GenAI based e-motor design G. Cossio1, JP. Berro Ramirez1, SO. Esparza1, 1: Altair Engineering France; 28 Rue de Châteaux de rentiers, 75013, Paris Abstract: This paper presents the first investigation expensive. In recent studies, generative AI has been...
More
GenAI based e-motor design G. Cossio1, JP. Berro Ramirez1, SO. Esparza1, 1: Altair Engineering France; 28 Rue de Châteaux de rentiers, 75013, Paris Abstract: This paper presents the first investigation expensive. In recent studies, generative AI has been into leveraging generative artificial intelligence for shown to work well in engineering design tasks [1, 2, electrical motor design. We develop a conditioned 10, 13, 15], often defined as the creation of diffusion model trained on diverse electric motor geometries respecting several constraints and designs to generate novel configurations that optimizing certain criteria. Evaluating these criteria is maximize torque output. Custom datasets created usually computationally expensive, which when through parametric modelling are specifically paired with the iterative nature of many optimization structured to enhance training efficiency. algorithms results in enormous computational costs. Comparative analysis between unconditioned and In a
Less
