Mathematical modelling of chemical engineering systems by finite element analysis using PDE/PROTRAN (Q914349)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Mathematical modelling of chemical engineering systems by finite element analysis using PDE/PROTRAN |
scientific article |
Statements
Mathematical modelling of chemical engineering systems by finite element analysis using PDE/PROTRAN (English)
0 references
1988
0 references
The PDE/PROTRAN program package is presented (PROTRAN being a powerful preprocessor), theoretical algorithm aspects as well as practical references on two interesting engineering problems. This package can be used to solve two-dimensional problems governed by partial differential equations, like steady-state or time-dependent, as well as eigenvalue problems. In all cases the boundaries must be well defined. Both, in steady-state or time-dependent problems, the piecewise polynomials of linear expansions, depending on space coordinates only or also on time, can be quadratic, cubic or quartic degree polynomials, on triangular finite elements. The iterative solution of the following equation system is carried out by the damped Newton-Raphson method, whilst the proper solution is either a Gaussian, or a frontal respectively Lanczos bi-conjugate gradient method. An error criterion is also given, but the approximation of the Galerkin finite element method can be improved by varying the triangle density. In time-dependent problems a variable time-step can also be used. Nevertheless, solution discontinuities may occur, which can be avoided by tricks like artificial diffusion. The first explanatory example is related to diffusion and reaction in a partially wetted catalyst (a nonlinear mixed boundary value problem). Triangulation criteria are given, as well as modalities to put the boundary conditions. The second example refers to incompressible flow over a step. The results can be expected as independent from local effects, if the number of selected triangles is at least 400. Appendices A and B show the input codes for the two examples.
0 references
numerical examples
0 references
PDE/PROTRAN program package
0 references
preprocessor
0 references
algorithm
0 references
two-dimensional problems
0 references
steady-state
0 references
time-dependent
0 references
eigenvalue problems
0 references
piecewise polynomials
0 references
finite elements
0 references
damped Newton-Raphson method
0 references
0 references
0 references
0 references