Coal Gas Adsorption Research Articles

Modeling of coal swelling induced by water vapor adsorption

Gas adsorption-induced coal swelling is a well-know phenomenon. Coal swelling or shrinkage by adsorption or desorption of water vapor has not been well understood but has significant implications on gas drainage process for underground coal mining and for primary and enhanced coalbed methane production. Decreased matrix moisture content leads to coal shrinkage and thus the change of cleat porosity and permeability under reservoir conditions. Unlike gas adsorption in coal which usually forms a single layer of adsorbed molecules, water vapor adsorption in the coal micropores forms multilayer of adsorbed molecules. In this work, a model has been developed to describe the coal swelling strain with respect to the amount of moisture intake by the coal matrix. The model extended an energy balance approach for gas adsorption-induced coal swelling to water vapor adsorption-induced coal swelling, assuming that only the first layer of adsorbed molecules of the multilayer adsorption changes the surface energy, which thus causes coal to swell. The model is applied to describe the experimental swelling strain data measured on an Australian coal. The results show good agreement between the model and the experimental data.

Application of a New Multicomponent Gas Adsorption Model to Coal Gas Adsorption Systems

Summary A new multicomponent adsorption model is proposed for application to coal gas adsorption systems. The model is derived by combining the vacancy solution and Dubinin-Polanyi theories. Applications of the new adsorption model include the modeling of multicomponent adsorption processes associated with primary and enhanced coalbed methane recovery (ECBM). In the new model, the adsorbed phase in the single-component (pure) adsorption system is treated as a binary mixture of a singlecomponent gas with a hypothetical "vacancy" species, which also occupies adsorption space. The adsorption system is modeled as equilibrium between the gaseous phase and the adsorbed-phase vacancy solution. The Dubinin-Astakhov (D-A) equation is used to generate activity coefficients, as a function of the degree of porefilling, for the pure component gas in the binary (adsorbate+vacancy) adsorbed-phase mixture. The Wilson equation is chosen to fit pure component (D-A-derived) activity coefficient curves by optimizing the binary interaction coefficients. These binary interaction coefficients are then used to predict multicomponent adsorption equilibrium, although only the case of binary adsorption is modeled here. The adsorbed phase mixture for binary gas adsorption is treated as a ternary mixture of the two pure component adsorbates and the hypothetical vacancy species. Binary gas adsorption equilibrium is described by equilibrium between the gaseous components and the components in the adsorbed phase solution. Adsorbed-phase activity coefficients are calculated from the Wilson equation, with the binary interaction coefficients obtained from pure component adsorption data. Thus, only pure component adsorption data are required to make binary and multicomponent adsorption predictions with the new model. Two binary (CH4+CO2) gas adsorption experimental data sets with coal as the adsorbent and one binary (CH4+C2H6) gas adsorption data set with activated carbon as the adsorbent are used to test the predictions of the new model. In most cases the new model is able to predict binary gas adsorption accurately. The poor fit of the Wilson equation to the D-A-derived activity coefficients for some pure component data suggests that some improvement in model predictions could be made with the choice of a different activity coefficient equation. A unique feature of the current model is the ability to predict multicomponent gas adsorption at different temperatures from the pure component adsorption data collected at a single temperature. The temperature independence of pure component "characteristic" curves, as demonstrated in Dubinin-Polanyi theory, allows pure component adsorption to be predicted for a range of temperatures. These pure component data can then be used in modeling binary or multicomponent adsorption data at various temperatures. This is demonstrated for one experimental binary gas adsorption data set.