On the use of infrared spectroscopy and statistical learning with sparsity to characterize hydrocarbon fuels
Abstract/Contents
- Abstract
- Two strategies (named strategy 1 and 2) on using infrared absorption spectra and sparse statistical models to characterize hydrocarbon fuels are proposed and demonstrated in this dissertation. Specifically, linear models and generalized linear models with Lasso and grouped-Lasso regularization are utilized to select meaningful wavelengths from the mid-infrared spectra for estimating the average structure and various physical and combustion properties of hydrocarbon fuels. Three applications of the two strategies are demonstrated in this dissertation. Strategy 1 has two demonstration uses: estimation of derived cetane number (DCN) using a novel spectroscopic predictor, i.e. ratio of room-temperature absorbance of unreacted fuel vapor at 3.41 and 3.39 um, termed as the "Absorption Ratio", and estimation of 15 different physical and combustion properties such as molecular weight and lean blow-out using the full 3300-3550 nm spectra. Strategy 2 has one demonstration use: estimation of average fuel structure, represented by number of functional groups, using the full 3300-3550 nm spectra, and the further estimation of physical and combustion properties using linear and non-linear additive models. In the first application of strategy 1, i.e. estimation of DCN, the wide availability and applicability of the Absorption Ratio are demonstrated for a range of pure hydrocarbons, mixtures of pure hydrocarbons, and distillate and synthetic jet fuels. Quasi-linear calculation methods are provided for practical use. Spectroscopic and kinetic interpretations are provided based on the fraction of CH2 hydrogen relative to all carbon-bonded hydrogen. In addition, the correlations between the proposed predictor and ignition delay time and C2H4 yield are presented and discussed to exhibit the predictor's potential as a fuel screening tool. In the second application of strategy 1, i.e. estimation of physical and combustion properties, the concept of a compact and economical analyzer based on Fourier-transform infrared (FTIR) spectroscopy for estimating the properties of hydrocarbon fuels with small amounts of fuel sample is proposed. The high correlations between mid-infrared FTIR absorption spectra of fuel vapor in the range 3300 to 3550 nm and 15 physical and chemical properties, such as density, initial boiling point, surface tension, kinematic viscosity, number of carbon and hydrogen atoms per average molecule, and derived cetane number, for 64 hydrocarbon fuels are demonstrated. Linear models with Lasso regularization, based on linear combinations of absorption cross sections at selected wavelengths, are built for each of these physical and chemical properties, yielding accurate estimations. In the application of strategy 2, i.e. characterization of average fuel structure and estimation of properties, we use a generalized linear model with grouped-Lasso regularization to characterize the average fuel structure in terms of the fraction of each functional group type and provide a new strategy to approach the property estimation problem. The robustness of this structure characterization method against low spectral resolution and high multiplicative noise in FTIR spectra are empirically studied and presented. Two property estimation models, i.e. a linear and a nonlinear additive model, are presented as demonstrations of estimating properties from functional group numbers. The dissertation is concluded with a demonstration of the potential gain in estimation performance by utilizing the full 2-11 um (i.e. 2000-11000 nm) infrared spectra using molecular weight as an example, and a discussion of future research directions.
Description
| Type of resource | text |
|---|---|
| Form | electronic resource; remote; computer; online resource |
| Extent | 1 online resource. |
| Place | California |
| Place | [Stanford, California] |
| Publisher | [Stanford University] |
| Copyright date | 2020; ©2020 |
| Publication date | 2020; 2020 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Author | Wang, Yu |
|---|---|
| Degree supervisor | Hanson, Ronald |
| Thesis advisor | Hanson, Ronald |
| Thesis advisor | Bowman, Craig T. (Craig Thomas), 1939- |
| Thesis advisor | Wang, Hai, 1962- |
| Degree committee member | Bowman, Craig T. (Craig Thomas), 1939- |
| Degree committee member | Wang, Hai, 1962- |
| Associated with | Stanford University, Department of Mechanical Engineering |
Subjects
| Genre | Theses |
|---|---|
| Genre | Text |
Bibliographic information
| Statement of responsibility | Yu Wang. |
|---|---|
| Note | Submitted to the Department of Mechanical Engineering. |
| Thesis | Thesis Ph.D. Stanford University 2020. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2020 by Yu Wang
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...