Research And Application Of Geostatistics In Reservoir Prediction

Reservoir prediction has been very concerned about in geologicalreservoir research ,which includes two important work: studying thereservoir heterogeneity and improving prediction accuracy. The methods ofgeostatistics in reservoir prediction have been studied , that focus on thereservoir heterogeneity research and improving reservoir prediction accuracy.Firstly, calculation methods, fitting methods and nested methods ofvariogram have been studied. Through practical application, calculatetwo-dimensional variogram and variogram of different directions, determinethe orientation of reservoir parameters and heterogeneity of reservoirparameters distribution by observing all directions of the variogramparameters. Then apply linear programming method to fit the variogram,analyze linear programming method and compare it with conventionalvariogram fitting method. According to fitting results and orientation ofreservoirparameters,thevariogramhavebeennested.Secondly, the reservoir parameters(porosity and thickness)oftwo-dimensional horizon have been predicted by using the methods ofOrdinary Kriging, Co-kriging, Collocated Co-kriging and sequential Gaussiansimulation. And use cross-validation method for analyzing feasibility and accuracy of reservoir prediction. Also study the impact factors of krigingmethods, such as the variogram models, the scope of the search and so on.Also the intuitive relationships between kriging estimate and these factors canbe built. These studies can help to choose reasonable kriging parameters andimprove reservoir prediction accuracy. The reservoir parameters (porosity andclay content) of crosshole profile have been predicted by using fractal krigingmethod and heterogeneity of inter-well reservoir parameters have beenannualized.Finally, characteristics and conditions of geostatistical methods andmeans of raising accuracy in reservoir prediction are summarized.