No abstract

ammerschmidt (1 934) developed an empirical method for predicting the hydrate formation temperature for natural gas H constituents in the presence of alcohol. ('Alcohol' is used here to describe alcohols and polyhhydroxy alcohols, or glycols.) Subsequently, several other models have been suggested, for example, Anderson and Prausnitz (1 986) and Moshfeghian and Maddox (1 993).In the presence of electrolytes, the method of Englezos and Bishnoi (1 988) is applicable. Their procedure uses the activity model developed by Pitzer and Mayorga (1 973) to predict the water activity coefficient in the presence of electrolytes. Their model is not recommended for hydrate forming gases with high solubility in the water phase. Javanmardi and Moshfeghian (1996) and later Javanamardi et al. (1 998) took a different approach and incremented the pure water hydrate formation temperature to predict conditions of incipient hydrate formation in the presence of electrolytes. Englezos (1 992), by using flash calculation algorithms and the fugacity model of Aasberg-Peterson et al.(1 991), predicted the temperature for incipient hydrate formation in single aqueous electrolyte solutions. Javanmardi and Moshfeghian (2000) used the same activity model to predict the hydrate formation temperature in mixed aqueous electrolyte solutions. Yousif and Young (1994), based on modification of the Hammerschmidt model, developed a simple model to predict the hydrate temperature suppression of aqueous mixtures of salts and a single alcohol. Their correlation is restricted to the prediction of the hydrate condition for a specified gas mixture; however, we used it to predict hydrate formation conditions for other systems. In the Nasrifar et al.( 1 998) work, the hydrate formation temperature in the presence of mixed electrolytes and alcohols is calculated by adding a correction to the hydrate formation temperature in the presence of inhibitors. Therefore, their method is a two-step procedure.In the present work, a new method is suggested for direct prediction of hydrate temperature in the presence of both aqueous electrolyte solutions and a single alcohol. The model is based on the extension of the water activity model to the above systems. The suggested method is A new thermodynamic model for calculating the hydrate formation temperature of different hydrate formers in aqueous solutions of both electrolytes and a single alcohol has been presented. This method uses a generalization of the Aasberg-Petersen model for water activity. For calculation of water activity in the presence of electrolytes, the effect of alcohols was taken into account without using any new fitting parameters. The results are in good agreement with published experimental data.

Carbon dioxide and hydrogen sulfide were interacted with anhydrous diglycolamine (DGA), di-2propanolamine (DIPA), methyldiethanolamine (MDEA), and dimethylethanolamine (DMEA) and the reaction products dissolved in DCC13 for NMR analysis. The resultant NMR spectra are compared with the NMR spectra of the unreacted amines at similar concentrations. Evidence for a Lewis acid-base complex involving H2S and DIPA was found. NMR absorptions attributable to the formation of protonated aminexarbamate salts were found for both DGA and DIPA. Although no NMR evidence was obtained for reaction products from DGA, MDEA, or DMEA using either H2S or C02, pressure measurements suggest that Lewis acid-base adducts are reversibly formed in these systems and decompose prior to NMR analysis.

Carbonyl sulfide and methyl mercaptan were interacted with anhydrous monoethanolamine (MEA), diglycolamine (DGA), diethanolamine (DEA), di-2-propanolamine (DIPA), methyldiethanolamine (MDEA), and dimethylethanolamine (DMEA). The reaction products were analyzed by * Author to whom correspondence should be addressed.

A simple thermodynamic model is proposed for predicting hydrate-forming conditions for natural-gas components in the presence of single-component or mixed-electrolyte solutions. The parameters required for use of the model are developed and presented. The model is quite accurate with average deviations between calculated and experimental values of less than 0.5 °C for systems not included in the model-parameter determination.