The Technologic Research Of Stochastic Modeling About Reservoir Development

By the end of Dec, 2007, ShengLi Oil field has the synthesis water cut of 90.3%, degree of reserve recovery of 28.6%, and oil production rate of remain recoverable reserves of 10.1%, which has completely accessed producing period of"three high". It is more difficult to promote the potential capacity. The reservoir is badly flooded in plane direction, and the distribution of remaining oil is particularly complicated. The exploitation objects have already been changed from early oil layer groups to individual bed and sand body. The former geological model with intrafacies, channel deposit and nonchannel deposit can not adapt to the requirement of development. More accurate geological model is in great need to be established for requirement of fine reservoir numerical simulation and remaining oil research.By means of survey of reservoir stochastic modeling technology and development situation of the corresponding software, on the base of the research of three-dimensional geologic structure modeling and reservoir stochastic modeling, multidisciplinary theory and methods have been used, such as computer science, stratum rupture processing method, geostatistics, interpolation algorithm, grid scaling up technology, OpenGL integration, select and certification of model, numerical reservoir simulation. Many kinds of information datum about reservoir geology, earthquake interpretation, well drilling, well log interpretation, sublayer comparison and production datum have been used comprehensively. Process of stochastic geological modeling with facies control has been improved basing on technology analysis of different stochastic simulation methods and evaluation of their adaptation. And small test software (SLDZSM) for reservoir 3D geology stochastic modeling has been developed by absorbing virtues of popular geology modeling software. The software function includes figure subdivision, interpolation and 3D visualization and so on. It can realize three dimensional geology modeling with any complicated faults conditions during reservoir development, and put the results to numerical reservoir simulation model. Besides, a method has been creatively proposed to process stratum rupture using cross-ring. Bridge-edge and stratum area decompounding method has been used to solve reverse fault issue, and the applying problem basing on state machine model has been resolved by using a function class mechanism. It supplies a new idea and new method for reservoir stochastic modeling. At the same time, the geology stochastic modeling theory system has been enriched and improved. It has important academic and practical meaning in deepening reservoir development geology stochastic modeling theory and increasing remaining oil research extent.SLDZSM software has been used to research reservoir stochastic modeling in pilot site which locates on Shengli Gudao oil field middle area of middle II production area. The production mode of the pilot site would be changed form water flooding to thermal recovery. The model not only coincides with well point data completely, but also has stratum structure simulation results which essentially accords with earthquake interpretation structure, and its sedimentary facie distribution form agrees with geology knowledge. In model section, the tip of sand body of channel facies is planar and its bottom is raised。The river channel slices downwards, and there is lenticular sand body. These typical river facies sedimentary characters completely accord with real geology situations. According to verifying 11 vacuating wells parameters such as structure and sand body thickness that have been predicted by model, it shows high coincidence rate of more than 90% to real datum. Moreover, the permeability variation in pilot site accords with water flooding situation. The computed parameters such as reserves, cumulative oil production of single well and area, water cut have small relative error of less than 2%, the matching accuracy degree is high. All the results demonstrate the stochastic modeling software has high accuracy degree and can offer a good guidance field production and remaining oil research.