That combined slug becomes the C6-plus group, a single peak representing everything hexane and heavier as one lumped quantity with an assigned average property. Reporting the heavies this way is standard practice for custody analysis, because it captures their contribution to heating value and density without spending analysis time resolving components that the contract does not require individually. The reward is a dramatically shorter cycle time, which is what lets the analyzer deliver a fresh composition every few minutes.
The moment of backflush is set by a valve that switches carrier flow direction at a precise point in the run. That timing has to fall in the narrow window after the last light component of interest has left the precolumn but before the heavy group has migrated past the point where reversing flow can still sweep it back. Getting this window right is the whole game, and it depends on stable column temperature, stable carrier flow, and correct valve actuation.
When the timing is early, the valve reverses flow before a component that should have been measured individually, such as one of the pentanes, has cleared the precolumn, so that component gets swept back into the C6-plus group instead of appearing where it belongs. When the timing is late, part of the heavy group has already moved forward and elutes as its own peak while the rest is backflushed, splitting the C6-plus into a piece that is measured and a piece that is lumped. Either way the composition is wrong, and because the heavies carry a large share of the heating value per mole, even a small misallocation moves the BTU result noticeably.
What makes this failure insidious is that the chromatogram can still look reasonable. The analysis completes, the peaks are present, and the mole percents may still sum near one hundred, so a casual glance sees nothing wrong even though the split between measured heavies and the C6-plus group has shifted. The error hides inside a plausible-looking result, and it persists on every analysis until the valve timing is diagnosed and corrected, biasing the billed energy the entire time.
Because a mistimed backflush does not announce itself, the way to catch it is to watch the analyzer's timing behavior rather than only its composition output. Two signals are especially telling. The overall analysis cycle time should be stable run to run, so a cycle that grows or shrinks is a hint that flow or temperature has shifted and the backflush window has moved with it. The retention times of the boundary peaks around the backflush point are the more direct indicator, since they show whether the switch is still landing in the right place relative to the eluting components.
A SCADA layer that collects these diagnostics turns them into alarms rather than after-the-fact discoveries. By trending the cycle time and the relevant retention times from each analysis, a host can flag the moment the timing starts to wander, before the C6-plus split has degraded enough to matter on the statement. It can also watch for the appearance of an unexpected extra peak where the heavy group has begun to split, which is a direct symptom of a late backflush.
This monitoring pairs naturally with the composition checks a host already runs. When the pre-normalization sum, the response factors, and the backflush timing are all trended together, a shift in one corroborates a shift in another, and the operator gains a coherent picture of analyzer health instead of a single ambiguous number. The valve timing that no one can see from the field becomes a monitored parameter, and a drifting backflush is caught while it is still a maintenance item rather than a billing dispute.
Because it would make each analysis far too slow to keep up with a live custody stream. The heaviest components move slowly through the column and are present in small amounts, and the contract usually only needs their combined contribution to energy and density. Backflushing them out as a single C6-plus group captures that contribution while letting the analyzer finish a full composition every few minutes.
The split between individually measured heavies and the lumped C6-plus group shifts. An early switch sweeps a component that should have been measured back into the C6-plus group; a late switch lets part of the heavy group elute separately while the rest is backflushed. Either misallocation biases the composition and, because heavies carry a large share of the heating value, noticeably skews the BTU result.
By trending the analysis cycle time and the retention times of the peaks near the backflush point, and watching for an unexpected extra peak where the heavy group has begun to split. Because a mistimed backflush leaves the chromatogram looking plausible, these timing signals are the reliable early warning, and a host can alarm on them before the composition error grows large enough to affect billing.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.