This content is not included in your SAE MOBILUS subscription, or you are not logged in.
Machine Learning Approach to Predict Aerodynamic Performance of Underhood and Underbody Drag Enablers
ISSN: 0148-7191, e-ISSN: 2688-3627
Published April 14, 2020 by SAE International in United States
This content contains downloadable datasetsAnnotation ability available
Implementing stringent emission norms and fuel economy requirement in the coming decade will be very challenging to the whole automotive industry. Aerodynamic losses contribute up to 13% to 22 % of overall fuel economy and aerodynamicists will be challenged to have optimum content on the vehicle to reduce this loss. Improving Aerodynamic performance of ground vehicles has already reached its peak and the industry is moving towards active mechanisms to improve performance. Calibrating or simulating these active mechanisms in the wind tunnel or in Computational Fluid Dynamics (CFD) would be very challenging as the model complexity increases. Computationally expensive CFD models are required to predict the transient behaviors of model complexity. To balance these complexities and to reduce cost, the objective of this study is to explore the feasibility of statistical data analytics and machine learning methods and come up with predictive meta-models with the least amount of data, which can help to make quick or fast technical decisions.
Machine Learning (ML) is the scientific study of algorithms and statistical models that computer systems use to perform a specific task without using explicit instructions, relying on patterns and inferences instead. ML algorithms are used in a wide variety of applications, where it is difficult or infeasible to develop a conventional algorithm for effectively performing the task.
To build the ML models to predict the aerodynamics Coefficient of Drag (Cd), the training data is generated through space filling Design of Experiments (DOE) using CFD simulations. In this paper, various ML algorithms are used to generate models for predicting the aerodynamic drag. These models consider three enablers as continuous variables (Airdam, TireDam and Belly-pan) and remaining three enablers as three discrete variables (Duckbill, Lower Bumper Stiffener and Engine Panel). Accuracy of various machine learning algorithms like Kriging, decision trees, linear regression, random forest, neural network etc. are discussed in this paper. The ML models can be used to predict a drag for any point in the experimental space. Effect of all the enablers and interactions between different aero enablers are also discussed. Components of Analysis of Variance (ANOVA) such as main effect plots and interaction effect plots are discussed in detail.
CitationDube, P. and Hiravennavar, S., "Machine Learning Approach to Predict Aerodynamic Performance of Underhood and Underbody Drag Enablers," SAE Technical Paper 2020-01-0684, 2020, https://doi.org/10.4271/2020-01-0684.
Data Sets - Support Documents
|[Unnamed Dataset 1]|
- Mehta, U.B. and Kutler, P. , “Computational Aerodynamics and Artificial Intelligence,” NASA-TM-85994, A-9807,NAS 1.15: 85994, doi: June 1, 1984.
- Bradley, L.W. and Stephen, C.-Y. , “Machine Learning in Engineering Automation - The Present and the Future,” IMS '91--Learning in IMS, Computers in Industry 17:91-100, 1991.
- Bradley, L.W. and Stephen, C.-Y. , “Lu  in their Artificial Intelligence in Engineering,” 5(4):189-198, October 1990.
- Dube, P., Hiravennavar, S., and Ali, Z.A. , “Statistical Model to Predict Air Side Pressure Drop for Heat Exchangers,” SAE Technical Paper 2018-01-0081, 2018, https://doi.org/10.4271/2018-01-0081.
- Pedregosa et al., “Scikit-Learn: Machine Learning in Python, JMLR 12, 2825-2830, 2011. Source code, binaries, and documentation can be downloaded from http://scikit-learn.sourceforge.net.
- Mehrotra, K., Mohan, C.K., and Ranka, S. ,” “Elements of Artificial Neural Networks,” 0-262-13328-8.