Bolted connections fail more often than the vessels and pipework they join. Anyone running a written scheme of examination under PSSR 2000 knows this, even where the flange management procedure sitting in a filing cabinet somewhere doesn't reflect it. A gasket seated wrong, a bolt pattern tightened out of sequence, a joint left disturbed after a modification and never re-verified: these are the paths loss of containment actually takes, far more often than a corroded wall thickness or a cracked weld.
That's not a reason to relax the rest of the process safety programme. It's a reason to treat joint integrity management as its own discipline, with its own procedures, competency requirements, and records, rather than as a subset of general maintenance that gets attention only when a joint has already started to weep.
Where loss of containment actually starts
A flange isn't a single component, it's an assembly: bolts, nuts, gasket, and two flange faces, each with tolerances that only work together if the joint is built and tightened correctly. Get the gasket selection wrong for the process fluid and temperature, or apply torque unevenly across the bolt pattern, and the joint can look fully made up while leaving a leak path that only shows itself under pressure or thermal cycling. On assets with a high volume of breakouts during a shutdown or turnaround, that risk multiplies simply through the number of joints disturbed and remade in a short window.
ASME PCC-1 sets out accepted practice for bolted flange joint assembly, including bolt tightening sequence and torque or tension verification. It isn't a legal requirement in the way PSSR 2000's written scheme of examination is, but where an operator has adopted it as its flange management standard, deviating from it without a documented reason is the kind of thing a written scheme review would reasonably want explained.
What a flange management procedure has to cover
A defensible procedure names the gasket type and material for each joint class, the target torque or tension value, the tightening sequence and number of passes, and who is competent to carry out the work. It also needs to say what happens when a joint is reworked, because a joint that's been broken and remade twice in a maintenance campaign carries a different risk profile to one that's been undisturbed for years.
Where this breaks down in practice isn't usually the procedure itself. It's the record of what was actually done to a specific joint, on a specific date, by a specific technician, and whether that record is still retrievable when someone needs it six months or three years later.
Torque and tension verification is evidence, not paperwork
A torque wrench reading or hydraulic tensioner log only has value if it's tied to the joint, the bolt, and the person who did the work, and if it can be produced on demand rather than reconstructed from memory or a technician's own notebook. Where torque records are kept on paper or in a spreadsheet separate from the maintenance system, verifying that a critical joint was made up correctly after a breakout can mean chasing down whoever did the work and hoping they remember it.
Safety-critical elements with a defined performance standard under the Safety Case Regulations 2015 raise the stakes further. If a flanged connection sits on a safety-critical line, its joint integrity record needs to demonstrate the performance standard was met, not just that a torque wrench was used at some point.
Where joint integrity meets control of work
Breaking a flange to carry out inspection or repair almost always sits inside a permit to work, often alongside an isolation certificate. The permit authorises the work; it doesn't, on its own, prove the joint was remade correctly. That link between the permit that authorised the breakout and the joint integrity record showing how it was remade is where a lot of the practical value sits, and where a lot of systems fall down, because the permit lives in one system and the torque record lives in another, or in someone's site notebook.
We've written before about why isolation is usually where this breaks down in permit to work, and the same pattern applies to joints: the control document and the physical evidence of the work drift apart unless something actively holds them together.
What to check before trusting a joint integrity record
A few practical tests are worth running, whether on an existing procedure or when evaluating a system to support one:
- Can you retrieve the complete torque and tension history for a specific joint across multiple breakouts, not just the most recent one
- Can you link a joint's record back to the permit that authorised breaking it, and forward to the technician's competency and authorisation at the time
- Can the system show which joints were disturbed during a given shutdown, rather than requiring someone to search job by job
- Does the gasket and bolt specification travel with the joint record, or does it live in a separate engineering document that has to be cross-referenced manually
None of these are exotic requirements. They're the difference between a joint integrity programme that can answer a question quickly and one that needs a week of searching to answer the same question.
Joint integrity as part of the wider assurance picture
Joint integrity rarely sits in isolation from the rest of an asset's operational assurance work. It touches the maintenance backlog, particularly where safety-critical joints fall behind their inspection interval, a topic we've covered in what OEUK's KPI guide means for maintenance backlog. It touches control of work every time a joint is broken under permit. And it touches process safety leadership more broadly, in the sense captured by OEUK's process safety leadership principles, which treat bolted joint management as one strand of a wider assurance discipline rather than a standalone maintenance task.
When designing Joint Integrity workflows, we've generally found the systems that work best are the ones that stop asking users to re-enter the same joint identity, gasket spec, or competency record in three different places. If your current flange management procedure relies on cross-referencing a permit system, a maintenance system, and a set of paper torque logs to answer a straightforward question about a single joint, that's worth looking at before the next shutdown, not during it.
If you're reviewing how joint integrity records connect to your control of work programme, we're happy to walk through where the gaps typically sit and what a more connected record actually needs to hold.