Advanced Process Simulation PhD course, Dept. of Chemical Engineering www.chemeng.lth.se

KET021F, PhD course at Chemical Engineering, Lund University
last time: spring 2012

Course description: Simulation of process behaviour based on mathematical models is probably the most important general tool that is available in chemical engineering science. The course discusses methods and techniques for the solution of typical problems in process engineering. These are often based on differential-algebraic equations from lumped or distributed systems.

The course will focus on simulation of distributed systems, i.e. models based on partial differential equations. Current techniques result in large and very large equation systems. The course presents solvers for solution of large linear and nonlinear algebraic equation systems and large differential equation system. The purpose is to give the PhD students insight in the behaviour and complexity of modern PDE solvers in order to become an advanced user of computers tools for simulation of distributed systems, like CFD tools.

Schedule, April 10-12
: KET library -1

1.   10/4, 10-12 Course introduction pdf,

Numerical solution of transport problems, pdf

2.   10/4, 13-15 Introduction to FVM, pdf, Finite difference approximations, Finite volume paradigm
Exercise I
Hand-In of exercise I, 20/4

3.   11/4, 9-11 Introduction to FEM, pdf, Method-of-weighted-residuals, finite element, Galerkin and orthogonal collocation,
Exercise II
Hand-In of exercise II, 27/4

4.   11/4, 13-15 Convection problems, pdf, upwind, streamline-diffusion, TVD and flux limiters
Exercise III
Hand-In of exercise III, 4/5

5.   12/4, 9-12 Solution of large systems,
A: dynamic and nonlinear systems
, pdf, implicit ODE/DAE-solvers, Newton and quasi-Newton methods
B: linear systems, pdf, LU, sparseness, iterative methods
Exercise IV
and extra exercise V+
Hand-In of exercise IV, 11/5

6.   25/5, 9-12, Seminar on course exercises

Tools:

The course use MATLAB and COMSOL in the exercises.
Other tools will be presented during the course, like Fluent

Some m-files: Ex1: FVMdisc2nd, FVMdisc1st, FVMdiscBV, FVM_Jpattern, Ex2: jacobi2, lagrange, Ex4: jacobi, gseidel, sor

Reference literature:
Beers, Numerical Methods for Chemical Engineering, Cambridge (2007)

Majumdar, Computational Methods for Heat and Mass Transfer, Taylor&Francis (2005)

Pepper&Heinrich, The Finite Element Method, Taylor&Francis (2006)

Examination: 3 exercise hand-ins (+1 optional), participation at seminar
Points
: 5 hp

Contact: prof. Bernt Nilsson (046 – 222 80 88) bernt.nilsson@chemeng.lth.se