裂缝性地层声波全波列测井时频特征研究

发布时间：2018-01-05 15:33

本文关键词：裂缝性地层声波全波列测井时频特征研究　出处：《吉林大学》2016年博士论文　论文类型：学位论文

【摘要】：声波测井是地球物理测井中的主要方法之一,其核心是运用声波在岩层中的各种传播规律,测量所钻地层的地质和岩石物理参数,从而获取地层的油、气藏的存在与岩性等特征。早期的声波测井只是简单的利用一个声源和一个接收器测量沿着井壁传播的首波(即纵波)到达的时间或幅度。随着科学技术的进步,为了提高测量精度及适应不同地层的需求,各大测井公司陆续推出了:单发双收声波测井仪、双发双收声波测井仪、长源距声波测井仪,阵列声波测井仪等。现代阵列声波测井仪同时具有单极子声源和偶极子声源以及多个接收探头。其中,2种声源可以分别以不同的振动方式激发声波信号,而接收器阵列则以不同的组合方式接收声波信号。相比于早期的声波测井,其探测深度更大,测量结果更精准,并且可以接收多种不同类型的波。但是,阵列声波测井信号并不能直接提供太多有用信息,其需要经过一定的数学方法进行处理。目前,现有的阵列声波测井的解释流程中,核心思想是提取各组分波的慢度,通过慢度随井深变化的曲线来判断井眼附近地层某些性质的变化。而慢度提取的方法则主要包括时间域方法与频率域方法两类。时间域方法通常从信号时间域的波形入手,并且能够充分利用阵列声波测井重复采集同一深度信息这一特点,包括门限法、慢度-时间相关法等。频率域方法则通常要对声波测井信号进行Fourier变换,最终可以求得各组分波的频散曲线,解决了斯通利波和偶极横波这类频散波的慢度提取问题,包括频谱相位分析法、Prony方法、Matrix Pencil算法等。众所周知,信号通常具备3个最基本的物理量:时间、频率和幅度(能量)。阵列声波测井信号亦是如此,但其慢度仅仅反映其中一个方面的信息,即各组分波到达时间的早晚。也就是说,常规的处理方法对测井信号的频谱特征的重视程度不足。信号的频谱可以很容易地由Fourier变换求出,但是,Fourier变换是一种全局式的变换,得到信号频率域信息的同时就会失去时间域信息。既然,即要获取信号的频谱特征,又不能放弃信号时间域信息,那么,就有必要引入时频分析方法。时频分析方法的优点在于可以将时间、频率和幅度联系起来,获得信号的幅度在时间-频率坐标系上的分布情况。时频分析方法种类繁多,主要分为线性与双线性两类,其中,线性的方法包括短时Fourier变换、小波变换、Hilbert-Huang变换、分数阶Fourier变换等;双线性的方法包括主要包括Cohen类双线性时频分布和Affine类双线性时频分布等。然而,若将单一的时频分析方法应用于阵列声波测井信号处理,效果并不理想。这主要有两方面的原因:第一,阵列声波测井信号往往由多种分波组成,各分波之间时间和频率较为接近,有时难以区分;第二,各分波之间幅度相差很大,这使得时频分布图很难将幅度较低的分波展现出来。为此,本文考虑联合分数阶Fourier变换和Choi-Williams分布,利用分数阶Fourier变换的旋转特性对阵列声波测井信号的Choi-Williams分布进行时频域滤波,分别提取纵波、横波的时频分布,再结合原始Choi-Williams分布中较易分辨的斯通利波的时频分布,研究纵波、横波和斯通利波的时频特征,探索一种新的阵列声波测井数据处理及解释的方法。本文所进行的工作主要包含三个部分:第一,引入分数阶Fourier变换和Choi-Williams联合分布的方法,并对相关程序进行编写;第二,通过相应的处理,获取阵列声波测井信号时频特征;第三,建立阵列声波测井信号时频特征与地层裂缝之间的联系。下面简述各部分的工作内容。1.引入分数阶Fourier变换和Choi-Williams联合分布的方法,继而探讨其物理意义及实践意义。将编程的工作分为两部分:井段处理和单点处理。井段处理的目的是计算某个深度范围乃至整口井内所有阵列声波测井信号的时频分布(可用于获取斯通利波特征),然后提取纵波和横波的时频分布,并按照深度顺序与原始波列一一对应排列,从而分别获取各组分波的时频特征随深度变化而变化的宏观规律。对于井段处理的部分,本文利用VB.NET与Matlab混合编程的方式进行面向对象的程序设计。单点处理的目的则是通过制作由声波测井全波列曲线,原始信号时频分布,纵波时频分布和横波时频分布4个部分组成的图像,更加精确地获取某些地层中具有代表性的深度点处各组分波的时频信息。对于单点处理的部分,本文利用Matlab对程序进行编写,继而利用Matlab中的GUI设计工具,制作可视化界面。2.选取来自不同地区、不同井孔的裂缝性地层阵列声波测井实测数据,对其进行数据解编,再进行去增益和均衡化处理,然后选取多个深度范围内的数据进行井段处理,分别得到原始测井信号、纵波及横波的时频分布随深度变化的图像。从中选取某些具有代表性的信号进行单点处理,得到该深度点原始测井信号、纵波及横波的时频分布,继而对各组分波进行时频定位,并确定其幅度的大小。3.利用井段处理的结果,探索各组分波时频特征随某些地层性质变化的规律;再利用单点处理的结果,更清晰的认识这些规律。继而结合其它测井及地质资料,从地学及物理学角度对这些规律进行解释,建立阵列声波测井信号时频特征与多种不同类型裂缝性地层之间的关系。经过上述研究工作以后,本文得到了如下结论:1.在阵列声波测井信号的Choi-Williams分布图中,只有斯通利波较为明显,而纵波和横波均不易识别。而分数阶Fourier变换和Choi-Williams联合分布的方法则能将二者加以提取。本文利用该方法处理及分析裂缝性地层的阵列声波测井数据的过程中,获得了较好的效果。因此,该方法具备一定的实践意义。2.对于致密性地层而言,在时频分布图中,阵列声波测井原始信号的Choi-Williams分布图比较规则,斯通利波幅度远高于纵波和横波。相对与中低角度张开缝或网状缝发育的地层而言,各组分波的幅度衰减不明显,波峰时间较早,纵波和横波主频均较高。3.对于半充填的裂缝性地层而言,在时频分布图中,各组分波的波峰时间和主频不会发生明显改变,而幅度衰减则介于致密性地层和中低角度裂缝性地层之间。4.对于中低角度张开缝发育的地层而言,较之致密性地层,在时频分布图中,纵波的波峰时间相对推迟,而主频相对降低,幅度衰减更加严重;横波的波峰时间有所推迟,主频相对降低,幅度衰减同样更加严重;斯通利波波峰时间相对推迟,其幅度衰减十分明显,但幅度值仍明显高于纵波和横波。5.对于高角度张开缝发育的地层而言,相对于致密性地层,在时频分布图中,纵波幅度略微有一些衰减,横波幅度变化不大,斯通利波幅度出现一定的衰减,而各组分波的波峰时间和主频则无明显变化。6.对于网状裂缝发育的地层而言,相对于致密性地层,在时频分布图中,纵波的波峰时间相对推迟,而主频相对降低,幅度衰减更加严重;横波的波峰时间有所推迟,主频相对降低,幅度衰减更加严重;斯通利波波峰时间相对推迟,幅度衰减更加严重。网状裂缝对阵列声波测井信号时频特征的影响较为复杂,无法准确探究其中的规律,但将其与致密性地层区分比较容易。7.对于某些常规测井曲线未明显反映的裂缝,仍可通过本文所介绍的方法获取阵列声波测井信号时频特征,继而对裂缝进行识别。8.将上述规律反过来利用,通过本文介绍的方法做出各井段原始信号及纵波和横波的时频分布图之后,若各组分波波峰时间较早,幅度较高,且纵波和横波主频较高,则可将其判定为致密性地层,并作为参照;若斯通利波幅度相对出现了一定的下降,则说明地层中存在裂缝,且裂缝未被完全充填;若斯通利波幅度严重衰减,同时,纵波幅度也出现了一些衰减,而各组分波的波峰时间和主频及横波幅度变化不大,则判定裂缝具有较高的倾角;若斯通利波幅度严重衰减的同时,其波峰时间也有所推迟,且纵波的波峰时间相对推迟,主频相对降低,幅度衰减严重,横波的波峰时间相对推迟,主频相对降低,幅度衰减严重,则判定裂缝的倾角为中低角度或裂缝呈网状。
[Abstract]:Acoustic logging is one of the main methods in geophysical logging, its core is the law of various communication uses sound waves in the rock, measuring the drilled formation geological and petrophysical parameters, so as to obtain the formation of oil and gas reservoir and lithology. Acoustic logging early simply by a sound source and a receiver to measure the first wave spread along the wall (i.e. longitudinal) time of arrival or amplitude. With the progress of science and technology, in order to improve the precision and adapt to different stratum needs, the logging company launched a single double acoustic logging instrument, BHC sonic log, long spaced acoustilog instrument. Array acoustic logging instrument. Modern array acoustic logging instrument also has the monopole and dipole source and a plurality of receiving probe. Among them, 2 kinds of sound source can be based on the different type of acoustic excitation vibration Wave signal, and the receiver array with different combinations of receiving acoustic signals. Compared to the early detection of acoustic logging, the depth, the measurement results more accurate, and can receive a variety of different types of waves. However, array acoustic logging signal does not directly provide much useful information, the need to go through some mathematical methods for processing. At present, array acoustic logging interpretation of the existing process, the core idea is to extract the components of wave slowness, the slowness curve changing with the depth to determine some properties of the formation of the near wellbore. And method of extracting slowness include time domain method and frequency domain method two. Time domain method usually starts from the signal time domain waveform, and can make full use of array acoustic logging repeated collection of the characteristics of the same depth information, including threshold method, relative slowness time Method. The frequency domain method is Fourier transform on the acoustic logging signal can be obtained, the final components of wave dispersion curves, solving Stone wave and dipole shear wave dispersion of this kind of slow wave extraction, including phase spectrum analysis method, Prony method, Matrix Pencil algorithm and so on. As everyone knows, the signal is usually have 3 basic physical quantities: time, frequency and amplitude (energy). So is the array acoustic logging signal, but its slowness just reflects one aspect, namely each component wave arrival time. That is to say, the lack of emphasis on the spectral characteristics of logging signal processing methods the signal spectrum can be easily obtained by Fourier transform, however, the Fourier transform is a global transformation, get the signal frequency domain information at the same time will lose time domain information. Now, to get the letter Spectral characteristics of signals, and can not give up the signals in the time domain information, so it is necessary to frequency analysis method is introduced. The advantages of the method of time-frequency analysis that can be linked to the time, frequency and amplitude, obtain the distribution of the amplitude of the signal in time - frequency coordinate system. Many kinds of time-frequency analysis methods, mainly divided into two kinds of linear and bilinear, the linear methods include short-time Fourier transform, wavelet transform, Hilbert-Huang transform, fractional Fourier transform; bilinear method includes Cohen bilinear time-frequency distribution and Affine bilinear time-frequency distribution. However, if the time-frequency analysis method is applied to a single array acoustic logging signal processing, the effect is not ideal. There are two main reasons: first, the array acoustic logging signal is often divided by a variety of waves, the wave between time and frequency is. Near, sometimes it is difficult to distinguish between the various points; second, wave amplitude varies greatly, which makes the time-frequency distribution is difficult to be the low amplitude wave show. Therefore, we consider the joint fractional Fourier transform and Choi-Williams distribution, the rotation characteristics of fractional Fourier transform of array acoustic logging signal Choi-Williams distribution when the frequency domain filtering, were extracted from the time-frequency distribution of compressional wave, shear wave, study of time-frequency distribution, combined with the relatively easy to distinguish the original Choi-Williams in the distribution of Stone wave wave, time-frequency characteristics of shear wave and Stone wave, to explore a new method of array acoustic logging data processing and interpretation of this work. Mainly includes three parts: first, introduce the method of fractional Fourier transform and Choi-Williams distribution, and the procedures for the preparation of; second, through the corresponding processing, access array The frequency characteristics of acoustic logging signal; third, relationship between the frequency characteristics and the formation of cracks of array acoustic logging signal. Methods are briefly described below content of each part of the.1. the introduction of fractional Fourier transform and Choi-Williams distribution, and then discuss its physical meaning and practical significance. The programming work is divided into two parts: wells period of treatment and single point processing. Wells is treated with the aim of time-frequency distribution to calculate a depth range and even the whole well all the array acoustic logging signal (which can be used to obtain the Stone wave characteristics), and then extract the time-frequency distribution of P-wave and S-wave, and in accordance with the depth order and the original wave corresponding arrangement, macroscopic law in order to obtain the time-frequency characteristics of each component wave change with depth respectively. For wells processing part, using mixed programming of VB.NET and Matlab way of object-oriented Single point processing program design. The purpose is through the production by full wave acoustic logging curve, the original signal time-frequency distribution, the time-frequency distribution of P-wave and S-wave time-frequency distribution of 4 parts of the image, to acquire more accurate time-frequency information depth representative of some strata at each component for wave. Single point processing part, this paper uses the Matlab to program to write, and then use GUI design tools of Matlab, making the visual interface of.2. selected from different regions, different wells fractured formation of array acoustic logging data, the data decoding, then to gain and equalization, and then select a depth range of the data processing section, the original logging signals were obtained, the longitudinal wave and transverse wave time-frequency distribution with depth image. Select some representative signals from single point The depth of processing, the original point of logging signal, time-frequency distribution of longitudinal wave and transverse wave, then the components of wave time-frequency localization, and determine the magnitude of the.3. by wells processing results, explore the components of wave time-frequency characteristics with certain formation changes in the nature of law; then using the single point processing results, a clearer understanding of these laws. Then combined with other logging and geological data, for the interpretation of these rules and to learn from the perspective of physics, the relationship between frequency characteristics and different types of fractured formation. The establishment of array acoustic logging signal after the above research work, this paper obtained the following conclusions: 1. in array acoustic logging the Choi-Williams signal distribution, only Stone wave is obvious, and the longitudinal and transverse waves are not easy to identify. The method of the joint distribution of the fractional Fourier transform and Choi-Williams will be two To the process of extraction. The array acoustic logging data processing and analysis of fractured formation by the method of obtaining good results. Therefore, this method has certain practical significance for dense.2. formation, in time-frequency distribution, Choi-Williams distribution map of the original signal array acoustic logging are regular. Stoneley wave amplitude is much higher than that of P-wave and S-wave. With low relative angle of open fracture and fractured reticular formation, components of wave amplitude attenuation is not obvious, the peak time is earlier, the P-wave and S-wave frequency were higher for.3. fractured strata semi filling, in time-frequency distribution, each wave crest time and frequency does not change obviously, while the amplitude attenuation between the fracture formation between dense formation and low angle.4. for low angle fractured formation, compared with Dense formation, in time-frequency distribution, the relative delay time and frequency of the wave crest, lower amplitude attenuation is more serious; the wave peak time delay, frequency is relatively low, the attenuation is also becoming more serious; the relative delay time Stone wave wave, the amplitude attenuation is very obvious, but the magnitude is still significantly higher than the P-wave and S-wave.5. for high angle fractured formation, with dense formation, in the time-frequency distribution of P-wave amplitude slightly some attenuation of wave amplitude changes little, amplitude of the certain attenuation of Stoneley wave, and the components of time and frequency did not change significantly for.6. mesh cracks the formation, relative to the density of the formation in time-frequency distribution, the relative delay time and frequency of the wave crest, lower amplitude attenuation more serious; shear wave Time delay, the frequency is relatively low, the attenuation is more serious; relatively delayed Stone wave peak time, amplitude attenuation is more serious. Effects of reticular cracks on the frequency characteristics of array acoustic logging signal is complex, unable to accurately explore the law, but with the compactness of the stratigraphic easier.7. for certain conventional logging curve crack not obviously reflects the characteristics of frequency method can through this acquisition of array acoustic logging signal, and then to identify the crack.8. the rules in turn use, make the time-frequency distribution of each section of the original signal and the P-wave and S-wave through the method of this paper introduces the components, if the peak time earlier. The amplitude is higher, and the longitudinal and transverse wave high frequency, it can be judged as dense formation, and as a reference; Ruositong and amplitude of the relative is certain Fall is indicating crack formation, and the crack is not completely filled; and Ruositong amplitude attenuation, wave amplitude at the same time, there are also some attenuation, peak time and frequency and wave amplitude change of wave components is determined with high dip fracture; Ruositong and amplitude of the severe attenuation at the same time the peak time, has also been postponed, delayed peak time and the relative wave frequency, amplitude attenuation is relatively low, relatively serious, postponed the peak time of shear wave, the frequency is relatively low, the attenuation of serious cracks is determined for low dip angle or reticular cracks.

【学位授予单位】：吉林大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：P631.81

【相似文献】