Application of Machine Learning technique to incorporate manufacturing and Testing variation for Robust BIW design for Crash performance

In vehicle development CAE plays crucial role in arriving at optimum structural design to meet various vehicle performance targets in different domain such as Crash, NVH, Durability etc. Accurate CAE methodology can aid in reducing the number of physical tests & reducing overall vehicle development time. However, there are instances where there are gaps observed between test results and CAE predictions. These gaps get amplified in crash simulations as the event is highly dynamic and non-linear behavior simulation is always challenging. In order to enhance CAE methodology, it was decided to incorporate the effect of manufacturing and testing variations in crash CAE simulations. Manufacturing process accounts for variations due to inherent variation in material properties, spot weld nugget diameter, manufacturing processes such as stamping etc. whereas Physical Testing houses variation in barrier positions, test speed etc. within specified tolerance defined by regulatory bodies. These variations affect structural performance and negating these issues in early design phase will help to arrive at robust structural design.