Plants are modified for forty years by people in a hurry. The as-built drawings record maybe half of it. The estimator who prices the tie-in from those drawings is pricing a plant that does not exist.
This is not a criticism of anyone. It is the natural history of every operating facility. The question is not whether the as-builts are wrong, but where, by how much, and what that uncertainty should cost in the estimate.
Why as-builts drift
The drift has honest causes. Construction never quite follows the drawings: lines get rerouted around obstructions in the field, and the markups that were supposed to flow back to the record set die in someone’s site trailer. Then operations takes over, and forty years of maintenance swaps, capacity tweaks and weekend repairs happen at the speed of production, not the speed of documentation. A pump gets replaced with a different model; a line gets re-routed during a turnaround; a tray fills with cables nobody logged. Each change was small and urgent. The sum is a plant whose paper describes its youth.
The last piece is people. The operators and maintenance leads who remember why that valve is there are retiring, and their knowledge was never on a drawing to begin with.
Find out first
The response is not better guessing. It is going and looking, in four specific ways.
- Laser scan the tie-in areas. It costs a fraction of one rework event, and it converts arguments about what is there into measurements of what is there.
- Walk down every tie-in point with operations, and photograph it. The walk-down finds what the scan cannot: which lines are live, which supports are corroded, which access route is fiction.
- Pull the isolation and permit history. It tells you what has actually been opened, isolated and worked on, which is the closest thing to a maintenance diary the plant keeps.
- Ask maintenance what they have replaced. It will not be on a drawing, and it will be cheerfully told to anyone who asks before the estimate is issued.
A worked example
Take a tie-in scope with three work areas in an operating plant. Scanning the three areas costs, call it, 40,000 dollars, and a week of walk-downs with two people adds, call it, 15,000 more. On this illustrative job, the survey catches three things the as-builts did not show: a process line rerouted a metre off its drawn position, sitting exactly where the new pipe rack was designed to land; a cable tray, live, crossing the planned excavation; and a pump replaced years ago with a model whose nozzles face the other way.
Each discovery is now a design change on paper, costing engineering hours. Undiscovered, each one was a crew standing down inside the work window while the redesign happened live, at premium time, with the schedule bleeding. On a job whose outage day is worth hundreds of thousands of dollars, any one of the three finds pays for the survey several times over. What a lost day inside the window actually costs is its own subject: see Treat the shutdown as its own project.
Price the residual
Even after the survey, some things stay unknown: what is behind the insulation, under the slab, inside the duct. Those get a named allowance each, with a basis, rather than a general brownfield percentage. “Allowance, corrosion under insulation at the tie-in headers, based on the condition found at the two points opened” can be reviewed, challenged and closed out. “Brownfield allowance, ten percent” can only be argued about.
The named residuals also do the productivity factor a favour: they keep discovery risk out of the labour norms, where it otherwise hides and double-counts. How the norms themselves should carry the live-plant burden is its own subject: see The productivity factor on live plants.
A drawing is a record of an intention. A scan is a record of a fact.
Say it in the basis
The basis of estimate lists which areas were surveyed and which were priced from drawings alone. That single list quietly allocates the discovery risk: when a surprise arrives in construction, the list decides whether it is a risk realised, already carried in a named allowance, or a change to the basis, priced as a change. Both the owner and the contractor are protected by knowing which, in writing, from day one. Without the list, every surprise becomes a negotiation, and negotiations in month fourteen cost more than surveys in month one.
Where this goes wrong
Trusting the revision date. A drawing replotted last year looks current, but replotted is not resurveyed. The revision block records the last time someone touched the file, not the last time anyone verified the plant.
Surveying the easy areas. The scan covers the open, accessible spots and skips the congested rack where the actual tie-ins happen, because access was hard to arrange. The survey budget got spent where the risk was not.
The general percentage. One brownfield allowance across the estimate, covering everything and explaining nothing. When it runs out, nobody can say which unknown consumed it, and the next estimate learns nothing.
The unused scan. The point cloud gets collected, admired and archived, and the design proceeds from the old drawings anyway. The survey only pays when the model is checked against it, clash by clash, before the drawings are issued.
The walk-down without operations. Engineers walking the plant alone see what is there. Operations sees what is live, what is fragile and what will be in the way in March. The second list is the expensive one, and it only shows up if they are invited.
Look before you price
Brownfield estimating rewards a simple sequence: survey what can be surveyed, name what cannot, write both in the basis. It is cheaper than the alternative every single time, and the alternative gets discovered at the worst possible moment, by the most expensive possible method.
Emerald Group estimates brownfield and tie-in work with the survey scope, named residuals and basis documented for review, for owners, engineering firms and contractors. If your next estimate leans on forty-year-old drawings, get in touch.