The high-power neutral beam injection system is a key auxiliary heating mechanism for achieving steady-state operation in the Experimental Advanced Superconducting Tokamak and the future Chinese Fusion Engineering Test Reactor. Due to the long-term operation of ion sources in high-voltage environments, frequent inter-electrode breakdown (arcing) phenomena pose severe challenges to equipment safety. Although traditional FPGA-based hard-wired protection schemes offer rapid response, they suffer from limitations such as rigid logic and difficult parameter tuning. To address these challenges, this paper proposes and designs a high-precision timing and protection system based on a general-purpose computing architecture. Based on the JYTEK PXIe hardware platform, the system utilizes the deterministic scheduling mechanism of the Linux real-time (Linux RT) operating system to achieve nanosecond-level synchronization for digital outputs and microsecond-level synchronization for analog outputs. Building upon this foundation, software-defined automatic arc detection and re-ignition logic was developed. Experimental results demonstrate that the system possesses rapid fault response capabilities, with the latency from detecting the arc signal to cutting off both the digital output and analog output remaining at the microsecond level. Furthermore, the stability of the re-ignition logic was validated under three typical operating conditions: normal recovery, multiple retries, and emergency stop, ensuring timing integrity and pulse width consistency during the fault recovery process. In addition, a remote communication module based on the UDP protocol achieves reliable parsing and closed-loop mapping of complex experimental parameter arrays, meeting the requirements for remote automated control.
Building similarity graph...
Analyzing shared references across papers
Loading...
Ao Liu
Liang Zhang
Xiancai Meng
AIP Advances
Institute of Plasma Physics
National Science Centre
Building similarity graph...
Analyzing shared references across papers
Loading...
Liu et al. (Fri,) studied this question.
www.synapsesocial.com/papers/69fed071b9154b0b82877922 — DOI: https://doi.org/10.1063/5.0324918