Bayesian uncertainty quantification of the ignition behavior of a HEFA-SPK surrogate
Abstract
Developing ad-hoc sustainable aviation fuel (SAF) surrogates to foster the understanding of composition effects on combustor operability figures of merit (FoM) is of crucial importance. Nevertheless, most proposed SAF surrogate mixtures in the literature exhibit a trivial chemical composition, resulting from the restricted palette of candidate compounds and the exclusive focus on gas-phase combustion properties. In this contribution, we develop a four-component surrogate mixture for the POSF-6152 HEFA-SPK fuel using Bayesian inference with the BayeSAF framework, targeting nine physicochemical properties and encompassing 321 candidate compounds, while accounting for experimental uncertainty. Both the maximum-a-posteriori (MAP) surrogate and the pushforward distributions of key properties satisfactorily replicate the reference data. Relying on the MAMOX++ software for automatic kinetic mechanism generation, we carry out a forward uncertainty quantification (UQ) analysis on the ignition delay time (τign) predictions for stoichiometric, pre-vaporized HEFA-SPK/air mixtures at 6 atm from 1050 to 1250 K. While the pushforward distribution of τign agrees closely with shock-tube measurements, we observe an average variability of 10%. Lastly, we extend the UQ analysis to preferential vaporization effects on HEFA-SPK ignition. Specifically, we investigate the DCN compositional dependence across distillation cuts, with preferential vaporization effects particularly notable near the initial boiling point. Correspondingly, the confidence interval for the ignition delay of the 1% volume distillation cut is larger than the variability observed under pre-vaporized conditions or along the MAP distillation curve. These results underscore the suitability of the BayeSAF-based surrogate formulation strategy for robustly predicting SAF ignition behavior, including preferential vaporization.
Related
- Online Bayesian Model Averaging with Joint Uncertainty Quantification for Models and Regression Coefficients in Binary Regression
- Enhanced Uncertainty Quantification Variational Autoencoders for the Solution of Bayesian Inverse Problems
- Uncertainty Quantification of State Variables Trajectories in the Context of Inverse Problems: An Approach from Bayesian Inference and FDA
- Quantum Bayesian Time Series Forecasting with Uncertainty Quantification
- Bayesian Uncertainty Quantification for fMRI Functional Connectivity via Simulation-Based Inference
BibTeX
@article{liberatori2026bayesian,
title = {{Bayesian uncertainty quantification of the ignition behavior of a HEFA-SPK surrogate}},
author = {Jacopo Liberatori and Salvatore Iavarone and Ronan Vicquelin},
journal = {Proceedings of the Combustion Institute},
year = {2026},
month = sep,
doi = {10.1016/j.proci.2026.106140},
url = {https://hal.science/hal-05765174v1},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Bayesian uncertainty quantification of the ignition behavior of a HEFA-SPK surrogate
AU - Jacopo Liberatori
AU - Salvatore Iavarone
AU - Ronan Vicquelin
PY - 2026
DA - 2026/09/25
JO - Proceedings of the Combustion Institute
DO - 10.1016/j.proci.2026.106140
UR - https://hal.science/hal-05765174v1
AB - Developing ad-hoc sustainable aviation fuel (SAF) surrogates to foster the understanding of composition effects on combustor operability figures of merit (FoM) is of crucial importance. Nevertheless, most proposed SAF surrogate mixtures in the literature exhibit a trivial chemical composition, resulting from the restricted palette of candidate compounds and the exclusive focus on gas-phase combustion properties. In this contribution, we develop a four-component surrogate mixture for the POSF-6152 HEFA-SPK fuel using Bayesian inference with the BayeSAF framework, targeting nine physicochemical properties and encompassing 321 candidate compounds, while accounting for experimental uncertainty. Both the maximum-a-posteriori (MAP) surrogate and the pushforward distributions of key properties satisfactorily replicate the reference data. Relying on the MAMOX++ software for automatic kinetic mechanism generation, we carry out a forward uncertainty quantification (UQ) analysis on the ignition delay time (τign) predictions for stoichiometric, pre-vaporized HEFA-SPK/air mixtures at 6 atm from 1050 to 1250 K. While the pushforward distribution of τign agrees closely with shock-tube measurements, we observe an average variability of 10%. Lastly, we extend the UQ analysis to preferential vaporization effects on HEFA-SPK ignition. Specifically, we investigate the DCN compositional dependence across distillation cuts, with preferential vaporization effects particularly notable near the initial boiling point. Correspondingly, the confidence interval for the ignition delay of the 1% volume distillation cut is larger than the variability observed under pre-vaporized conditions or along the MAP distillation curve. These results underscore the suitability of the BayeSAF-based surrogate formulation strategy for robustly predicting SAF ignition behavior, including preferential vaporization.
SN - 1540-7489
SN - 1540-7489
ER -