自适应对数螺旋路径萤火虫算法反演瑞雷面波频散曲线  

Inversion of Rayleigh wave dispersion curve via adaptive logarithmic Spiral-Lévy firefly algorithm

作　　者：刘国鑫 沈鸿雁 车晗 王鑫 李庆春[4] 李欣欣 赵静 LIU GuoXin;SHEN HongYan;CHE Han;WANG Xin;LI QingChun;LI XinXin;ZHAO Jing(School of Earth Sciences and Engineering,Xi'an Shiyou University,Xi'an 710065,China;Shaanxi Key Laboratory of Petroleum Accumulation Geology,Xi'an 710065,China;College of Geoexploration Science and Technology,Jilin University,Changchun 130012,China;College of Geology Engineering and Geometics,Chang'An University,Xi'an 710054,China)

机构地区：[1]西安石油大学地球科学与工程学院,西安710065 [2]陕西省油气成藏地质学重点实验室,西安710065 [3]吉林大学地球探测科学与技术学院,长春130012 [4]长安大学地质工程与测绘学院,西安710054

出　　处：《地球物理学报》2023年第7期3011-3025,共15页Chinese Journal of Geophysics

基　　金：国家自然科学基金项目(41874123,42004110);陕西省自然科学基础研究计划项目(2017JZ007,2021JQ-587,2021JQ-589);陕西省重点研发计划项目(2022GY-148);西安石油大学研究生创新与实践能力培养计划项目(YCS21112056)联合资助。

摘　　要：瑞雷面波技术是浅地层地质勘探的重要手段之一,通过瑞雷面波频散曲线反演可有效获得地下横波速度模型.然而,瑞雷面波频散曲线反演具有多参数、多极值和非线性等特点,面对复杂地震-地质条件下的瑞雷面波资料处理,传统方法难以快速、精确地对地层参数进行反演和重建.本文将自适应对数螺旋路径萤火虫算法(Adaptive Logarithmic Spiral-Lévy Firefly Algorithm,简称ALSL-FA)引入到瑞雷面波频散曲线反演中,有效解决了经典萤火虫算法(Firefly Algorithm,简称FA)精于探索,疏于开发的缺点.ALSL-FA集成了对数螺旋引导萤火虫路径,并在搜索过程中通过自适应切换因子实现了全局搜索与局部开发自适应切换,在增强局部开发能力的同时,确保了全局搜索的能力.通过Rastrigin函数求解,对比分析了FA、Lévy飞行萤火虫算法(Lévy Flying Firefly Algorithm,简称LF-FA)和ALSL-FA的运算性能;通过速度递增和含高速硬夹层地质模型的瑞雷面波频散曲线反演,比较了人工蜂群算法(Artificial Bee Colony algorithm,简称ABC)、FA、LF-FA和ALSL-FA的反演结果,验证了本文方法的正确性和抗噪能力,并进一步应用于一套实际瑞雷面波资料反演处理.研究结果表明,本文方法有效弱化了对初始模型的依赖,并进一步提高了反演的精度,具有较强的全局寻优能力和局部开发能力,同时兼备良好的抗噪性能.Rayleigh wave technique is one of the important means of shallow surface geological survey.The underground shear wave velocity model can be effectively obtained through the inversion of Rayleigh wave dispersion curve.However,Rayleigh wave dispersion curve inversion has the characteristics of multi-parameter,multi-extreme and nonlinear.Facing the Rayleigh wave data processing under complex seismic-geological conditions,the traditional methods are difficult to invert and reconstruct formation parameters quickly and accurately.In this paper,the Adaptive Logarithmic Spiral-Lévy Firefly Algorithm(ALSL-FA)was introduced into the inversion of Rayleigh wave dispersion curve.ALSL-FA effectively solves the shortcomings of the classical Firefly Algorithm(FA)that is good at exploration and poor at development.ALSL-FA integrates log spiral to guide the path of firefly,and realizes the adaptive switching between global search and local development through adaptive switching factor in the search process,which enhances the ability of global search while enhancing the ability of local development.The computational performance of FA,Lévy Flying Firefly Algorithm(LF-FA)and ALSL-FA was compared and analyzed by solving the Rastrigin function.The inversion results of Artificial Bee Colony algorithm(ABC),FA,LF-FA and ALSL-FA were compared through the inversion of Rayleigh wave dispersion curve of velocity increasing and high-velocity interlayer models while the correctness and anti-noise ability of our method were verified.And then,it was further applied to a real Rayleigh wave dataset inversion processing.Our research results show that the method effectively weakens the dependence on the initial model,and further improves the inversion accuracy.It also has strong global optimization ability and local development inversion accuracy.It has strong global optimization ability and local development ability as while as good anti-noise ability.

关 键 词：瑞雷面波 频散曲线 自适应对数螺旋路径 萤火虫算法 非线性反演 横波速度 浅地层结构 

分 类 号：P631[天文地球—地质矿产勘探]