Tabaka Gözenekliliği Ve Akışkan Cinsinin Sismik Yansıma Genliklerinin Açı İle Değişimine Etkisi

Abstract

Sismik yansıma genliklerinin ofsete veya geliş açısına (Amplitude Versus Offset, AVO; Amplitude Versus Angle, AVA) bağlı değişiminin incelenmesi yeraltı hidrokarbon aramalarında kullanılan en temel yöntemlerden biridir. AVO ve AVA verilerinin analizi, düz ve/veya ters çözüm yaklaşımları ile belirlenen genlik anomalileri bölgeye ait kuyu bilgileri ile ilişkilendirilerilerek petrol ve doğalgaz alanlarının bulunmasında önemli katkılar sağlar. AVO/AVA verilerinin düz çözüm yaklaşımı ile olası hidrokarbon kapanı içinde ne tür bir akışkan olduğu ve bu akışkanın miktarı hakkında kestirimlerde bulunulabilir. Elde edilen düz çözüm sonuçları ile saha verileri karşılaştırılarak yeraltı hakkında daha detaylı yorumlamalar yapılabilir. Bu tez çalışmasında AVA verileri ile düz çözüm modellemesi gerçekleştirilmiştir. Bu amaçla petrol ve doğalgaz kapanı oluşturacak şekilde üstte geçirimsiz şeyl tabakası altta da gözenekli çökel kayaçlardan oluşan hazne kayaç olduğu varsayılmıştır. Oluşturulan üç adet kapan modelinin tümünde üstteki sıfır gözenekliliğe sahip şeyl tabakası sırasıyla kireçtaşı, kumtaşı ve killi kumtaşı tabakası ile yatay bir arayüzey oluşturmuştur. Yatay bir arayüzeye sıfır ve daha büyük açılarla P dalgası gelmesi durumunda arayüzeyde gerçekleşen enerji paylaşımını açıklayan Zoeppritz Denklemleri AVA düz çözümünde matematik model olarak kullanılmıştır. Oluşturulacak düz çözüm modelleri hakkında daha doğru yorum yapabilmek için şeyl-kireçtaşı modeli kullanılarak tabakalara ait yoğunluk, P ve S dalga hızlarının yansıma genlikleri üzerindeki etkileri ayrı ayrı incelenmiştir. Düz çözüm modellemesinde kullanılan yoğunluk ve sismik hızlar kayaç numunelerine laboratuvar ortamında yapılan deneylerle saptanan gerçek değerlerden oluşmaktadır ve tez için yapılan literatür araştırması ile belirlenmiştir. Her modele ait atanan gözeneklilik oranı ve ilişkili akışkan cinsine göre yeni kayaç hızı ve yoğunluk değerleri Gassmann denklemleri kullanılarak hesaplanmıştır. Zeoppritz denklemlerinin matematik model olarak kullanıldığı bu düz çözüm yaklaşımında seçilen üç kapan modelinde şeyl örtü kayacı ve hazne kayaç arasındaki arayüzeyde yansıyan P dalga genlikleri geliş açısının fonksiyonu olarak hesaplanmış ve grafiklenmiştir. Modellerde arayüzeyin altındaki hazne kayaca %10, %20 ve %30 seviyelerinde olmak üzere farklı gözeneklilik oranları atanmıştır ve hazne kayacın bu gözeneklilik oranlarında doğalgaz, petrol ve tuzlu suya doygun olmalarına göre yansıma genliklerinin açı ile değişimleri incelenmiştir. Bu tez çalışmasında AVA verisi oluşturmak için Zoeppritz denklemleri kullanılmasının yanısıra 2-boyutlu dalga yayınım yazılımı kullanılarak da yapay sismik veriler oluşturulmuş ve Zoeppritz katsayıları ile karşılaştırmalar yapılmıştır. Modelleme sonuçlarından 2 ve 3 boyutlu yansıma genliği grafikleri oluşturularak hazne kayaç gözenekliliği ve akışkan türünün sismik genlikler üzerindeki etkileri incelenmiştir. Oluşturulan AVA eğrilerinde yansıma genliklerinin gözenekliliğin artmasıyla birlikte tüm açılarda azaldığı görülmektedir. Her modelde gözeneklilik oranlarındaki değişimin yansıma eğrilerinde belirgin farklar meydana getirebildiği görülmüştür. Modellemeler sonucunda oluşturulan eğrilerin açı ile değişimini belirlemede kayaç türü ve gözeneklilik oranının, gözenek akışkanına kıyasla daha etkin olduğu görülmüştür. Gözeneklilik oranının artması P dalgası hızlarında ve yoğunluklarda belirgin bir düşüşe neden olarak yansıma genliklerinin azalmasındaki temel nedendir. Örtü ve hazne kayacın sismik dalga hız ve yoğunluk değerleri arasındaki farklarının az olması durumunda yansıma genlikleri çok küçük değerler alabilir ve bu yapılar hidrokarbon içermesine rağmen değerlendirme aşamasında gözden kaçırılabilirler. Kayaç gözeneklerindeki akışkanların türlerinin değişmesi ise sadece yoğunlukta ve S dalgası hızlarında küçük bir değişime neden olabilmiştir. Kayaç gözeneğindeki akışkanların değişmesi P dalgası hızlarında aynı oranda değişime neden olmamıştır.

Amplitude Versus Offset (AVO) or Amplitude Versus Angle (AVA) approaches are the main methods used in hydrocarbon exploration. The methods study increase or decrease of the reflection amplitudes over offset ranges or angles to predict lithological information of the rocks. The theory of these methods was first studied by German scientist Zoeppritz. Although there were other studies about offset or angle dependent reflectivity, the considerable applications in hydrocarbon exploration have mainly been performed since 1980s. Researches on gas sands show that angle dependent reflectivity could be used to identify hydrocarbons. The results of these studies showed that pre-stack data could be used for interpretation as well as post-stack data. Nowadays the method is used also for understanding the microfracture orientation or stress measuring in shale plays. But the usage of this method on field seismic data itself is not sufficient for interpretation. Forward AVO/AVA modelling approaches of the traps can be encountered in hydrocarbon exploration as well as interpretation of the actual seismic data acquired in the field. Well logs or core samples obtained from other studies or for a specific type of formation can be used to have a beter-constrained model data. AVA or AVO analysis studies indicate that exploration studies always include forward modelling using well logs. This forward model can give us insight about the parameters that cannot be estimated from field data like attenuation or porosity. Also it could be used to identify the fluid type of the reservoir. The parameters required for forward AVO/AVA modeling are P wave velocity, S wave velocity and density. These are the basic physical parameters that effect seismic properties. P and S wave velocity also depend on other elastic parameters such as bulk and shear moduli and also density. Bulk and shear moduli represent the compressibility and rigidity of a rock respectively. P and S wave velocities are directly proportional to these parameters whereas they are indirectly proportional to the density. Therefore using these parameters is a better way to forward modeling. These parameters require the usage of Gassman equations to obtain P and S wave velocities of a saturated rock. Forward modelling for AVA analysis can be done in two seperate ways. First method for forward modelling is an approach by using Zoeppritz equations that give a quick calculation of what the reflection coefficients are. But modelling by Zoeppritz equations applications can only be done for two half-space layers on a horizontal interface. Moreover, Zoeppritz equations do the calculation on the assumption of plane waves whereas the real data acquired in the field is the result of a spherical wave generated by the source. In this thesis matrix form of the Zoeppritz equations are used for modelling. The second method is wavefield-modelling programs. These programs create seismic records that can be seen on a real dataset. Besides, unlike the Zoeppritz equations modelling can be done with various amounts of layers and interfaces. Thus, a wavefield modelling program can sample subsurface as close to as it can be. However, wavefield modelling requires substantial amount of computational power. Beacuase the modelling time of a model increases with the amount of detail defined in the model. The stability of the program is also an issue for wavefield modelling because it uses numerical methods for calculation. For wavefield modelling fledmodc program is used in this thesis. In this thesis, amplitude versus angle gathers were created by using the matrix form of the Zoeppritz equations and analysed for possible trap models in hydrocarbon exploration. In this manner, shale – limestone, shale – sandstone and shale – shaly sandstone (sandstone with clay minerals) interfaces were used for modelling. Each interface is modelled with assumption of reservoir rock has porosities of 10%, 20% and 30%. Seismic wave velocities and density of the reservoir rock (limestone) in shale-limestone model are changed one at a time to inspect their effects on reflection amplitudes. These physical parameteras are increased and decreased by 10% and 20% to have wide range of values. The resulting reflection amplitudes versus incidence angle graphs are used to interpret the models having different porosity ratios. Classifications of AVA or AVO anomalies give a certain insight to the interpretation of the reflection data. There are four types of anomalies for interpretaion. Class-1 anomalies indicate a well-cemented reservoir rock which starts with strong positive amplitudes and decreases as the incidence angle increases. Amplitudes can change polarity if there is sufficient amount of offset. Class-2 anomalies indicate a reservoir rock that is very similar in physical properties to the overlying seal. This class of anomalies has 2 types of AVA curves. The first type starts with a very small positive amplitude and changes polarity in medium angles and then increases in the negative direction. The second type of anomaly starts with very small negative amplitude and increases in the negative direction with increasing incidence angles. Class-3 anomalies are the most desired anomalies because of their distinctive nature. This class of anomalies starts with high negative amplitudes and increases as the incedence angle increases. After classification of these three classes of anomalies, more studies that are presented Class-4 as anomalies. Class-4 anomalies start with high negative amplitudes like Class-3 but in this case, as the incidence angle increases, the amplitudes decrease. These classes help the interpreter to isolate certain types of anomalies. To implement different porosity ratios on the velocity values of the rocks, empirical relations for limestone, sandstone and shaly sandstone are used as well as critical porosity equations. Because of the variety of fluids can be found in hydrocarbon traps; gas, oil and brine are used to model the AVA anomalies. The calculations using different porosity values and different fluid types of reservoir rocks have been implemented using Gassmann equations. The rock types that are chosen from core samples have a certain percentage of porosity. To model AVA anomalies at 10%, 20% and 30% porosity of the reservoir rock, empirical relations and Gassman equations are used in combination. First the bulk and shear moduli of the dry rock (with porosity values from the core sample) is calculated using the velocities and densities of the rocks. After that, using empirical relations and critical porosity formula bulk and shear moduli of the rock in mineral form with no porosity is obtained. These moduli of the mineral are used to form rocks with the porosities of 10%, 20% and 30%. To perform mentioned above, the bulk and shear modulei of the dry rock (porosity 10%, 20% or 30%) are obtained by using the empirical relation and the critical porosity formula. Finally, the Gassman equations are used to calculate the bulk and shear moduli of the fluid (gas, oil, brine) saturated rock for all the porosity values. These moduli are then used to calculate the velocities that are required for forward modelling. For a better interpretation of modelling results, P and S wave velocity versus porosity of the reservoir rock and; density versus porosity of the reservoir rock graphs are plotted. Moreover, three dimensional graphs that show the reflection coefficients versus incident angle and porosity are prepared in MATLAB. Thus, a general idea about the effects of porosity on reflection amplitudes is obtained. Wavefield modeling results are only given as reflection amplitude versus angle graphs. To have an amplitude versus angle graph in wavefield modeling, multiple shots in a line have been performed. Some processing steps are required on these shots so that the amplitudes of the reflecting interface can be read easily. Indeed, processing for AVA analysis must be done carefully because amplitudes can be effected in a way that can give false anolmalies. The first step in proccessing procedure is defining geometry of the line for source, receiver and common midpoint locations. Also sphreical divergence correction is done in common shot gathers to compensate for the energy decrease at far offsets. After that, the shot records are sorted into Common Midpoint Gathers (CMP). Thus CMP gathers are formed as including a range of offsets (or incidence angles) and amplitudes. Finally the common midpoint gather is corrected for Normal Move Out (NMO) because the far offsets arrive later in time than near offsets. After NMO correction, the traces effected from the surface waves are removed from the CMP gather. Normally unwanted surface waves can be removed by f-k filtering and reflections concealed in these traces can be obtained. Unfortunately this approach smears all of the amplitudes in the seismic record this processing step was not applied. NMO correction results in the aligning of the reflections from the same point and allows for an easier reading of the amplitudes of the anomaly. The AVA data produced from the Zoeppritz equations and wavefield modelling are different in scale. Zoeppritz equations give a coefficient of the layers at various angles whereas the modelling programs gives a amplitude value for the angles. To be able to compare the results, amplitudes are normalized by the first reflection in their own data so that both modelling results represent the change of amplitudes with relation to the first reflection amplitude. Increasing porosity in all the models reduces the amplitudes from positive values to negative values. In shale-limestone model despite the limited range of incidence angles, the effects of critical angle can be seen around 40 degrees. In every trap model increasing porosity ratios change the AVA curve significantly and thus the AVA class type can change especially in shale–sandstone trap model. Lithology and porosity are the main factors effecting on the angle dependent reflectivity in these models. In certain conditions, seismic wave velocities and densities of the seal and the reservoir rock can be similar and reflections can become so weak that even there is hydrocarbon in the trap, the results could be overlooked. However the reservoir rock sometimes can be well-cemented and its physcical parameters can be much higher than the seal. The amplitudes from this trap model will be stong and may increase with higher incidence angles, but it will include the effects of critical angle. P wave velocity and density of the saturated rocks are the biggest factors for effecting the amplitudes. Change in the rock fluid type is less effective for the AVA anomalies in these models. Rock fluids mainly effect the density of the rock which occurs as a big decrease from brine saturated models to gas saturated model. In contrast, S wave velocities are increased by a small amount because of the change from brine to gas. But the fluid change is not as effective on the P wave velocities in these models as they are on S wave velocities and densities.