Muiti-scale Joint Modeling And Joint Inversion

The information of wave impedance is an important kind of physical parameters for rockwhich is good at characterizing rock property. It plays an important part in present reservoirprediction and evaluation. Wave impedance can effectively distinguishes reservoir andnon-reservoir and determine s the favorable oil-gas locations. It plays a very important role inthe oil-gas exploration and exploitation, especially in the prediction of lithologic hydrocarbonreservoirs and subtle reservoirs. This article firstly estimates the initial wave impedancemodel from multi-scale seismic data through the redundant dictionary inversion method, thenoptimizes the initial model quickly with oriented simulated annealing algorithm and finallyobtains high-resolution and high-accuracy wave impedance data.Conventional wave impedance inversion methods depend on the statistical theory orinterpolated well data to construct initial model, the accuracy is limited. According to theredundant dictionary algorithm, seismic data can be directly translated to wave impedanceinformation which is loyal to the actual data. This method can be applied to multi-scaleseismic data to obtain high-resolution and wideband inversion data which provides effectivebasis for high-accuracy reservoir exploration and exploitation.In the research, the first step is to create a certain redundant dictionary from Gammafunction, then to decompose and estimate each scale of seismic data and obtain correspondinginitial wave impedance information. Estimated results merge according to logging dataspectrum distribution. Optimization process quickly improves the initial model with theoriented simulated annealing algorithm and the constraint of seismic data according to therelevant restraint function.This inversion algorithm just needs the seismic wavelet and certain geologicalknowledge. It can be widely used in the area where exploratory wells are rare. The methodapplied in a foreign work area,the inversion result and actual well logging data fit well whichshows that the method is applicable to the actual process. This method can provide reliableinformation of wave impedance for reservoir prediction and evaluation.