Gradients 1 Model Equations

The gradient model has been combined with two equations of state to successfully model the temperature dependence of the surface tension of polar and nonpolar fluids [54]. Widom and Tavan have modeled the surface tension of liquid He near the X transition with a modified van der Waals theory [55]. 

When developing the dusty gas model flux relations in Chapter 3, the thermal diffusion contributions to the flux vectors, defined by equations (3.2), were omitted. The effect of retaining these terms is to augment the final flux relations (5.4) by terms proportional to the temperature gradient. Specifically, equations (5.4) are replaced by the following generalization... 

As shown in this chapter, in the simulation of systems described by partial differential equations, the differential terms involving variations with respect to length are replaeed by their finite-differenced equivalents. These finite-differenced forms of the model equations are shown to evolve as a natural eonsequence of the balance equations, according to the manner of Franks (1967). The approximation of the gradients involved may be improved, if necessary, by using higher order approximations. Forward and end sections can... 

The nonlinear programming problem based on objective function (/), model equations (b)-(g), and inequality constraints (was solved using the generalized reduced gradient method presented in Chapter 8. See Setalvad and coworkers (1989) for details on the parameter values used in the optimization calculations, the results of which are presented here. 

The dimensionless model equations are used in the program. Since only two boundary conditions are known, i.e., S at X = l and dS /dX at X = 0, the problem is of a split-boundary type and therefore requires a trial and error method of solution. Since the gradients are symmetrical, as shown in Fig. 1, only one-half of the slab must be considered. Integration begins at the center, where X = 0 and dS /dX = 0, and proceeds to the outside, where X = l and S = 1. This value should be reached at the end of the integration by adjusting the value of Sguess at X=0 with a slider. 

The CSTR model, on the other hand, is based on a stirred vessel with continuous inflow and outflow (see Fig. 1.2). The principal assumption made when deriving the model is that the vessel is stirred vigorously enough to eliminate all concentration gradients inside the reactor (i.e., the assumption of well stirred). The outlet concentrations will then be identical to the reactor concentrations, and a simple mole balance yields the CSTR model equation ... 

The study of the intra-phase mass transfer in SCR reactors has been addressed by combining the equations for the external field with the differential equations for diffusion and reaction of NO and N H 3 in the intra-porous region and by adopting the Wakao-Smith random pore model to describe the diffusion of NO and NH3 inside the pores [30, 44]. The solution of the model equations confirmed that steep reactant concentration gradients are present near the external catalyst surface under typical industrial conditions so that the internal catalyst effectiveness factor is low [27]. 

When only the control is discretized as described in section 5.7, integration of the model equations is required to evaluate the performance index and to obtain the gradients. The evaluation of such gradients will consume a significant part of the total computational time needed to solve the optimisation problem. 

With the results of the forward integration of the model Equation 5.4 and then by integrating backward the adjoint equations (Equation 5.18) the gradients can be determined from ... 

The model equation (Equation 5.4) is integrated only once simultaneously with the trajectory sensitivity equation (Equation 5.22) for each of the elements of y, to determine the final state sensitivity for all the gradients in equation (Equation 5.26), i = 0, 1,.,S. Changes in (zj) in te[(tj.], tj), do not affect the state trajectory for any t smaller than tj h because (zj) acts during the interval te[(tph tj). Therefore, the gradient dG, Jdzj is obtained by integrating the respective trajectory sensitivity... 

In addition to the above-mentioned problem, numerical difficulties may arise. The system (model equations) describing the multicomponent off-cut recycle operation needs to be reinitialised at the end of each main-cut and off-cut to accommodate the next off-cut to the reboiler. To optimise these initial conditions (new mixed reboiler charge and its composition) it is essential to obtain the objective function gradients with respect to these initial conditions. 

Solution of optimisation problems using rigorous mathematical methods have received considerable attention in the past (Chapter 5). It is worth mentioning here that these techniques require the repetitive solution of the model equations (to evaluate the objective function and the constraints and their gradients with respect to the optimisation variables) and therefore computationally can be very expensive. 

The model equations are based on the assumption that the disproportionation of EB, the present case study, is carried out in a gradientless recycle reactor (Fig. 26), where neither concentration gradients... 

The anisotropy introduces two new features (i) equations (6.305) and (6.306) cannot in general be transformed into each other, as the drift term V D may not be a gradient field. Equation (6.306) can describe systems where the directions of the principal axes depend on the spatial position, (ii) Detailed balance implies that the diffusion flow J vanishes everywhere in the stationary state. However, this is not automatically satisfied for anisotropic systems and one needs to exercise extra care in the modeling of such systems. Inhomogeneity does not affect the detailed balance, (iii) The diffusive part of the diffusion flow must be represented by J = VD73, while the drift is represented by (PV D). 

The axial dispersion flow model can be valid when we do not have the gradient of the property with respect to the normal flow direction. In other words, for this direction, we have a perfect mixing state. When this last condition is not met, we have to consider a flow model with two dispersion coefficients a coefficient for the axial dispersion and another one for the radial dispersion. In this case, the flow model equation becomes ... 

The selection and design of a reactor for bench-scale kinetic experiments should be considered case by case. It is important to stress, however, that one should not try to build a bench-scale replica of what is believed to be or is the industrial reactor. Industrial reactors are designed to operate a process in a profitable way, which is not the case for experimental reactors. In industrial reactors heat, mass and momentum transport has to occur in an economically justifiable way, leading in general to temperature, concentration and/or pressure gradients inside the reactor. Also, the hydrodynamics can be rather complicated. Fluidized beds, bubble columns and trickle-flow reactors require model equations that involve several physical parameters, besides the intrinsic kinetic parameters. Empirical... 

Circulation models are based on the equations of motion of the geophysical fluid dynamics and on the thermodynamics of seawater. The model area is divided into finite size grid cells. The state of the ocean is described by the velocity, temperature, and salinity in each grid cell, and its time evolution can be computed from the three-dimensional model equations. To reduce the computational demands, the model ocean is usually incompressible and the vertical acceleration is neglected, the latter assumption is known as hydrostatic approximation. This removes sound waves in the ocean from the model solution. In the horizontal equations, the Boussinesq approximation is applied and small density changes are ignored except in the horizontal pressure gradient terms. This implies that such models conserve... 

Open Boundary Conditions The regional Baltic Sea model has a western open boundary in the Skagerrak. At open boundaries, the model equations for the calculation of horizontal gradients are not complete and boundary conditions are needed to close the numerical scheme. The open boundary conditions implemented in the Baltic Sea version of MOM-3.1 have the following three main components ... 

The latter process also has to be parameterized because the model equations are based on the hydrostatic approximation. Convection is of special importance in the surface layer during the autumn and winter, and it is responsible for the formation of winter water layers in the deep basins of the Baltic Sea. Surface salinity well below 24 implies a maximum density at temperatures above the freezing point and a stable re-stratification before the freezing point is reached. Hence, a good approximation for the equation of state is required to simulate such processes, even if the existence of large density gradients could suggest that a simple and numerically cheap approximation could be sufficient. 

This relation is referred to as the Maxwell-Stefan model equations, since Maxwell [65] [67] was the first to derive diffusion equations in a form analogous to (2.302) for dilute binary gas mixtures using kinetic theory arguments (i.e., Maxwell s seminal idea was that concentration gradients result from the friction between the molecules of different species, hence the proportionality coefficients, Csk, were interpreted as inverse friction or drag coefficients), and Stefan [92] [93] extended the approach to ternary dilute gas systems. It is emphasized that the original model equations were valid for ordinary diffusion only and did not include thermal, pressure, and forced diffusion. 

---