Exact Discrete Formulation

We shall use Eq. (4.3) as a base for the following approximate discrete formulations. 

The five steps of formulation considered in the preceding chapters for differential for-mulation are now applied to a finite-difference formulation  

Step 1. Assign a difference system with length Ax to each node, as shown in Fig. 4.1. Let the system be centered around the node. 

Step 2. State the first law, Eq. (1.16), in terms of the system assigned to each node. Since all terms of Eq. (1.16) other than the net heat flux are absent, the first law reduces to a heat balance. For node i, we have in terms of Fig. 4.2, 

Step 3. For the system around node i, express Fourier s law  

---

There are a number of approaches to discrete formulation some are illustrated here in terms of a steady one-dimensional fin problem (recall Ex. 2.9). For an infinitely long fin, with a specified base temperature To, transferring heat with a coefficient h to an ambient at temperature T00, an exact solution for temperature is... 

Table 4.1 compares the results of the foregoing discrete formulations for a typical inner node. Since Eqs. (4.3), (4.7), and (4.9) differ only in the coefficient of 0,-, these coefficients, their percent error relative to the exact coefficient, and their series expansions for small values of B are given in Table 4.1. Figure 4.4 shows the coefficients for larger values of B. As the number of grids n(= t/Ax) is increased, the cell Biot number B decreases, and all formulations approach the exact one.

A typical sequence of steps is presented below. For each step, the current objective value (in discrete space) and a measure of its infeasibility (shortfall in demands met in m /s) is obtained and these are collated in Table 1. Due to the highly constrained nature of the discrete formulation, an exactly feasible solution is unlikely to be achieved. However, the aim is not so much to solve the problem directly with the visualization tool but to provide good initial solutions for the rigorous optimization procedure. 

A mathematical formulation based on uneven discretization of the time horizon for the reduction of freshwater utilization and wastewater production in batch processes has been developed. The formulation, which is founded on the exploitation of water reuse and recycle opportunities within one or more processes with a common single contaminant, is applicable to both multipurpose and multiproduct batch facilities. The main advantages of the formulation are its ability to capture the essence of time with relative exactness, adaptability to various performance indices (objective functions) and its structure that renders it solvable within a reasonable CPU time. Capturing the essence of time sets this formulation apart from most published methods in the field of batch process integration. The latter are based on the assumption that scheduling of the entire process is known a priori, thereby specifying the start and/or end times for the operations of interest. This assumption is not necessary in the model presented in this chapter, since water reuse/recycle opportunities can be explored within a broader scheduling framework. In this instance, only duration rather start/end time is necessary. Moreover, the removal of this assumption allows problem analysis to be performed over an unlimited time horizon. The specification of start and end times invariably sets limitations on the time horizon over which water reuse/recycle opportunities can be explored. In the four scenarios explored in... 

On a lattice, so-called crankshaft moves are trivial implementations of concerted rotations [77]. They have been generalized to the off-lattice case [78] for a simplified protein model. For concerted rotation algorithms that allow conformational changes in the entire stretch, a discrete space of solutions arises when the number of constraints is exactly matched to the available degrees of freedom. The much-cited work by Go and Scheraga [79] formulates the loop-closure problem as a set of algebraic equations for six unknowns reducible... 

The analysis of polymer processing is reduced to the balance equations, mass or continuity, energy, momentum and species and to some constitutive equations such as viscosity models, thermal conductivity models, etc. Our main interest is to solve this coupled nonlinear system of equations as accurately as possible with the least amount of computational effort. In order to do this, we simplify the geometry, we apply boundary and initial conditions, we make some physical simplifications and finally we chose an appropriate constitutive equations for the problem. At the end, we will arrive at a mathematical formulation for the problem represented by a certain function, say / (x, T, p, u,...), valid for a domain V. Due to the fact that it is impossible to obtain an exact solution over the entire domain, we must introduce discretization, for example, a grid. The grid is just a domain partition, such as points for finite difference methods, or elements for finite elements. Independent of whether the domain is divided into elements or points, the solution of the problem is always reduced to a discreet solution of the problem variables at the points or nodal pointsinxxnodes. The choice of grid, i.e., type of element, number of points or nodes, directly affects the solution of the problem. 

In these equations fi is the coluirm mass of dry air, V is the velocity (u, v, w), and (jf) is a scalar mixing ratio. These equations are discretized in a finite volume formulation, and as a result the model exactly (to machine roundoff) conserves mass and scalar mass. The discrete model transport is also consistent (the discrete scalar conservation equation collapses to the mass conservation equation when = 1) and preserves tracer correlations (c.f. Lin and Rood (1996)). The ARW model uses a spatially 5th order evaluation of the horizontal flux divergence (advection) in the scalar conservation equation and a 3rd order evaluation of the vertical flux divergence coupled with the 3rd order Runge-Kutta time integration scheme. The time integration scheme and the advection scheme is described in Wicker and Skamarock (2002). Skamarock et al. (2005) also modified the advection to allow for positive definite transport. 

Mathematical formulations of various boundary conditions were discussed in Section 2.3. These boundary conditions may be implemented numerically within the finite volume framework by expressing the flux at the boundary as a combination of interior values and boundary data. Usually, boundary conditions enter the discretized equations by suppression of the link to the boundary side and modification of the source terms. The appropriate coefficient of the discretized equation is set to zero and the boundary side flux (exact or approximated) is introduced through the linearized source terms, Sq and Sp. Since there are no nodes outside the solution domain, the approximations of boundary side flux are based on one-sided differences or extrapolations. Implementation of commonly encountered boundary conditions is discussed below. The technique of modifying the source terms of discretized equation can also be used to set the specific value of a variable at the given node. To set a value at... 

This is the familiar formulation of Bragg s law for a three-dimensional point lattice. It says that the Fourier transform of a point lattice is absolutely discrete and periodic in diffraction space, and that we can predict when a nonzero diffraction intensity will appear for any family of planes hkl, and what the angle of incidence and reflection 0 must be in order for an intensity to appear. Bragg s law, notice, is completely independent of atoms, or molecules, or unit cell contents. The law is imposed by the periodicity of the crystal lattice, and it strictly governs where we may observe any nonzero intensity in diffraction space. It tells us when the resultant waves produced by the scattering of all of the atoms in the many individual unit cells, each represented by a single lattice point, are exactly in phase. 

There are several reasons for observing differences between the computed results and experimental data. Errors arise from the modeling, discretization and simulation sub-tasks performed to produce numerical solutions. First, approximations are made formulating the governing differential equations. Secondly, approximations are made in the discretization process. Thirdly, the discretized non-linear equations are solved by iterative methods. Fourthly, the limiting machine accuracy and the approximate convergence criteria employed to stop the iterative process also introduce errors in the solution. The solution obtained in a numerical simulation is thus never exact. Hence, in order to validate the models, we have to rely on experimental data. The experimental data used for model validation is representing the reality, but the measurements... 

When linear programming relaxations are not exact, it is important to make them as sharp an approximation as possible. Different formulations of discrete models as (/LP)s can produce quite different linear programming relaxations, even though the models have the same discrete solutions. 

When a discrete optimization problem is in the polynomially solvable complexity category, it is usually clear how to proceed with its analysis. A clever and efficient algorithm is at hand. Often an exact Unear programming formulation is also known, and a strong duality theory is available for sensitivity studies. Very complete analysis should be possible, unless limits on the time available for solution (e.g., in a real-time setting) mandate quicker methods. 

So far we have considered the homogeneous case for which the waiting time density is independent of the position of the particles or their state. Let us formulate the general equations describing a random walk with discrete states in continuous time for which the waiting time PDF depends on the current state. (CTRWs with space-dependent waiting time PDFs have been studied in [75].) We introduce the mean density of particles Pmit) in state m and the density of particles j (t) arriving in state m exactly at time t. The balance equations can be written as... 

To avoid the evaluation of these quantities, the continuous phase equilibrium function is expressed in terms of the single-space basis e.g. K T,p, F ) G Vn- Using the fast Wavelet transform T the single-scale representation of the trial solution is obtained and the residuum of (1.3) can be formulated in terms of the single-scale basis The residuum is thus expressed by means of scaling functions of common level n. A subsequent fast wavelet transform provides the residuum of (1.3) in the multiscale representation, whose inner products with the weight functions can easily be evaluated due to orthogonality. This approach exactly recovers the discrete model without any deviations due to the continuous model formulation if the Haar basis is used. 

It is possible to make simplified calculations to determine approximate interior pressure conditions by assuming uniform vent flow and essentially working the problem in steps over discrete time intervals. The purpose of the present paper, however, is to set up exact thermodynamic and flow equations. A critical parameter in the results arises in connection with the net effect of the four processes mentioned above. After the thermodynamic formulation, the magnitude of these effects will be discussed. 

The explanation for these regular series lies in the existence of discrete, quantized energy levels. In 1913 Niels Bohr was able to derive the formula for these series in terms of the ad hoc quantum assumptions of the BOHR THEORY. In the mid-1920s the formula was derived in a deductive way from quantum mechanics. In the wave mechanics formulation of quantum mechanics it is possible to derive the formula because the Schrodinger equation can be solved exactly for the hydrogen atom. 

---