Application Of Multi-point Pattern Optimization And Compression In Reservoir Modeling

In this paper, through the research of multiple point geostatistics, while select the classic Single Normal Equation Simulation(SNESIM) to study.we also choose several important factors to analyse. Further the memory consumption in the modeling process is also in-depth studied, and puts forward a optimal scheme, through the “compact search tree“ pattern optimization and compression, coding the scheme. And using the updated algorithm to modeling work area, Final record the efficiency of modeling, analysis the record, and to evaluate the effect of modeling.Target proportion and servosystem correction parameters, scanning template node number, the number of multiple grids, minimum repealicate number factors on the process of modeling play an important role. At the beginning of the algorithm analysis, through the target proportion and servosystem parameter modification experiments of overall modeling,which results in various simulation results in a proportional quantitative control. These two parameters affect the direction of the river through the complementary action.Influence of scanning template node number on the simulation speed and effection plays a key role. Based on the experimental and analysis, which find that the number of nodes in the template is dependent on the simulation work area size, larger area needs corresponding increase, but can not be too large, otherwise it will affect the speed, but is not improve on the simulation results. The number of multiple grids is proposed mainly to solve, the node number is certain,on large scale data structure reproduction problem.through the application of multiple grids, in a smaller template node number can make a wide range of spatial structure of reproduction. Minimum repealicate number is to assure that the extracted pattern is effective.Through the minimum repealicate number exclude the lesser important pattern, making the simulation results more accurate.The main innovation of the SNESIM algorithm is introducing the search tree for the original multiple point geostatistics modeling approach which makes the algorithm available. Search tree can greatly improve the memory consumption,indirectly making simulation speed accelerated, and update the modeling process,as a result,The multiple point geostatistics go from theoretical research to the real practical phase. The SNESIM provides a practical and reliable method for establishment of three-dimensional model. But with the development of oil exploitation, the amount of data generated rapidly grown. Due to the large three-dimensional modeling needing more memory. Through careful study of the search tree structure, proposes a compact search tree optimization scheme, mainly through extract the optimal selection of pattern from training image and compression.The memory consumption is greatly reduced, and coding, which makes the mode optimization can be achieved, the computation speed is enhanced.Finally through the application of work area. Based on a variety of training images,we apply the multiple modeling. We found this method in the guarantee of information integrity, capable of rapid modeling of 3D model.