There is a question we ask early on retrofit projects, and the reaction to it tells us more than the answer does.
Give me the P&ID. Then sign the bottom of it and tell me it is exactly as it appears.
In thirty years of brownfield work, nobody has ever signed it.
What comes back instead is some version of well, I think that’s what we have. Which is an honest answer from someone doing their job properly. It is also not a basis you can engineer against.
This is not a records problem
It is tempting to treat drawing drift as an administrative failure — somebody should have updated the file. That framing is wrong and it leads to the wrong fix.
Plants change faster than records do, and almost every change is legitimate. A valve gets added during an outage because the isolation was inadequate. A line gets rerouted eighteen inches to clear new equipment. An instrument is replaced with a different type because the original is obsolete. A temporary spool is installed on a Friday and is still there four years later because it works.
Each of those is a small, defensible decision made by someone competent under time pressure. None of them individually justifies stopping to revise a drawing set. The drift is the sum of a hundred reasonable calls.
And the institutional memory that explains it degrades on a different schedule. The person who remembers why the line was rerouted retires. The contractor who installed the spool is not on the site anymore. Ten years of context walks out the door in ones and twos.
So the drawing is wrong, and the only record of why it is wrong is gone.
What that costs when engineering starts
A retrofit basis inherits every undocumented change at once.
Pipe stress analysis computed against a routing that is not the routing. Structural loading assessed on framing that has had equipment added to it. Tie-in geometry taken from a drawing rather than measured. Clearances that exist on paper and not in the building.
None of this surfaces during design review, because design review checks the calculation against the basis. If the basis is wrong, the review confirms that the arithmetic is correct.
It surfaces at fit-up. Material is on site. A crane is scheduled. A crew is standing in the building. And the flange is two inches from where the model said it would be.
At that point every available option is expensive. Field-modify the fabricated piece and accept the quality and schedule consequences. Send it back. Or re-engineer the tie-in on the spot with the crew waiting, which is how genuinely bad decisions get made.
The cost is not the error. The cost is when the error is discovered.
The alternative is not a better drawing
The fix people reach for first is to mark up the existing set. Walk the plant, redline the drawing, update the file.
That carries the original errors forward. You are checking a document against itself, and you only find the discrepancies you thought to look for. The ones that hurt are the ones nobody suspected.
A verified basis means something narrower and more useful: the conditions that the engineering depends on have been physically measured, not inherited.
On a retrofit that is four things.
Where the tie-in points actually are. Not where they were drawn. Measured position, orientation, and elevation of every connection the new work has to land on.
What the existing structure is carrying today. Not what it was designed for in 1974. What has been added to it since, and where the load path actually runs.
What the access route will allow. Whether the assembly can physically be delivered, rigged, and bolted up through the openings and clearances that exist — rather than the ones shown on the general arrangement.
What the surrounding geometry leaves you. The envelope you are designing into, including the things nobody drew because they were installed later.
Every one of those is measurable. None of them is reliably documented anywhere.
Why the timing is commercial, not just technical
Long-lead procurement commits to the basis you set at the beginning.
Order a skid, a fan, a section of ductwork or a structural assembly against an assumed dimension, and the assumption ships with it. Every package downstream inherits the same basis. The error propagates quietly through procurement, fabrication and delivery, and the first person to encounter it is holding a wrench.
This is why verifying conditions belongs at the front of the schedule rather than treated as a documentation task to catch up on later. It is not a premium service or a nice-to-have. It is the step that makes everything after it defensible.
A package built on verified conditions can be executed. A package built on a drawing nobody would sign is a bet, and the odds are set by how much the plant has changed since the drawing was drawn.
The question worth asking about your own site
Not whether the P&IDs are accurate — the answer is always a qualified yes, and the qualification is where the risk lives.
Ask when the set was last verified against the field, rather than against the previous revision. And ask whether anyone would know which drawings were stale before they needed them.
If the honest answer to either is uncomfortable, that is not a records problem to schedule. It is the first item on the next retrofit.
Acensium is an industrial engineering firm for complex brownfield retrofits — bulk material handling, high-energy piping, and equipment replacement in operating facilities. Founded in 2003, 1,300+ projects delivered. We connect verified field conditions with practical design, analysis, fabrication control and installation planning.
Engineer the retrofit around reality.
→ Retrofit engineering · Scan-driven engineering