BP Neural Network Method on the Frequency Stabilization System of the Portable Electron Linear Accelerator for Nondestructive Inspection

As the power source of portable electron linear accelerator for nondestructive inspection, the magnetron with the operating frequency at 9 730 MHz has the main disadvantage of bad oscillating frequency stabilization by itself, the phenomena of striking and the dispersion and the slight jump of the spectrum by any cause during the operating process. As a result, the accelerator will be unstable. In order to solve the above problems efficiently a frequency stabilization system, based on the BP neural network method is presented in the paper.