|
" Structure of complex mixtures through characterization, reaction, and modeling "
D. M. Trauth
M. T. Klein
| Document Type
|
:
|
Latin Dissertation
|
| Language of Document
|
:
|
English
|
| Record Number
|
:
|
1113464
|
| Doc. No
|
:
|
TLpq304068206
|
| Main Entry
|
:
|
D. M. Trauth
|
|
|
:
|
M. T. Klein
|
| Title & Author
|
:
|
Structure of complex mixtures through characterization, reaction, and modeling\ D. M. TrauthM. T. Klein
|
| College
|
:
|
University of Delaware
|
| Date
|
:
|
1993
|
| student score
|
:
|
1993
|
| Degree
|
:
|
Ph.D.
|
| Page No
|
:
|
260
|
| Abstract
|
:
|
Most fossil fuels, including coal, petroleum, tar sand, oil shale, and bitumen, along with many industrial process streams may be considered complex mixtures. This complexity is a direct result of the enormous number of different molecules which comprise the mixture. However, it is this diversity of molecules which often renders complex mixtures difficult to characterize. Additionally, difficulties arise during analytical characterization since many complex mixtures, such as coal, asphaltene, and petroleum resid, are either nonvolatile or insoluble. Thus complex mixtures are seldomly thoroughly understood at the molecular level. The subject of this investigation, petroleum resid, was selected due to its industrial relevance. Over the past twenty years the average resid content of petroleum has increased significantly. Since most refineries were built before this trend was established, they were not designed to cope with the unexpected higher quantities of resid. Thus refinery capacity is often limited by the ability to convert resid. The goal of this thesis is the elucidation of resid structure at the molecular level. A detailed, precise understanding of resid structure coupled with model compound kinetics may lead to fundamental insights which will increase resid conversion efficiencies. This thesis is organized along three main thrusts. First, a series of thermal pyrolysis experiments on four resids and their isolated asphaltenes were performed to determine feedstock differences and reaction environment effects. Based on these results it was hypothesized that the reactivity differences between feedstocks could be attributed to structural differences. Second, a molecular level, statistical-based resid construction algorithm was developed. Resid molecules are stochastically constructed by sampling structural attributes. Each attribute is represented using a statistical distribution. Thirdly, optimization of the resid construction algorithm resulted in a unique set of structural attribute parameters. This general approach was utilized to determine the optimal structures of four industrial resids: Hondo, Maya, Arabian Light, and Arabian Heavy. A sensitivity analysis indicated that 10 molecules was the optimal simulation sample size providing a balance between reproducibility and computational demand. This analysis also showed that the structural attributes were dependent. Thus a global optimization technique was necessary. The optimization routine was found to yield optimal parameter results with a small enough variance to uniquely define each of the four resids. Optimal structural attribute distributions fit all analytical data well. The manner in which the structural attributes are distributed helps to explain the often conflicting interpretations of resid structure previously reported.
|
| Subject
|
:
|
Applied sciences
|
|
|
:
|
Chemical engineering
|
|
|
:
|
petroleum
|
|
|
:
|
resid
|
| |