Direct finite element method cases

Neither the finite element method nor a discretised integration will "catch" eddy or other motion below a certain scale determined by the choice of mesh. Small-scale motion in many cases may be better described as random, in which case the transport of the quantity A is called diffusion. Diffusion can be described by Pick s law, assuming that the flux density /, i.e., the number of "particles" (here a small parcel of a gas), passing a unit square in a given direction, is proportional to the negative gradient of particle concentration n (Bockris and Despic, 2004),... 

The pairwise Brownian dynamics method has several advantages over numerical methods based on Smoluchowski s [9] approach (e.g., finite element method), and we discuss these here. The primary advantage of the method is the ease of mathematical formulation even for cases involving complex reaction site geometries, hydrodynamic interactions, charge effects, anisotropic diffusion and flow fields. Furthermore the method obviates the need to solve complex diffusion equations to obtain the concentration field from which the rate constant is calculated in the Smoluchowski method. In contrast, the rate constant is obtained directly in the pairwise Brownian dynamics method. The effective rate constants for different reaction conditions may be obtained from a single simulation this is not possible using the finite element method. 

In view of the solution of the potential model with non-linear boundary conditions an alternative Newton-Raphson iteration process was constructed in the case of the finite element method and an original one was obtained for the direct boundary element method. 

The major difference between the modeling of a problem with the finite element and boundary element method is related to the discretization, cf. O Fig. 26.17. In the case of the finite element method, the entire domain must be discretized, whereas the boundary element method requires only the discretization of the boundary. In addition, a symmetry plane does not require a discretization. As a direct result, the size of the obtained system of equations can be much lower than in finite element approaches. 

There are a few points with respect to this procedure that merit discussion. First, there is the Hessian matrix. With elements, where n is the number of coordinates in the molecular geometry vector, it can grow somewhat expensive to construct this matrix at every step even for functions, like those used in most force fields, that have fairly simple analytical expressions for their second derivatives. Moreover, the matrix must be inverted at every step, and matrix inversion formally scales as where n is the dimensionality of the matrix. Thus, for purposes of efficiency (or in cases where analytic second derivatives are simply not available) approximate Hessian matrices are often used in the optimization process - after aU, the truncation of the Taylor expansion renders the Newton-Raphson method intrinsically approximate. As an optimization progresses, second derivatives can be estimated reasonably well from finite differences in the analytic first derivatives over the last few steps. For the first step, however, this is not an option, and one typically either accepts the cost of computing an initial Hessian analytically for the level of theory in use, or one employs a Hessian obtained at a less expensive level of theory, when such levels are available (which is typically not the case for force fields). To speed up slowly convergent optimizations, it is often helpful to compute an analytic Hessian every few steps and replace the approximate one in use up to that point. For really tricky cases (e.g., where the PES is fairly flat in many directions) one is occasionally forced to compute an analytic Hessian for every step. 

---