Predicting BTEX emissions in refrigeration plants - April 2012 GPA presentation

Process Ecology presented a paper at the April 2012 GPA conference in New Orleans – “BTEX Emissions from Ethylene Glycol Circulation in Natural Gas Refrigeration Plants: A Process Simulation Study”. The main simulators used in industry to predict benzene and BTEX emissions in refrigeration plants differ widely in their results. As a result of our review, we can provide some guidance regarding the simulation predictions and expected pitfalls.

Monoethylene Glycol (EG) is widely used as a hydrate inhibitor in low-temperature refrigerated gas plants.  One of the main environmental issues with the operation of glycol units is the release of BTEX (benzene, toluene, ethylbenzene, xylenes) emissions, which are controlled by federal and/or local regulatory agencies. Although the vast majority of glycol units use Triethylene Glycol (TEG) to remove water from natural gas, there are a significant number of dewpoint control and refrigeration plants that rely on EG to ensure hydrate-free operation. An earlier report¹ showed that the main process simulation tools used in the sector to predict BTEX emissions from TEG systems provide reasonably similar results and adequately match the available experimental data. EG systems, however, have not been extensively studied and it is well-known among practitioners that the simulation software tools generally used can produce widely different results, and in some cases even fail to produce results at all. The main tools compared in the paper are Aspentech’s HYSYS, BR&E’s ProMax and Gas Technology’s GRI-GLYCalc.

• The Peng-Robinson property package in HYSYS (HYSYS-PR) should not be used for the prediction of BTEX emissions in refrigeration plants. The Peng-Robinson property package predicts negligible amounts of BTEX in the aqueous phase of low temperature separators, resulting (incorrectly) in predicting zero emissions.
• For the prediction of BTEX emissions in refrigeration plants, the Glycol property package in HYSYS (HYSYS-Glycol) is recommended by Aspentech and is significantly better than HYSYS-PR for this purpose. Of the three reviewed simulators, HYSYS-Glycol tended to predict the highest BTEX emissions. However, HYSYS-Glycol should be used with caution at lower LTS pressures (less than 2,500 kPag).
• BR&E has published much information related to TEG dehydration facilities and BTEX emissions but very little related to EG (refrigeration). Of the three simulators reviewed in this study, ProMax consistently predicted BTEX emissions higher than GLYCalc but lower than HYSYS-Glycol.
• It was found that GLYCalc often fails to predict a hydrocarbon liquid phase in the low temperature separator, resulting in very high predicted BTEX emissions in those cases. The user must be careful when using GLYCalc for EG (refrigeration) units. When GLYCalc calculates “reasonable” results, it consistently predicts the lowest amount of BTEX emissions among the three simulators reviewed.
• While VLE data has been published for BTEX-hydrocarbon systems at cold separator conditions, LLE data is not available in the literature, which would be crucial in understanding how much BTEX would be soluble in the hydrocarbon liquid and aqueous phases, respectively.
• Field data was obtained and the results of the simulators were compared to this data. For two of the three units, HYSYS-Glycol had the closest results to the data for the combined BTEX predictions. For these units, the results of ProMax were about 50-60% of the original data, and GLYCalc significantly overpredicted the results due to the inability to properly predict a hydrocarbon liquid phase in the cold separator. For the third unit, HYSYS significantly overpredicted BTEX emissions, understood to be a thermodynamic issue at lower pressures (less than 2,500 kPag). In this third case, the ProMax results were very close to the data, while GLYCalc results were about 40% of the original data. While these results were observed for BTEX overall, individual component predictions were much less comparable to the field data.
• Some effort was made to fit thermodynamic parameters to field data; while it was found that a reasonable match could be made to the field data, it did not extrapolate well to the wide range of conditions required.

While GLYCalc has been successfully used in industry to calculate BTEX emissions in EG units, there are limitations. Customizable simulation platforms such as HYSYS or ProMax can be used to improve thermodynamic calculations as well as improve on data transfer and modeling aspects, especially when combined with a knowledge of the range of applicability of each simulator.

To read the full study please download the pdf file.

1. Holoboff, J. and Khoshkbarchi, M. Prediction of BTEX Emissions and Water Content in TEG Dehydration Units, January 2009. Presented at 2009 GPA Conference.

GPA 2012 paper (Process Ecology).pdf

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