基于深度学习的激光放大器特性数学建模研究  

Mathematical modeling of laser amplifier characteristics based on deep learning

作　　者：郭欢 GUO Huan(College of Basic Sciences,Zhengzhou Institute of Technology,Zhengzhou 450044,China)

机构地区：[1]郑州工程技术学院基础科学学院,郑州450044

出　　处：《激光杂志》2021年第9期166-169,共4页Laser Journal

基　　金：河南省科技厅重点研发与推广专项(科技攻关)(No.202102210156)。

摘　　要：当前数学建模方法无法准确描述激光放大器的特性,导致激光放大器特性分析精度低,无法满足激光放大器实际应用的要求。为了改善激光放大器特性数学建模的效果,提出了基于深度学习的激光放大器特性数学建模方法。首先采集影响激光放大器特性的因素数据,然后将它们作为深度学习算法的输入,激光放大器输出功率作为深度学习算法的输出,通过深度学习算法的训练,、建立拟合激光放大器特性和影响之间变化关系的数学模型,最后在Matlab 2018平台进行了激光放大器特性数学建模方法性能的测试实验,结果表明,深度学习的激光放大器特性数学建模精度高于93%,建模误差控制在激光放大器实际应用有效范围内,相对于当其激光放大器特性数学建模方法,本方法具有十分明显优势,获得了更优的激光放大器特性分析结果。The current mathematical modeling methods can not accurately describe the characteristics of laser amplifiers,resulting in the low accuracy,which cannot meet the requirements of practical application of laser amplifiers. In order to improve the effect of mathematical modeling,a mathematical modeling method of laser amplifier characteristics based on deep learning is proposed. Firstly,the data of the factors that affect the characteristics are collected,and then they are used as the input and the output power of the laser amplifier is taken as the output of the deep learning algorithm. deep learning algorithm is used to establish the mathematical model which can fit relationship between characteristics of the laser amplifier and the influence. Finally,performance of laser amplifier characteristics mathematical modeling method was tested on the Matlab 2018 platform. The results show that the precision of deep learning is higher than93%,and the modeling error is controlled within the effective range of practical application of laser amplifier. Compared with the mathematical modeling method of laser amplifier characteristics,the proposed method has obvious advantages and obtains better laser analysis results of amplifier characteristics.

关 键 词：深度学习 激光放大器 特性分析 数学建模方法 光功率 

分 类 号：TN247[电子电信—物理电子学]