One-Dimensional Beam-Wave Interaction Prediction Regression Model for Helix TWT

This paper proposes a one-dimensional beam-wave interaction regression prediction model for helix traveling wave tube (TWT), which is based on neural network and successfully applied to the design and optimization of broadband helix TWT. In this paper, the operating frequency and input power are chosen as key input variables and the forward mapping relationship between them and TWT output power is established by training the convolutional neural network (CNN) model and multilayer perceptron (MLP) model. Based on these prediction models, the TWT output power under different operating frequency and input power can be quickly predicted, providing effective support for beam-wave interaction design and performance optimization.