Onshore Pipeline - Cost Estimation

Ask three estimators to price the same 120 km trunkline and you will get three numbers. The spread between them rarely traces back to the line pipe — tonnage is arithmetic. It traces back to terrain assumptions, how crossings were carried, whether escalation reached the midpoint of construction, and what got left out entirely. The discipline below is built to surface exactly those things.

01Set the estimate class before anything else

The first decision is not a number — it is a class. The AACE framework ties the maturity of your project definition to the accuracy you are allowed to claim, and that in turn governs how much contingency you carry and how much of the estimate can stay parametric rather than take-off based.

Most over- and under-runs are not estimating mistakes at all. They are a Class 5 estimate being quoted with Class 2 confidence.

02Build it on a complete WBS

A pipeline estimate is only as trustworthy as its work breakdown structure. Price from a structured list of buckets, not from a blank page, and the forgotten items have nowhere to hide. The five top-level buckets:

• Materials / supply — line pipe, coating, fittings, valves, CP, pig traps.
• Construction / installation — the largest single bucket, typically 40–55% of the total.
• Crossings — priced individually, never as a blanket allowance (see §05).
• Land & right-of-way — acquisition, easements, permits.
• Indirects — engineering/PMC, owner's cost, escalation, and contingency.

03Materials — the easy part, if the basis is right

The line-pipe tonnage is the one number you can compute exactly. From the outside diameter and the governing wall thickness — the value driven by design pressure, plus any uprating for cased or crossing sections, not a placeholder — the steel mass per metre follows directly:

Around that core figure sit the supply items that are simple to list and costly to forget: external coating (3LPE / 3LPP / FBE) on an applied-area basis, field-joint coating per joint, internal flow coat where specified, induction bends and fittings, block valves with actuators, the cathodic protection system, insulating joints, and the pig launcher / receiver pair. None are difficult. All are easy to leave off a hurried take-off.

04Construction — and the terrain swing factor

Construction is usually the biggest line in the estimate, and within it the single largest source of variance is the ground itself. The same welded, lowered-in, backfilled metre of pipe can cost two to three times more through rock or sabkha than across open desert. Estimating mainline construction at one blended rate per kilometre buries that swing where no reviewer can see it.

Instead, split the route by ground type and price each band against its own unit rate:

The activity list inside each band is the familiar sequence — ROW survey, clearing and grading, stringing, bending, line-up, welding, NDT (radiography or AUT), trenching, lowering-in, padding and backfill, tie-ins, hydrotest, and reinstatement. One of those deserves a callout of its own.

05Crossings — estimate every one individually

Roads, rail, rivers, wadis, and foreign utilities are where a tidy estimate goes quietly wrong. A bored or HDD crossing can cost an order of magnitude more per metre than open-cut, and a handful of them can carry 10–20% of the entire construction cost. A single blanket "crossings allowance" is the fastest way to be wrong by a number you cannot defend.

Build a crossings register instead — type, count, method, and unit cost for each — and let it total itself. The discipline of listing them forces the conversation about HDD versus bored versus open-cut that should happen at FEED anyway.

06Indirects, escalation and contingency

With direct cost assembled, the wrap-around items finish the estimate:

• Engineering & PMC and owner's cost — as a percentage of direct cost.
• Escalation — and crucially, escalated to the midpoint of construction, not to the award date. On a multi-year line that distinction moves real money.
• Contingency — sized to the estimate class from §01. A Class 3 line does not carry Class 1 contingency, and pretending otherwise is how the number gets quietly tightened past what the definition supports.

07The items most often under-scoped

Before the estimate leaves your desk, run it against the list that catches the recurring omissions:

• Crossings estimated individually, not as a blanket allowance
• Terrain split reflects the actual route survey — rock, sabkha, side-slope
• Hydrotest water sourcing, treatment and disposal carried explicitly
• Escalation taken to the construction midpoint, not the award date
• AGIs and block-valve stations counted
• Logistics and lay-down allowed for remote sites
• Customs and import duties where applicable
• Contingency sized to the stated estimate class

08The cross-check that catches what you missed

A bottom-up estimate validates against a top-down one. Once you have a total installed cost, back-calculate two parametric metrics and hold them against a recent comparable line in the same region and vintage:

Cost per inch-km is the more useful of the two, because normalising by diameter lets you compare lines of different sizes directly. If your bottom-up figure and the analogue disagree by more than the Class band allows, something is missing — and in practice it is almost always one of three things: the crossings, the terrain factor, or the escalation. Unit rates are always regional and dated, so state the basis of the analogue you compare against; a benchmark from another country or another year is worse than no benchmark at all.