Electromagnetic and Thermomechanical Design Method of Graded Dielectric Streamlined Radome Accelerated by Artificial Intelligence

Streamlined radomes on hypersonic vehicles demand high electromagnetic (EM) performance and must possess robust thermomechanical performance to resist thermal loads. However, an efficient multidisciplinary design method considering these aspects is lacking. This communication proposes an EM and thermomechanical design method for streamlined radomes using a genetic algorithm (GA) accelerated by an artificial intelligence (AI) model. The method establishes thermomechanical performance as a constraint, with EM performance as the objective. The design cases of graded dielectric streamlined radomes with nine layers of silicon nitride ceramics at arbitrarily designated frequencies are used to demonstrate its effectiveness and efficiency. The results show that the designed radome has an average insertion loss less than 0.60 dB and a boresight error (BSE) within 3.00 mrad at any antenna scan angle for designated frequency bands. Meanwhile, the induced thermal stress remains within the safety threshold value. Notably, the utilization of AI accelerates the design process by approximately 90 times.