Machine Learning to Predict Quasi TE₀₁₁ Mode Resonance of Dielectric Resonators in Free Space

Dielectric resonators, which are used to create standing waves in the microwave band, have been an important topic of research for decades due to their usefulness as antennae, filters, enhancing electron spin resonance, and a variety of other applications. For all these uses it is important to know the resonant frequency of a given resonator, but predicting this frequency is difficult, as only approximate formulas exist and performing precise numerical calculations to determine the frequency is computationally expensive. In previous work, we demonstrated that a neural network can be trained to accurately predict the quasi TE₀₁₁ mode frequency and field modes of a pair of dielectric resonators inside a cylindrical microwave category, with the training data created for a variety of different configurations using the finite element method. Here, we have extended this method to similarly determine the frequency and field modes for the quasi TE₀₁₁ mode for a single dielectric resonator in free space, by applying the appropriate modifications to our boundary conditions in our finite element modeling.