P2-21: Optimization of waveguide coupler for coupled-cavity TWT using Artificial Neural Network

In this present work, optimization of a Ku-band waveguide coupler for coupled-cavity traveling-wave tube waveguide coupler has been carried out using an Artificial Neural Network (ANN). The ANN model takes the cavity physical dimensions as the input and the VSWR as the output. The training data for the network has been taken from the numerical simulation using the 3D electromagnetic simulation software MAFIA. The ANN uses a 3 layer feed-forward network consisting of 60 neurons for each layer. The ANN was trained so that the simulated error will be less than 10 %.