Refining Report Complexity index indicates refinery capability, value

Daniel Johnston
Daniel Johnston & Co. Inc.
Dallas

Refinery size usually is measured in terms of distillation capacity. Relative size, however, can be measured using refinery complexity-a concept developed by W.L. Nelson in the 1960s.

A review of complexity calculations, and an explanation of how indices have changed, provide a simple means of determining the complexity of single refineries or refining regions. The impact of complexity on product slate also will be examined.

The Nelson index

Nelson developed the complexity index to quantify the relative cost of components that make up a refinery. It is a pure cost index that provides a relative measure of the construction costs of a particular refinery based on its crude and upgrading capacity.

The Nelson index compares the costs of various upgrading units-such as a fluid catalytic cracking (FCC) unit or a catalytic reformer-to the cost of a crude distillation unit. Computation of the index is an attempt to quantify the relative cost of a refinery based on the added cost of various upgrading units and the relative upgrading capacity.

Nelson assigned a factor of 1 to the distillation unit. All other units are rated in terms of their cost relative to this unit.

For example, assuming a crude distillation unit costs $400/b/cd to construct, a 50,000 b/cd unit would cost $20,000,000. If another component costs $1,200/b/cd to build, this unit would have a complexity factor of 3.

The complexity rating of a refinery is calculated by multiplying the complexity factor for each downstream unit by the percentage of crude oil it processes, then totaling these individual factors. To illustrate, consider the case of a refinery with 50,000 b/cd of crude capacity and 30,000 b/cd of vacuum distillation capacity.

The throughput of the vacuum tower relative to the crude distillation capacity is 60%. Given a vacuum unit complexity factor of 2, the contribution of the vacuum unit to overall refinery complexity is 2 x 0.6, or 1.2.

Economy of scale

The optimum size for a U.S. refinery is at least 100,000 b/cd of atmospheric distillation capacity. This is slightly larger than the U.S. average of about 91,000 b/cd. The closing of many small refineries has caused this average to increase from its late-1980s level of 80,000 b/cd.

Smaller refineries, especially those with less than 50,000 b/cd capacity, have a tough time competing. Larger plants have greater economy of scale, and can survive longer in lean periods, when margins are narrow. Small refineries usually are able to survive these periods only when geographically or politically insulated.

Nelson was careful to point out that the cost of a 50,000 b/cd refinery with a complexity of 12 would not necessarily be the same as that of a 100,000 b/cd refinery with an index of 6. Many other factors are involved.

The information normally used to calculate refinery complexity does not indicate the number of units that comprise each refiner's capacity of a particular process. In other words, total capacities usually are given, rather than the number of units and each unit's capacity.

Small units have relatively higher construction costs per unit. Nelson estimated that a duplication of units, such as two 40,000 b/cd units instead of one 80,000 b/cd unit, would increase construction costs by a factor of 25%. Four units rather than one would increase costs by a factor of 60%.

Nelson reported that the amount of duplication for larger (300,000+ b/cd) refineries averaged 2.7 units for each process. This value compares to an industry average of 1.5 units for each process.

Table 1 [20929 bytes] shows the general indices for the various refining processes. The categories in this table coincide with those published annually by Oil & Gas Journal (bibliography).

As an example, the complexity of Kuwait National Petroleum Co.'s Mina Abdulla refinery is calculated in Table 2 [30863 bytes]. This refinery has a rated capacity of 242,000 b/cd and a complexity index of 7.

This technique can be used to analyze the complexity of a single refinery, or the complexity of a country or region. Table 3 [36057 bytes] outlines the processing capacities and overall complexity index for Japan's refineries.

Worldwide refining complexity is summarized in Table 4 [39346 bytes].

Off sites

The complexity calculation outlined here does not account for off site facilities, such as storage tanks, land, pipelines, terminals, and utilities. This information seldom is published.

Nelson evaluated the relationship between different refineries and the associated off site facilities and developed a relationship for empirically estimating total refinery complexity, including processing complexity and off sites.

In this work, Nelson assigned a multiplier to a series of complexity values:

• For a complexity of 3, use a 3.25 multiplier, for a total complexity of 9.8.

• For a complexity of 4, use a 2.70 multiplier, for a total complexity of 10.8.

• For a complexity of 6, use a 2.26 multiplier, for a total complexity of 13.6.

• For a complexity of 10, use a 1.96 multiplier, for a total complexity of 19.6.

• For a complexity of 16, use a 1.77 multiplier, for a total complexity of 28.3.

The total complexity is rarely used or quoted. Most of the time, the focus is on processing or "regular" complexity, as Nelson called it.

This calculation implies that, for a grassroots refinery with an index of 10, the cost of the off sites would rival that of the distillation and upgrading facilities.

EDC

Equivalent distillation capacity is another means of comparing refinery costs. In this calculation, the atmospheric distillation capacity of a refinery is multiplied by its overall complexity rating.

For example, a 75,000 b/cd refinery does not necessarily have half the processing capability of a 150,000 b/cd refinery. If the smaller refinery has a complexity index of 9 and the larger plant an index of 4, the EDC for the 75,000 b/cd refinery is greater than that of the 150,000 b/cd refinery.

The smaller refinery has an EDC of 675,000 (75,000 x 9), and the larger refinery has an EDC of 600,000 (150,000 x 4). In other words, the 75,000 b/cd refinery is "larger," in terms of construction costs, than the 150,000 b/cd refinery.

The approach outlined here for evaluating refinery cost complexity is helpful for estimating refinery values, but should be treated as a starting point. The Nelson complexity index (NCI) is a convenient and useful number, and, certainly, any estimate of refinery value should address the issue of complexity.

Refinery value, however, should be considered within the context of other data. For example, two similar refineries, one operating at 75% capacity and the other at 90% capacity, will have significantly different values.

Changes in indices

The generalized complexity indices, shown in Table 1 [20929 bytes], have not changed dramatically over the years. Building an FCCU still costs 5-7 times as much, on a $/b/cd basis, as building an atmospheric distillation unit. The generalized indices provide a quick, close estimate.

Detailed analyses of refinery complexity can be done with specific indices for 50 or so refining processes. An excellent study of refinery consruction costs was given by Maples (bibliography).

There are eight separate processes for catalytic hydrotreating, for example, all of which are treated in Table 1 [20929 bytes] as though they have an index of 2. Based on the construction costs, however, a 30,000 b/cd hydrotreater can range in cost from $16 million to $37 million, depending upon the feed.

The indices for these units range from 1.4 to 2.2. A weighted average of 2, however, seems to work well.

Thermal processes, on the other hand, are broken into five categories, as shown in Table 1 [20929 bytes]. Although the complexity of these units ranges from 2 to 6, they are given a weighted average complexity of 5.

The distribution of these units is:

• Thermal cracking (NCI = 3.0) and visbreaking (NCI = 2.5) comprise 6% of U.S. thermal processing capacity and 68% of non-U.S. thermal processing capacity.

• Delayed and fluid coking (NCI = 6) and other thermal processes (NCI = 6) comprise 94% of U.S. thermal capacity and 32% of non-U.S. thermal capacity.

  It is clear from these data that the configuration of U.S. refineries is very different from non-U.S. refineries. This difference illustrates how complexity calculations can be used for regional comparisons.

Refinery cost, value

The average complexity rating for a U.S. refinery is 9.5. A refinery with a complexity of 12 therefore should cost about 26% more to build than a similar-sized refinery with a complexity equivalent to the U.S. average, all other things being equal. All other things, however, are seldom equal. Even if they were, the relationship is not perfect.

The NCI provides insight into not only complexity, but also refinery replacement costs and values. The relationship between complexity index, replacement value, and fair market value is illustrated in Fig. 1 [42745 bytes].

These value estimates generally are consistent with the refinery transactions and valuation appraisal trends of the late 1980s and early 1990s in the U.S. Part of the decrease in refinery values over this period was a result of the Clean Air Act Amendments of 1990. The other key reason is that current and anticipated margins are too slim.

Refinery utilization, however, has been increasing. This increase has had a positive effect on refinery value. Capacity utilization can be expected to continue to increase in the U.S. because it is hard to keep pace with demand by expansion alone, and grassroots construction is unlikely.

Fig. 1 [42745 bytes] does not take into account capacity utilization, nor does it consider off site facilities, cash flow, earnings history, or forecasts. These are important considerations that are beyond the scope of this graph.

Estimates of refinery value should be considered an educated guess. The difference between fair market value and replacement costs, for U.S. refineries, illustrates the reason for lack of grassroots refinery construction.

Product slate

The advantage of higher-conversion refineries is that they turn out more high-value products. Fully 50% of a typical U.S. high-conversion refinery output is gasoline. Furthermore, because of cracking, the volumes of petroleum products can exceed total crude oil runs by as much as 5-10%.

By contrast, a refinery with an index of 3-5 likely will have volumetric contraction: 100,000 b/cd in and 95,000 b/cd out, for example. Added value and volumetric expansion are the bases for justifying the additional capital and operating expenditures required for high-conversion facilities.

Typical yields from low, medium, and high-conversion refineries are:

• Low-conversion (NCI = 2-3)-20% gasolines, 35% middle distillates, 30% fuel oil, 10% other products (including refinery gas, LPG, solvents, coke, lubes, wax, and bitumen), and 5% loss

• Medium-conversion (NCI = 5-6)-30% gasolines, 30% middle distillates, 30% fuel oil, 15% other products, and 5% gain

• High-conversion (NCI = 9-10+)-50% gasolines, 30% middle distillates, 15% fuel oil, 15% other products, and 10% gain.

Future

The total world refining capacity of about 74 million b/cd has an overall complexity index of 5.9. The trend of increasing conversion capability will pull up the complex index as world demand for lighter products increases.

Refining capacity utilization (less than 70% in the early 1980s) reached 87% in the early 1990s. Yet margins still are too tight to encourage much grassroots refinery expansion.

Bibliography

Burdik, D.L., and Leffler, W.L., Petrochemicals in Non-technical Language, PennWell Publishing Co., Tulsa, 1990.

Farrar, G.L., "Interest reviving in complexity factors," Oil & Gas Journal, Oct. 2, 1989, p. 90.

Gill, D., "Jack and the Integrated Bean Stalk," Oil & Gas Investor, March 1990, pp. 64-70.

Leffler, W.L., Petroleum Refining for the Non-technical Person, 2nd ed., PennWell Publishing Co., Tulsa, 1985.

Maples, R.E., Petroleum Refinery Process Economics, PennWell Publishing Co., Tulsa, 1993.

Nelson, W.L., "How to describe refining complexity," Oil & Gas Journal, Mar. 14, 1960, p. 189.

Nelson, W.L., "How complexity of a refinery affects costs of processing," Oil & Gas Journal, Sept. 26, 1960, p. 216.

Nelson, W.L., "How to compute refinery complexity," Oil & Gas Journal, June 19, 1961, p. 109.

Nelson, W.L., "Complexity-1: The concept of refinery complexity," Oil & Gas Journal, Sept. 13, 1976, p. 81.

Nelson, W.L., "Complexity-2: Process unit complexity factors examined," Oil & Gas Journal, Sept. 20, 1976, p. 202.

Nelson, W.L., "Complexity-3: How to apply complexity," Oil & Gas Journal, Sept. 27, 1976, p. 83.

Petzet, A.G., "Surge of non-U.S. joint ventures, acquisitions eases in U.S. refining," Oil & Gas Journal, Aug. 7, 1989, p. 15.

"Creating the high performance international petroleum company: Dinosaurs can fly," Treat, J.L., ed., PennWell Publishing Co., Tulsa, 1994.

Williamson, Michelle, Worldwide Refining Survey, Oil & Gas Journal, Dec. 18, 1995, p. 47.

The Author