Modeling of Gasification Processes. Final Report: Distinguished Scientist/Engineer Pgoram

摘要

The major effort was involved with modelling of single particle char burning and gasification, although some work on fluidized bed combustion was also carried out. It was discovered early that the treatment in the literature of single particles and how they burned or gasified was, in general, superficial and certainly not systematic, so that attempts to model complex reactor geometries are frustrated by not knowing how to handle the individual particles. The view of what mathematical modelling is supposed to accomplish is a strong function of the viewer. There are three main ideas about modelling. First, anc certainly the most common view, is that a valid model should aid in design. Second, models are devised to explain some previously observed or pathological behavior in a process: thermal run aways, difficult start-ups of plants, or completely different behavior than what had been anticipated from a less rational approach. These modelling efforts are always after the fact and are related to the third kind. This kind of modelling is related to learning models, i.e., those models from which one hopes to learn as much in a qualitative way about the process as possible. Theese models should predict the gross qualitative structure and when tuned with the right parameters be quantitatively correct. This is not the kind of modelling most engineers are attuned to but should be carried out early in the conceptual stages of a process since it should elucidate what the important parameters of a process might be depending upon the sophistication of the model builder. In this report we stress learning models mostly on single particle char burning and gasification. We consider models from the simple to the complex and try to compare the results from the spectrum of model assumptions. (ERA citation 07:045493)