A custody gas chromatograph is an expensive instrument, and installing one on every meter run at a station with several runs would be costly and hard to maintain. A common alternative is to share a single GC among several streams, connecting it to each meter run in turn through a stream selector so one analyzer measures many points. This saves money but introduces complications: each stream must be flushed clean before it is measured to avoid contamination from the previous one, each stream is only sampled intermittently, and the host has to keep straight which analysis belongs to which meter run. This guide explains how a shared custody GC serves several runs through a selector, why each stream needs a stabilization purge, how stale analysis on a non-selected stream affects its energy calculation, and how a host maps the right analysis to the right meter run with correct timestamps.
GC Stream Switching in one line: GC stream switching is the arrangement in which a single custody gas chromatograph is shared across several meter runs, connecting to each stream in turn through a stream selector or sequencer rather than dedicating an analyzer to each run. Between streams the analyzer flushes the sample path with the new stream so the previous gas is purged out and does not contaminate the next analysis, a step called the stabilization purge. Because each stream is measured only when it is selected, its energy calculation runs on the most recent analysis until a fresh one arrives, and the host must map each analysis to the correct meter run with an accurate timestamp so the right composition is applied to the right volumes.
In a multi-stream setup, sample lines from several meter runs are brought to a stream selection valve manifold ahead of a single chromatograph. A sequencer steps through the streams on a schedule, connecting one at a time to the analyzer, letting it be measured, then moving to the next. Over a full cycle every stream is analyzed once, and then the sequence repeats. The result is that one costly instrument, with one set of columns, one carrier supply, and one calibration to maintain, serves the whole station instead of one GC per run.
The economics are compelling where several runs share similar gas and a single analyzer's cycle time is short enough to visit each run often enough for the application. Rather than buying, housing, and maintaining multiple chromatographs, an operator maintains one and multiplexes it, which reduces both capital cost and the ongoing burden of carrier gas, calibration standards, and spares. For stations with modest run counts and stable gas, this is a well-established and sensible design.
The trade for that saving is that no stream is measured continuously; each is sampled only during its slot in the sequence. The more streams share the analyzer, the longer between visits to any one of them, so the analyzer's availability is divided across the runs. This intermittency is the root of the two main complications of stream switching: the need to purge the path clean when moving from one stream to another, and the need to run each stream's energy calculation on an analysis that is, by design, not the current instant's gas but the most recent measurement of that stream.
When the selector switches from one stream to another, the sample path, the tubing, the valves, and the column front, still contains residual gas from the stream just measured. If the analyzer measured the new stream immediately, that residual would mix with the incoming gas and bias the result, effectively contaminating the new stream's analysis with the composition of the previous one. This carryover is the central hazard of a shared analyzer, because it corrupts a custody measurement with gas that does not belong to that run.
The defense is the stabilization purge. After switching to a new stream, the analyzer flows the new gas through the sample path for a period before taking the measurement, flushing out the previous stream's residual so that what reaches the detector is genuinely the selected stream. The purge time has to be long enough to clear the dead volume of the sample system and let the flow and pressure settle, and it is chosen with margin because an under-purged stream carries a contamination error that a custody measurement cannot tolerate. The purge is therefore not optional overhead; it is the step that makes a shared analyzer trustworthy.
The purge also costs time, and this feeds back into the intermittency problem. Every stream in the sequence spends part of its slot purging and only part of it measuring, so adding streams or lengthening purges stretches the interval between fresh analyses of any given run. There is a genuine tension between purging long enough to guarantee no carryover and cycling fast enough to keep each stream's analysis current, and the design has to balance the two. Skimping on the purge to speed the cycle trades a visible timing problem for an invisible contamination one, which is the worse of the two.
Because each stream is measured only intermittently, its energy calculation between measurements runs on the last analysis taken for that stream. That is acceptable when the gas is stable, because the composition does not change much between visits, but it means the composition applied to a run's volumes is always somewhat aged, and the older it gets the more it can diverge from the gas actually flowing if that gas has shifted. A stale analysis on a non-selected stream applied to changing gas introduces an error that grows with the time since the last measurement, which is why cycle time and gas stability have to be considered together.
The most important discipline is correctly associating each analysis with the stream it came from and timestamping it accurately. A shared analyzer produces a series of results that belong to different runs, and if the host attaches an analysis to the wrong run, it applies one stream's composition to another stream's volumes, a serious custody error. Equally, the host needs to know when each analysis was valid so it can apply the right composition over the right period and recognize when a stream's analysis has aged past a freshness threshold. Getting the stream identity and the timestamp right on every result is the backbone of a defensible multi-stream measurement.
A cloud SCADA platform such as Merobix is well placed to manage this mapping and to make the intermittency visible. It can record each analysis against the correct meter run with its timestamp, apply that run's most recent valid composition to its volumes, and flag when a stream's analysis has grown too old, which can happen if the sequencer stalls on a stream or skips one. It can also watch the purge and cycle behavior indirectly by trending how current each stream's analysis is, so an operator can tell whether every run is being visited on schedule. Presenting per-stream analysis age and the stream-to-run mapping in one view is what keeps a shared analyzer honest across all the runs it serves.
When the selector switches streams, the sample path still holds residual gas from the previous stream, and measuring immediately would let that residual contaminate the new analysis with the previous stream's composition. The stabilization purge flows the new gas through the path for a period first, flushing out the residual so the detector sees only the selected stream. The purge must be long enough to clear the sample system's dead volume and let flow and pressure settle, because an under-purged measurement carries a carryover error a custody point cannot tolerate.
It runs on the most recent analysis taken for that stream until a fresh one arrives, because a shared analyzer only measures each stream during its slot in the sequence. That is fine when the gas is stable, but if the gas shifts between visits, applying an aged composition introduces an error that grows with the time since the last measurement. This is why cycle time and gas stability have to be considered together, and why a host should flag a stream whose analysis has grown too old.
The sequencer tells the host which stream is selected for each measurement, and the host records each analysis against the correct meter run with an accurate timestamp. This mapping is critical, because attaching an analysis to the wrong run would apply one stream's composition to another stream's volumes, a serious custody error. The timestamp also lets the host apply the right composition over the right period and recognize when a stream's analysis has aged past a freshness threshold.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.