Abstract
Deep learning approaches were applied to predict soot emissions under transient conditions in a diesel engine using the worldwide harmonized light vehicles test procedure (WLTP) cycles. The accuracies of deep neural networks (DNN) and long short-term memory (LSTM) models were compared to predict emissions. The accuracy of the LSTM model had an R2 value of 0.9761, which was higher than that of the DNN model, with an R2 value of 0.9215. The mean absolute errors (MAEs) of the WLTP cycles predicted by the LSTM model were between 0.30 %–1.47 % compared with the maximum measured values in WLTP cycles. For local prediction, the LSTM model followed the fluctuations in the data and local peak values well. However, the calculation time of the LSTM model is longer than that of the DNN model. Researchers should consider the purpose of the model based on the accuracy-time trade-off relationship between DNN and LSTM models.
Similar content being viewed by others
Abbreviations
- ANN :
-
Artificial neural network
- BSFC :
-
Brake-specific fuel consumption
- CFD :
-
Computational fluid dynamics
- DNN :
-
Deep neural networks
- DPF :
-
Diesel particulate filter
- ECU :
-
Engine control unit
- ELU :
-
Exponential linear unit
- LSTM :
-
Long short-term memory
- MAE :
-
Mean absolute error
- NO x :
-
Nitrogen oxides
- RDE :
-
Real driving emissions
- RMSE :
-
Root mean squared error
- RNN :
-
Recurrent neural network
- SOF :
-
Soluble organic fraction
- STD :
-
Standard deviation
- WLTP :
-
Worldwide harmonized light vehicle test procedure
References
Continental-Automotive, Worldwide Emission Standards and Related Regulations–Passenger Cars/Light and Medium Duty Vehicles, Continental, Hanover, Germany (2019).
Z. Meng et al., Particle emission characteristics of DPF regeneration from DPF regeneration bench and diesel engine bench measurements, Fuel, 262 (2020) 116589.
C. Barro, P. Obrecht and K. Boulouchos, Development and validation of a virtual soot sensor: part 1: steady-state engine operation, Int. J. of Engine Research, 15(6) (2014) 719–730.
C. Barro, P. Obrecht and K. Boulouchos, Development and validation of a virtual soot sensor: part 2: transient engine operation, Int. J. of Engine Research, 16(2) (2015) 127–136.
X. Tauzia, H. Karaky and A. Maiboom, Evaluation of a semi-physical model to predict NOx and soot emissions of a CI automotive engine under warm-up like conditions, Applied Thermal Engineering, 137 (2018) 521–531.
T. Q. Pan et al., Particulate matter emission modelling based on soot and SOF from direct injection diesel engines, Energy Conversion and Management, 48(2) (2007) 510–518.
H. Karaky et al., Semi-empirical 0D modeling for engine-out soot emission prediction in DI diesel engines, SAE Technical Paper (2016) 2016-01-0562.
G. Vishwanathan and R. D. Reitz, Application of a semi-detailed soot modeling approach for conventional and low temperature diesel combustion-part I: model performance, Fuel, 139 (2015) 757–770.
P. P. Duvvuri et al., Modeling the effect of parametric variations on soot particle size distribution in a diesel engine, J. of Energy Resources Technology, 142(3) (2020) 032201.
Y. Lecun, Y. Bengio and G. Hinton, Deep learning, Nature, 521(7553) (2015) 436.
R. K. Mehra et al., Application of artificial neural network (ANN) to predict the performance and emission of hydrogen enriched compressed natural gas (HCNG) engines, CICTP 2018: Intelligence, Connectivity, and Mobility, American Society of Civil Engineers, Reston, VA, USA (2018) 2330–2343.
S. H. Hosseini et al., Artificial neural network modeling of performance, emission, and vibration of a CI engine using alumina nano-catalyst added to diesel-biodiesel blends, Renewable Energy, 149 (2020) 951–961.
T. Van Hung et al., Prediction of emission characteristics of a diesel engine using experimental and artificial neural networks, Applied Nanoscience, 13 (2023) 433–442.
S. Shin et al., Deep learning procedure for knock, performance and emission prediction at steady-state condition of a gasoline engine, Proc. of the Institution of Mechanical Engineers, Part D: J. of Automobile Engineering, 234(14) (2020) 3347–3361.
S. Lee et al., Proposal of a methodology for designing engine operating variables using predicted NOx emissions based on deep neural networks, Journal of Mechanical Science and Technology, 35(4) (2021) 1747–1756.
H. Zheng et al., Modeling and prediction for diesel performance based on deep neural network combined with virtual sample, Scientific Reports, 11(1) (2021) 1–17.
A. Coraddu et al., Marine dual fuel engines monitoring in the wild through weakly supervised data analytics, Engineering Applications of Artificial Intelligence, 100 (2021) 104179.
R. Wang, H. Chen and C. Guan, A Bayesian inference-based approach for performance prognostics towards uncertainty quantification and its applications on the marine diesel engine, ISA Transactions, 118 (2021) 159–173.
D. Giordano et al., Dissecting a data-driven prognostic pipeline: A powertrain use case, Expert Systems with Applications, 180 (2021) 115109.
C. M. Le Cornec et al., Modelling of instantaneous emissions from diesel vehicles with a special focus on NOx: insights from machine learning techniques, Science of The Total Environment, 737 (2020) 139625.
R. Pillai et al., Modeling and predicting heavy-duty vehicle engine-out and tailpipe nitrogen oxide (NOx) emissions using deep learning, Frontiers in Mechanical Engineering, 8 (2022).
N. Planakis et al., Recurrent and time-delay neural networks as virtual sensors for NOx emissions in marine diesel power-trains, SAE Technical Paper (2021) 2021-01-5042.
Y. Wang, Y. Yu and J. Li, Predicting the transient NOx emissions of the diesel vehicle based on LSTM neural networks, 2020 IEEE Conference on Telecommunications, Optics and Computer Science (TOCS), IEEE (2020).
X. Miao et al., Prediction and control of knock at high load boundary for HCCI engine based on neural network, Fuel, 333 (2023) 126421.
S. Shin et al., Predicting transient diesel engine NOx emissions using time-series data preprocessing with deep-learning models, Proc. of the Institution of Mechanical Engineers, Part D: J. of Automobile Engineering, 235(12) (2021) 3170–3184.
S. Hochreiter et al., Gradient Flow in Recurrent Nets: The Difficulty of Learning Long-Term Dependencies, IEEE Press (2001).
S. Shin et al., Deep neural network model with Bayesian hyperparameter optimization for prediction of NOx at transient conditions in a diesel engine, Engineering Applications of Artificial Intelligence, 94 (2020) 103761.
Keras, EarlyStopping, Keras Documentation, https://keras.io/api/callbacks/early_stopping/ (2019).
S. Ioffe and C. Szegedy, Batch normalization: accelerating deep neural network training by reducing internal covariate shift, arXiv:1502.03167 (2015).
D. A. Clevert, T. Unterthiner and S. Hochereiter, Fast and accurate deep network learning by exponential linear units (elus), arXiv:1511.07289 (2015).
X. Glorot, A. Bordes and Y. Bengio, Deep sparse rectifier neural networks, Proc. of the Fourteenth International Conference on Artificial Intelligence and Statistics, USA (2011) 315–323.
D. P. Kingma and J. Ba, Adam: A method for stochastic optimization, arXiv:1412.6980 (2014).
S. Shin et al., Designing a steady-state experimental dataset for predicting transient NOx emissions of diesel engine via deep learning, Expert Systems with Applications, 198 (2022) 116919.
Acknowledgments
This study was supported by Sejong University and Korea Railway Research Institute (PK2201C1). The results of this study were also supported by “HPC Support” Project, organized by ‘Ministry of Science and ICT’ and NIPA.
Author information
Authors and Affiliations
Corresponding author
Additional information
Seunghyup Shin obtained his B.S. (2009), M.S. (2011) and Ph.D. (2021) in Department of Mechanical and Aerospace Engineering from Seoul National University. He is currently an Assistant Professor in Department of Artificial Intelligence at Sejong University, Seoul, Korea.
Minjeong Kim obtained her B.S. (2010), M.S. (2012) and Ph.D. (2016) in Department of Environmental Engineering from KyungHee University, Korea. She is currently a Senior Researcher in Artificial Intelligence Railroad Research Department at Korea Railroad Research Institute (KRRI).
Rights and permissions
About this article
Cite this article
Shin, S., Won, JU. & Kim, M. Comparative research on DNN and LSTM algorithms for soot emission prediction under transient conditions in a diesel engine. J Mech Sci Technol 37, 3141–3150 (2023). https://doi.org/10.1007/s12206-023-0538-y
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s12206-023-0538-y