Machine-learning guided optimization of laser pulses for direct-drive implosions

Wu, Feiyang; Yang, Xiaohu; Ma, Yunlong; Zhang, Qi; Zhang, Zhe; Yuan, Xiaohui; Liu, Hao; Liu, Zhengdong; Zhong, Jiayong; Zheng, Jian; Li, Yutong; Zhang, Jie

doi:10.1017/hpl.2022.4

“…The waveform of compressing lasers shown in Fig. 1(b) was designed by Multi-1D code [16], and has been optimized with artificial intelligence algorithm to achieve a quasi-isentropically compression [17]. Simulated results from Multi-1D accorded quite well with the experimental results on the shell velocity of the plasma jets ejected from the vertex.…”

Section: Methodsmentioning

confidence: 66%

Diagnosing the fast-heating process of the double-cone ignition scheme with X-ray spectroscopy

Dai,

Gu,

Fang

et al. 2024

High Pow Laser Sci Eng

1

0

View full text Add to dashboard Cite

In the double-cone ignition scheme of inertial confinement fusion, the head-on collision of two compressed fuel jets from the cone-tips forms an isochoric plasma, which is then heated suddenly by an MeV relativistic electron beam produced by ultra-intense picosecond laser pulses. This fast-heating process was studied experimentally at the Shenguang II upgrade laser facility. By observing temporal resolved x-ray emission and spatial resolved x-ray spectrum, the colliding process and heating process are carefully

show abstract

“…Recently, data-driven methods [20][21][22][23][24][25][26][27] have shown great potential in ICF research. For example, the evolutionary algorithms can automate the exploration of pulse shapes and target geometries, and generate new classes of implosion designs [21][22][23].…”

Section: Introductionmentioning

confidence: 99%

“…Recently, data-driven methods [20][21][22][23][24][25][26][27] have shown great potential in ICF research. For example, the evolutionary algorithms can automate the exploration of pulse shapes and target geometries, and generate new classes of implosion designs [21][22][23]. Regression and classification algorithms can identify the complex correlations between the experimental inputs and outputs [21,24,25], and help increase the experimental neutron yield [26].…”

Section: Introductionmentioning

confidence: 99%

Laser pulse shape designer for direct-drive inertial confinement fusion implosions

Tao

¹

,

Zheng

²

,

Jia

³

et al. 2023

High Pow Laser Sci Eng

Self Cite

8

0

View full text Add to dashboard Cite

Pulse shaping is a powerful tool for mitigating implosion instabilities in directdrive inertial confinement fusion. However, the high-dimensional and nonlinear nature of implosions makes the pulse optimization quite challenging. In this research, we develop a machine learning pulse shape designer to achieve high compression density and stable implosion. The facility-specific laser imprint pattern is considered in the optimization, which makes the pulse design more relevant. The designer is applied to the novel doublecone ignition scheme, and simulation shows that the optimized pulse increases the areal density expectation by 16% in 1-D, and the clean-fuel thickness by a factor of 4 in 2-D. This pulse shape designer could be a useful tool for direct-drive ICF instability control.

show abstract

“…It becomes important to investigate the formation of a hot spot in a more realistic configuration with the progress of fast ignition schemes, such as the cone-in-shell fast-ignition [22,23], indirect-drive fast-ignition [24] and double-cone ignition (DCI) [25][26][27]. In the DCI scheme, drive laser beams first drive the fuels embedded in the two head-on cones, leading to a head-on collision of two high speed plasma jets from the tips of the cones.…”

Section: Introductionmentioning

confidence: 99%

Formation of hot spots at end-on pre-compressed isochoric fuels for fast ignition

Xu,

Wu,

Jiang

et al. 2023

Nucl. Fusion

Self Cite

3

0

View full text Add to dashboard Cite

It is crucial to form a hot spot that can drive a self-heating nuclear burn propagation into the surrounding cold fuel in ignition process of laser fusion. In this paper, we discuss the formation of a hemispherical hot spot located at the edge of an end-on precompressed isochoric fuel instead of a central spherical hot spot surrounded by cold fuel in a corona plasma shell. This configuration leads to a new energy loss mechanism named hot spot collapse, which originates from mass loss of the hot spot through the interface with the vacuum. A semi-analytical model is proposed including α particle induced burning, hot spot collapse and other energy loss processes. Then a modified ignition criterion is derived. The results show that the competition between the hot spot collapse and the burn propagation is crucial in the formation of the hot spot. The formation of such a hot spot at end-on precompressed isochoric fuels would require a somewhat higher initial temperature of T h = 9 keV for a hot spot with an initial areal density ρ d = 0.6 g c m − 2 . Simulations are performed with a 3D hybrid particle-in-cell/fluid code for the heating process by fast electrons and a 3D radiation hydrodynamics code for the burning process. Simulation results agree with the model. Our optimized result shows a 10 ps heating laser pulse with an energy as low as 39 kJ can lead to a fast ignition in the ideal case. This study provides an effective reference for design and evaluation of experiments planned for fast ignition schemes.

show abstract

Machine-learning guided optimization of laser pulses for direct-drive implosions

Cited by 30 publications

References 30 publications

Diagnosing the fast-heating process of the double-cone ignition scheme with X-ray spectroscopy

Diagnosing the fast-heating process of the double-cone ignition scheme with X-ray spectroscopy

Laser pulse shape designer for direct-drive inertial confinement fusion implosions

Formation of hot spots at end-on pre-compressed isochoric fuels for fast ignition

Contact Info

Product

Resources

About