Reservoir Prediction Based On Seismic Sensitive Parameter Template

The fluvial facies reservoir is the reservoir of widely developed reservoir in china, the main feature are the lithological changes, poor continuity, etc. The8th member of59block of Sulige gas field belongs to the typical reservoir of fluvial facies, the distribution of gas reservoir is chiefly influenced by changing of horizontal sandstone and physical property. In addition, the reservoir has a lot of features, such as the reservoir of this work area is thin and vertical sandstone overlaps each other, strongly anisotropic, complicated method of distribution of gas-water. It is the biggest challenge for us to figure out space distribution of drainage line, method of gas-water, favorable place for prospecting and preferred site through reservoir prediction.This thesis starts from the study of the reservoir prediction technique. Then, it extracts seismic attributes and studies a variety of mathematical method of attribute optimization technology and neural network reservoir prediction technology as well as discuss emphatically larger errors which may be caused by the sample of reservoir prediction distribution or strongly anisotropic. At the same time, in combination with seismic attribute analysis technology and the method of comprehensive analysis of multiple attributes, it creatively proposes seismic sensitive parameter template. At last, it tests the seismic sensitive parameter template through the theory of model of forward modeling and application in actual work area shows that the method is to improve the precision of reservoir prediction.For the purposes of the thesis to research work area target reservoir, from the limited information we have, geological characteristics, reservoir characteristics, exploration and so on. Firstly, it extracts the corresponding seismic attributes with a appropriate time window to find out favorable exploration area through the application of a variety of methods combination optimization sensitive attributes and neural network prediction. But the neural network has certain requires to the sample distribution, the predication accuracy will be influenced especially in less well area or interwell.Combine with the characteristics of target reservoir, such as strongly anisotropic, horizontal change quickly and complicated distribution of gas-water, it finds out the favorable exploration area and select the target well location to increase the drilling success rate as well as reduce the exploration risk with the help of the seismic sensitive parameter template to solve the errors of neural network in less well area, finding out the favorable exploration area and select the target well location to increase the drilling success rate as well as reduce the exploration risk. At present, the method has been applied in the actual work area. The method can effectively improve the precision of reservoir prediction through the real drilling. It is conclude that it has strongly practical value and research potential.