Imaging solutions for 4D quantitative interpretation

Abstract

During the production of a geomechanically active reservoir, massive pressure depletion happens giving rise to geomechanical changes which can lead to significant time-lapse signals across the reservoir and its surrounding. Therefore, geomechanical characterisation of the reservoir and monitoring are very important for this type of reservoir. In this thesis, I use pre-stack time-lapse time-shifts observed between 4D seismic surveys for the geomechanical characterisation of the Ekofisk field which is a geomechanically active field in the North Sea. This thesis consists of three parts. Before using pre-stack time-shifts, post-stack time shifts can be a valuable guide toward the geomechanical activities of the reservoir. In the first part, I estimate the post-stack time-shifts using various methods. Then, I evaluate the advantages and disadvantages of each method in terms of their performance in revealing the local time-lapse signals such as time-strains. I have found that all the time-shift methods can successfully measure time-shifts. Among them, NLI is the most outstanding method as it gives smooth time-shifts with relatively good accuracy and the time-strains derived from there are more stable and interpretable. In the second part, I use the reflectivity and velocity models of the Ekofisk field and perform a finite-difference simulation to generate synthetic seismic data, followed by imaging the generated data. Migrating baseline and monitor datasets with baseline velocity model caused considerable mispositioning in the overburden resulting in false amplitude-differences in the overburden. The analysis of the images shows that it is not simply a matter of mispositioning that contaminates the seismic images. A more serious problem caused by migration with an erroneous velocity model is the defocusing of amplitudes. This problem cannot be solved by warping and requires a more sophisticated remedy to correct the monitor’s migration velocity model. In the third part, which is the major development of this thesis I measure the pre-stack time-shifts and design a tomographic approach to utilise them for estimating the time lapse changes. First, I show how to measure the pre-stack time-shifts and discuss the practical aspects of the process. Second, I design a ray-based tomography customised for 4D application in order to utilise the pre-stack time-shifts and invert for velocity changes that cause the time-shifts. Finally, I extend the tomography method into an anisotropic inversion where both the time-lapse velocity changes and the ratio of lateral-to-vertical strains are inverted in a two-step inversion process. The two products of the inversion can be used extensively in the geomechanical model calibration of the reservoirs. Overall, my PhD research has successfully measured the time-lapse velocity changes and the ratio of lateral-to-vertical strains. The anisotropic time-lapse tomography is a new paradigm in the pre-stack time-lapse seismic analysis and will be an integrated part of the geomechanical characterisation of the reservoirs.