欢迎访问《核聚变与等离子体物理》官方网站,今天是 分享到:

核工业西南物理研究院 ›› 2023, Vol. 43 ›› Issue (4): 469-476.DOI: 10.16568/j.0254-6086.202304016

• 等离子体物理 • 上一篇    下一篇

激光诱导荧光诊断的波长标定方法的研究 

马铭键 1 ,叶孜崇(Yip Chi-shung)*2,江 堤 2 ,毛世峰 1 , 张 炜 2 ,靳琛垚 2 ,叶民友 1    

  1. (1. 中国科学技术大学核科学技术学院,合肥 230026;2. 中国科学院等离子体物理研究所,合肥 230031) 
  • 收稿日期:2022-02-28 修回日期:2023-05-12 出版日期:2023-12-15 发布日期:2023-12-07
  • 作者简介:马铭键(1997-),男,河南平顶山人,硕士研究生,主要从事激光诱导荧光诊断研究。
  • 基金资助:
    国家磁约束核聚变能发展研究专项(2019YFE03030004);国家自然科学基金(11975232) 

Investigation of wavelength calibration method for laser induced fluorescence diagnostics

MA Ming-jian1 , YIP Chi-shung2 , JIANG Di2 , MAO Shi-feng1 , ZHANG Wei2 , JIN Chen-yao2 , YE Min-you1   

  1. (1. School of Nuclear Science and Technology, University of Science and Technology of China, Hefei 230026) (2. Institute of Plasma Physics, Chinese Academy of Sciences, Hefei 230031) 
  • Received:2022-02-28 Revised:2023-05-12 Online:2023-12-15 Published:2023-12-07

摘要: 激光诱导荧光作为一种非侵入式的诊断技术,通过激光扫频得到粒子速度分布函数来计算其速度和温度。为得到较高信噪比的粒子速度分布函数,需要对多次采集的数据进行平均处理,这需要标定每次扫频的激光波 长,以确定每次扫频的相同波长范围。为此发展了利用碘分子吸收光谱来进行波长标定的技术。为探究有效的光谱 对齐算法,分别使用最小均方误差法、误差梯度下降算法和卷积神经网络在模拟碘分子吸收光谱数据集上进行了测试。结果显示,最小均方误差法具有很好的对齐精度,且对齐平均误差小于两个采样间隔。利用最小均方误差法对 齐碘分子吸收光谱,对 Ar+ 激光诱导荧光实验结果的激光波长进行了标定,平均了多组 Ar+ 速度分布函数。相较于利用法布里-珀罗干涉仪共振峰标定以及激光触发曲线标定,通过碘吸收光谱标定得到的离子温度更准确。

关键词: 等离子体诊断, 激光诱导荧光, 速度分布函数, 波长标定, 卷积神经网络

Abstract: Laser Induced Fluorescence is a non-invasive diagnostic technology that obtains the velocity and temperature through the particle velocity distribution function obtained by laser scanning. In order to obtain a particle velocity distribution function with a higher signal-to-noise ratio, data from multiple scans needs to be averaged. This requires calibration of the laser wavelength of each scan to ensure that the wavelength range of each set of data is basically the same. For this reason, a scheme for wavelength calibration using iodine spectrum is developed. The minimum mean square error method, the error gradient descent algorithm and the convolutional neural network are tested respectively on the simulated iodine spectrum dataset in order to probe into the effective spectrum alignment algorithm. The results show that, minimum mean squared error has good alignment accuracy, and the average error of alignment is less than two sampling intervals. The laser wavelength of laser induced fluorescence experiment of Ar+ was calibrated by aligning iodine spectrum using the minimum mean squared error method. Multiple sets of velocity distribution function of Ar+ were averaged. Compared with Fabry-Perot etalon trigger peak calibration and laser scan synchronization calibration, the ion temperature obtained by iodine spectrum calibration is more accurate.  

Key words: Plasma diagnostics, Laser induced fluorescence, Velocity distribution function, Wavelength calibration, Convolution neural network

中图分类号: