【作者】 蒋婵君；

【导师】 周竹生；

【作者基本信息】 中南大学 ， 地球探测与信息技术， 2009， 硕士

【摘要】 瑞雷面波作为一种新的弹性波勘探方法,具有分辨率高、受场地影响小、检测设备简单、检测速度快等优点,因此广泛地应用于浅表层岩土工程勘察和灾害地质调查、无损检测等领域。瑞雷面波的频散特征与介质的紧密相关性是瑞雷面波勘探的物理基础。瑞雷面波资料处理的效果直接影响到地下介质性质及参数的准确性,所以如何从原始地震记录中准确提取频散曲线,并利用频散曲线来反演地下介质参数是瑞雷面波探测技术的关键问题。本文从面波提纯、提取频散曲线和反演解释这三个关键环节对面波资料处理进行了系统的研究。本文采用τ-p正、反变换提取瑞雷面波。τ-p变换能将在x-t域内相互交叉干涉的复杂波场在τ-p域内各自分离,从而达到滤波的目的,而且它是线性变换,具有较好的保真性、易于实现。实例证明,对原始地震记录进行τ-p正、反变换,剔除了其它震相,得到了同相轴清晰的瑞雷面波波形剖面,为准确计算瑞雷面波相速度奠定了基础。本文介绍了三种提取频散曲线的方法:互相关法、相移法和小波分析方法。互相关法先利用互功率谱来求取相位差Δφ,再根据c_f=2πfΔx/Δφ式计算相速度c_f。相依法巧妙利用e^i2πfx/c作为积分变换的核,在某一频率下进行速度扫描,当c=c_f,说明此时相位差已抵消,叠加能量最大。小波分析法是根据小波分析得到的时频图求取时间差Δt,进而计算相速度。通过理论模型对比,发现互相关法效果欠佳;小波分析方法作为时频分析方法,在理论上虽然优于由于常规傅立叶变换,但在运用小波分析提取频散曲线时,Δt作为分母,极小的Δt误差也会给相速度的计算结果带来较大的误差,所以提取的频散曲线不稳定,实际应用效果并不比相移法理想,却比相移法更费时,所以相移法提取频散曲线的效果最好。本文采用标量传递算法正演计算频散曲线,采用遗传算法反演频散曲线。遗传算法（Genic Algorithm）是模拟生物在自然环境中的遗传和进化过程而形成的一种自适应的全局优化随机检索的算法,利用选择、交换、突变等遗传操作,使优胜者繁殖,劣败者消失,一代一代重复这样的操作,最终找出最优解。与广义逆反演和最小二乘反演相比,遗传算法具有精度高、不依赖初始模型和不需要偏导数等优点。综上所述,本文研究了面波资料处理的每一个环节,找到了每一个环节的最优方法,得到了一整套最优资料处理流程。更多还原

【Abstract】 Rayleigh wave exploration is a new elastic wave exploration method. It has many advantages,such as high resolution, small space affection, simple equipment has been widely used in geotechnical Engineering Investigation and high working speed. So rayleigh wave exploration has been widely used in geological disaster survey and nondestructive testing in shallow. Dispersion characteristic of Rayleigh wave is closely related to the underground media parameters, which is the physical basis of Rayleigh wave exploration. Therefor,it is how to extract accurate dispersion curve from the original seismic records and invert the underground medium parameters from the dispersion curve that is the key problem. In this paper, surface wave data processing are systematically studied from surface wave purification, dispersion curve extraction and inversion.At first ,τ-p transforms are used topurify surface wave data. These transforms can separate Rayleigh wave data from the whole wave data. The transforms is linear and reliable because of their good fidelity. In addition,the transforms can be easily realized with computer.τ-p transforms is proofed to be feasible and effective by examples.To extract the dispersion curve of Rayleigh wave, crosscorrelation method, phase-shist method and wavelet analytical method are introduced and compared in this paper. The effect of crosscorrelation method is the most unsatisfactory. The effect wavelet analytical method is not better than phase-shist method,while wavelet analytical method cost more time than phase-shist method. So the phase-shist method is more effective than the others.This paper introduced Genic Algorithm , using the scalar transmission algorithm to forward the frequency dispersion curve. Genic Algorithm is global optimization algorithm .Genic Algorithm simulates Genetic and biological evolution In the natural environment .Compared with generalized counter algorithm and the least squares method inversion, this method has higher accuracy , does not lay on the initial model and does not require partial derivatives.Overall,this paper researched each step of Surface wave data processing, found the best method in every step and got optimal surface wave data processing flow.更多还原