Predictive Capability Assessment of Probabilistic Machine Learning Models for Density Prediction of Conventional and Synthetic Jet Fuels
Abstract
Machine Learning (ML) models are increasingly applied in the field of jet fuel property predictions due to their ability of modeling a high number of complex composition–property relationships directly on measurement data. Their applicability is still limited as for safety relevant use cases like synthetic fuel approval or aircraft design consequences of prediction errors might be too severe to be acceptable. For Machine Learning algorithms, the predictive capability strongly depends on the data utilized for the training of the models. Predictions for fuels that differ from the training data might have uncertainties that need to be systematically considered. We present an approach of utilizing the probabilistic ML algorithm Gaussian Process Regression to model jet fuel properties and estimate the uncertainty that results from limited training data. We apply this approach for the example of density over a range of −40 to 140 °C. To assess the influence of synthetic fuels on the predictive capability, two models are studied, one trained exclusively on conventional fuels data and the other one trained on the same conventional fuels and additional synthetic fuels. To quantify the predictive capability of the models, we introduce three metrics that measure the accuracy and precision of the prediction as well as the validity and reliability of the estimated prediction interval. Results show that prediction intervals can correctly be estimated by both models and a valid estimation of the predictive capability is possible. Furthermore, the addition of synthetic fuels data drastically improves the accuracy, reduces the uncertainty, and is necessary to achieve adequate predictions for the considered hold-out fuels.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The DLR model is a Gaussian Progress Regression algorithm that is trained on conventional and synthetic fuels.
“The DLR model [30,31] is a Gaussian Progress Regression algorithm that is trained on conventional and synthetic fuels.”
From Sustainable aviation fuel prescreening tools and procedures (2021) · cites this paper as [31]
Cites
Gaussian Processes in Machine Learning
Rasmussen, C. E. (2004) · Lecture notes in computer science · doi:10.1007/978-3-540-28650-9_4 · Cited by 5,734 · see citing works
Verification and Validation in Scientific Computing
Oberkampf, W. L., & Roy, C. J. (2010) · doi:10.1017/cbo9780511760396 · Cited by 1,556 · see citing works
KATRITZKY, A. R., Kuanar, M., Slavov, S. H., Hall, C. D., Karelson, M. M., Kahn, I., & Dobchev, D. A. (2010) · Chemical Reviews · doi:10.1021/cr900238d · Cited by 576 · see citing works
Self-citationNieto‐Draghi, C., Fayet, G., Creton, B., et al. (2015) · Chemical Reviews · doi:10.1021/acs.chemrev.5b00215 · Cited by 161 · see citing works
Prediction of Density and Viscosity of Biofuel Compounds Using Machine Learning Methods
Saldana, D. A., Starck, L., Mougin, P., Rousseau, B., Ferrando, N., & Creton, B. (2012) · Energy & Fuels · doi:10.1021/ef3001339 · Cited by 129 · see citing works
Yang, H. S., Ring, Z., Briker, Y., et al. (2002) · Fuel · doi:10.1016/s0016-2361(01)00121-1 · Cited by 105 · see citing works
Shi, X., Li, H., Song, Z., Zhang, X., & Liu, G. (2017) · Fuel · doi:10.1016/j.fuel.2017.03.073 · Cited by 85 · see citing works
A review of aviation turbine fuel chemical composition-property relations
Vozka, P., & Kilaz, G. (2020) · Fuel · doi:10.1016/j.fuel.2020.117391 · Cited by 80 · see citing works
Jet fuel density via GC × GC-FID
Vozka, P., Modereger, B. A., Park, A. C., Zhang, W. T. J., Trice, R. W., Kenttämaa, H. I., & Kılaz, G. (2018) · Fuel · doi:10.1016/j.fuel.2018.08.110 · Cited by 76 · see citing works
Calculation of jet and diesel fuel properties using carbon-13 NMR spectroscopy
Cookson, D., & Smith, B. E. (1990) · Energy & Fuels · doi:10.1021/ef00020a004 · Cited by 71 · see citing works
Liu, G., Wang, L., Qu, H., Shen, H., Zhang, X., Zhang, S., & Mi, Z. (2007) · Fuel · doi:10.1016/j.fuel.2007.02.023 · Cited by 63 · see citing works
Investigation of the chemical basis of kerosene (jet fuel) specification properties
Cookson, D. J., Lloyd, C. P., & Smith, B. E. (1987) · Energy & Fuels · doi:10.1021/ef00005a011 · Cited by 62 · see citing works
Pang, J., Liu, D., Yu, P., et al. (2018) · Sensors · doi:10.3390/s18040967 · Cited by 54 · see citing works
Evaluation of Uncertainty Quantification in Deep Learning
Ståhl, N., Falkman, G., Karlsson, A., & Mathiason, G. (2020) · Communications in computer and information science · doi:10.1007/978-3-030-50146-4_41 · Cited by 43 · see citing works
Rocha, W. F. D. C., & Sheen, D. A. (2019) · Fuel · doi:10.1016/j.fuel.2018.12.126 · Cited by 29 · see citing works
Kramer, K., Morris, R. A., Rose‐Pehrsson, S. L., Cramer, J. A., & Johnson, K. (2007) · Energy & Fuels · doi:10.1021/ef700403s · Cited by 25 · see citing works
Cramer, J. A., Hammond, M. H., Myers, K. M., Loegel, T. N., & Morris, R. A. (2014) · Energy & Fuels · doi:10.1021/ef4021872 · Cited by 22 · see citing works
Practice for Evaluation of New Aviation Turbine Fuels and Fuel Additives
D02 Committee (2019) · doi:10.1520/d4054-19 · Cited by 8 · see citing works
Specification for Aviation Turbine Fuel Containing Synthesized Hydrocarbons
D02 Committee (2018) · doi:10.1520/d7566-18 · Cited by 5 · see citing works
Flagged first, then by citations
Cited by
Sorted by citations
Sustainable aviation fuel prescreening tools and procedures
Heyne, J., Rauch, B., Le Clercq, P., & Colket, M. (2021) · Fuel · doi:10.1016/j.fuel.2020.120004 · Cited by 170 · see citing works
Self-citation
- Authors
- Hall, C., Rauch, B., Bauder, U., Le Clercq, P., & Aigner, M.
- Affiliations
- Deutsches Zentrum für Luft- und Raumfahrt e. V. (DLR) (DE)
- DLR, German Aerospace Center, Institute of Combustion Technology, MAT, 70569 Stuttgart, Germany
- Corresponding author
- Hall, Clemens
- Venue
- Energy & Fuels
- Publisher
- American Chemical Society (ACS)
- Published
- 2021-01-21
- Volume
- 35
- Issue
- 3
- Pages
- 2520-2530
- ISSN
- 0887-0624 (ISSN-L)
- 1520-5029
- OpenAlex
- W3122603573
- Cited by
- 31
- References
- 19
- Citations by year
- 2026: 2 · 2025: 3 · 2024: 6 · 2023: 9 · 2022: 7 · 2021: 3 · 2020: 1
- FWCI
- 1.90
- Type
- journal article
- Language
- en
- Open access
- closed
- Licence
- Locations
- Energy & Fuels (publishedVersion)
- elib (German Aerospace Center) (acceptedVersion)
- Funders
- Horizon 2020 Framework Programme (723525)
- Topics
- Physical Sciences › Materials Science › Materials Chemistry › Machine Learning in Materials Science
- Physical Sciences › Engineering › Computational Mechanics › Heat transfer and supercritical fluids
- Physical Sciences › Computer Science › Artificial Intelligence › Gaussian Processes and Bayesian Inference
- Keywords
- synthetic aviation fuel, Gaussian process regression, probabilistic machine learning, forecast uncertainty, prediction intervals, out-of-distribution prediction, limited training data, fuel property prediction
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
