Relief valves are the turnaround scope everyone underestimates. A mid-size unit outage can put several hundred valves through the shop in a few weeks: pull, as-received pop test, teardown, repair, reassembly, set-pressure test, reinstall. Each valve is small work. The population is a logistics operation, and the paperwork behind it is a regulatory record that auditors read line by line.
The question an auditor asks is simple: for this valve, show me the as-received test, the repair, the final test, who certified it, and that the valve on the stand is the valve that went back on the nozzle. Simple questions are the ones that expose weak records.
Where valve records go wrong
The failure points repeat across sites:
- Identity breaks in transit. A valve leaves its nozzle with a tag, passes through transport, cleaning, and a crowded shop floor, and comes back to a reinstall crew working from a list. Every handoff is a chance for the serial number and the service location to part company. A certified valve on the wrong nozzle is a failure that tests as a success.
- As-received results vanish. The pre-test, whether the valve popped at set pressure as found, is the single most valuable data point for setting future intervals. It is also the one most often lost, because it happens early, in a hurry, before the "real" record starts.
- The shop and the site keep separate books. The repair shop, often a contractor, documents in its own system and delivers certificates at the end. The site tracks progress on a spreadsheet. Reconciling the two after demobilization is where gaps are discovered, too late to fix.
- Certificates without linkage. A box of test certificates organized by shop job number, not by tag and service location, is technically complete and practically useless.
One record per valve, from pull to reinstall
The structure that works treats each valve as a record that travels with the physical valve through every station:
- Pull: the record opens against the tag and service location, photo of the installed valve, condition noted, transport begins.
- As-received test: result recorded before teardown, pass or fail against set pressure, on the valve's record. This is the interval-setting data. It gets captured now or never.
- Teardown and repair: parts replaced, seat condition, spring data, machining performed. The shop technician records against the same record the site opened.
- Final test: set pressure verified, certificate generated from the record, tester identified, results attached.
- Reinstall: the crew confirms tag against location at the nozzle, photo, sign-off. The loop closes where it opened.
Run this way, the population status is a live count: how many valves at each station, which ones are stalled in the shop, which systems can close out. The expediting conversation happens during the outage, when it can change something.
The as-received data is the prize
Beyond compliance, the as-received test results across the whole population are what let a pressure relief program defend its intervals. A population that overwhelmingly passes as-received supports extension arguments. A cluster of as-received failures in one service is a process problem announcing itself. Neither signal exists if the pre-test results were scribbled and lost.
Audit-ready is a side effect
When each valve carries its own complete record, the auditor's question takes thirty seconds to answer, for any valve, years later. Nobody assembled an audit package. The execution discipline produced one.
intelliSPEC handles valve populations this way, one mobile record per valve from pull to reinstall, with tests, certificates, and sign-offs attached in sequence and the as-received results kept for interval analysis. If your valve program still reconciles a shop's books against a site spreadsheet, the single-record structure is the change that matters.
