The Mixture Models

In this section the classical continuum theory of mixtures is reviewed [ 17]. In this concept the multiphase mixture is treated as a single homogeneous continuum. Thereby the balance principle can be applied to derive conservation laws for the macroscopic pseudo-fluid in analogy to the single phase formulation examined in Chap. 1. Approximate constitutive equations are postulated for the expected macroscopic behavior of the phases. 

Consider any quantity V m associated with a mixture of phases. For an arbitrary material macroscopic control volume ym t) bounded by the surface area Am(t), a generalized integral balance can be postulated stating that the time rate of change is equal to the molecular transport flux plus the volumetric production [200]  

To relate the classical mixture theory to the more familiar volume averaging method we may assume that the mixture CV, which is larger than a phase element but smaller than the characteristic domain dimension, coincides with the averaging volume used in the volume averaging approach. 

Since (3.423) must be satisfied for any Vm, the macroscopic equation can be transformed into its differential form  

The basic mixture model equation (3.424) contains several undetermined quantities which have to be determined. These variables are conventionally postulated as the sum of the volume fraction weighted property values of the phases in the mixture  

To determine the dispersed phase velocities as occurring in the phasic continuity equations in both formulations, the momentum equation of the dispersed phases are usually approximated by algebraic equations. Depending on the concept used to relate the phase k velocity to the mixture velocity the extended mixture model formulations are referred to as the algebraic slip-, diffusion- or drift flux models. 

These mixture models have been applied to investigate chemical engineering problems by [17, 163, 202, 182, 145, 165, 230, 85, 193, 19], among others. 

In this notation, n is the outward directed normal unit vector to the external mixture surface closing the mixture volume lAi(f), D/Dt is the 

---

Computationally, the use of pseudocomponents improves the conditioning of the numerical procedures in fitting the mixture model. Graphically, the expansion of the feasible region and the rescaling of the plot axes allow a better visualization of the response contours. 

The parameter estimation for the mixture model (Equation 5.25) is based on maximum likelihood estimation. The likelihood function L is defined as the product of the densities for the objects, i.e.,... 

For computational reasons, we used a different program as in Section 5.3.2 to fit the mixture models. This program is based on the ideas of Hastie and Tibshirani (1996) to use several prototype objects representing each group in order to obtain more accurate predictions. 

The tangent plane in M for a response surface fitted with In P can also be described by using the adjusted mixture model (i.e. the restrictions of the mixture models are under consideration 0< x< 1 x,+ X2+ 3= 1) (equation (3) becomes equation (21a)) ... 

Uniformist, Average Models. We divide the current water structure models into two major categories. The first treats water essentially as an unstructured liquid while the second admits the simultaneous existence of at least two states of water—i.e., the structural models which Frank has termed the mixture models. ... 

Mixture Models Broken-Down Ice Structures. Historically, the mixture models have received considerably more attention than the uniformist, average models. Somewhat arbitrarily, we divide these as follows (1) broken-down ice lattice models (i.e., ice-like structural units in equilibrium with monomers) (2) cluster models (clusters in equilibrium with monomers) (3) models based on clathrate-like cages (again in equilibrium with monomers). In each case, it is understood that at least two species of water exist—namely, a bulky species representing some... 

We have discussed some examples which indicate the existence of thermal anomalies at discrete temperatures in the properties of water and aqueous solutions. From these and earlier studies at least four thermal anomalies seem to occur between the melting and boiling points of water —namely, approximately near 15°, 30°, 45°, and 60°C. Current theories of water structure can be divided into two major groups—namely, the uniformist, average type of structure and the mixture models. Most of the available experimental evidence points to the correctness of the mixture models. Among these the clathrate models and/or the cluster models seem to be the most probable. Most likely, the size of these cages or clusters range from, say 20 to 100 molecules at room tempera-... 

Starting from the mixture model, the structural behavior of water in the presence of dissolved simple ions is discussed from the point of view of defect formation and lattice distortions at interfaces. The observed behavior of the ions and the water lattice is applied to a number of unsolved biological problems in an attempt to elucidate the specific interface phenomena that are characteristic of such systems. 

Component-based approach. This is an option if the mixture composition can be determined, for example, by means of chemical analysis, and if a mixture model is available that can predict the mixture effects. The mixture model can either be simple, for example, summation of PEC/PNEC ratios over all compounds into a hazard index (HI) moderately complex, for... 

While the thermodynamic evidence may favor the mixture models, the diffraction studies from the static crystalline state tend to support the continuum model. The water molecules in the ices and high hydrates are always four-coordinated. 

Clustering can be viewed as a density estimation problem. The basic premise used in such an estimation is that in addition to the observed variables (i.e., descriptors) for each compound there exists an unobserved variable indicating the cluster membership. The observed variables are assumed to arrive from a mixture model, and the mixture labels (cluster identifiers) are hidden. The task is to find parameters associated with the mixture model that maximize the likelihood of the observed variables given the model. The probability distribution specified by each cluster can take any form. Although mixture model methods have found little use in chemical applications to date, they are mentioned here for completeness and because they are obvious candidates for use in the future. 

The most widely used and most effective general technique for estimating the mixture model parameters is the expectation maximization (EM) algorithm. " It finds (possibly suboptimally) values of the parameters using an iterative refinement approach similar to that given above for the k-means relocation method. The basic EM method proceeds as follows ... 

The mixture model in the high dimensional feature space [Pg.190]

In problems in which the dispersed phase momentum equations can be approximated and reduced to an algebraic relation the mixture model is simpler to solve than the corresponding multi-fluid model, however this model reduction requires several approximate constitutive assumptions so important characteristics of the flow can be lost. Nevertheless the simplicity of this form of the mixture model makes it very useful in many engineering applications. This approximate mixture model formulation is generally expected to provide reasonable predictions for dilute and uniform multiphase flows which are not influenced by any wall effects. In these cases the dispersed phase elements do not significantly affect the momentum and density of the mixture. Such a situation may occur when the dispersed phase elements are very small. There are several concepts available for the purpose of relating the dispersed phase velocity to the mixture velocity, and thereby reducing the dispersed... 

This form of the mixture model is called the drift flux model. In particular cases the flow calculations is significantly simplified when the problem is described in terms of drift velocities, as for example when ad is constant or time dependent only. However, in reactor technology this model formulation is restricted to multiphase cold flow studies as the drift-flux model cannot be adopted simulating reactive systems in which the densities are not constants and interfacial mass transfer is required. 

Maiminen M, Taivassalo V, Kallio S (1996) On the mixture model for multiphase flow. Technical Research Center of Finland VIT Publications, Espoo. 

The mixture model is assessed for a hypothetical example and an example for WinBUGS code is given in Figure 5.2. Evaluation of the individual model predictions for one- and two-compartment models is shown on lines 5 and 7,... 

Due to the limited information that is available describing the use of mixture models, we have exemplihed the process in a brief series of simulations. The full analysis is available elsewhere (44). To assess the performance of the mixture model, a hypothetical data set was constructed. Concentrations at predetermined time points were simulated from a two-compartment (2-c) model with bolus input ... 

Here the mixture model is set to describe a steady linear decline when patients worsen over time and an model for patients who improve. The selection of the Eniax model was based on the fact that the disease severity score had a natural maximum. 

Note that the rest of the code is still needed to run NONMEM successfully, but the parts of the code relevant to the mixture model are presented in this chapter.)... 

Since theta(5) is a probability it must be constrained to the interval [0,1] in the THETA block. Also note that changing from the nonmixture model to the mixture model required the addition of two new theta parameters. One was used to control the probability partition, and the other to specify how the two subpopulations differed. For now, note that neither can be entered into the model uniquely. They must both go into the model together, or be removed from the model together (the designated driver system), and this leads to issues regarding the hypothesis testing for the presence of a mixture (see Section 28.5). Two control stream/report/output table pairs (C2. txt/r2. TXT/T2A. TXT and C3. txt/r3. txt/t3a. txt) can be... 

Inclusion of the posthoc option instructs NONMEM to obtain the Bayesian post hoc ETA estimates when the first-order method is used. These effects and other relevant parameters can be output into a table using the table record. Thereafter, the distribution of the effects can be characterized, including skewness if present. Both the mixture model and the nonmixture models need to be reestimated with the first-order method, as one cannot compare the mofs in a meaningful way between models differing only in estimation method. The mof has dropped 676 points between the nonmixture model (see r5.txt) and the mixture model (r4.txt). Furthermore, the mixture model run has now concluded with a successful covariance step. A choice has to made whether to make two plots (one for each subpopulation) or one (after all, the etas all share the same distribution). The latter approach is shown in Figure 28.2. Similar plots can be generated for each subpopulation. 

For the mixture model a bit more work is needed to get the job done. Because averaging pred defined data items is problematic when a mixture is employed, one might be tempted to use two problems concatenated. The first might generate the expected predictions under the mixture model and table them, and the second could... 

Flere, bias is seen in the plot from the nonmixture model, supporting the notion that the mixture model might be superior. 

T. Shiiki, Y. Hashimoto, and K. Inui, Simulation for population pharmacodynamic analysis of dose-ranging trials usefulness of the mixture model analysis for detecting nonresponders. Pharm Res 19 909-913 (2002). 

We start with detailed definitions of the singlet and the pair distribution functions. We then introduce the pair correlation function, a function which is the cornerstone in any molecular theory of liquids. Some of the salient features of these functions are illustrated both for one- and for multicomponent systems. Also, we introduce the concepts of the generalized molecular distribution functions. These were found useful in the application of the mixture model approach to liquid water and aqueous solutions. 

---