Efficient Data-Driven Simulation of Microwave Interaction With Complex Plasma Profiles

Microwave interactions reveal electromagnetic properties of complex materials like plasma that aid in under-standing plasma’s wide range of applications in manufacturing, aerospace, and more. The interactions generally comprise transmission, absorption, and scattering of microwaves with the dense plasma profile. Traditional computational methods face a challenge in simulating the complex interaction between the microwave and asymmetric plasma profiles due to the trade-off between the number of computations and the accuracy of the result, which encourages exploring other techniques. Data-driven deep learning techniques due to their ability to decipher hidden pattern from input data finds application in solving various advanced scientific/engineering problems. The technique has been applied to investigate complex microwave plasma interactions. The proposed deep learning model, trained on different asymmetric plasma profiles and corresponding scattered microwave E-field patterns, achieved 300-500 times faster and accurate predictions that are within acceptable limits. Thus, affirms the model’s applicability in accelerating future studies on plasma dynamics.