Study On Reservoir Seismic Prediction Techniques For Thin Sand And Its Application In Pai10Xi Area

With the development of oil and gas exploration and development,lithologic reservoir and other hidden reservoirs,which have replaced structural reservoir,are the main source of oil and gas resources.And so,geophysicists have to transfer their focus on the exploration and development of lithologic reservoir.Thin sand reservoir has some typical characteristics,such as strong lateral-heterogeneity,low vertical-resolution complex sedimentogenesis and so on.How to describe and predict the thin sand reservoir accurately is difficult and challenging.There are many types of reservoirs in pai10 xi area,apart from bright-spot reservoir,which alao includes dim-spot reservoir.But in this paper,the meaning of dark spot is different from the one we known,and it refers to the weak reflection of the reservoir,which caused by the strong reflection shield of the overlying.In addition,high water-bearing possibility is also a serious problem,so the identification of false bright spots(water-bearing sand)also crucial.First of all,this paper can use strata slicing technique to macroscopically know the sedimentary characteristics of thin sand in pai10 xi area.And next this paper analyses the reflection characteristics of bright spots,which,on the seismic profile,show the “double peak”feature of upper trough and lower crest.Then,this paper analyses seismic trace integration and 90° phase shift theoretically and apply it into actual data,and this paper can find that compared with the original profile and the three instantaneous attributes,trace integration can describe and predict the thin sand thickness more accurately,which is a helpful technique for the prediction of bright-spot thin sand reservoir.Furthermore,on the seismic profile,dim spots show the “two peak and one trough” feature and the upper peak amplitude is larger than the lower one.In order to improve the reservoir prediction accuracy,this paper uses MP algorithm to strip the strong shield caused by the overlying to strengthen the reflection of the reservoir.Finally,this paper uses seismic attributes to identify the false bright spots,in which dominant frequency amplitude ratio attribute and 50 Hz single frequency attribute have a better effect on the recognition of oil-bearing and water-bearing sand.K-means algorithm can be used to seismic attributes clustering,which can identify the relationship between oil and water better,but this algorithm has its limitations which can cause to “hard” clustering results.PCA is a method to optimize the seismic attributes,but compared with the original attributes,the optimization result is not obvious.PPCA is a probabilistic model added to PCA,whose optimization result is good at the prediction of the false bright spots.