Computational Modeling of Multicomponent Diffusion Using Fortran
Final rept. 13 Apr 2003-31 Dec 2005
CONNECTICUT UNIV STORRS
Pagination or Media Count:
The objective of this thesis was to develop a Fortran software package, using three modules, in order to extract diffusion data from concentration profiles and to predict future concentration profiles. The first module will be a finite difference code that uses the multicomponent form of Ficks First Law, and the time evolution of concentration will be calculated using the implicit Crank-Nicholson method. The second module will take into account movements of boundaries between regions in the interdiffusion zone, via equations that assume local equilibrium, and take into account that mass must be conserved. The first and second module will be used to predict how measured concentration profiles will change with time. The purpose of the third module will be to extract diffusivity data from measured concentration profiles. This module will use a matrix inversion method to calculate the diffusivities. In conclusion, the original objective of this project was not met to its full completion. Several factors contributed to this shortcoming, but the primary obstacle was the correlation between the software and the input data. While the software ran successfully with many different known solutions, it did not perform well using actual concentration profile data from the U.S. Army Research Laboratory. Most likely, this is due to the limited amount of species data, the accuracy of the data itself, and the spacing between each data point. All efforts should be taken to obtain more accurate, smoother input data which will allow the software to run with fewer obstacles and, in turn, produce cleaner output data. Once the input data is appropriate, the programmer should return to fine tune the individual software programs to allow them to work with the new data. This may require using a filtering subroutine in order to accept only worthy data from the input stream.
- Theoretical Mathematics
- Computer Programming and Software