ETD system

Electronic theses and dissertations repository


Tesi etd-08302016-115207

Thesis type
Tesi di laurea magistrale
Algorithmic Seismic predictions of overpressure in AVA in fluid line
Corso di studi
relatore Prof. Mazzotti, Alfredo
Parole chiave
  • Overpressure
Data inizio appello
Riassunto analitico
The subject of detection of abnormally high pressured zones from seismic has received a great deal of attention in exploration and production geophysics because of increasing activities in frontier areas or offshore and a need to lower cost without compromising safety and environment, and managing risk and uncertainty associated with very expensive drilling. An estimation of pore pressure can be obtained from seismic velocity as well as from well logs. Pore pressure is one of the most important parameters for drilling plan and for geomechanical and geological analyses. If the pore pressure is higher than the hydrostatic pressure (normal pore pressure), it is abnormal pore pressure. When pore pressure exceeds the normal pressure, it is overpressure. In the present study, amplitude variation with angle (AVA), will be evaluated. AVA analysis can provide important information about reservoir rocks such as lithology, porosity, and pore fluids that can be used to reduce hydrocarbon exploration risk. Seismic amplitudes reflecting from an interface change when the angle increases between the source and receveir points at the surface. AVA analysis is normally carried out in a deterministic way to predict lithology and fluids from seismic data. Amplitude variation with angles (AVA) interpretations may be facilitated by crossplotting the AVA Intercept (A) and Gradient (B). In the present study an algorithm has been developed, with the goal to discover wheter it was possible to predict overpressure phenomenons from the seismic data. It considers and compares an AVA Analytical and Experimental response. In the AVA Experimental response, well logs of the seismic velocities and density have been exploited to derive the synthetic seismograms (CMP) using the convolutional method, and the Intercept and Gradient values extracted from synthetic seismograms. In the AVA analytical response well logs of the seismic velocities have been used, and the Intercept and Gradient values extracted from the velocities (Vp, Vs) and density well logs. The strategy followed to tackle the problem has been Shuey’s approximation. Shuey’s approximation was used for AVA crossplot analysis. The primary attributes extracted for AVA analysis are Intercept (A) and the Gradient (B), which are obtained from velocity corrected CMP records. This Gradient (“slope”) AVA attribute is calculated from a least square regression analysis of the amplitudes for angles from 0-30 degrees, using an X-axis of sine squared theta (where theta is the incidence angle), and the Intercept is zero-offset amplitude determined (using an Y-axis the amplitude) by extrapolating the AVA Gradient. This yields two AVA attributes, basically the Slope and Intercept of a straight line, which describes, in simpler terms how the amplitude behaves with angle of incidence. Each point in the AVA crossplot is mapped using the amplitude of Intercept (A) and Gradient (B) in a time window. The extraction provides band-limited information on which attempt to discover anomalies caused by overpressure phenomens. AVA crossplotting can play a significant role in minimizing the risk associated with an exploration play. The stability of the Background Trend (Vp/Vs) can have an impact on what is being interpretated as anomalous, be it fluid or lithology induced outliers. The third and fourth chapters deal with the direct problem. We shall start with an initial model in which the Vp/Vs ratio is known, after which I shall apply Shuey’s approximation, and through the methodology of least squares and the “ Singular Value Decomposition” method, I obtain two values of Vp/Vs to predict two empirical equations, Costant Density and Gardner Density. Several examples were evaluated. The first example, it started from a Vp/Vs model that varied linearly. The second example added a bit of random noise to the initial model (Vp/Vs). The third example is considered to show the correct relationship of Vp/Vs. The fourth example considered a ratio Vp/Vs that varied linearly from 4 to 2 and then returned again to a ratio of Vp/Vs = 4, with the addition of random noise. As previously explained, it is considered an AVA Analytical and Experimental response. Using an Analytical AVA response, the Vp/Vs ratio that was predicted was very close to the Vp/Vs ratio of the initial model. Using an Experimental AVA response, the predicted Vp/Vs ratio deflected from the Vp/Vs ratio of the initial model, especially when we consider the Gardner Density equation. This present study, attempts to define a low resolution profile of Vp/Vs ratio and not its local variations of high frequency. It was necessary to study many strategies which make the method robust and reliable. The present study is partitioned in four chapters. The first chapter attempts to analyze the problems of overpressure from the theoretical and practical points of view. The second chapter focuses on the use of AVA seismic attributes, and how that seismic attribute can provide information about possible presence of overpressure from seismic data. The third and fourth chapters illustrate the algorithm developed in Matlab and provide information about the possible presence of overpressure.