Research On Seismic Prediction Of Deep Water Turbidite Reservoir

The lateral continuity of seismic data is very strong,and a large amount of useful information can be improved in the process of proving underground structures.Therefore,it is feasible to use seismic reservoir prediction technology during the exploration and production of oil and gas.Reservoir prediction is mainly carried out from three aspects: formation elastic parameters,reservoir physical parameters and oil and gas reservoir parameters.The seismic data can be used to obtain formation elastic parameters through inversion methods.Probabilistic inversion based on Bayesian theory can predict physical property parameters.In addition,the rock physical model can build a bridge between reservoir parameters and formation elastic parameters.These are the main methods for reservoir prediction.In this paper,combined with the characteristics of the actual work area as a turbidite sand reservoir,the following studies were carried out:Firstly,because turbidite sandstone is a special type of clastic rock,turbidite has a complex type,sand bodies intersect and overlap,lateral changes are fast,turbidite diagenesis is weak,and the difference between the longitudinal wave impedance of the reservoir and the surrounding rock is small Therefore,we established a petrophysical model suitable for this type of rock for this type of rock,and also analyzed the sensitive geophysical characteristic parameters of this area.In addition,although seismic inversion has been widely recognized as an effective means of reservoir prediction,the accuracy of inversion results that rely solely on seismic data is often low.If logging data is added to the inversion process,the inversion can be greatly improved The accuracy of the result.However,in actual work,the logging data is often missing,and the accuracy of the obtained inversion parameters cannot meet the ideal requirements.We can use the genetic optimization algorithm of global optimization to invert on the prestack angle gather seismic data to construct a virtual well,obtain a reliable initial model,and carry out the subsequent inversion process.Finally,there is a certain relationship between elastic parameters and physical property parameters in the formation,but this relationship is not linear,so it is more difficult to directly retrieve the physical property parameters using seismic data.In this paper,the seismic data is used to obtain the physical property parameter information directly based on the Bayesian principle,which can not only give the prediction result of the physical property parameter,but also obtain the confidence probability of the prediction result,which can be used as a reference for uncertainty evaluation.