W. Y. Zhai
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
W. Y. Zhai has a total of 62 co-authors. Here are the top co-authors by number of shared publications:
- W. L. Zhong
Southwestern Institute of Physics
3 shared publications C H Liu
Southwestern Institute of Physics
2 shared publicationsG L Xiao
Southwestern Institute of Physics
2 shared publicationsH Funaba
National Institute for Fusion Science
2 shared publicationsI Yamada
National Institute for Fusion Science
2 shared publications- J B Wu
Southwestern Institute of Physics
2 shared publications
Latest Publications
- 29 October 2026
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (43) , 436106
Weighted sparse template fitting for stray-light separation in Thomson scattering diagnostics
Delayed stray-light-like waveform interference was observed in the HL-3 Thomson scattering (TS) system, where a laser-synchronous delayed background overlaps the scattering signal and biases the channel-integrated intensity. A weighted sparse template fitting (WSTF) method is developed for waveform-level correction. Using a measured delayed-background waveform as a template prior, WSTF constructs a time-shifted template dictionary and determines non-negative sparse coefficients while down-weighting the scattering integration window. The fit is therefore constrained mainly by the background-dominated region, aiming to preserve the scattering-window integral rather than minimize the full-waveform residual. The method is evaluated using semi-experimental synthetic waveforms with a known scattering integral, over scans of signal-to-noise ratio (SNR), temporal overlap, and amplitude ratio A scat / A bg . Under a representative strong-overlap high-SNR baseline with A scat / A bg ≈ 1.1 and Δ t sep = 0 , WSTF gives an integral error of about 1.0 % , lower than the other tested methods. In a representative five-channel polychromator T e lookup model, the scattering-integral error magnitudes are introduced as perturbations to the fifth channel and propagated to the inferred T e . At T e = 0.5 keV , the resulting propagated relative T e errors are approximately 3.8 % , 12.6 % , 13.6 % , and 0.8 % for FTS, DG, TPTF, and WSTF, respectively.
- 29 October 2026
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (43) , 436106
Weighted sparse template fitting for stray-light separation in Thomson scattering diagnostics
Delayed stray-light-like waveform interference was observed in the HL-3 Thomson scattering (TS) system, where a laser-synchronous delayed background overlaps the scattering signal and biases the channel-integrated intensity. A weighted sparse template fitting (WSTF) method is developed for waveform-level correction. Using a measured delayed-background waveform as a template prior, WSTF constructs a time-shifted template dictionary and determines non-negative sparse coefficients while down-weighting the scattering integration window. The fit is therefore constrained mainly by the background-dominated region, aiming to preserve the scattering-window integral rather than minimize the full-waveform residual. The method is evaluated using semi-experimental synthetic waveforms with a known scattering integral, over scans of signal-to-noise ratio (SNR), temporal overlap, and amplitude ratio A scat / A bg . Under a representative strong-overlap high-SNR baseline with A scat / A bg ≈ 1.1 and Δ t sep = 0 , WSTF gives an integral error of about 1.0 % , lower than the other tested methods. In a representative five-channel polychromator T e lookup model, the scattering-integral error magnitudes are introduced as perturbations to the fifth channel and propagated to the inferred T e . At T e = 0.5 keV , the resulting propagated relative T e errors are approximately 3.8 % , 12.6 % , 13.6 % , and 0.8 % for FTS, DG, TPTF, and WSTF, respectively.
- Open Access
- 1 May 2026
- Published by AIP Publishing in Physics of Plasmas
- Vol. 33 (5)
Effects of biasing on current drive and estimation of the parallel extent of the biasing current in the HL-3 tokamak
A wall mounted biasing plates system by eight biasing plates distributed toroidally on the upper first wall is applied to the HL-3 tokamak. Plasma configuration has a significant impact on the amplitude of bias-driven current. We also find that bias-driven current will affect the parameters in the scrape-off layer (SOL). It is also found that changes in magnetic fluctuations measured by the Mirnov probe are consistent with the spatial distribution of the bias-driven current. This indicates that the biasing plate system can drive the SOL current filament in HL-3 and provide the possibility of edge localized mode control.
- 1 May 2026
- Published by AIP Publishing in Physics of Plasmas
- Vol. 33 (5)
Open AccessEffects of biasing on current drive and estimation of the parallel extent of the biasing current in the HL-3 tokamak
A wall mounted biasing plates system by eight biasing plates distributed toroidally on the upper first wall is applied to the HL-3 tokamak. Plasma configuration has a significant impact on the amplitude of bias-driven current. We also find that bias-driven current will affect the parameters in the scrape-off layer (SOL). It is also found that changes in magnetic fluctuations measured by the Mirnov probe are consistent with the spatial distribution of the bias-driven current. This indicates that the biasing plate system can drive the SOL current filament in HL-3 and provide the possibility of edge localized mode control.
- 1 May 2026
- Published by IOP Publishing in Plasma Physics and Controlled Fusion
- Vol. 68 (5) , 055021
A microsecond-latency real-time Thomson scattering system on the HL-3 Tokamak
Real-time (RT) electron temperature ( T e ) diagnostics are fundamental for active profile control in tokamaks. This paper present a field-programmable gate array-based RT Thomson scattering processor deployed on the HL-3 tokamak. Driven by strict RT constraints, the fully pipelined architecture integrates Z -score standardization with a lookup-table inversion. This removes the explicit amplitude degree of freedom in the mapping definition, fundamentally mitigating sensitivity to global signal scaling. The hardware achieves a deterministic trigger-to-output latency of 22.4 μ s . To guarantee RT data reliability for prospective closed-loop control, a firmware-level gating mechanism explicitly flags low-confidence samples, bounded by a Monte Carlo baseline error assessment. Benchmarked against an offline non-linear least-squares reference via a shared-trigger setup, the system demonstrates exceptional agreement across 142 valid pulses, yielding a linear fit of y = 1.00 x − 0.002 ( R > 0.999 ). Validated under 30 Hz single-point operation, this architecture establishes a definitive microsecond-latency foundation for future multi-point RT upgrades.
- 1 May 2026
- Published by IOP Publishing in Plasma Physics and Controlled Fusion
- Vol. 68 (5) , 055021
A microsecond-latency real-time Thomson scattering system on the HL-3 Tokamak
Real-time (RT) electron temperature ( T e ) diagnostics are fundamental for active profile control in tokamaks. This paper present a field-programmable gate array-based RT Thomson scattering processor deployed on the HL-3 tokamak. Driven by strict RT constraints, the fully pipelined architecture integrates Z -score standardization with a lookup-table inversion. This removes the explicit amplitude degree of freedom in the mapping definition, fundamentally mitigating sensitivity to global signal scaling. The hardware achieves a deterministic trigger-to-output latency of 22.4 μ s . To guarantee RT data reliability for prospective closed-loop control, a firmware-level gating mechanism explicitly flags low-confidence samples, bounded by a Monte Carlo baseline error assessment. Benchmarked against an offline non-linear least-squares reference via a shared-trigger setup, the system demonstrates exceptional agreement across 142 valid pulses, yielding a linear fit of y = 1.00 x − 0.002 ( R > 0.999 ). Validated under 30 Hz single-point operation, this architecture establishes a definitive microsecond-latency foundation for future multi-point RT upgrades.
- Open Access
- 23 April 2026
- Published by IOP Publishing in Plasma Physics and Controlled Fusion
- Vol. 68 (5) , 055005
Identification of mode numbers and wave numbers by multiple harmonics of instabilities
Harmonic phenomena of plasma instabilities are frequently observed in tokamak plasmas. The harmonics of instabilities have higher toroidal mode numbers than their base modes, leading to proportionally larger Doppler shifts. This paper presents a method for determining the toroidal mode number based on the frequency differences between harmonics. Based on Doppler shift effect, the toroidal mode number can be expressed as n = (f l -f 1 )/((l-1)f vϕ ), where l is the harmonic order and f vϕ is the toroidal rotation frequency. Compared to conventional techniques for mode number identification, the proposed approach offers advantages such as simplified data processing and lower requirement for the number of diagnostic channels. The harmonic components in the frequency spectrum can also be used to retrieve the instability frequency in the plasma frame. In this paper, it is demonstrated that the harmonics of plasma instabilities are not generated by intrinsic mode coupling, but are attributed to the nonlinear distortion of waveform. Diagnostic measurements, including soft X-ray and microwave interferometry, confirm that higher-order harmonics of plasma instabilities exist as real structural features inside the plasma and can interact with microwaves when the Bragg condition is satisfied. This property can be utilized to determine the perpendicular wave numbers of the instabilities. A simple physical model of the distribution of an instability and its harmonics is also proposed.
- 23 April 2026
- Published by IOP Publishing in Plasma Physics and Controlled Fusion
- Vol. 68 (5) , 055005
Open AccessIdentification of mode numbers and wave numbers by multiple harmonics of instabilities
Harmonic phenomena of plasma instabilities are frequently observed in tokamak plasmas. The harmonics of instabilities have higher toroidal mode numbers than their base modes, leading to proportionally larger Doppler shifts. This paper presents a method for determining the toroidal mode number based on the frequency differences between harmonics. Based on Doppler shift effect, the toroidal mode number can be expressed as n = (f l -f 1 )/((l-1)f vϕ ), where l is the harmonic order and f vϕ is the toroidal rotation frequency. Compared to conventional techniques for mode number identification, the proposed approach offers advantages such as simplified data processing and lower requirement for the number of diagnostic channels. The harmonic components in the frequency spectrum can also be used to retrieve the instability frequency in the plasma frame. In this paper, it is demonstrated that the harmonics of plasma instabilities are not generated by intrinsic mode coupling, but are attributed to the nonlinear distortion of waveform. Diagnostic measurements, including soft X-ray and microwave interferometry, confirm that higher-order harmonics of plasma instabilities exist as real structural features inside the plasma and can interact with microwaves when the Bragg condition is satisfied. This property can be utilized to determine the perpendicular wave numbers of the instabilities. A simple physical model of the distribution of an instability and its harmonics is also proposed.
- 1 August 2022
- Published by AIP Publishing in Review of Scientific Instruments
- Vol. 93 (8) , 083512
Visible imaging system with changeable field of view on the HL-2A tokamak
A new visible imaging system characterizing a flexible optical design and delivering high resolution frames is established on the HL-2A tokamak. It features a modular configuration, consisting of a front-end imaging lens, a set of bilateral telecentric relay lenses, and a camera. To avoid the effects of plasma radiation (x and gamma-rays) and magnetic field variation on the camera, it should be away from the coils. Therefore, the length of the relay lenses determines the total size of the imaging system. The main feature of this imaging system is to realize the variation of field of view (FOV) by interchanging the front-end prime lenses or by using a zoom lens directly rather than designing the optical system afresh, which lowers the cost drastically. The primary purpose of varying FOV is to enrich the versatility of this system, i.e., focusing on a narrow FOV such as gas puff imaging or a wide FOV such as the plasma cross sections. During the HL-2A experiments, this visible imaging system is used to provide high quality pictures of the plasma–wall interaction, divertor detachment, pellet injections, and so on. The frames confirmed that a strong radiation close to the X point is correlated with the completely detached inner target.
- 1 August 2022
- Published by AIP Publishing in Review of Scientific Instruments
- Vol. 93 (8) , 083512
Visible imaging system with changeable field of view on the HL-2A tokamak
A new visible imaging system characterizing a flexible optical design and delivering high resolution frames is established on the HL-2A tokamak. It features a modular configuration, consisting of a front-end imaging lens, a set of bilateral telecentric relay lenses, and a camera. To avoid the effects of plasma radiation (x and gamma-rays) and magnetic field variation on the camera, it should be away from the coils. Therefore, the length of the relay lenses determines the total size of the imaging system. The main feature of this imaging system is to realize the variation of field of view (FOV) by interchanging the front-end prime lenses or by using a zoom lens directly rather than designing the optical system afresh, which lowers the cost drastically. The primary purpose of varying FOV is to enrich the versatility of this system, i.e., focusing on a narrow FOV such as gas puff imaging or a wide FOV such as the plasma cross sections. During the HL-2A experiments, this visible imaging system is used to provide high quality pictures of the plasma–wall interaction, divertor detachment, pellet injections, and so on. The frames confirmed that a strong radiation close to the X point is correlated with the completely detached inner target.
- 1 April 2021
- Published by AIP Publishing in Review of Scientific Instruments
- Vol. 92 (4) , 043542
Simulation of stray laser light of Thomson scattering systems in the HL-2M tokamak
Stray laser light is a serious problem that interrupts the measurement of electron temperature and density in Thomson scattering (TS) systems. This paper presents a ray-tracing simulation of stray laser light in HL-2M TS systems. A model including (i) a simplified laser-beam injection system, (ii) the scattered-light collection systems for central-point TS (CPTS) and edge TS (ETS), and (iii) the HL-2M vessel is built using the TracePro and CATIA software packages based on measurements of the bidirectional scattering distribution function at a wavelength of 532 nm. The simulation results show that no stray laser light reaches the injection surface of the collection lens of the CPTS system, and only a few stray laser rays of all the rays reaching the injection surface of the collection lens of the ETS system can be collected when the energy threshold per ray is 1 × 10−16. The stray laser rays that reach the scattered-light collection systems are mainly from the first and second parts of the slides of the beam dump, so decreasing the roughness of those parts could be effective in reducing the level of stray laser light.
- 1 April 2021
- Published by AIP Publishing in Review of Scientific Instruments
- Vol. 92 (4) , 043542
Simulation of stray laser light of Thomson scattering systems in the HL-2M tokamak
Stray laser light is a serious problem that interrupts the measurement of electron temperature and density in Thomson scattering (TS) systems. This paper presents a ray-tracing simulation of stray laser light in HL-2M TS systems. A model including (i) a simplified laser-beam injection system, (ii) the scattered-light collection systems for central-point TS (CPTS) and edge TS (ETS), and (iii) the HL-2M vessel is built using the TracePro and CATIA software packages based on measurements of the bidirectional scattering distribution function at a wavelength of 532 nm. The simulation results show that no stray laser light reaches the injection surface of the collection lens of the CPTS system, and only a few stray laser rays of all the rays reaching the injection surface of the collection lens of the ETS system can be collected when the energy threshold per ray is 1 × 10−16. The stray laser rays that reach the scattered-light collection systems are mainly from the first and second parts of the slides of the beam dump, so decreasing the roughness of those parts could be effective in reducing the level of stray laser light.