Font Size: a A A

Studies On Prediction Reservoir Of ES3 And ES4 In BO XING Depression

Posted on:2010-03-17Degree:MasterType:Thesis
Country:ChinaCandidate:C X ZhangFull Text:PDF
GTID:2120360278961305Subject:Earth Exploration and Information Technology
Abstract/Summary:PDF Full Text Request
The Es3 and Es4 members in boxing sag have the characteristics of complex place connection, serious sandstone thin interbedding, and acute physical properties changes in vertical and horizontal directions, few differences of feature of seismic reflection, which generate difficulties in identifying favorable reservoir. In order to research the development and distribution of the late sedimentary and to study the assemblage of reservoirs and capping formations, we have used the ancient landform recovering method based on seismic data .For several technique key problem of this: how to select equitime base surface, correct the divergence structure and process large irregular data.Aiming at the problem of sandstone thin interbedding for geological interpretation, the spectral inversion is applied to seismic data in this area to predict the thickness of thin layer which is lower than tuning thickness. The spectral inversion based on the notches in-thin-bed. To calculate the input data of inversion, we take short-time Fourier transform to process the seismic data, then to spectral inversion with the simulated annealing algorithm. We validated the spectral inversion by the model. We get some achievement in complex model by research the length of time windows, time windows shift and the connections of amplitude spectrum and real and imaginary part. Thin bed can be to recognition using the spectral inversion, in real seismic data.The spectral inversion is based on optimization parameters of seismic data inversion and used SA algorithm for calculating extremely minimum of target function. The simulated annealing algorithm is a serial optimization algorithms and its speed is so slow. Based on classic simulated annealing algorithm, researched the generate function was created by Cauchy function rely on temperature. The new generate function could have a wide range of search in high temperature and disturbance in near the current model in low temperature. We also improved annealing and sampling process, and added memory function in these process. It gets more efficient and more precise result.
Keywords/Search Tags:prediction reservoir, ancient landform recovering, spectral inversion, thin interbed body, simulated annealing
PDF Full Text Request
Related items