In this dissertation, different space marching implementations of the Mollification method are introduced to numerically recover the temperature and heat flux histories on a bounded two-dimensional rectangular body when the initial sample data are collected on one side of the body. We combined the mollification method with a singular perturbation scheme to obtain a stable algorithm. A reliable set of parameter values is experimentally determined by numerical tests to guarantee the accuracy and stability of the algorithm.
Theoretical and experimental contributions to the study of the behavior of beds of granular desiccants in removing moisture from a flowing gas stream are presented. A differential equation for the performance of such a bed is developed under various limiting assumptions regarding the mechanism of the process. Rigorous and approximate solutions to the equation, together with a method of testing experimental data for conformity to it, are developed. Critical analysis of previous contributions to this theory is undertaken.