基于多项式混沌的激光大气传输湍流效应不确定度量化  

Uncertainty Quantification for Turbulence Effect of LaserPropagation in Atmosphere Based on Polynomial Chaos

作　　者：杜配冰 刘钰[1] 陈志华[1] 关奇[1] 夏洪富[1] 蔡利兵[1] DU Pei-bing;LIU Yu;CHEN Zhi-hua;GUAN Qi;XIA Hong-fu;CAI Li-bing(Northwest Institute of Nuclear Technology,Xi’an 710024,China)

机构地区：[1]西北核技术研究所,西安710024

出　　处：《现代应用物理》2021年第2期29-34,共6页Modern Applied Physics

基　　金：国家自然科学基金资助项目(61907034)。

摘　　要：针对激光大气传输湍流效应不确定度量化中采用蒙特卡罗(MC)方法计算效率低的问题,提出了一种基于多项式混沌(PC)的不确定度量化方法。该方法采用拉丁超立方抽样随机抽取少量样本,经过数值模拟后通过多项式混沌构造代理模型,并依据代理模型进行不确定度量化。以准直Gauss光束的地空垂直上行传输为例,不确定度计算中考虑了激光器出口光束质量和大气相干长度对远场功率密度的影响,并以大样本量MC方法的不确定度量化结果作为基准,对PC方法进行了验证。结果表明,PC方法只需对少量样本进行数值模拟计算,便可快速构建远场功率密度的代理模型,得到的不确定度量化结果与MC方法的量化结果吻合较好。Monte Carlo(MC)method is the most commonly used method to calculate the uncertainty quantification for turbulence effect of laser propagation in atmosphere.It has the characteristics of simple operation and strong applicability.However,the mean convergence rate of MC method is too slow,which requires a large number of samples for numerical simulation to get more accurate results,so the calculation cost is too high.In this paper,we propose a method to quantify the uncertainty by using Latin hypercube sampling to randomly select a small number of samples and constructing a surrogate model through polynomial chaos(PC)after numerical simulation.Taking the vertical uplink propagation of collimated Gauss beam as an example,the influence of beam quality of the laser and atmospheric coherence length on the far-field power density is considered in the uncertainty calculation,and the PC method is verified by the uncertainty quantization results of MC method.The results show that the PC method can quickly build the surrogate model of far-field power density with only a small number of samples,and the quantitative results of uncertainty are in good agreement with those of MC method.

关 键 词：激光大气传输 湍流 多项式混沌 不确定度量化 代理模型 拉丁超立方抽样 

分 类 号：TN012[电子电信—物理电子学]