The huge and growing gasoline demand dictates that refiners maximize gasoline volume. After maximizing component volumes and meeting octane and other specs with minimum give-away, what else can be done?
One way to get a few more barrels is to optimize blending to get the highest octane blending values possible while meeting all the specs. This will allow slightly lower reforming severity, giving a slight increase in volume. However, the blending linear program will not achieve this goal if inaccurate blending values are used.
Use of refinery blend data is almost guaranteed to produce inaccurate blending values relative to octane (incremental blending values). One reason is narrow composition ranges which multiply extrapolation errors. Another is inaccuracy of octane numbers which can be expected in refinery operation. A third reason is high correlation between component percentages, which leads to large compensating errors. This results in misleadingly good prediction accuracy, when the incremental blending value is too high for component A and too low for component B.
Accurate incremental blending values can be obtained in well-conducted laboratory blending studies. If appropriate people are not available to plan and carry out such a study, a generalized octane blending program is available.1
When there is space in the gasoline for normal butane, a slight increase in volume might be available by maximizing use of normal butane in the low aromatic grade while using isopentane in the high aromatic grade.2
William E. Morris
Consultant
Wilmington, Del.
- www.gasolineblendingplus.com.
- OGJ, Mar. 3, 2008, p. 56.
