Abstract:In the field of magnetic confinement plasma control, the accurate feedback of plasma position and shape primarily relies on calculations derived from magnetic measurements through equilibrium reconstruction or matrix mapping method. However, under harsh conditions like high-energy neutron radiation and elevated temperatures, the installation of magnetic probes within the device becomes challenging. Relying solely on external magnetic probes can compromise the precision of EFIT in determining the plasma shape. To tackle this issue, we introduce a real-time, non-magnetic measurement method on the HL-3 tokamak, which diagnoses the plasma position and shape via imaging. Particularly, we put forward an adapted Swin Transformer model, the Poolformer Swin Transformer (PST), to accurately and fastly interpret the plasma shape from the Charge-Coupled Device Camera (CCD) images. By adopting multi-task learning and knowledge distillation techniques, the model is capable of robustly detecting six shape parameters under disruptive conditions such as a divertor shape and gas injection, circumventing global brightness changes and cumbersome manual labeling. Specifically, the well-trained PST model capably infers R and Z within the mean average error below 1.1 cm and 1.8 cm, respectively, while requiring less than 2 ms for end-to-end feedback, an 80 improvement over the smallest Swin Transformer model, laying the foundation for real-time control. Finally, we deploy the PST model in the Plasma Control System (PCS) using TensorRT, and achieve 500 ms stable PID feedback control based on the PST-computed horizontal displacement information. In conclusion, this research opens up new avenues for the practical application of image-computing plasma shape diagnostic methods in the realm of real-time feedback control.

Time-frequency Analysis of Non-Stationary Fusion Plasma Signals Using an Improved Hilbert-Huang Transform

Time-Frequency Analysis for Microwave Reflectometry Data Processing in the Hl-2a Tokamak

Microwave Diagnostics System in the HL-2A Tokamak

Dynamic Optimization Of Tokamak Plasmas Via Control Parameterization And The Time-Scaling Transformation

Kinetic Investigations of Nonlinear Electrostatic Excitations in Quantum Plasmas.

Phase tracking with Hilbert transform and nonlinear wave-wave coupling analysis on the HL-2A tokamak

Spatiotemporal Distribution Characterization for Terahertz Waves Generated From Plasma Induced by Two-Color Pulses

Experimental Study of Turbulence and Zonal Flow in Edge Plasmas of the HL-2A Tokamak

Spectral characteristics of zonal flows in the edge plasmas of the HT-7 tokamak

Spline-Kernelled Chirplet Transform for the Analysis of Signals with Time-Varying Frequency and Its Application

Study on signal pre-processing method for Hilbert-Huang transform

Tessellation-based analysis of impurity clustering in the edge plasma of tokamaks

Turbulence and Zonal Flows in Edge Plasmas of the HL-2A Tokamak

MHD-FiT: MHD-based dynamic reconstruction of tokamak plasma configuration

Adapted Swin Transformer-based Real-Time Plasma Shape Detection and Control in HL-3

Real-time equilibrium reconstruction by multi-task learning neural network based on HL-3 tokamak

A Comprehensive Analysis of the Gravitational Wave Events with the Hilbert-Huang Transform: From Compact Binary Coalescence to Supernova

Real-time equilibrium reconstruction by neural network based on HL-3 tokamak

Characteristics of Geodesic Acoustic Mode Zonal Flow and Ambient Turbulence at the Edge of the Hl-2a Tokamak Plasmas

Time–frequency Data Fusion Technique with Application to Vibration Signal Analysis

Infra-red thermographic inversion in ST40