AI Reduced Order Model of Induction Machine

Induction motors commonly used in electric drive, need precise efficiency calculations across various operating conditions from the early design stages. Electromagnetic finite-element analysis (FEA) offers high accuracy but is computationally demanding. Simpler models, like the T-equivalent circuit, are well-documented but challenging to determine due to dependencies on frequency, voltage, and slip. While reduced order models based on finite elements have been proposed, they often lack accuracy across different conditions. AI can be a good alternative but so far is primarily used for control and monitoring. This article presents a hybrid approach that combines FEA data with Reduced order Modeling leveraging AI, achieving a good compromise between accuracy and time computation.