Dynamic Performance Investigation of Acceleration Grid Power Supply for NNBI System With RBF-Type-III Compensator

Acceleration grid power supply (AGPS) rated 200 kV/28 A is an important part of the negative neutral beam injection (NNBI) system. Dynamic performance is the focus of the design of the closed-loop system for AGPS. For special power applications, the single voltage mode of the control system is determined in AGPS. On this basis, a type-III compensator is introduced to optimize the control system. For the problem of the type-III compensator parameter tuning, the traditional
$k$
-factor method is introduced in this article. However, it is inadequate to deal with the system of parameter variation. Given this problem, an improved type-III compensator based on the radial basis function (RBF) neural network is proposed in this article. The algorithm can identify the system and train the optimal compensator parameters, which enhances the system’s robustness. In this article, the performance of the traditional type-III and the RBF-type-III compensator is compared under experimental conditions of different bus voltages. Simulation and experimental results validate the efficacy of this design.