New calculation method alters tank emission reports

Michael E. Anderson, Jeffrey H. Siegell
Exxon Research & Engineering Co.
Florham Park, N.J.

Recent changes to the calculation method for estimating hydrocarbon emissions from floating roof tanks will change reported emissions from refineries, terminals, and production facilities.

The changes consolidate the previously separate formulas for external floating roof and internal floating roof tanks. The method now explicitly includes equations for covered external floating roof tanks. In addition, the new method adds a wind speed correction factor used for calculating external floating deck fitting losses.

This article compares the impact of the changes on calculated emissions for two operating facilities. Using data from a gasoline marketing terminal and from a refinery, tank emissions were compared according to the new calculation method^{1 2} and the previous method.^{3 4}

For the two sites, the new calculations resulted in a decrease of about 17% for calculated emissions from external floating roof tanks and an increase of about 5% for calculated emissions from internal floating roof tanks. The refinery experienced about a 10% decrease in overall tank emissions, and the gasoline marketing terminal experienced about a 1% decrease.

New calculation methods

The new method for calculating emissions from floating roof tanks was developed by the Environmental Technology Advisory Group of the American Petroleum Institute (API), and has been adopted by the U.S. Environmental Protection Agency (EPA). The new method more closely reflects emissions for actual plant conditions than the previous method.

The two major changes affect calculation of rim seal losses and deck fitting losses:

• The new equation for calculating the rim seal loss factor, K_r, as a function of wind velocity is K_r = K_ra + K_rb(V)^m, where K_ra is the zero wind speed rim seal loss factor, K_rb is the wind-dependent rim seal loss factor, V is the locally reported wind velocity, and m is the wind speed exponent.

The old calculation did not include a zero wind speed loss factor. The introduction of this factor allows calculation of losses for an internal floating roof tank (i.e., the top fixed roof shelters the floating roof resulting in an effective wind speed of zero) and an external floating roof tank using the same equation. Also, the new calculation now allows direct calculation of covered floating roof tank emissions where the wind speed is not a factor.

• A new wind speed correction factor, K_v, is included in deck fitting loss calculations. K_v accounts for the shielding effect of the tank shell. When multiplied by the locally reported wind speed, it reduces the effective velocity of the air above the floating roof used in the calculation of deck fitting losses.

The currently accepted value of K _v is 0.7. The data analysis, however, indicated a value of 0.5 was possibly more representative of actual conditions. For comparison purposes, the emissions calculated using both values of the wind speed correction factor have been included.

External floating roof tanks

Of the tank types, external floating roof tanks (EFRTs) demonstrated the largest change in calculated emissions with a reduction of about 17% ( Table 1 [50,502 bytes]). The largest part of this reduction is due to the new wind speed correction factor, K _v, which reduced deck fitting losses. For EFRTs, deck fitting losses dropped over 25% with K _v equal to 0.7, and dropped about 54% with K _v equal to 0.5.

For EFRTs, the new equation for the rim seal loss factor, K_r, also contributed to the reduction in calculated emissions. Total rim seal emission declined by 8.6%.

The magnitude and direction of each change in calculated emissions depend on the number of each seal type present at the site. For example, the average EFRT equipped with a vapor-mounted primary seal and a rim-mounted secondary seal experienced almost a 16% decrease in calculated emissions. However, the average EFRT equipped with only a mechanical shoe primary seal experienced a 16% increase in calculated emissions.

Internal floating roof tanks

Calculated emissions from internal floating roof tanks (IFRTs) increased 4.5% overall ( Table 2 [31,257 bytes]). Because the effective wind speed for an IFRT is zero, the wind speed correction factor has no effect on evaporative loss calculations. Deck-fitting emissions increased only slightly, about 1%, due to some minor changes in the deck-fitting factors.

The most significant difference in IFRT emission calculations results from changes in the rim seal loss equation for an IFRT with a mechanical shoe primary seal. In the previous calculation method, an IFRT fitted with a mechanical shoe primary seal is treated the same as an IFRT with a liquid-mounted, resilient primary seal. The new equations for rim-seal loss, however, include a separate zero wind speed factor for a mechanical shoe primary seal that is nearly twice the value of the old factor. This factor is now much closer to that of a vapor-mounted primary seal than a liquid-mounted primary seal. Thus, IFRTs equipped with mechanical shoe primary seals without secondary seals will almost double the calculated rim seal emissions previously reported.

The magnitude of rim seal loss factors with an IFRT equipped with only a vapor-mounted primary seal does not change.

Covered floating roof tanks

External floating roof tanks that have been retrofitted with fixed roofs are traditionally called covered floating roof tanks (CFRTs). These tanks were previously treated as external floating roof tanks.

In the past, the EPA Bulletin Board recommended an effective wind speed of 2 mph be used to calculate emissions. With the new method, calculated emissions from CFRTs decreased by about 13% (Table 3 [28,086 bytes]). Rim seal calculated emissions nearly doubled, but deck fitting-calculated losses reduced to almost half. As with IFRTs, the effective wind speed on the floating roof of a covered-floating roof tank is zero. Thus, changing the wind speed correction factor has no effect on estimated emissions.

Refinery

For the refinery, total calculated tank emissions decreased about 10% using the wind speed factor of 0.7 in the new deck fitting loss calculations. A wind speed factor of 0.5, if adopted, would have decreased the total evaporative loss calculations by almost 20% ( Table 4 [25,157 bytes]). Calculated IFRT emissions increased by about 4% while calculated EFRT emissions decreased by almost 17%.

Marketing terminal

The marketing terminal used in this study mainly consisted of IFRTs and contained no EFRTs. Hence, the large reduction in estimated emissions found for the refinery was not repeated. A slight decrease, about 1%, in total estimated tank emissions was found ( Table 5 [21,999 bytes]). This reduction is mainly due to the revised calculation method for covered floating roof tanks.

References

1. Compilation of Air Pollutant Emission Factors, AP-42, Vol. I: Stationary Point and Area Sources, Chap. 7, Supplement A, U.S. Environmental Protection Agency, Research Triangle Park, N.C., Feb. 1996.
2. Manual of Petroleum Measurement Standards, Chap. 19-Evaporative Loss Measurement, Section 2-Evaporative Loss from Floating Roof Tanks, American Petroleum Institute, Washington, D.C., April 1997.
3. API Evaporative Loss from External Floating Roof Tanks, Publication 2517, 3rd edition, February 1989.
4. API Evaporative Loss from Internal Floating Roof Tanks, Publication 2519, 3rd edition, June 1983.