Estimating Piping Work for a Gas Plant or GOSP at FEED Stage

The hardest number to produce on any process project is the first one.

At FEED — or worse, at concept select — someone asks how much piping the plant will need. How many tonnes of pipe. How many isometrics. How many engineering hours to design it all, and how many people that means on the org chart by Q3. You don't have a 3D model. You don't have a line list. You barely have firm P&IDs. And yet a defensible figure has to go into the schedule, the proposal, or the owner's cost book — usually by Friday.

This is the problem parametric estimating exists to solve, and it is worth doing properly, because the alternative — a guess wearing the costume of a number — has caused more grief on capital projects than bad welds ever have. What follows is the method I use, why each step works, and the part most people skip: being honest about how accurate the answer can possibly be.

Start from what you actually know

Early in a project, the piping discipline knows almost nothing for certain. But the project as a whole usually knows one thing fairly well: the major equipment count. Process flow diagrams and the equipment list mature long before piping does. By the time anyone is asking you for a piping estimate, the vessels, columns, exchangers, pumps, and compressors are largely defined — even if their nozzles aren't.

So the equipment count becomes the anchor. Everything else is derived from it.

This is also where the single biggest estimating error hides, and it has nothing to do with the factors. It is scope completeness. The estimate is only ever as good as the equipment list it is built on. A forgotten knock-out drum, an off-plot utility package left off the count, a flare system someone assumed was "someone else's scope" — these omissions move the answer far more than any debate over whether the right figure is eight lines per equipment or nine. Before you tune a single ratio, make sure the count in front of you is the whole plant.

The cascade — from equipment to everything else

Once you trust the equipment count, the piping quantities fall out of it in a chain. Each major equipment item connects a roughly predictable number of pipe lines. From the line count, almost everything else follows: lines × average length gives total pipe length, and length × a blended pipe weight gives tonnage; lines drive valves and flanges; length drives fittings, pipe supports, and welds; and lines × isometrics-per-line gives the isometric count, which in turn sets the spool count. A percentage of those lines are stress-critical, which fixes the number of formal CAESAR II analysis packages.

Typical planning anchors sit around six to ten lines per major equipment item, an average line length in the region of 25–35 m, and somewhere between one and a half and two and a half isometrics per line once long runs are split for drafting. None of these are sacred. They are starting points, and I'll come back to where the real numbers come from.

Don't forget the man-hours — and don't confuse them

Two traps swallow inexperienced estimators here.

The first is estimating the material and stopping there. Tonnage and valve counts are only half the question. The project also needs to know the engineering effort — the design, modelling, drafting, material take-off, support engineering, stress analysis, and checking hours that turn a P&ID into a construction-ready isometric. That number sets your resource plan and a large slice of the engineering budget, and it cascades from exactly the same driver: the isometric count.

The second trap is more subtle, and I see it even among experienced people. Construction man-hours and engineering man-hours are not the same thing. The classic references — Page's Estimator's Piping Man-Hour Manual foremost among them — tabulate the labour to fabricate and erect a weld, a flange, a fitting in the field. They are superb for construction estimating. They tell you nothing about the hours to design the line in the first place. Engineering man-hour rates are a different animal, drawn from contractor productivity history and benchmarking outfits like IPA, and they are largely proprietary. Keep the two buckets separate, label them clearly, and never let a construction norm masquerade as a design norm.

A practical note: stress analysis is labour-heavy out of all proportion to its line count. A stress-critical line that needs a formal CAESAR II package and report can consume an order of magnitude more engineering time than a routine small-bore line. If your plant runs hot, large, or sour, the stress-critical percentage climbs, and the man-hours climb with it.

Gas Plant versus GOSP — why the factors differ

A gas plant and a gas-oil separation plant are not the same estimating problem, and using one set of factors for both will quietly mislead you.

A gas plant — with its acid gas removal, sulphur recovery, dehydration, NGL recovery, and compression trains — is piping-dense. More alloy and stainless, more large-bore headers, more sour service with its attendant material and routing constraints, and a higher fraction of hot, large, stress-critical lines. The density per equipment item is higher, and the engineering hours per isometric are higher still.

A GOSP is built around separation trains, produced-water handling, and gas gathering. It is real work, but generally a more moderate piping density per equipment item, with fewer of the exotic-material and high-temperature complications that make a gas plant expensive to engineer. The two deserve different presets, and any tool worth using lets you switch between them — or set your own.

A worked example

Take a mid-sized gas plant of 80 major equipment items at typical complexity. Running it through the equipment-factored cascade gives, in round terms:

These are sanity-check figures, not committed quantities. But notice what they give you immediately: a feel for whether a contractor's proposed team is plausible, a first pass at tonnage for procurement to react to, and a number of isometrics that tells the squad lead how many drafters they'll need and for how long.

The honesty problem — know your class

Here is the part nobody likes to put in writing.

An equipment-factored estimate is, by the global standard for this work — AACE International Recommended Practice 18R-97 — a Class 4 estimate. Its honest accuracy is roughly −30% to +50%. That is the nature of the method, and no amount of careful factor-tuning changes it.

The most damaging thing you can do is present a Class 4 number with Class 2 confidence.

There is a trap in the project phase, too. By definition maturity, FEED corresponds to Class 3 (−20% / +30%). But maturity and method are different things. If, at FEED, all you have done is apply equipment factors, your output is still a Class 4 estimate — you reach Class 3 accuracy only with a model-based material take-off, not with ratios. State the class. State the range. Let the reader set their expectations honestly.

One more lever deserves a warning. The biggest single swing in your tonnage is the average pipe weight you assume — that is, your guess at the size mix. A blended 35 kg/m and a blended 45 kg/m differ by nearly 30% on the same length of pipe. Large-bore headers pull that average up fast. So state your assumed size mix explicitly; it is doing more work than any other input.

Calibrate against your own jobs

The textbook factors get you into the right postcode. Your own completed projects get you to the front door.

If you have closed out a comparable gas plant or GOSP, you are sitting on the best estimating data that exists for your next one: the actual line list, the real isometric count, the as-built MTO, and the timesheets that recorded what the engineering actually took. Back-calculate your own lines-per-equipment, isos-per-line, and MH-per-isometric from that history and you will out-estimate any published norm. This is precisely why a credible tool exposes every factor instead of hiding them — your calibrated ratios are the asset, and the defaults are only there to get you started.

How to use it well

Used in the right spirit, an equipment-factored estimate is one of the most useful things a piping lead can produce at FEED. Treat it as a sanity check on a contractor's proposed quantities and team, an early input to resource loading, a screen on a bid, and a basis for an honest conversation about scope. Re-run it as the design matures, expect the number to move, and — most importantly — track why it moved. The drift between your equipment-factored estimate and the eventual model-based MTO is itself a lesson worth keeping.

What it is not is a committed quantity. Anyone who treats a Class 4 estimate as a firm number has misunderstood the tool, and usually finds out the hard way.