Engineering study phases: scoping, prefeasibility, feasibility and FEED, the gates between them, and the design reviews that catch problems early
The engineering study phases from scoping to detailed design, the AACE estimate class and accuracy each typically supports, what each phase delivers, the decisions at the gates between them, FEED and the scope freeze, the design reviews held along the way, and the maintainability review that decides decades of operating cost. With a drawing of estimate accuracy by phase.
Key takeaways
- A capital project moves through scoping, prefeasibility, feasibility and FEED before detailed design, and each phase answers one question: is it worth studying, which option, is it ready to fund, is it ready to build.
- Each phase supports a class of cost estimate. AACE 18R-97 gives typical ranges from about minus 50 to plus 100 per cent at scoping, narrowing to about minus 10 to plus 30 per cent at a feasibility estimate.
- The gates between the phases are decisions, not milestones. A gate that is passed on the calendar rather than on what the phase delivered is the start of most budget overruns.
- Design reviews catch problems while they are still cheap to move. The operability and maintainability review is the one most often skipped and the one that decides decades of operating cost.
- Operations and maintenance belong in the study team from prefeasibility, because the option selected there fixes most of the cost of running the plant.
A new crushing circuit is approved on a feasibility study and comes in a long way over budget. Looking back, the study was not wrong about the crusher. It assumed a layout that the maintenance team would never have accepted, with no crane access to the crusher and no room to pull the conveyors, and the redesign was done in detailed design at several times the cost of doing it in FEED. Nobody from maintenance had seen the model until the 90 per cent review.
This guide sets out the engineering study phases from scoping to detailed design, the estimate class and accuracy each typically supports, what each phase delivers, the decisions at the gates between them, and the design reviews that catch problems while they are still cheap to fix. Phase names vary by industry and by owner, and the accuracy ranges are the typical ranges published by AACE International. The project's own risk analysis sets its real range.
Definition against accuracy
Each phase adds definition to the project, and the estimate gets narrower as the definition grows. AACE International Recommended Practice 18R-97 sets out five estimate classes for the process industries, from class 5 at concept screening to class 1 near the end of design, and defines each by the maturity of the project definition behind it. Accuracy follows from definition, not the other way round. An estimate cannot be made more accurate by spending more time on the estimate. Only more engineering does that.
The ranges are wide at the start. A scoping estimate may fall anywhere from about half the eventual cost to about twice it, which is enough to decide whether a project is worth studying and nothing more. By feasibility and FEED the typical range is about minus 10 to minus 20 per cent at the low end and plus 10 to plus 30 per cent at the high end, which is the basis on which most projects are funded.
Accuracy comes from definition. The only way to narrow an estimate is to do more engineering.
What each phase delivers
Each phase answers one question and hands a defined set of deliverables to the next. The names change between industries. Mining tends to use scoping, prefeasibility and definitive feasibility, and under the JORC Code publicly reported ore reserves must be supported by a study at prefeasibility or feasibility level. Oil, gas and process plants tend to use front end loading or appraise, select and define. The questions are the same.
| Phase | The question | What it delivers | Estimate |
|---|---|---|---|
| Scoping study | Is there a project worth studying? | Order of magnitude capital and operating cost, a rough schedule, the main risks, and a list of options worth comparing | Class 5 |
| Prefeasibility study | Which option should go forward? | The options compared on one basis, one selected with the reasons recorded, a block flow diagram and a preliminary layout | Class 4 |
| Feasibility study and FEED | Is the selected option ready to fund? | Process flow diagrams, P&IDs issued for design, equipment list and data sheets, layouts, an execution plan, and the operating and maintenance strategy | Class 3 |
| Detailed design | Is it ready to buy and build? | Drawings and specifications issued for construction, procurement packages, the commissioning plan, and the data for the CMMS and spares | Class 2 |
The gates between the phases
A gate is a decision to fund the next phase, send the work back or stop. Its value is in stopping weak projects early, when little has been spent, and in refusing to start expensive engineering on a scope that is still moving. A gate passed because the date arrived, rather than because the phase delivered what the decision needs, is where most overruns begin. Gate reviewers should be independent of the study team, so that someone asks the questions the team has stopped asking.
| Gate | The decision | What the reviewers check |
|---|---|---|
| Gate 1, after scoping | Fund a prefeasibility study, or stop | Is the opportunity real, is it worth the study cost, and does anything rule it out early? |
| Gate 2, after prefeasibility | Select one option and fund feasibility, or stop | Were the options compared fairly, is the selected one recorded with its reasons, and has the owner's operating team agreed it? |
| Gate 3, after feasibility and FEED | Final investment decision, or stop | Is the scope frozen, does the estimate meet the accuracy the decision needs, and are the execution and operating plans credible? |
| Gate 4, before handover | Accept the asset into operation | Is it complete, tested and documented, and is the operation ready to run and maintain it? |
FEED and the scope freeze
Front end engineering design is the basic engineering on the selected option. It produces the process flow diagrams, the P&IDs issued for design, the equipment list and data sheets, the layouts, the main specifications, a class 3 estimate and the execution plan. It also produces the control system documents at philosophy level, which decide how the plant will be operated.
FEED ends with a scope freeze. After the final investment decision, every change costs more than it would have before, because it moves work that has already been done and orders that have already been placed. Changes after the freeze are managed through a formal change process with their cost and schedule effect stated before they are approved.
The design reviews
Design reviews are where people outside the design team look at the design before it is fixed. The model reviews at 30, 60 and 90 per cent are the backbone, each one a chance to see the plant as a whole while it is still cheap to move things. Hazard studies sit alongside these on their own schedule and are not covered here.
| Review | When | Who | What it catches |
|---|---|---|---|
| Basis of design review | End of prefeasibility | Owner, operations and the study team | Design criteria, throughput, operating hours and standards settled before detailed work begins |
| Model reviews at 30, 60 and 90 per cent | Through FEED and detailed design | Design team, operations, maintenance and construction | Access, clashes, layout and the parts of the plant nobody has looked at yet, while they are still cheap to move |
| Constructability review | FEED and early detailed design | Construction and design teams | Lift studies, laydown, sequence and modules sized for transport |
| Operability and maintainability review | FEED and the 60 per cent model | Operators, maintainers and the reliability engineer | Access for people and lifting gear, isolation points, removal paths, spares commonality and condition monitoring access |
| Value engineering | Late prefeasibility and FEED | Owner and design team | Function delivered for less, without moving cost into operation |
| Gate review | At each gate | People independent of the study team | Whether the phase delivered what the gate needs, before money is committed to the next |
The maintainability review
The review most often skipped is the one that sets the cost of running the plant for decades. Operators and maintainers walk the model with the design team and ask how each piece of equipment will be inspected, isolated, removed and replaced. A crusher with no crane access, a pump that can only come out by removing a pipe rack, or a conveyor pulley with no room to pull the shaft will cost far more in downtime over the life of the plant than the design change would have cost in FEED.
The review also settles things the CMMS and the spares holding will need later: equipment tagging on a single convention, spares commonality across similar duties, access for condition monitoring, and the critical spares that should be bought with the project. These are the inputs to operational readiness, and they are far cheaper to settle in the design than after handover.
- Lifting points and crane access for every heavy component, with the lift path drawn.
- Isolation points reachable and lockable from a safe standing position.
- Removal paths and laydown space for the largest component of each machine.
- Common spares across duplicate duties, rather than three makes of the same pump.
- Access for vibration, thermography and oil sampling without scaffolding.
- Instrument and valve positions reachable without working at height.
Where studies go wrong
- Options compared on different bases, so the selection reflects the assumptions rather than the options.
- A gate passed on the date rather than on what the phase delivered.
- The estimate class quoted without its accuracy range, so the decision makers read a single number as certain.
- Scope still moving after the final investment decision, with each change treated as small.
- Operations and maintenance first shown the design at the 90 per cent review.
- The operating cost studied far less carefully than the capital cost, so the cheapest plant to build is selected over the cheapest to own.
- The handover data for the CMMS, spares and documents left to the end of execution.
Common questions
What are the phases of an engineering study?
The common sequence is a scoping or concept study, a prefeasibility study, a feasibility study, then detailed design and execution. In mining these are often called scoping, PFS and DFS. In oil, gas and process plants the same stages are called front end loading, FEL 1 to FEL 3, or appraise, select and define. Front end engineering design, FEED, is the engineering that brings the selected option to the definition needed for a final investment decision.
What is the difference between a prefeasibility and a feasibility study?
A prefeasibility study compares the options and selects one. A feasibility study develops the selected option in enough detail to decide whether to fund it, with a more accurate estimate, a defined scope and an execution plan. In mineral reporting, the JORC Code describes a prefeasibility study as a comprehensive study of a range of options and a feasibility study as a comprehensive study of the selected option, detailed enough to be the basis for a decision to proceed or to finance.
What is FEED in engineering?
Front end engineering design is the basic engineering done on the selected option before the final investment decision. It typically produces the process flow diagrams, P&IDs issued for design, the equipment list and data sheets, layouts and the plot plan, the main specifications, a class 3 estimate and the execution plan. FEED is the last point where the design can change cheaply, which is why it is also where operations and maintenance input matters most.
How accurate is a feasibility study estimate?
AACE International Recommended Practice 18R-97 gives a typical range for a class 3 estimate, the class usually prepared at feasibility or FEED, of minus 10 to minus 20 per cent at the low end and plus 10 to plus 30 per cent at the high end. The actual range depends on the technology, the complexity of the project and the quality of the data, and should come from a risk analysis rather than being assumed from the class.
What is a stage gate process in a project?
A stage gate process divides a project into phases separated by decision points, the gates. At each gate an owner or review panel checks that the phase delivered what the decision needs, then decides to fund the next phase, send the work back, or stop. The value is in stopping weak projects early, when little has been spent, and in fixing the scope before the expensive phases begin.
Key terms
Plain-language definitions from our glossary for the concepts this article leans on.
Standards and further reading
- AACE International Recommended Practice 18R-97, cost estimate classification system as applied in engineering, procurement and construction for the process industries (AACE International)
- The JORC Code, 2012 Edition: Australasian code for reporting of exploration results, mineral resources and ore reserves (JORC)
- IEC 62337:2012 Commissioning of electrical, instrumentation and control systems in the process industry: specific phases and milestones (IEC)
Related case studies and tools
- Operational readiness for a new plant asset (case study)
Related reading
Working through something like this?
See how we approach these initiatives, or tell us what you are dealing with.