What Is a FEED Study and Why Does It Matter?

What Is FEED?

Front End Engineering Design — universally abbreviated to FEED — is the engineering phase that bridges conceptual design and detailed engineering. It is where an idea becomes a project: defined well enough to estimate costs reliably, schedule realistically, and identify the major technical risks before significant capital is committed.

In the upstream oil and gas project lifecycle, FEED typically sits between a Pre-FEED (or Conceptual Study) and Full Field Development. It is the last engineering phase before a Final Investment Decision (FID). The studies a FEED contains — process simulation, flare and relief loads, HAZID and the early process safety work — are where the technical risk of the project is actually pinned down.

The Project Lifecycle Context

Understanding FEED requires understanding where it sits in the broader sequence:

Conceptual Study / Pre-FEED Screening-level engineering. The goal is to evaluate technical feasibility and produce a rough order of magnitude (ROM) cost estimate — typically accurate to ±40%. Multiple concepts may be compared at this stage.

FEED Defined scope engineering. The goal is to produce a complete engineering package that supports procurement, contracting, and a bankable cost estimate — typically ±15 to ±20% (Class 3 AACE). Only one selected concept is taken through FEED.

Detailed Engineering (EPC/EPCM) Full engineering to construction. The goal is to produce every drawing, document, and specification needed to build, procure, and commission the facility.

FEED is the critical decision gate. It is expensive enough to require real investment, but cheap enough — relative to full development — that the cost of a bad FEED is recoverable. The same cannot be said of cost overruns discovered in detailed design or, worse, during construction.

What a FEED Package Contains

A complete FEED package typically includes:

Process Documentation

Process Design Basis (PDB)
Block Flow Diagrams (BFDs)
Process Flow Diagrams (PFDs) with mass and energy balances
Piping and Instrumentation Diagrams (P&IDs) — developed, not necessarily IFC
Utility Flow Diagrams (UFDs)
Equipment List with preliminary datasheets
Chemical consumption summary

Engineering Studies

Process simulation model (steady-state, and dynamic where required)
Equipment sizing calculations
Flare and relief load study
Safety concept (process safety philosophy)
HAZID study report

Layout and Civil

Plot plan and equipment arrangement drawings
Civil and structural design basis
Tie-in list (brownfield projects)

Cost and Schedule

Capital Cost Estimate (±20%)
Project Execution Strategy
Master Schedule (pre-FEED to first production)
Long Lead Items register

Why FEED Quality Matters

The accuracy of a FEED study determines the quality of the investment decision. A poorly scoped FEED — one that misses process complexity, underestimates equipment counts, or overlooks tie-ins — produces a cost estimate that will not survive contact with the real world. The consequences appear months later, during detailed engineering, as scope growth and budget overruns.

The most common causes of poor FEED quality are:

Scope definition failure. The process design basis is not agreed with the client before engineering begins, leading to repeated redesign as requirements change.

Junior delivery. FEED is frequently staffed with mid-level or junior engineers supervised at a distance. This saves money in the short term and creates problems in the long term. A FEED process package written by someone who has never commissioned a separator will contain assumptions that an experienced engineer would never make.

Optimism in estimate. ROM cost estimates are sometimes produced to satisfy commercial requirements rather than to accurately reflect scope. A low estimate that enables project sanction is not a good estimate — it is a liability.

Inadequate simulation. A process simulation model that has not been validated against vendor data or operating cases is not a reliable basis for equipment sizing.

When Do You Need a FEED?

A FEED study is appropriate when:

The project is large enough to justify the investment in definition (typically above $10M capital cost)
The client intends to seek competitive tenders for EPC or EPCM execution
A financial institution or partner requires a defined scope and estimate as a condition of funding
The project has significant technical complexity that cannot be adequately characterised at conceptual level

For smaller projects — modular early production facilities, package equipment, retrofit modifications — a Pre-FEED or Conceptual Study may be sufficient to proceed directly to procurement.

FEED for Fast-Track Projects

A common misconception is that fast-track projects cannot afford a proper FEED. The opposite is usually true: a fast-track project that skips FEED trades a short-term schedule gain for long-term execution risk. The engineering definition that FEED provides is exactly what enables fast contracting, parallel working, and early long-lead equipment procurement.

For fast-track EPFs and MOPU topsides, FEEC has executed FEED studies in four to eight weeks — compressed, but complete. The key is parallel working between disciplines and experienced engineers who do not need to learn on the job.

What FEED Is Not

Almost as important as knowing what a FEED delivers is knowing where its boundary sits — because misunderstanding that boundary is how scope and budget disputes start.

FEED is not detailed engineering. P&IDs are developed, not Issued-For-Construction. Datasheets are preliminary, sized for procurement enquiry, not final fabrication. Isometrics, detailed structural design, and full 3D model clash detection belong to the EPC/EPCM phase. A client expecting IFC deliverables out of FEED has misjudged the gate.
FEED is not a guarantee of the final cost. A Class 3 estimate is ±15–20% by design. It is accurate enough to sanction and to tender competitively, but the residual range is real — it narrows to ±10% or better only when detailed engineering has priced the bulks and locked the equipment.
FEED is not where you re-open concept selection. The single selected concept enters FEED; relitigating MOPU-vs-fixed-platform or onshore-vs-offshore mid-FEED wrecks the schedule and the estimate. Those trade-offs belong in Pre-FEED.
FEED is not a substitute for the safety lifecycle. A HAZID and the relief/flare philosophy are part of FEED, but the full HAZOP, LOPA, and SIL work largely follow once the design is frozen in detailed engineering. FEED sets the safety concept, not the closed-out safety case.

Holding this boundary firmly is what keeps a FEED from quietly sprawling into a half-done detailed-engineering exercise that is neither one thing nor the other.

Worked Example — How a FEED Estimate Earns Its ±20%

Scenario (illustrative): a 15,000 bopd onshore oil EPF taken through a four-week fast-track FEED. The question the FID committee asks is blunt: how confident are you in the number? A FEED estimate earns its AACE Class 3 ±20% label only if the scope behind it is genuinely defined. Walk the chain:

Process basis fixes the equipment count. A simulation converged against the agreed design basis produces a heat-and-material balance, which sets the duty on every vessel, pump, and compressor. From the separator sizing through to the relief and flare loads, each major item now has a preliminary datasheet — the foundation of a priced equipment list.
The equipment list anchors the estimate. Major-equipment costs come from budget quotes against those datasheets. Bulk materials (pipe, valves, cable, steel) are factored from the equipment cost using factors calibrated to the facility type. Installation, commissioning, and indirects are layered on. Each layer carries a defined accuracy band.
The contingency reflects what is still open. A Class 3 estimate typically carries 10–15% contingency. If the produced-water spec is still unconfirmed, or a long-lead compressor has only a single budget quote, that uncertainty is named in the basis-of-estimate and sized in the contingency — not buried.

The contrast that proves the point. Run the same project at Pre-FEED (ROM, ±40%) and the equipment list is a count with allowances, not datasheets; the estimate is a parametric scale-up. The reason FEED costs real money is that it does the engineering that converts allowances into priced, defined scope — and that is precisely what shrinks the band from ±40% to ±20%. Skip it and the ±40% number goes to FID dressed up as a ±20% number, which is how projects sanction on estimates that do not survive detailed design.

Common Pitfalls

Unagreed design basis. Starting engineering before the process design basis is signed off guarantees rework as requirements drift. Lock the basis first.
Junior delivery, supervised at a distance. A FEED package written by engineers who have never commissioned the equipment they are sizing contains assumptions an experienced engineer would never make — and they surface as scope growth in detailed design.
Optimistic estimate to enable sanction. A low number that gets the project approved is a liability, not an estimate. The basis-of-estimate and contingency must be honest about what is still open.
Unvalidated simulation. A process model not checked against vendor data or operating cases is not a reliable basis for equipment sizing — and every downstream cost rides on that sizing.
Skipping FEED to save schedule. A fast-track project that skips FEED trades a short-term schedule gain for execution risk; the definition FEED provides is exactly what enables fast contracting and parallel working.
No early safety input. Deferring HAZID and the relief/flare philosophy past FEED means safety-driven scope (extra relief, segregation, materials) arrives as a surprise in detailed design, when it is most expensive.

Conclusion

FEED is the engineering foundation on which a project is built. Its quality determines the accuracy of the investment decision, the competitiveness of EPC tenders, and the smoothness of detailed design execution. Invest in a FEED study that is thorough, properly resourced, and delivered by engineers with the experience to know what they do not know — and what questions to ask before committing capital.