4.7 Article

Predicting cetane numbers of hydrocarbons and oxygenates from highly accessible descriptors by using artificial neural networks

Journal

FUEL
Volume 207, Issue -, Pages 344-351

Publisher

ELSEVIER SCI LTD
DOI: 10.1016/j.fuel.2017.06.104

Keywords

Cetane number; Renewable fuels; Artificial neural network; Quantitative structure property relationship

Funding

  1. Innovation Grant from Singapore Institute of Technology [P09000-1/R-MNR-E103-A005]
  2. Billion Miles Pte. Ltd.

Ask authors/readers for more resources

Artificial neural network (ANN) was applied to the prediction of cetane numbers (CN) of hydrocarbons and oxygenates. We built an ANN model based on the quantitative structure property relationship strategy. Descriptors of pure compounds, that can be easily obtained, were selected based on the knowledge of evaporation and combustion process of fuels inside compression-ignition engines. The homogeneity of these descriptors were also considered, in order to develop a general model. Our optimized ANN models, obtained by screening all possible configurations, showed higher accuracy than that using multiple linear regression. After performing ANN regressions on two subgroups of the dataset, average absolute errors of 6.5 and 4.4 were achieved for compounds with and without rings, respectively. Ranking the impact of descriptors on CN values hinted that evaporation relevant features and bond energy could be more important than molecular structure related features. (C) 2017 Elsevier Ltd. All rights reserved.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.7
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available