Measurement of BTEX Emissions from Glycol Dehys

A significant amount of work was performed by both researchers and industry in the early 1990s regarding the measurement and estimation of BTEX emissions from glycol dehydration units. Many of the results reported from these efforts were captured during the 1992 and 1995 Gas Research Institute Air Toxic Conferences as well as a number of reports for the USEPA. 

Since then, little published information is available on this topic; however, the conclusions from this research are important in understanding the challenges associated with directly measuring BTEX emissions.

Most of the authors discussed the significant difficulties associated with direct measurement of BTEX emissions from glycol dehydrators and present as some of the major issues the following [1, 4]:

• High moisture content in the stack vent.
• Low, cyclical, and highly variable vent exhaust flow rate.
• Small stack diameter and vent line wall effects.

Since the vent flow rate is essential for an accurate emissions estimate, the preferred test method for accurate estimation was determined to be total capture via condensation and further chromatographic analysis of liquids and gas fractions. This method however is not practical as a routine approach for emissions control since it is usually expensive to perform and it could only handle smaller units. Facing this challenge, the next options for emissions estimation include simulation (GLYCalc or HYSYS) and the Atmospheric Rich/Lean method that is based on a mass balance around the glycol stream.

The ARL method has its own challenges given that the sampling protocols must be followed carefully:

• the Ventura County report [1] indicates that the ARL method is acceptable only for units equipped with flash tanks.
• A report by Radian Corporation [2] notes that the gas entrained in the rich glycol in units without flash tanks can introduce significant errors. Although they provided correction factors, some authorities decided to rely on total capture for these units.
• A similar situation is found for units using stripping gas.

There are some authors that question the validity of the ARL method, particularly with respect to the dilution required for the samples [3] and the option of pressurized vs atmospheric samples [1]. There is evidence presented from both fields to justify their positions; however a general conclusion that may be drawn from these reports is that simulation remains the most economical option for estimating emissions and that a logical sequence of actions may involve the use of ARL only for units that appear close or around the emissions limits (identified from simulation studies) and that a final confirmation would require total capture (if feasible). Clearly there is significant room for improvement on all these tools and further work should continue in order to further validate the simulation models.

References

[1] S. Cowen and F. Nunnery (1995) “Regulation and Permitting Requirements for Glycol Dehydrators in Ventura County, California”. Paper presented at the 1995 Gas Research Institute Glycol Dehydrator/ Gas Processing Air Toxics Conference. November 5-8, 1995

[2] D. L. Reif, C.O. Reuter and D. B. Myers (1996) “Atmospheric Rich/Lean (ARL) Method for Determining Glycol Dehydrator Emissions”. Topical Report prepared for the Gas Research Institute

[3] P.L Grizzle (1992) “A Multi-Task Approach to Addressing the Emissions from Glycol Dehydration Units” Paper presented at the 1992 Gas Research Institute Glycol Dehydrator/ Gas Processing Air Toxics Conference.

[4] Texaco Corporation (1992) “Texaco’s Dehydrator Air Emissions” Paper presented at the 1992 Gas Research Institute Glycol Dehydrator/ Gas Processing Air Toxics Conference

Do you have questions or comments regarding this article? Click here to contact us.