【作者】 张超；

【导师】 胡恒山；

【作者基本信息】 哈尔滨工业大学 ， 固体力学， 2022， 博士

【摘要】 随钻声波测井的特点是钻井和测井同时进行,其优势在于测量结果准、测量速度快,且能实时指导钻井方向,这使得随钻声波测井在大斜度井、水平井和海上钻井平台中被广泛应用。但目前在声场模拟和数据解释方面仍然存在着一些尚未解决的问题,如软地层横波速度不易获取;模拟波形和实验波形存在较大差异,接收信号的力学量类型不甚明确;各分波的反演慢度频散曲线难以提取;反演慢度频散曲线中伪模式对真模式的干扰难以去除。因此,针对随钻声波测井中的上述问题,本文分析了单极、偶极和四极源激发的各模式波特点,实现了随钻声场的解析与数值模拟,反演了慢度和幅值频散曲线。主要工作和研究进展如下:针对单极源随钻声波测井模拟波形和实验波形存在差异的问题,对比了小模型井实验中的阵列波形与同参数下理论模拟的声压和径向位移波形,分析了声压和径向位移信号中各模式波的特性及模式波能量的径向分布。结果表明单极随钻声波测井中径向位移信号更能反映换能器接收信号的特征。此外,针对偶极源在软地层中没有激发出明显横波波群的问题,考察了舒尔特波的速度频散特征以及舒尔特波速度对井孔和地层参数的敏感度,提出了通过偶极舒尔特波来反演软地层横波速度的方法。该方法利用了舒尔特波速度与地层横波速度的一一对应关系,以及舒尔特波速度对地层横波速度的高敏感度。与前人只模拟换能器压电效应或者只模拟测井声波传播过程的有限元模型不同,本文建立了综合考虑换能器压电效应和声波传播过程的有限元模型,探究了随钻声波测井换能器接收到的力学信号类型。计算了单极子随钻声波测井的力学响应和输出电压,进而利用压电方程以及换能器的极化关系,给出了电压与力学量之间的显式表达式,并进一步考察了换能器记录的输出电压与径向位移、声压、径向应变和轴向应力的关系,结果表明径向位移、径向应变和轴向应力对电压中钻铤波的贡献大。同时研究了钻铤刻槽对波形的影响,结果显示刻槽对电压波形的影响更接近于对径向位移波形的影响。针对慢度频散曲线反演过程中难以分辨给定频率下计算出的多个慢度值属于哪个频散模式的问题,本文通过密度聚类方法将本应属于同一分波模式的频率-慢度点划分为一类,从而自动分辨出了各分波模式的慢度。在此基础上,结合激发强度曲线,阐明了某些模式波反演慢度不准确的原因。此外,针对反演慢度计算中黎曼叶的错误选择导致伪模式对真模式干扰的问题,分析了伪模式产生的原因并提出了去除伪模式的方法。该方法利用幅值谱的单值性,通过对幅值谱进行密度聚类分离了真模式和伪模式,并根据频散曲线的连续性将伪模式移动到正确的黎曼叶上使其成为真模式,从而解决了伪模式对真模式的干扰问题。更多还原

【Abstract】 Acoustic logging while drilling(LWD)is characterized by simultaneous drilling and logging,accurate measurement results and fast measurement,which can guide drilling direction in real time.It can also be applied to highly deviated wells,horizontal wells and offshore drilling platforms.However,there are still some unsolved problems in acoustic field simulation and data interpretation.For example,the shear wave velocity of soft formation is not easy to be measured;There are differences between simulated waveform and experimental waveform;The signal types received by acoustic LWD are not clear;It is difficult to extract the inversion slowness dispersion curve of each individual wave;The aliased mode interferes with the true mode in the inversion slowness dispersion curve.To solve these problems,this paper analyses the characteristics of various acoustic LWD mode waves excited by monopole,dipole and quadrupole sources,simulates the acoustic field of LWD,and obtains the slowness and amplitude dispersion curves from waveform array.The main work and research progress are as follows:To explain the difference between simulated waveforms and experimental waveforms of monopole acoustic LWD,the waveform in the experiment is compared with the waveforms of acoustic pressure and radial displacement under the same parameters.It can be seen that the radial displacement signal can better reflect the signal characteristics received by the transducer in monopole acoustic LWD.In addition,to solve the problem of obtaining shear wave velocity in soft formation,we investigate the dispersion and sensitivity characteristics of Scholte wave and propose a method to obtain the shear wave velocity in soft formation by dipole Scholte wave.This method utilizes the one-to-one correspondence between Scholte wave velocity and shear wave velocity,and the highly sensitivity of Scholte wave velocity to shear wave velocity.Different from the previous finite element models which only simulate the piezoelectric effect of the transducer or the acoustic propagation process in logging,this paper establishes a finite element model which comprehensively considers the piezoelectric effect of the transducer and the propagation process of acoustic wave.We investigate the types of mechanical signals received by the transducer and calculate the mechanical responses and output voltage of the monopole acoustic LWD.Using the piezoelectric equation and the polarization relation of the transducer,we give an explicit expression of the voltage and mechanical quantities,and investigate the relationship between the output voltage recorded by the transducer and the radial displacement,pressure,radial strain and axial stress.The results show that the radial displacement,radial strain and axial stress contribute greatly to the voltage.At the same time,we study the effect of collar grooves on simulated waveform.The results show that the influence of grooves on the voltage waveform is closer to that of radial displacement waveform.To determine which dispersion mode the slowness values calculated at a given frequency belong to in the process of slowness dispersion curve inversion,we utilize density clustering method to classify frequency-slowness points belonging to an individual mode into a class,so as to extract the slowness of the individual wave.On this basis,combining with the intensity excitation curve,we elucidate the reasons for the inaccuracy of inversion slowness of some mode waves.Besides,the wrong choice of Riemannian sheet of inversion slowness leads to the interference of aliased mode to true mode.This paper explains that the reason for the aliased mode and proposes a method to remove the aliased mode.In this method,the true mode and the aliased mode are separated by density clustering of the amplitude spectrum.The aliased mode is moved to the correct Riemannian sheet according to the continuity of the dispersion curve to make it become the true mode,which fundamentally solves the interference of the aliased mode to the true mode.更多还原