基于前馈神经网络的等离子体光谱诊断方法  

Plasma optical emission spectroscopy based on feedforward neural network

作　　者：王彦飞[1] 朱悉铭 张明志[3] 孟圣峰 贾军伟[3] 柴昊[3] 王旸[1] 宁中喜[1,2] Wang Yan-Fei;Zhu Xi-Ming;Zhang Ming-Zhi;Meng Sheng-Feng;Jia Jun-Wei;Chai Hao;Wang Yang;Ning Zhong-Xi(Harbin Institute of Technology,Harbin 150001,China;Key Laboratory of Aerospace Plasma Propulsion,Ministry of Industry and Information Technology,Harbin 150001,China;Beijing Orient Institute of Measurement and Test,Beijing 100086,China)

机构地区：[1]哈尔滨工业大学,哈尔滨150001 [2]工信部“航天等离子体推进”重点实验室,哈尔滨150001 [3]北京东方计量测试研究所,北京100086

出　　处：《物理学报》2021年第9期149-160,共12页Acta Physica Sinica

基　　金：国家自然科学基金(批准号:11775063);国防基础科研计划(批准号:JCKY2018203B029);国防计量基础课题(批准号:JSJL2016203B017)资助的课题。

摘　　要：光谱诊断在等离子体刻蚀、材料处理、等离子体设备和工艺开发,以及航天等离子体推进等领域得到了广泛的应用.光谱诊断依赖的碰撞辐射模型会受到碰撞截面等基础物理数据所含偏差的影响,导致诊断结果出现误差.针对这一问题,本文开发了一种基于前馈神经网络的等离子体光谱解算方法.通过对比新方法与以往常用的最小二乘诊断方法的误差特性,发现神经网络诊断方法能够通过辨识光谱向量的主要特征,减小基础数据偏差向诊断结果的传递.对实验光谱数据的分析进一步印证了这一点.本文还对神经网络算法对抗基础数据偏差的机理进行了分析.这种方法在等离子体参数在线监测、成像监测海量数据处理等领域具有良好的应用前景.Optical emission spectroscopy(OES)has been widely applied to plasma etching,material processing,development of plasma equipment and technology,as well as plasma propulsion.The collisional-radiative model used in OES is affected by the deviation of fundamental data such as collision cross sections,thus leading to the error in diagnostic results.In this work,a novel method is developed based on feedforward neural network for OES.By comparing the error characteristics of the new method with those of the traditional least-square diagnostic method,it is found that the neural network diagnosis method can reduce the transmission of basic data deviation to the diagnosis results by identifying the characteristics of the spectral vector.This is confirmed by the experimental results.Finally,the mechanism of the neural network algorithm against fundamental data deviation is analyzed.This method also has a good application prospect in plasma parameter online monitoring,imaging monitoring and mass data processing.

关 键 词：等离子体光谱诊断 前馈神经网络 碰撞辐射模型 

分 类 号：O53[理学—等离子体物理] TP183[理学—物理]