University of Tennessee
Department of Chemical Engineering


ChE 415 Course Website



Project Description
In this project, you will be asked to perform a multivariate nonlinear global optimization of a model to experimental data. The code must be run on a multi-processor machine and must make efficient use of all processors available to it. The theoretical model will contain 5 or 6 parameters, which will be fit to approximately 250 data points of self-diffusivities of mixtures of methane and ethane across a broad range of compositions (from pure methane to pure ethane), temperatures (from 300 to 700 K), and pressures (from 10-2 to 103 atm).

Project Materials
You will need the following materials.

  • Hand-out: A Primer for Parallel Implementation of Molecular Dynamics Simulations (This contains information on how to compile and execute code on our cluster.)
  • Background: A recent paper describing how this data was generated and showing some plots of the data. (This contains useful background information.)
  • Theory: A brief description of the theoretical model.
  • FORTRAN Code:This code performs multivariate nonlinear global optimization of an arbitrary objective function with respect to an arbitrary number of parameters.
  • Data File:This text file contains the self-diffusivity data for methane and ethane as a function of composition, temperature, and molar density.
  • Numerical Recipes:This text book provides a version of Nelder and Mead's Downhill Simplex Method for multivariate, nonlinear optimization. See Chapter 10 Section 4 of Volume 1 (Fortran 77). Please note that this version provides a single-processor version of the code.


Contact:  Instructor
Link:  Back to my Research Homepage

Notice: All lecture notes and programming codes in this site are the property of David Keffer and are protected by copyright.
These notes may not be reproduced for public consumption without prior consent by David Keffer.
Currently, only registered students of ChE 415 at UTK have this consent.