Machine learning is relatively new in the CAE and testing world despite being used in a wide range of industries and applications.
Traditional physics-based FEA (Finite Element Analysis) is an essential aspect of product development in the transportation industry.
Advancements have been made in the last few decades, improving FEA-based product design development processes. Improvements in CPU technologies have reduced simulation time from a week to a day, and currently only a few hours. Iterative studies such as if-then, reliability, sensitivity, and optimization are still time-consuming. With increased focus on digital twin and industry 4.0, there is also a perception that FEA normally comes late in the product development cycle. It’s more advantageous to use machine learning before design maturation to asses performance (without running any simulations).
This is where machine learning and artificial intelligence can help. They provide an intelligent way to integrate design, CAE, test data, and historical knowledge into a model. The model then serves as a basis for what’s called “predictive engineering analytics.” Machine learning can benefit manufacturing as well. It can identify key parameters early in the design cycle, guiding product development or use image processing to identify quality issues during manufacturing.
Soon initial designs will be produced by machine learning using previous results and new packaging requirements. The design will include predictive CAE results, all without performing any simulation.