Oct-2009

Analysis of hydrate conditions and property predictions in acid gas injection systems

This paper will focus on the hydrate formation conditions for acid gas mixtures commonly found in acid gas injection systems. Many studies have analysed the hydrate and solid formation temperatures of CO2 systems.

Cory Hendrick, Vicente Hernandez, Michael Hlavinka and Gavin McIntyre
Bryan Research & Engineering, Inc

Article Summary

However, very little is known about hydrate conditions in varying acid gas mixtures as experienced in many injection systems. Herein, the available published data will be presented and compared with predictions from the process simulation software, ProMax. In particular, the hydrate conditions for both saturated and under-saturated systems will be explored for comparison on these injection systems.

Introduction
Hydrates are solid formations similar to ice. They are composed of water molecules combined with other smaller molecules such as CO2 and H2S. Many studies have analysed hydrate and solid formation temperatures of CO2 and H2S systems. These studies have been historically taken on as a response to interest in enhancing oil recovery through gas injection. However, the industry’s interest in hydrates now grows due to stricter regulations on CO2 emissions. CO2 capture has now become necessary in many geographical areas. As a result, there is increasing interest in acid gas injection as a method of capturing and storing CO2.

As natural gas demand increases and regulations for a sweeter gas become stricter, the focus on properly handling these acid gas injection streams becomes more important. Engineers must properly predict the vapour-liquid 
equilibrium and the hydrate formation temperatures of these CO2/H2S/H20 streams to ensure operating conditions will not lead to complete or even partial hydrate formation.

The gas sweetening process typically found in sour gas plants produces a CO2/H2S rich gas stream. An acid gas stream is the overhead product of a solvent regeneration tower and is thus saturated with water. In preparation for injection, this stream enters a series of compressors, coolers and knock-out drums. As an acid gas stream is compressed, cooled and fed to a knock-out drum, water is removed. As the acid gas stream is further compressed, it can then be described as under-saturated.

Unfortunately, much less is known about the vapour-liquid equilibrium and solids formation conditions of the water and acid gas in injection systems. Specifically, very few data sets are available for H2S/CO2 hydrate formation in under-saturated conditions. Engineers challenged with designing these injection systems typically rely on process simulation to estimate these hydrate formation conditions. Further, process simulators rely on existing data and thermodynamic equations to interpolate and extrapolate solid formation at various conditions.

This paper reviews known data sets in various acid gas conditions and compares them to the predictions given by the ProMax simulation software. Proper vapour-liquid equilibrium and property predictions will be verified first. Then hydrate predictions can be compared to those experimentally determined in trusted data sets. These trusted data sets all occur in saturated conditions. Once simulations are trusted with hydrate conditions, we can then review hydrate temperature predictions in under-saturated conditions and compare to a single data set.

First vapour-liquid equilibrium data sets and molar densities for CO2, H2S and CO2/H2S systems will be compared to simulation results. Next, experimentally determined hydrate temperatures will be compared to simulation result for CO2, H2S and CO2/H2S systems. Finally, predictions of injection acid gas systems in under-saturated conditions will be reviewed and benchmarked against experimental data.

Vapour-liquid equilibrium
Proper determination and interpretation of hydrate formation boundaries requires understanding of the vapour-liquid equilibrium and the variables, which can affect their boundaries. Certainly, composition, temperature and pressure of a system play an important role in vapour-liquid equilibrium and hydrate formation. Moreover, the effect of composition on hydrate potential is magnified as the operating pressure increases. At pressures found in acid gas injection systems, simple gas gravity correlations to determine water content must be abandoned for experimental data or more rigorous VLE calculations. Therefore, any simulation predictions must be compared to trusted data sets of vapour-liquid equilibrium for general agreement.

Presented in Figure 1 is a graph displaying the water content, in grams of H20 per m3, of a CO2 system at varying temperatures as a function of pressure. Data from GPA report RR-991 at 25°C are plotted along with simulation predications at various temperatures. This graph is similar in design to that provided by Strickland2 and includes the minimum water content line at 43°C condition. The purpose of this graph is to verify that composition predictions agree with published and trusted data.

Obviously, pressure and temperature both influence the composition of a gas stream. The 43°C line contains a trough near 70 bar. The bottom of this trough represents the minimum amount of water the CO2 gas can hold at 43°C at that corresponding pressure. As Strickland pointed out, much of the water can be removed at the trough of the curve during the first or second compression stage of injection systems through a knock-out drum. Once the water is dropped out in a knock-out drum, any increase in pressure will result in under-saturated conditions at that same temperature.

Figure 2 is an equivalent graph considering an H2S and H2O only system. Again, ProMax results at varying temperatures are plotted along with results from GPA RR-483 at 37.8°C. A minimum water content line is also included for the system at 43°C.

Figure 2 shows that water content of H2S gas systems is more sensitive to pressure in the range of 0 to 50 bar than that of the CO2 gas systems. The H2S system in this case does not show a smooth trough but rather a “cliff.” This “cliff” signifies a phase change.