Reservoir Parameter Predicted Method Research And Primary Application Basing On Seismic Attribute

The seismic attribute is the important means of horizontal reservoir prediction. They arekey problem which is decide to keep the reservoir prediction success that how to accuratewithdraw a purpose layer seismic attribute and how to carry out seismic attribute optimizationand How build up the relation of reservoir parameter.In this thesis we write many programs to predict reservoir parameter, such as unaryregression analysis program, multiple regression analysis program, principal componenttransform program, seismic attribute picked-up in well position program, reservoir predictionprogram using multiple regression analysis method reservoir prediction program using artificialneural network method We pick up seismic attribute along reflector using GEOFRAM software.We choose pick-up method carefully and control time windows' length to assure quality ofseismic attribute. We use a lot of method to optimize seismic attribute such as expertoptimization, relativity analysis optimization and principal component transform optimization,we also attempt use many method together and we obtain satisfied result. In this thesis ,we usemultiple regression analysis method and artificial neural network method to draw up function ofseismic attributes and reservoir parameter in the wells location respectively, then we predict thereservoir parameter in the whole area using this function.Reservoir fracture distribute extensively in the Hailaer basin. It is very important toexploration and development of the reservoir that how predict distributing of fracture exactly.We use curvature method to compute several kinds spatial curvature of structural surface.Multiple regression analysis method is used to draw up function of spatial curvatures andfracture density on drilling core, then we predict the reservoir fracture density in the whole areausing this function and achieved good result.