现代应用物理2025，Vol.16Issue(1)：133-137,5.DOI:10.12061/j.issn.2095-6223.202412026
基于随机森林的直接驱动惯性约束聚变内爆压缩代理模型
Inertial Confinement Fusion Implosion Compression Surrogate Model Based on Random Forest
摘要
Abstract
A direct-driven inertial confinement fusion(ICF)implosion compression surrogate model based on random forests is established,which can be used to predict the implosion compression performance and assess the stability of the implosion compression under a given laser profiles and target designs.Using a three-pulse laser profile and a three-layer fusion target design,300 sets of simulated data are obtained through a radiation hydrodynamics programs.The data are pre-processed using shockwave compression theory,and the original parameters are combined with the pre-processed parameters to jointly train the random forest surrogate model.To validate the efficiency and accuracy of the model prediction,cross-validation is employed to reconstruct the prediction data.The correlation coefficients of most predicted parameters exceed 0.94,and the maximum error of the surrogate model is 12.59％.This method can be applied to the construction of high dimensional radiation hydrodynamics surrogate models and experimental surrogate models in the future.关键词
惯性约束聚变/随机森林/代理模型/机器学习/辐射流体力学Key words
inertial confinement fusion/random forest/surrogate model/machine learning/radiation hydrodynamics分类
信息技术与安全科学引用本文复制引用
李择,杨晓虎,田佳乐,张国博,马燕云,邵福球..基于随机森林的直接驱动惯性约束聚变内爆压缩代理模型[J].现代应用物理,2025,16(1):133-137,5.基金项目
国家自然科学基金资助项目(12175309,12475252,12275356) （12175309,12475252,12275356）
中国科学院战略重点研究项目(XDA25050200,XDA25010100) （XDA25050200,XDA25010100）
国防工业技术发展计划(JCKYS2023212807) （JCKYS2023212807）
湖南省研究生科研创新项目(CX20230005) （CX20230005）
