Nonequilibrium, Rate-Based Model

Although the equilibrium-based model, modified to incorporate stage efficiency, is adequate for binary mixtures and for the major components in nearly ideal multicomponent mixtures, that model has serious deficiencies for more general multicomponent vapor-liquid mixtures. Murphree himself stated clearly the limitations 

When the equilibrium-based model is applied to multicomponent mixtures, a number of problems arise. Values of EMG differ from component to component and vary from stage to stage. But at each stage, the number of independent values of EMG must be determined so as to force the mole fractions in the vapor phase to sum to 1. This introduces the possibility that negative values of EMG can result. This is in contrast to binary mixtures for which the values of EMG are always positive and are identical for the two components. 

Krishna et al. (1977) showed that when the vapor mole-fraction driving force of a component (call it A) is small compared to the other components in the mixture, the transport rate of A is controlled by the other components, with the result that Emg for A is anywhere in the range from minus infinity to plus infinity. They confirmed this theoretical prediction by conducting experiments with the ethanol/ferf-butanol/water system and obtained values of EMG for /er/ butanol ranging from -2978% to +527%. In addition, the observed values of EMG for ethanol and water sometimes differed significantly. 

In the rate-based models, the mass and energy balances around each equilibrium stage are each replaced by separate balances for each phase around a stage, which can be a tray, a collection of trays, or a segment of a packed section. Rate-based models use the same m-value and enthalpy correlations as the equilibrium-based models. However, the m-values apply only at the equilibrium interphase between the vapor and liquid phases. The accuracy of enthalpies and, particularly, m-values is crucial to equilibrium-based models. For rate-based models, accurate predictions of heat-transfer rates and, particularly, mass-transfer rates are also required. These rates depend upon transport coefficients, interfacial area, and driving forces. It 

The general forms for component mass-transfer rates across the vapor and liquid films, (JAf v, JV. L), respectively, on a tray or in a packed segment are as follows, where both diffusive and bulk-flow contributions are included  

---

Two different approaches have evolved for the simulation and design of multicomponent distillation columns. The conventional approach is through the use of an equilibrium stage model together with methods for estimating the tray efficien -cy. An alternative approach, the nonequilibrium, rate-based model, applies rigorous multicomponent mass- and heat-transfer theory to distillation calculations. This non-... 

The rate-based models suggested up to now do not take liquid back-mixing into consideration. The only exception is the nonequilibrium-cell model for multicomponent reactive distillation in tray columns presented in Ref. 169. In this work a single distillation tray is treated by a series of cells along the vapor and liquid flow paths, whereas each cell is described by the two-film model (see Section 2.3). Using different numbers of cells in both flow paths allows one to describe various flow patterns. However, a consistent experimental determination of necessary model parameters (e.g., cell film thickness) appears difficult, whereas the complex iterative character of the calculation procedure in the dynamic case limits the applicability of the nonequilibrium cell model. 

Real distillation processes, however, nearly always operate away from equilibrium. In recent years it has become possible to simulate distillation and absorption as the mass-transfer rate-based operations that they really are, using what have become known as nonequilibrium (NEQ) or rate-based models [Taylor et ah, CEP (July 28, 2003)]. 

In recent years a new approach to the modeling of distillation and absorption processes has become available the nonequilibrium or rate-based models. These models treat these classical separation processes as the mass-transfer rate governed processes that they really are, and avoid entirely the (a priori) use of concepts such as efficiency and HETP [Krishnamurthy and Taylor, AZChE/., 31, 449-465 (1985) Taylor, Kooijman, and Hung, Comput. Chem. Engng., 18, 205-217 (1994)]. 

Transport models employing the bicontinuum-sorption formulation, with one domain equilibrium controlled, were presented by Selim et al. (1976) and Cameron and Klute (1977), while Selim et al. (1976) also presented a model where both domains were rate limited. The one-site model mentioned previously is a special case of the two-site model, where all sorption sites are assumed to be of the time-dependent class (Selim et al., 1976 van Genuchten, 1981). The bicontinuum-based model has generally been able to represent nonequilibrium data much better than has the one-site model. 

For gaseous flames, the LES/FMDF can be implemented via two combustion models (1) a finite-rate, reduced-chemistry model for nonequilibrium flames and (2) a near-equilibrium model employing detailed kinetics. In (1), a system of nonlinear ordinary differential equations (ODEs) is solved together with the FMDF equation for all the scalars (mass fractions and enthalpy). Finite-rate chemistry effects are explicitly and exactly" included in this procedure since the chemistry is closed in the formulation. In (2). the LES/FMDF is employed in conjunction with the equilibrium fuel-oxidation model. This model is enacted via fiamelet simulations, which consider a laminar counterflow (opposed jet) flame configuration. At low strain rates, the flame is usually close to equilibrium. Thus, the thermochemical variables are determined completely by the mixture fraction variable. A fiamelet library is coupled with the LES/FMDF solver in which transport of the mixture fraction is considered. It is useful to emphasize here that the PDF of the mixture fraction is not assumed a priori (as done in almost all other flamelet-based models), but is calculated explicitly via the FMDF. The LES/FMDF/flamelet solver is computationally less expensive than that described in (1) thus, it can be used for more complex flow configurations. 

The results of the present 10-level model studies have shown that nonequilibrium rates are in general quite different from rates calculated based on constant equilibrium distributions of reactant molecules. In particular, for exothermic reactions, the nonequilibrium rates will decrease to a value of zero. It is possible that the nonequilibrium rates may approach the equilibrium rate within a reasonable amount of time, but only for some slightly endothermic cases or for cases where both reactants and products are almost in the same energy state and consist of the same number of energy levels such as the previous studies of the four-level model. Further studies of this treatment for a larger number of energy levels and for complicated but realistic systems would be of interest. 

The simplest model for droplet evaporation is based on an equilibrium uniform-state model for an isolated droplet [28-30]. Miller et al. [31] investigated different models for evaporation accounting for nonequilibrium effects. Advanced models considering internal circulation, temperature variations inside the droplet, and effects of neighboring droplets [30] may alter the heating rate (Nusselt number) and the vaporization rates (Sherwood number). For the uniform-state model, the Lagrangian equations governing droplet temperature and mass become [28-30]... 

It should be stated that the number of equations in this approach is very large, and specialized computer programs have been developed far their solution. Our objective here is very limited, namely to provide an idea of the basis of these equations. The basic set of equations in such rate based or nonequilibrium models of a distillation plate/ stage is sometimes referred to as the MERSHQ equations (Taylor et id., 2003), with each letter representing equations for a particular aspect of the problem for any plate n of the n-component system. 

Artola-Garicano et al. [27] compared their measured removals of AHTN and HHCB [24] to the predicted removal of these compounds by the wastewater treatment plant model Simple Treat 3.0. Simple Treat is a fugacity-based, nine-box model that breaks the treatment plant process into influent, primary settler, primary sludge, aeration tank, solid/liquid separator, effluent, and waste sludge and is a steady-state, nonequilibrium model [27]. The model inputs include information on the emission scenario of the FM, FM physical-chemical properties, and FM biodegradation rate in activated sludge. 

Jardine et al. (1985b) employed a two-site nonequilibrium transport model to study Al sorption kinetics on kaolinite. They used the transport model of Selim et al. (1976b) and Cameron and Klute (1977). Based on the above model, Jardine et al. (1985a) concluded that there were at least two mechanisms for Al adsorption on Ca-kaolinite. It appeared that there were equilibrium (type-1) reactions on kaolinite that involved instantaneous Ca-Al exchange and rate-limited reaction sites (type-2) involving Al polymerization on kaolinite. The experimental breakthrough curves (BTC) conformed well to the two-site model. 

The outline of the remainder of this contribution is as follows. In Section 3.7.2, we discuss radical anion dissociation in solution, in which a conical intersection has an important impact on the ground state reaction barrier, rate constant and reaction path, all of which are also influenced by nonequilibrium solvation. The excited electronic state conical intersection problem for the cis-trans isomerization of a model protonated Schiff base in solution is discussed in Section 3.7.3, focusing on the approach to, and passage through, the conical intersection, and the influence of nonequilibrium solvation thereupon. Some concluding remarks are offered in Section 3.7.4. We make no attempt to give a complete discussion for these topics, but rather focus solely on several highlights. Similarly, the references herein are certainly incomplete. We refer the interested reader to refs [1-9] for much more extensive discussions and references. 

The relaxation, inside-out, and bomotopy-continuation methods are extensions of whole or part of the first four methods in order to solve difficult systems or columns. The nonequilibrium models are rate- or transport phenomena-based methods that altogether do away with the ideal-stage concept and eliminate any use of efficiencies. They are best suited for columns where a theoretical stage is difficult to define and efficiencies are difficult to predict or apply. 

The prediction and use of stage efficiencies are described in detail in Sec. 14. Alternative approaches based on mass-transfer rates are preferred, as described in the subsection below. Nonequilibrium Modeling. 

---