Mar-2002

Improved conceptual process design avoids revamp scope growth

For grassroots design, defining project scope is a straightforward exercise. Grassroots projects can follow optimum execution procedures for engineering and construction

Tony Barletta and Gary R Martin, Process Consulting Services
Adrie Visser, National Petroleum Refiners of South Africa (PTY)

Article Summary

Office-based conceptual process design (CPD) works well because there is no existing unit to present unique challenges and obstacles. For a revamp, however, CPD demands a much more detailed attack, because most equipment - with all its possibly hidden shortcomings - already exists. The only foolproof way of carrying out CPD for a revamp is to first establish a performance base line of the existing unit. This entails making a comprehensive test run to gather actual field data on temperatures, pressures and flows. Original equipment drawings, P&IDs and control room data are not sufficient. Once real-time field data is obtained, it is used for calibrating computer simulations in the design office. All major cost bottlenecks are now identified and alternative process flow schemes are evaluated to determine the lowest cost. Finally, equipment lists and cost estimates are developed. Admittedly, all this takes time and costs money, but unless such work is thoroughly and carefully carried out at this point, scope can grow enormously during front-end process design (FEED) and detailed engineering (DE).

This article presents data on a revamp undertaken for National Petroleum Refineries of South Africa (PTY), (NATREF, a joint venture of SASOL and Total Fina Elf) in which a study was begun to determine whether increasing crude capacity at its Sasolburg refinery was a sound investment. A NATREF multidisciplinary team coordinated engineering work with several contractors to carry through a staged process beginning with CPD, continuing through FEED and ending in DE.

Introduction
CPD is the most important activity in revamp design. Often, its significance is overlooked. As a result, minimum engineering effort is expended on it, with the expectations that the more detailed process design work can be performed during FEED or the beginning of DE. Consequently, many revamps start with superficial process work and little money assigned for CPD. Often looming behind these seemingly cost-effective CPD packages are scope growth and revamp cost escalation.

CPD largely determines revamp costs and whether or not the results will meet yield, run length and reliability objectives.^1,2,3,4 Failure to meet any one of these processing objectives can turn an otherwise profitable revamp into one that loses refiners millions of dollars due to poor performance or an unscheduled outage to correct revamp design flaws. When done properly, CPD will identify all significant process and equipment modifications, and scope growth will be minimal as engineering progresses.

By today’s standards, revamps are deemed successful if: throughput, yield and reliability objectives are achieved, and they are on schedule and under budget. Large overruns can wreck revamp economics. Volumes have been written about cost estimating, cost control, project management and scheduling. All of these are important activities that must be executed well for a successful revamp. However, if the CPD is poor quality or insufficient in detail, then no amount of cost estimating, cost control, project management and scheduling activities will prevent scope growth.

This article presents guidelines for revamp CPD. If these guidelines are executed properly, they will minimise revamp scope growth. While the examples presented are specific to crude/vacuum units and the approach used by NATREF, the techniques can be applied to other refinery unit revamps.

Revamp conceptual process design
Project schedules often demand fast-track revamp conceptual process design. Overall revamp durations can be as short as one year from beginning to startup for fast-track revamps (typically under $20 MM) or longer for larger investments. Whatever the case, CPD must be fast and efficient. There is no time for re-engineering.

CPD costs must also be controlled. At the conceptual design stage, the revamp has not yet received full funding (see Figure 1). If engineering costs are excessive and the revamp does not get funded, money is wasted. However, there is some minimum amount of engineering that must be performed to sensibly direct the capital expenditure. Otherwise, all major scope-related items may not be identified and the process flow scheme selected may not be minimum cost.⁵

Minimum CPD cost and sufficient engineering are always competing objectives during the conceptual design stage. The trend has been to reduce the cost of CPD by pushing essential process engineering evaluations into FEED and DE. In many revamps, this has resulted in either scope growth or scope rationalisation, where many pieces of equipment were removed to control costs. At this point, it may be well to ask if equipment can be removed to control costs without impairing the process scheme, and why the equipment was specified in the first place. In corollary, if the process scheme will be impaired, how can removal be justified? A more intelligent, efficient and thorough CPD is needed to satisfy the competing objectives.

Why is a different approach needed?
When a conventional office-based approach is applied to revamp design, it is common today for a major portion of the process engineering design to take place in the FEED stage. Only a superficial amount of process design is performed during the CPD, resulting in a cursory review of much of the equipment. Consequently, the revamp scope of work is poorly defined. Conventional office-based CPD focuses on scheduling and cost estimating, not on process design. But if scope is poorly defined, the estimate will not be accurate even if it is expertly prepared with much detail.

During CPD, all related revamp modifications must be identified so a cost estimate can be prepared. If the scope of work is well defined, costs can be estimated with a reasonable degree of accuracy. However, if the scope is incomplete, the estimate will not capture all revamp costs. Poorly defined scope is the number one cause of revamp cost escalation. It is the conceptual process design, not cost control or project management activities, that defines the flow scheme and therefore the revamp scope.