Abstract
Several compression-ignition and low-temperature combustion strategies require a fuel where the ignition-delay-time (IDT) is highly sensitive to the fuel-to-air equivalence ratio (ϕ). Quantitative prediction of ϕ-sensitivity (i.e., the change in IDT with respect to ϕ) would enable rapid screening of the numerous possible (bio)fuel candidates for this desired high ϕ-sensitivity characteristic. We propose a new ϕ-sensitivity metric (η), which is primarily a function of only temperature (T) and pressure (P). We assess the reliability of 0-D (perfectly homogeneous) simulation and state-of-the-art reaction mechanisms for two well-studied fuels, namely, iso-octane and a primary reference fuel (PRF80). 0-D simulation results are in good agreement with experimental constant volume IDT data at low- and intermediate-temperatures, while systematic deviations are observed at higher temperatures (where full 3-D computational fluid dynamics simulations are required for accurate prediction). We also perform a traditional single-parameter sensitivity analysis to determine the key reactions that affect ϕ-sensitivity. This is followed by a more rigorous Bayesian uncertainty quantification (UQ) analysis to elucidate the possible sources for the discrepancies at high T. Due to the computational cost of UQ, we train artificial neural networks to rapidly predict η for randomly perturbed sets of low- and intermediate-temperature reaction rate parameters. The primary implications of this study are that experimental ϕ-sensitivity data can be used to refine and validate proposed reaction mechanisms, while machine learning and uncertainty quantification of 0-D simulations are essential for quantitative prediction of ϕ-sensitivity in order to rapidly screen fuel candidates.
| Original language | English |
|---|---|
| Pages (from-to) | 103-115 |
| Number of pages | 13 |
| Journal | Combustion and Flame |
| Volume | 214 |
| DOIs | |
| State | Published - Apr 2020 |
| Externally published | Yes |
Funding
We would like to acknowledge Gina Fioroni and Earl Christensen of the National Renewable Energy Laboratory (NREL) for discussions regarding ϕ -sensitivity and Nicole Labbe of The University of Colorado Boulder for her invaluable insights into kinetic mechanisms. All of the 0-D and 3-D CFD simulations were run on Eagle, the National Renewable Energy Laboratory (NREL) high-performance computing system. This work was authored in part by Alliance for Sustainable Energy, LLC, the manager and operator of the National Renewable Energy Laboratory for the U.S. Department of Energy (DOE) under Contract no. DE-AC36-08GO28308. Funding provided by U.S. Department of Energy Office of Energy Efficiency and Renewable Energy Bioenergy Technologies Office. The views expressed in the article do not necessarily represent the views of the DOE or the U.S. Government. The U.S. Government retains and the publisher, by accepting the article for publication, acknowledges that the U.S. Government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this work, or allow others to do so, for U.S. Government purposes. This research was conducted as part of the Co-Optimization of Fuels and Engines (Co-Optima) project sponsored by the U.S. Department of Energy (DOE) Office of Energy Efficiency and Renewable Energy (EERE), Bioenergy Technologies and Vehicle Technologies Offices. We would like to acknowledge Gina Fioroni and Earl Christensen of the National Renewable Energy Laboratory (NREL) for discussions regarding ?-sensitivity and Nicole Labbe of The University of Colorado Boulder for her invaluable insights into kinetic mechanisms. All of the 0-D and 3-D CFD simulations were run on Eagle, the National Renewable Energy Laboratory (NREL) high-performance computing system. This work was authored in part by Alliance for Sustainable Energy, LLC, the manager and operator of the National Renewable Energy Laboratory for the U.S. Department of Energy (DOE) under Contract no. DE-AC36-08GO28308. Funding provided by U.S. Department of Energy Office of Energy Efficiency and Renewable Energy Bioenergy Technologies Office. The views expressed in the article do not necessarily represent the views of the DOE or the U.S. Government. The U.S. Government retains and the publisher, by accepting the article for publication, acknowledges that the U.S. Government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this work, or allow others to do so, for U.S. Government purposes. This research was conducted as part of the Co-Optimization of Fuels and Engines (Co-Optima) project sponsored by the U.S. Department of Energy (DOE) Office of Energy Efficiency and Renewable Energy (EERE), Bioenergy Technologies and Vehicle Technologies Offices.
| Funders | Funder number |
|---|---|
| Alliance for Sustainable Energy, LLC | |
| U.S. Department of Energy Office of Energy Efficiency and Renewable Energy BioEnergy Technologies Office | |
| U.S. Government | |
| Vehicle Technologies Offices | |
| U.S. Department of Energy | DE-AC36-08GO28308 |
| U.S. Department of Energy | |
| Office of Energy Efficiency and Renewable Energy | |
| National Renewable Energy Laboratory | |
| University of Colorado Boulder |
Keywords
- Kinetic mechanisms
- Machine learning
- Uncertainty quantification
- ϕ-sensitivity
