
Brownfield modelling starts with an awkward truth: Autodesk Plant 3D was built to design new plant. Clean lines, straight runs, square connections, everything laid out on a fresh grid from nothing. That is greenfield: a new site, or a new section, modelled before it is built. Your site is not new. It is brownfield: already […]

Brownfield modelling starts with an awkward truth: Autodesk Plant 3D was built to design new plant. Clean lines, straight runs, square connections, everything laid out on a fresh grid from nothing. That is greenfield: a new site, or a new section, modelled before it is built.
Your site is not new. It is brownfield: already built, already running, very likely twenty or forty years old, modified a dozen times since, and quietly sagging where the pipe has settled. The original drawings stopped matching it a long time ago.
So the real skill here is not modelling. It is modelling reality in software that assumes everything is tidy and new.
This is where the whole series comes to a point. We have said it in nearly every article, the design model and the as built plant are not the same thing, and brownfield modelling is where that stops being a principle and becomes the actual job.
On an old site, pipework is not straight. It sags, it has shifted, it was never installed quite where the drawing said. And when you model a long run of it from the point cloud, you face a genuine choice, because the software does not want to bend the way real pipe has.
There are two honest ways to model it, and which is right depends entirely on what you will use the model for. It is the same purpose first thinking from the earlier articles, applied to brownfield modelling at the level of the pipe itself.
Option 1 is to model all the movement, putting in the small one and two degree deviations to follow how the pipe actually runs. This gives you a model that is genuinely true to the as built site, matched to the point cloud, which is exactly what you want when the model will drive a design change and new work has to fit what is really there.
Option 2 is to keep the runs clean and straight, still positioned accurately to the point cloud but without chasing every wobble. You lose the last degree of fidelity, but you gain a great deal at the drawing stage, because Plant 3D drags those small deviations through into the isometrics, so clean runs give you far cleaner isometrics. This is the same reason the Compass keeps greenfield pipe tidy, applied in reverse to existing pipe.
Neither is wrong. We have even built the same site by more than one method on request, including translating existing isometrics into Plant 3D, so a proposed design could be checked to see whether it fitted or fouled the existing plant. The skill is choosing the method that serves the job, not insisting on one.

This is the part of brownfield modelling that decides whether the whole exercise was worth it. A brownfield model is a snapshot of a moving target. If your site is constantly changing and those changes are not captured back into the model, it slips, slowly at first, away from the plant, until the day your team no longer trusts it and stops using it. At that point the money was wasted.
So a brownfield model is not a one off deliverable, it is something you keep current. That is what turns brownfield modelling into the single source of truth we talked about right at the start of this series, rather than just another drawing that goes out of date.
The reasons for brownfield modelling are exactly the ones we covered in the purpose article, so I will not repeat them all here, beyond the one that is brownfield specific: recording your site as it actually stands gives every future design change a true starting point, instead of a drawing you already know is wrong. For the wider case, visual understanding, inspection isometrics, safety reviews, planning and the rest, that earlier article makes the argument in full.
This is the heart of what we do. We laser scan your site and model it from the point cloud, and we will talk to you first about which approach fits, the fully accurate as built of Option 1, or the cleaner runs of Option 2, depending on what you need the model for. Two things matter when you commission brownfield modelling. Scan with a firm that has actually worked on process sites, because they are not like other buildings, and have the finished model cross referenced against your P&IDs so the two agree. And plan from the start how the model will be kept current, because a brownfield model only holds its value while it still matches the plant. Keeping it matching is part of what we offer.
Brownfield modelling is the craft of capturing a messy, real, working site in software designed for a clean new one. Done well, you get a model that matches your plant. Kept current, you get one that still matches it next year. That is the difference between an asset and an expensive picture.
Next in this series: Exports and Piping Isometrics.


