Peter Laing explains how combining modern software packages with experienced engineers allows reliable estimates to be produced for around a fifth of the cost of traditional approaches
Front-end engineering and design (FEED) is the phase of a project during which a percentage of the overall engineering required for a project is completed, with the purpose of developing a sanction estimate. Capital cost has a major bearing on return on investment, so the sanction estimate is an important part of the investment decision.
FEEDs are associated with different sanction estimate accuracies (e.g. ±10%, ±20%) as estimate accuracy is proportional to the level of definition available. Many factors impact on cost but are difficult to accurately quantify without a preliminary design in place.
Projects in the process industries vary enormously in terms of cost makeup, but generally engineering costs somewhere between 10% and 20% of the total installed cost (TIC). FEEDs are often referred to in terms of shallow or deep, which relates to the quantity of engineering performed. Given shallow FEEDs start at 10% of engineering with deep FEEDs approaching 25%, a conventional FEED can cost in the region of 2%-3% of TIC.
An alternative approach to conventional FEED
Over the past decade the UK market for FEED has changed as smaller operating companies have limited access to funds. As an alternative to the conventional FEED, powerful, software-based estimating tools were developed. These made use of key data contained within commonly used engineering tools to generate more accurate estimates through the use of algorithms and datasets.
Process simulation software packages, used globally throughout the process industries, find the optimal conditions for a given process through the creation and manipulation of a heat and mass balance. The latest generation of estimating tools takes advantage of the fact that most simulators contain algorithms that provide initial estimates of equipment size.
Primary data is extracted from the process simulators and used as the basis from which to build a TIC estimate. The data extracted includes equipment types, equipment sizes, equipment connectivity, piping sizes and electrical loads.
Volumetric modelling for further front end engineering design definition
‘Volumetric modelling’ techniques are necessary to develop improved engineering definition to support the estimate. It starts with the allocation of a pre-determined template of ancillary piping and instrumentation for each of the main plant items. For example, if we take a distillation column fitted with a thermal reboiler, then information concerning the feed, bottoms liquid, overhead vapour and reboiler lines, as well as the equipment items themselves, are extracted from the simulation. The pre-determined template adds relief, blowdown and purge lines plus commensurate numbers of temperature, level and pressure instruments. Manual isolation valves, check valves, control valve bypasses, vents and drains plus allowances for tertiary steelwork and lighting are also included in the template. This operation is performed for each main plant item.
As process simulation exists only in a two-dimensional world, it is still necessary for the user to define the plot size and number of levels required for the new plant. Based upon typical process plant configurations and densities, the estimating software can then approximate piping and cabling meterage and the primary and secondary steelwork required for the main structures. The cost of foundations for the main structures and equipment items are selected from look up tables on the basis of equipment type, size and an assumed soil density. The requirement to pile foundations is optional.
The costs for process control, emergency shutdown and power distribution systems are estimated from the instrumentation count and number of electrical drives. Line temperatures are assessed and where necessary, an allowance for thermal insulation is calculated.
Using the above process the estimating software generates calculated quantities for equipment and bulk materials. Engineering and construction man hours are factored on the basis of these quantities. Costs for these services are then generated using rates contained within look up tables which are periodically updated.
Accuracy improvement for front end engineering design
Changes to the estimating model can progressively be made by the estimator to improve accuracy, for instance to take advantage of further engineering definition. The actual plant layout can be entered into the estimating model when known thereby improving the accuracy of calculated steelwork, piping and cabling quantities. Also, the estimated costs of main plant items can be replaced with manufacturer quotations.
Various other inbuilt options to improve accuracy are offered by the estimating packages including specific labour rates for different geographic locations, adjustments for local climatic and seismic conditions, various design codes, productivity rates and an annual update of the cost databases to reflect latest inflation.
How accurate are the new software estimating tools?
Clearly volumetric modelling represents no more than an approximation of quantities, nevertheless estimates generated using this technique have proved to be reasonably accurate. An underestimate in one area of the plant is often counteracted by an overestimate in another.
Experience indicates TIC estimates generated straight from a process simulator without any additional data or refinements made by the operator are 40% accurate or better. The input of the actual unit layout, vendor quotations for main equipment items, electrical distribution details, ground conditions, etc. will progressively improve the accuracy to 30% and better.
Advantages of the streamlined approach to front end engineering and design
The primary benefit of the streamlined approach is it provides the information required for an investment decision at a fraction of the cost of conventional FEED. Instead of spending 2%-3% of TIC expect to spend around 0.5% using the streamlined approach. Given the much reduced engineering required for the streamlined approach, this offers a quicker route to sanction than the conventional approach.
A less obvious benefit is the flexibility for optioneering. The costs for differing capacities, configurations, technologies or locations can be relatively quickly evaluated without substantial rework. This allows value engineering studies to be completed in order to improve the overall business return for the project.
Disadvantages of the streamlined approach to front end engineering and design
Infrastructure, offsites and utilities cannot be generated from the process simulation and as such allowances for these items must be added by the estimator. Nor do the new tools cater well for specialised items such as packaged equipment, fired heaters or large rotating machines.
As previously noted, the estimating algorithms are driven solely by the equipment items contained within the process simulation and as such the costs for the instrumentation and electrical systems and distribution are based on scaled factors rather than installed base. As a consequence, the costs generated for these items merit further scrutiny.
Finally, much of the documentation generated by the packages is generic, the piping and instrument diagrams (P&IDs) in particular are of little value beyond project sanction.
In summary, the current generation of software-based estimating tools offer operators a relatively cheap and quick way of assessing the viability of a project in terms of return on investment. They do not, however, provide a short cut for the engineering required for a project should it obtain sanction.
Peter Laing is operations manager at ABB Consulting.