摘要
针对非线性全局优化反演算法在Rayleigh波反演中存在固有的计算效率低、收敛迭代慢、内存占用高等问题,这里首先引入改进的Haskell-Thomson矩阵传递正演算法,用以提高频散曲线正演的计算效率;其次采用Monte carlo反演策略代替非线性全局优化反演,进而改善反演效率低及内存占用高等问题;最后在MASW技术中,发展了一套基于Monte carlo的Rayleigh波局部线性化反演流程。通过典型模型试算表明相对于经典的阻尼最小二乘反演算法,本方法不仅能够有效提升计算效率、减少编程难度,还兼具较高的收敛速度和反演质量,反演过程较为稳定。
As the nonlinear global optimization inversion algorithm exists inherent problems in the Rayleigh wave inversion,such as low computational efficiency,slow convergence iteration,and high memory usage,an improved Haskell-Thomson matrix transfer forward algorithm is introduced to improve the calculation efficiency of the dispersion curve forward at first in this paper;secondly,the Monte Carlo inversion strategy is used to replace the nonlinear global optimization inversion,thereby improving the problems of inversion efficiency and memory occupation;finally,a set of local linearization inversion procedures for Rayleigh waves based on Monte Carlo was developed in the MASW technology.The trial calculations of typical models show that compared with the Classic damped least squares inversion algorithm,this method not only can effectively improve the calculation efficiency and reduce the programming difficulty,but also has a higher convergence speed and inversion quality as well as more stable inversion process.
作者
胡昌荣
黄金强
刘宏
窦勇硕
何云川
HU Changrong;HUANG Jinqiang;LIU Hong;DOU Yongshuo;HE Yunchuang(Key Laboratory of Karst Georesources and Environment, Ministry of Education,Guiyang 550025,China;The College of Resource and Environmental Engineering,Guizhou University,Guiyang 550025,China)
出处
《物探化探计算技术》
CAS
2022年第3期286-301,共16页
Computing Techniques For Geophysical and Geochemical Exploration
基金
贵州省科技计划项目(黔科合基础[2020]1Y174)
贵州大学人才引进项目(贵大人基合字[2018]27)
贵州省人才基地项目(RCJD2018-21)。