Browse Publications Technical Papers 2024-28-0007
2024-10-17

Predicting Full Vehicle Drag Coefficient using a Convoluted Neural Network Approach 2024-28-0007

Artificial Intelligence (AI) and Machine Learning (ML) technologies have emerged as transformative forces across various domains, revolutionizing industries, and reshaping societal paradigms. In this work, we show how a CNN model was used in the field of CFD to predict drag coefficient of a full vehicle profile. A brief description is also provided of the data set used for training and fitting the prediction model. This was done using the integrated AI/ ML technology in the DEP MeshWorks tool focusing on quick design iterations and results generation. The design advisor function within MeshWorks is an intelligent system that comprehends the real-time prediction of the responses such as model stiffness, frequency and others related to durability, NVH or CFD domain, on a model building phase without needing to run the post-processing for every design iteration. The uniqueness of this application is with the MeshWorks tool as it enables the user to efficiently engineer the data on a finite element level with the help of various advanced functions and the parameterization technology housed within the same system. The user can create the required parameters with less effort and the design changes are treated within the finite element-based model thereby, erasing the need for involving a design engineer at every stage of design modification. Additionally, this design advisor model can also be trained to derive accurate results for topologically different structures, wherein the test data need not be of a similar data type as the training dataset rather the input data can be of any geometrical information format, such as a CAD model or even a scanned model data. In conclusion, the application of a design advisor from MeshWorks will aid an analysis engineer in handling the required design iterations and the model updates with a significant reduction in process time as well as the process flow required in the conventional approach.

SAE MOBILUS

Subscribers can view annotate, and download all of SAE's content. Learn More »

Attention: This item is not yet published. Pre-Order to be notified, via email, when it becomes available.
Members save up to 16% off list price.
Login to see discount.
Special Offer: Download multiple Technical Papers each year? TechSelect is a cost-effective subscription option to select and download 12-100 full-text Technical Papers per year. Find more information here.
X