Artificial Neural Network Based Group Contribution Method for Estimating Cetane and Octane Numbers of Hydrocarbons and Oxygenated Organic Compounds

Kubic, William L.; Jenkins, Rhodri W.; Moore, Cameron M.; Semelsberger, TA; Sutton, Andrew D.

doi:10.1021/acs.iecr.7b02753

Cited by 55 publications

(53 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…As a database for our model, we take RON (MON) values for 335 (318) species from literature. 2,10,12,32,[60][61][62][63][64][65][66][67] For all reported data (RON, MON, DCN), we take average values whenever multiple values have been reported for a single species.…”

Section: Data Basis For Fuel Ignition Qualitymentioning

confidence: 99%

“…above 40) exhibit a short ignition delay which is required in compression-ignition (CI) engines. 4,8,9 Experimental RON, MON, and (D)CN values are available for a range of different fuel molecules, [10][11][12] however, for many interesting molecules such data is not readily available. For these molecules predictive models are required that enable rapid estimation of fuel ignition quality.…”

Section: Introductionmentioning

confidence: 99%

“…For these molecules predictive models are required that enable rapid estimation of fuel ignition quality. 4,13 In the past decades, several models have been developed to predict (D)CN [12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28] and RON/MON 12,16,[29][30][31][32] of (oxygenated) hydrocarbons by utilizing quantitative structureproperty relationship (QSPR) modeling. In QSPR, the modeling process can be broken down into two steps: First, QSPR models introduce molecular descriptors D = [d 1 , d 2 , ..., d n ] T that depend on the structure of a molecule m. Second, a regression model F (D) : D →p is fitted that predicts a propertyp as a function of D. 33 The regression model is either linear or nonlinear, depending on the QSPR.…”

Section: Introductionmentioning

confidence: 99%

“…QSPR models for DCN, RON, and MON differ in the way they encode the molecular structure. Various descriptors have been used including structural group counts (e.g., in [12][13][14]21,[29][30][31][32] ), the number of aromatic bonds (e.g., in 13,27 ), and topological indices, such as the Wiener Index 37 or branching indices (e.g., in 20,26,31 ). Previous models have also used a variety of techniques for the regression step, e.g., linear or nonlinear regression, 13,14,21,26,29,30 or artificial neural networks (ANNs).…”

Section: Introductionmentioning

confidence: 99%

“…Previous models have also used a variety of techniques for the regression step, e.g., linear or nonlinear regression, 13,14,21,26,29,30 or artificial neural networks (ANNs). 12,[17][18][19]22,24,27,[30][31][32] Development of QSPR models, however, highly depends on the choice of informative descriptors, a selection process that requires domain knowledge and intuition.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Graph Neural Networks for Prediction of Fuel Ignition Quality

et al. 2020

View full text Add to dashboard Cite

Prediction of combustion-related properties of (oxygenated) hydrocarbons is an important and challenging task for which quantitative structure-property relationship (QSPR) models are frequently employed. Recently, a machine learning method, graph neural networks (GNNs), has shown promising results for the prediction of structure-property relationships. GNNs utilize a graph representation of molecules, where atoms correspond to nodes and bonds to edges containing information about the molecular structure. More specifically, GNNs learn physico-chemical properties as a function of the molecular graph in a supervised learning setup using a backpropagation algorithm. This end-to-end learning approach eliminates the need for selection of molecular descriptors or structural groups, as it learns optimal fingerprints through graph convolutions and maps the fingerprints to the physico-chemical properties by deep learning. We develop GNN models for predicting three fuel ignition quality indicators, i.e., the derived cetane number (DCN), the research octane number (RON), and the motor octane number (MON), of oxygenated and non-oxygenated hydrocarbons. In light of limited experimental data in the order of hundreds, we propose a combination of multi-task learning, transfer learning, and ensemble learning. The results show competitive performance of the proposed GNN approach compared to state-of-the-art QSPR models making it a promising field for future research. The prediction tool is available via a web front-end at www.avt.rwth-aachen.de/gnn. File list (2) download file view on ChemRxiv Schweidtmann-et-al_Manuscript_ChemRxiv_20200511.... (627.65 KiB) download file view on ChemRxiv Schweidtmann-et-al_SI-Data-DCN-RON-MON_ChemR... (120.02 KiB)

show abstract

Section: Data Basis For Fuel Ignition Qualitymentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Graph Neural Networks for Prediction of Fuel Ignition Quality

et al. 2020

View full text Add to dashboard Cite

show abstract

Graph machine learning for design of high‐octane fuels

et al. 2022

View full text Add to dashboard Cite

Fuels with high-knock resistance enable modern spark-ignition engines to achieve high efficiency and thus low CO 2 emissions. Identification of molecules with desired autoignition properties indicated by a high research octane number and a high octane sensitivity is therefore of great practical relevance and can be supported by computer-aided molecular design (CAMD). Recent developments in the field of graph machine learning (graph-ML) provide novel, promising tools for CAMD. We propose a modular graph-ML CAMD framework that integrates generative graph-ML models with graph neural networks and optimization, enabling the design of molecules with desired ignition properties in a continuous molecular space. In particular, we explore the potential of Bayesian optimization and genetic algorithms in combination with generative graph-ML models. The graph-ML CAMD framework successfully identifies well-established high-octane components. It also suggests new candidates, one of which we experimentally investigate and use to illustrate the need for further autoignition training data.

show abstract

Unraveling the reaction route and kinetics of 3‐methyl‐3‐penten‐2‐one synthesis for synthetic ketone fragrances

Peng

Kiss

et al. 2020

J of Chemical Tech & Biotech

View full text Add to dashboard Cite

BACKGROUND: 3-Methyl-3-penten-2-one (3M3P) is an indispensable raw material used for the synthesis of OTNE (octahydrotetramethyl acetophenone), a synthetic ketone fragrance known by commercial trade names, such as Iso E Super®, which is one of the most important ingredients in perfumes. This work is the first to unravel the reaction route and kinetics for the synthesis of 3M3P via the cross-aldol condensation of acetaldehyde with butanone, catalyzed by the heterogeneous acidic ion exchange resin (NKC-9). RESULTS: An effective method was developed to evaluate this complex reaction system: the reaction routes were determined experimentally from the product structures detected by Gas Chromatography Mass Spectrometry (GC-MS), and the significance of side reactions were evaluated by the Yoneda group contribution method. The major side reactions are the self-condensation of acetaldehyde (significant even under optimized reaction conditions) and the cross-aldol condensation of acetaldehyde with 3M3P (which could be effectively suppressed at high butanone-acetaldehyde molar ratio). The reaction kinetics was investigated at temperatures of 60-90°C, catalyst mass fraction of 5 to 15%, and molar ratio of 5-20. At optimized reaction conditions, the yield of 3M3P was up to 90.85%, which is 2.4 times higher than that for the classic homogeneous catalysts (yield <38%). CONCLUSION: Acidic ion exchange resin NKC-9 is a high-selective green catalyst for the cross-aldol condensation of acetaldehyde with butanone. Integrating experimental determination and thermomechanical analysis is an effective method to unravel the main reaction routes for a complex system.

show abstract

Artificial Neural Network Based Group Contribution Method for Estimating Cetane and Octane Numbers of Hydrocarbons and Oxygenated Organic Compounds

Cited by 55 publications

References 31 publications

Graph Neural Networks for Prediction of Fuel Ignition Quality

Graph Neural Networks for Prediction of Fuel Ignition Quality

Graph machine learning for design of high‐octane fuels

Unraveling the reaction route and kinetics of 3‐methyl‐3‐penten‐2‐one synthesis for synthetic ketone fragrances

Contact Info

Product

Resources

About