Page 262 - Fundamentals of Enhanced Oil and Gas Recovery
P. 262

250                                                                     Alireza Keshavarz et al.


                   Eq. (8.15) reveals that rather than pressure gradient, diffusion is dependent on con-
                centration gradient. Fick’s second law could be solved in terms of fractional adsorp-
                tion/desorption:

                                                  N                  !
                                     M t        6  X  1        2 2  Dt
                                         5 1 2         exp 2n π                       (8.16)
                                    M N        π 2   n 2           r 2
                                                  n51              p
                where M t is the amount of adsorbed/desorbed gas at time t, M N is the total amount
                of adsorbed/desorbed gas at equilibrium condition, D is the diffusion coefficient, r p is
                the mean radius of coal particle radius.
                   In order to estimate diffusion coefficient measure, D, at a given time, the experi-
                mental data is inserted into Eq. (8.16), thereby D is calculated. Since in this equation
                micropores are assumed to be monosized, this model is referred to as “unipore diffu-
                sion model.” Although unipore models have been applied to coals, they were proved
                to fit data only over restricted time intervals [44,50,51]. Thus, predicting gas flow in
                heterogeneous pore structures requires the consideration of different pore sizes in dif-
                fusion models.


                8.4.2.2 Bidisperse Model
                Considering the fact that the coal structure is highly heterogeneous, the unipore model
                usually does not predict the diffusion coefficient precisely [52 54]. Ruckenstein et al.
                proposed a bidisperse diffusion model to more realistically describe the pore size distri-
                bution and consequently more accurately predict the diffusion behavior [55]. Bidisperse
                model limits pore size distribution to two sizes: macropore and micropore. In this
                model, the adsorbent contains microporous spherical particles separated by inter-particle
                macropores. The bidisperse model was applied to coal for the first time by Smith and
                Williams [56,57]. In this approach, coal matrix is assumed as a double porosity medium
                with two distinct pore sizes, macropores indicating fast diffusion [Eq. (8.17)]and micro-
                pores characterized by slow diffusion, as observed in Eq. (8.18) [58,59].

                                                  N                  !
                                    M at       6  X  1         2 2  D a t
                                         5 1 2         exp 2n π                       (8.17)
                                    M aN       π 2   n 2           r 2
                                                  n51              pa
                                                  N                  !
                                     M it      6  X  1         2 2  D i t
                                         5 1 2         exp 2n π                       (8.18)
                                    M iN       π 2   n 2           r 2
                                                  n51              pi
                where M at and M it are the gas adsorption/desorption amount from macropores and
                micropores at time t, respectively; M aN and M iN are the total amount of adsorbed/
                desorbed gas in macropores and micropores at equilibrium condition, respectively; D a
                and D i are the macropores’ and micropores’ diffusion coefficients, respectively.
   257   258   259   260   261   262   263   264   265   266   267