Geostatistical Inversion And Its Application In Reservoir Prediction

In the context of China’s rapid economic development, the demand for oil and natural gas and other energy resources is increasing, and therefore, oil and gas exploration and production has become an important factor for economic development. Now, with the targets of oil and gas exploration gradually change from hidden conventional big tectonic oil and gas reservoirs to unconventional lithological subtle oil and gas reservoirs and small or discontinuous reservoirs, it becomes more and more difficult for reservoir prediction, and attendant, the requirements for precision of reservoir prediction result are gradually increased. Seismic inversion is an important means of reservoir prediction, and it’s application in reservoir prediction has got more and more attention. There are many kinds of seismic inversion methods, different inversion methods have different advantages and disadvantages. For some specific research area, a reasonable choice of seismic inversion methods will have an important influence on the final prediction.Among many kinds of inversion methods, the method that widely used relatively is conventional post-stack P-Impedance inversion, conventional P-Impedance inversion is a method that based on seismic data, due to the limit of frequency range of seismic data, the vertical resolution of inversion often very low, it can get a good result in prediction of reservoir that has relatively simple structure and uniform sand distribution with large-scale, but it is difficult reach the demand of reservoir prediction for unconventional lithological subtle reservoirs and small or discontinuous reservoirs. The geostatistical inversion method that based on geostatistical analysis and stochastic simulation, can fully integrate the lateral continuity of seismic data and the vertical high-resolution of logging data, and adding a priori geological information, to achieve high resolution inversion with accurate and detailed description of the reservoir characteristics, improve ability of inversion in thin and discontinuous sand recognition.This thesis mainly focuses on the following study:First, summarize the characteristics and using conditions of common seismic inversion methods, provide guidance to selection of seismic inversion methods under different conditions and different, select conventional constrained sparse spike inversion and geostatistical inversion method as the main analysis object. Second, the study focus on the principles and implementation process of geostatistical inversion method, including the calculating and fitting method for variogram, analyzing the relationship between variogram parameters and reservoir heterogeneity; it also includes study variety of Kriging interpolation method, analysis and test of important parameters such as sampling interval that influence the result of kriging interpolation. Third, it analyzes several important techniques in the inversion process such as seismic data processing, standardization of log data, establish deep and time relationships, wavelets selection, statistical analysis of reservoir parameters. It determine the technical requirements of all aspects the impact on the final results.The data used in this study came from an area of China. The study used constrained sparse spike inversion and geostatistical inversion to predict the reservoir. The results show that conventional constrained sparse spike inversion can generally describe the reservoir distribution in this area, but the resolution of inversion results is too low to predict the distribution of many thin sand bodies in the study area. The geostatistical inversion method fully integrate the vertical high-resolution of logging data and lateral continuity of seismic data, the resolution was significantly higher than conventional constrained sparse spike inversion, it can clearly identify the distribution of thin sand bodies and discontinuous small sand bodies, it realizes precise and fine prediction of reservoir.