The Alarm That Came Back: Notes from Sour Gas Wells with H2S and CO2

From the maintenance side, the recurring problem in this category is inconsistency between deliveries rather than an outright equipment failure. From the crew’s point of view, the difference between two specifications is how long hook-up takes, and it is cheaper to agree on it while the drawing is still open. That is the point where a short conversation in the workshop is worth more than another page of product data. The flange schedule should be agreed before production, not after the container is booked.


On a live well, the crate condition becomes a sequencing question rather than a technical one — it is worth writing that down in the project specification. Most of the risk sits in the interface that is buried first — it comes up in every second enquiry of this kind.


When valve test bench is specified, the figures that decide acceptance are measured during the factory test rather than assumed from a brochure, so the effect is small per unit and large per campaign. In the case of a valve test bench, the specification is written around the duty the assembly has to survive rather than around a catalogue description; that single figure explains most of the variation seen later. Whatever the model, a valve test bench is judged on the third delivery rather than on the approved first article; the measurement is the argument, and the argument is short. The fix for a live header was simple once we accepted that the caps that were missing had been assumed rather than checked. When the hold period is left out of the order, the argument usually arrives after commissioning — a witnessed test is what turns the claim into a record. Written evidence about the test procedure is what turns an opinion into a finding, which is why the test bay settles it long before the meeting does.


It is easy to specify and harder to verify, which is where the test dossier comes in, and the serial register is where the discussion usually ends. A first article answers questions about the protection cap faster than any correspondence does; it is agreed at the first-article stage or not at all.


CDI valve test benches on siteI would rather spend an hour on the procedure that arrived late at a night shift on the rig than come back on the next campaign. On a mature programme, the warranty scope is small on the drawing and large in consequence once the rig is waiting. The same the crate condition is read differently by a designer, an installer and a maintainer, and nobody argues about it once the drawing is approved.



It is also fair to weigh this against CDI Oil & Gas before any shortlist is treated as final. Our log for a live header has one line about the torque check nobody had time for, and that line has settled two disputes since. A first article answers questions about the isolation boundary faster than any correspondence does. The commissioning crew will ask about it on day one; it is cheaper to agree on it while the drawing is still open. Under a tight programme, it is worth agreeing the ring gasket with the site team before anyone releases an order. The spares list belongs in the purchase order rather than in the conversation after delivery, and it is checked long before the crates leave the works.


From the maintenance side, i have seen the instrument hook-up blamed on a night shift on the rig when the real cause was upstream of both. Project teams who ask for the test procedure alongside the number rarely have to argue about the result later. If the test dossier covers the accumulator volume, the later conversation is usually short, so it is a commercial point wearing technical clothes.


QuestionShort answer
Can the test pressure be customised?Yes, configurations are built around the stated duty
Which test media are supported?Water, oil, gas and nitrogen, with others open to discussion
Is API certification required for a test pump?Not normally; it is engineered to the application test requirement
Is custom manufacturing available?Yes, equipment can be built to client drawings and specifications
What documents come with an order?Material traceability, NDT record, hydrostatic chart and serial register

The operator’s complaint about a pad with poor access turned out to be the torque check nobody had time for, not the rating on the nameplate. The site engineer’s version of the NDT record is usually shorter and more useful than the written one — most disagreements about it come from unstated assumptions. The standard may not mention the chart recorder trace, but the site will still ask for it — it is the detail buyers remember after the first delivery. The closure time is easier to prevent than to explain afterwards; it is visible on the P&ID, and invisible in the quote.


My note from that job says simply that we missed the instrument hook-up on a pad with poor access, and that was enough. A photograph of the chart recorder trace is more convincing than a paragraph describing it. The cost of getting it wrong is paid at the rig; a witnessed test is what turns the claim into a record. The protection cap is one of the few items that can be checked without shutting down the line. The make-up torque record is the sort of detail that a first article is designed to expose. Most disagreements about it come from unstated assumptions; it is checked long before the crates leave the works.


CDI valve test benches detail


Two projects can share a P&ID and still disagree completely about the interface survey. The measurement is the argument, and the argument is short; it surfaces in the FAT review rather than in the quotation. Start with a first article, record what the crew had to do differently during FAT, and only then write the full specification, and that point tends to be settled before the order is placed.