##
Sıfır açılımlı tek kanal deniz sismiği verilerinin modellenmesi

Sıfır açılımlı tek kanal deniz sismiği verilerinin modellenmesi

##### Dosyalar

##### Tarih

1993

##### Yazarlar

Barghi, Alireza

##### Süreli Yayın başlığı

##### Süreli Yayın ISSN

##### Cilt Başlığı

##### Yayınevi

Fen Bilimleri Enstitüsü

##### Özet

Deniz sismiğinde elde edilen sismogramların yorumlarında çeşitli problemlerle karşılaşılır. Bunlar arasında tekrarlı yansımalar, hayalet yansımalar, saçılmalar ve benzeri problemler sayılabilir. Bu problemlere bir çözüm bulmak ve sismogramların yorumlamasında bazı hataları gidermek amacıyla bu çalışma gerçekleştirilmiştir. Bu çalışmada klasik ışın izleme yöntemi kullanılarak yapay sismik veri üretilmiştir. Bunun için Snell yasası ve trigonometrik bağıntılardan yararlanılarak gerekli denklemler çıkarılmış ve bilgisayar ortamında yapay sismik veriler üretilmiştir. Yapılan çalışmada yapay sismik veri üreten bilgisayar programlan yazılmış, daha sonra çeşitli modeller ele alarak kullanılan yöntemin ve çıkarılan bağıntıların çalışıp çalışmadığı denenmiştir.

The production of synthetic seismograms is one of the subjects studied by scientists and researchers, and recently, widely spoken about in geophysics. Applied geophysics and seismology studies the production of synthetic seismograms and their comparison with realistic seismograms, have put a light on several geophysical and seismological problems. Standard inversion methods are available only for some relatively simple models of the Earth's interior. Generally, the properties of the Earth's interior may deviate more or less from the basic assumptions inherent to the different inversion methods. To evaluate the accuracy of the results it is therefore necessary to compute the travel times for the model obtained and to compare them with the observed travel times. If appropriate methods are available for the computation of exact traveltimes for complex and realistic models, it is then possible to improve first order models iteratively by comparison öf computed and observed traveltimes. In a horizontally stratified medium, the ray tracing is performed by applying Snell's low. The physical basis of raytracing and traveltime calculations for arbitrary two-dimensional model is the same as in the case of a horizontally stratified medium. What is needed is the Snell's low. ix If a mathematical model is defined which attributes a velocity value to any point of the subsoil, any ray path can be computed stepwise, starting at the source of the seismic wave with corresponding angle of incidence and successively applying Snell's law. The computation is repeated and the traveltime increments are summed up, until the ray reaches the Earth's surface or any other boundary of the model. The problem is straight forward when the starting point and the initial direction are given. It becomes complicated for cases in which the starting point and the ending point are given. There are basically two approaches to this problem. The first one is shooting method, which uses trial values for the initial direction and adjust them to make the calculated raypaths go as near as to a given ending point. The second one is bending method, which uses trial ray paths with the right starting and ending point and then adjust the paths to satisfy the differential equation for the ray trajectory. The ray method is applicable only when the velocity distribution in the model and interfaces in the vicinity of the ray under consideration are sufficiently smooth. The ray method has some other limitations. In smooth media, it has a limited accuracy, or even not valid in the singular regions of the ray field. Those are caustic regions, critical regions (in the neighborhood of a critical angle for reflections), transition region between the shadow and the illuminated region. In its standard form, it cannot be used to investigate various diffracted waves, inhomogeneous waves, etc. Various methods in synthetic seismogram production are widely used by seismologists, geophysicists, and scientists. In the production of seismograms various ray tracing methods are used for various aims. Various techniques in the production of synthetic seismogram that have been developed recently are the classical ray tracing method, WKBJ seismograms (Chapmen and Drummond, 1982), Gaussian Beam method (Cerveny and others, 1982), Kirchoff integral method (Haddon and Buchon, 1981). Each one of the above mentioned methods have their advantageous and disadvantageous. Using ray tracing methods, wave propagation duration calculations have been secured in non-homogenous regions. Gebrande (1977), Cerveny and others (1977), Julian and Gubbis (1977), Whittall and Clowes (1979) made important contributions to this subject. As a result of those studies, it was understood that the traveltime of seismic waves is as important as their amplitudes. A single effective ray-tracing method was developed by Whittal and Clowes (1979) for regions of horizon variations. In this paper, the velocity model was expressed as large blocks in random boundaries. The velocity gradient was constant in each block; however, it was taken in random directions. The sufficiency and applicability of this method was proved on several refraction problems by Clowes and others (1981), Delando and Moon (1982), Ellis and others (1983), Green and others (1983), and Horn and others (1983). The terms in the series obtained from the asymptotic solution of elastodynamic wave equation provided the foundation of the ray tracing method. Whereas for the geometric raytracing method, the optimum solution obtained under monochromatic wave, the values of the first term in the ray series are considered (Cerveny and Ravindra, 1971; Aki and Richards, 1980). The applicability of the geometric ray theories are in accordance with the variations in wave front correlations and velocity gradients (Cerveny and others 1977). In this study, using the classical ray tracing techniques, synthetic seismograms are produced. There are two advantages of this method: the first advantage is the execution speed in computer environment, and the second XX advantage is the independent handling of each seismic event once at a time. The necessary equations have been derived using Snell's law and trigonometric relations. This thesis is to serve to seismic interpreters, especially for those who evaluate zero-offset, single-channel, high resolution analog seismic data. The following are some of the typical interpretation pitfalls the interpreter faces: ? Misinterpreted intrabed multiples, reverberations, and ghosts as possible layers, ? Mislocated sea bottom bathymetry and geological structures due to the migration of the tilted reflectors observed in the offset-time seismic sections, ? Wrongly delineated seismic horizons due velocity pull-ups or pull-downs, and erroneous layer thickness computations caused by the lack off correct velocity knowledge. For this purpose, three separate computer programs were written. A brief description of these programs are: ? The Scaling Program: This mouse controlled interactive program helps the interpreter to supply a scale so that the created synthetic seismogram fits the field data when they are superposed, ? The Geological Model Producer: This mouse controlled interactive program creates or modifies geological models in a fast and easy manner. ? The Synthetic Seismic Data Generator: This interactive program generates the seismic section based on a geological model produced in the previous step. xn The interpreters goal would be the optimal fit of the synthetic data with the real data. He/She will achieve this in trial basis; i.e., he/she will iteratively choose appropriate layer boundary and velocity values until an optimal match between real data and synthetical data in reached. The contribution of the present Thesis would be the development of a Seismic Modelling Software to assist the interpreter. However, when the acquired data is analog, and the digitization of recorded seismograms is practically impossible. Therefore, the conventional Seismic Data Processing techniques cannot be used. Another alternative would be the Seismic Modelling in computer environment. Such a modelling package should be easy to use, fast, and can be run in personal computers. But, at least a 486DX based microprocessor and 4 Mb memory are required. To speed up the computations, 2-D ray tracing technique (exploding reflectors) is used. The Seismic Modelling Software, besides generating the primary reflections, will also produce source/receiver-site ghosts, water-bottom multiples, intra-bed multiples, and diffractions from sharp corners (like faults). As a result, both sea bottom bathymetry and geological structure beneath the sea bottom may be mapped, and their velocities can be estimated. Misinterpretations of ghosts and multiples as seismic horizons can also be avoided.

The production of synthetic seismograms is one of the subjects studied by scientists and researchers, and recently, widely spoken about in geophysics. Applied geophysics and seismology studies the production of synthetic seismograms and their comparison with realistic seismograms, have put a light on several geophysical and seismological problems. Standard inversion methods are available only for some relatively simple models of the Earth's interior. Generally, the properties of the Earth's interior may deviate more or less from the basic assumptions inherent to the different inversion methods. To evaluate the accuracy of the results it is therefore necessary to compute the travel times for the model obtained and to compare them with the observed travel times. If appropriate methods are available for the computation of exact traveltimes for complex and realistic models, it is then possible to improve first order models iteratively by comparison öf computed and observed traveltimes. In a horizontally stratified medium, the ray tracing is performed by applying Snell's low. The physical basis of raytracing and traveltime calculations for arbitrary two-dimensional model is the same as in the case of a horizontally stratified medium. What is needed is the Snell's low. ix If a mathematical model is defined which attributes a velocity value to any point of the subsoil, any ray path can be computed stepwise, starting at the source of the seismic wave with corresponding angle of incidence and successively applying Snell's law. The computation is repeated and the traveltime increments are summed up, until the ray reaches the Earth's surface or any other boundary of the model. The problem is straight forward when the starting point and the initial direction are given. It becomes complicated for cases in which the starting point and the ending point are given. There are basically two approaches to this problem. The first one is shooting method, which uses trial values for the initial direction and adjust them to make the calculated raypaths go as near as to a given ending point. The second one is bending method, which uses trial ray paths with the right starting and ending point and then adjust the paths to satisfy the differential equation for the ray trajectory. The ray method is applicable only when the velocity distribution in the model and interfaces in the vicinity of the ray under consideration are sufficiently smooth. The ray method has some other limitations. In smooth media, it has a limited accuracy, or even not valid in the singular regions of the ray field. Those are caustic regions, critical regions (in the neighborhood of a critical angle for reflections), transition region between the shadow and the illuminated region. In its standard form, it cannot be used to investigate various diffracted waves, inhomogeneous waves, etc. Various methods in synthetic seismogram production are widely used by seismologists, geophysicists, and scientists. In the production of seismograms various ray tracing methods are used for various aims. Various techniques in the production of synthetic seismogram that have been developed recently are the classical ray tracing method, WKBJ seismograms (Chapmen and Drummond, 1982), Gaussian Beam method (Cerveny and others, 1982), Kirchoff integral method (Haddon and Buchon, 1981). Each one of the above mentioned methods have their advantageous and disadvantageous. Using ray tracing methods, wave propagation duration calculations have been secured in non-homogenous regions. Gebrande (1977), Cerveny and others (1977), Julian and Gubbis (1977), Whittall and Clowes (1979) made important contributions to this subject. As a result of those studies, it was understood that the traveltime of seismic waves is as important as their amplitudes. A single effective ray-tracing method was developed by Whittal and Clowes (1979) for regions of horizon variations. In this paper, the velocity model was expressed as large blocks in random boundaries. The velocity gradient was constant in each block; however, it was taken in random directions. The sufficiency and applicability of this method was proved on several refraction problems by Clowes and others (1981), Delando and Moon (1982), Ellis and others (1983), Green and others (1983), and Horn and others (1983). The terms in the series obtained from the asymptotic solution of elastodynamic wave equation provided the foundation of the ray tracing method. Whereas for the geometric raytracing method, the optimum solution obtained under monochromatic wave, the values of the first term in the ray series are considered (Cerveny and Ravindra, 1971; Aki and Richards, 1980). The applicability of the geometric ray theories are in accordance with the variations in wave front correlations and velocity gradients (Cerveny and others 1977). In this study, using the classical ray tracing techniques, synthetic seismograms are produced. There are two advantages of this method: the first advantage is the execution speed in computer environment, and the second XX advantage is the independent handling of each seismic event once at a time. The necessary equations have been derived using Snell's law and trigonometric relations. This thesis is to serve to seismic interpreters, especially for those who evaluate zero-offset, single-channel, high resolution analog seismic data. The following are some of the typical interpretation pitfalls the interpreter faces: ? Misinterpreted intrabed multiples, reverberations, and ghosts as possible layers, ? Mislocated sea bottom bathymetry and geological structures due to the migration of the tilted reflectors observed in the offset-time seismic sections, ? Wrongly delineated seismic horizons due velocity pull-ups or pull-downs, and erroneous layer thickness computations caused by the lack off correct velocity knowledge. For this purpose, three separate computer programs were written. A brief description of these programs are: ? The Scaling Program: This mouse controlled interactive program helps the interpreter to supply a scale so that the created synthetic seismogram fits the field data when they are superposed, ? The Geological Model Producer: This mouse controlled interactive program creates or modifies geological models in a fast and easy manner. ? The Synthetic Seismic Data Generator: This interactive program generates the seismic section based on a geological model produced in the previous step. xn The interpreters goal would be the optimal fit of the synthetic data with the real data. He/She will achieve this in trial basis; i.e., he/she will iteratively choose appropriate layer boundary and velocity values until an optimal match between real data and synthetical data in reached. The contribution of the present Thesis would be the development of a Seismic Modelling Software to assist the interpreter. However, when the acquired data is analog, and the digitization of recorded seismograms is practically impossible. Therefore, the conventional Seismic Data Processing techniques cannot be used. Another alternative would be the Seismic Modelling in computer environment. Such a modelling package should be easy to use, fast, and can be run in personal computers. But, at least a 486DX based microprocessor and 4 Mb memory are required. To speed up the computations, 2-D ray tracing technique (exploding reflectors) is used. The Seismic Modelling Software, besides generating the primary reflections, will also produce source/receiver-site ghosts, water-bottom multiples, intra-bed multiples, and diffractions from sharp corners (like faults). As a result, both sea bottom bathymetry and geological structure beneath the sea bottom may be mapped, and their velocities can be estimated. Misinterpretations of ghosts and multiples as seismic horizons can also be avoided.

##### Açıklama

Tez (Yüksek Lisans) -- İstanbul Teknik Üniversitesi, Fen Bilimleri Enstitüsü, 1993

##### Anahtar kelimeler

Jeofizik Mühendisliği,
Deniz sismiği,
Veri modelleri,
Geophysics Engineering,
Sea seismic,
Data models