Non-Ideal Behavior of Gases and Their Mixtures
- Abdus Satter (U. Of Oklahoma) | John M. Campbell (U. Of Oklahoma)
- Document ID
- Society of Petroleum Engineers
- Society of Petroleum Engineers Journal
- Publication Date
- December 1963
- Document Type
- Journal Paper
- 333 - 347
- 1963. Original copyright American Institute of Mining, Metallurgical, and Petroleum Engineers, Inc. Copyright has expired.
- 5.2.2 Fluid Modeling, Equations of State
- 5 in the last 30 days
- 445 since 2007
- Show more detail
- View rights & permissions
Reported herein are the results of a careful and detailed study of the non-ideal behavior of pure gases and their mixtures. Included are: (1) new data on five ternary systems composed of methane, ethane and H2S; (2) a simple compressibility factor correlation that is inherently superior to present correlations, particularly for gases containing H2S and CO2; and (3) a detailed study of combination rules and the effect of system composition on the choice thereof. This study makes use of the rather large mass of data already available in the literature. A complete re-examination of the data and ideas presented in the last 25 years was considered desirable as a prelude to our basic concern - the effect of diluents on gas behavior. A consideration of both the macroscopic and microscopic properties of gases provides a better insight which, in turn, gives a firmer basis for improved correlation techniques. Such a study has shown that expressing the compressibility factor Z as a function of acentric factor w, as well as reduced temperature and pressure, yields a correlation that is broader in scope. The study of various combination rules has shown that better results are obtained by "tailoring" the rule used to the system composition. To do so improves the basic reality of results by overcoming some of the anomalies often found when using Kay's rule alone. Tentative recommendations are made regarding the most reliable combination rule for use with a given class of gas. The data presented are useful for estimating the direction and magnitude of the expected deviation `when using a given rule. Although more work is needed, particularly around the critical region and with CO2 mixtures, the advantage of the classification scheme proposed is apparent.
When one attempts to write a PVT equation to fit the data for actual gases, greater precision is obtained by the use of a multiple number of empirical constants. This has lead to multiple- constant equations such as Benedict-Webb-Rubin, Beattie-Bridgman, Keyes, etc., which are capable of yielding very precise results for pure gases in a range for which data to bet the constants are available. As a matter of practicality, though, the use of such equations for gas mixtures is limited. Because of the infinite number of gas analyses available, any attempt to compile the constants needed requires a prohibitive amount of experimental data. This could be overcome by the use of a combination rule, but there is no real advantage in doing so because the end result offers no practical improvement over the Z factor correlation. The most widely used method of predicting the volumetric properties of pure gases is based upon the "theorem of corresponding states". According to this theorem, "all pure substances have corresponding molal volume at corresponding temperature and pressure if the reference point of correspondence is the critical point". Generalized compressibility charts for gases were prepared first by Cope and associates in 1931 and later by Brown and co-workers in 1932. However, the most commonly used charts are those of Dodge, Nelson and Obert, Hougen and Watson, and Standing and Katz. The work of Katz and co-workers has provided us with basic data for the hydrocarbons most widely used today. Their original chart was compared with a relatively large amount of multi- component data for gases consisting almost entirely of normal paraffin hydrocarbons. A deviation of only + 1.2 per cent was obtained. In the 20 years following publication of this work it has been found that the behavior of most mixtures of paraffin hydrocarbons could be predicted by this correlation within at least 5 per cent. Where difficulty has been encountered it has largely involved one or more of the following circumstances: pressures above 4,000 psig, mixtures containing large amounts of heavy ends and/or aromatics, systems in the critical region and mixtures containing polar compounds and/or CO2.
|File Size||1 MB||Number of Pages||15|