Research On Intelligent Recognition Method Of Seismic Attribute Sedimentary Microfacies

This paper,firstly,studies the classification,extraction methods and influencing factors of seismic attribute and further,understands the geological significance of each seismic attribute.Secondly,in the aspect of attribute optimization,the necessity of attribute optimization,preprocessing methods of seismic attribute,and advantages of random forest correlation algorithms are analyzed.The use of intelligent algorithms to optimize seismic attributes and perform sedimentary microfacies prediction is an important part of this paper.In practical applications,the microfacies sedimentary features of the Upper Tertiary in Chengdao area are studied.It is recognized that the resolution of the original seismic record can not meet the ability to identify the sedimentary microfacies,so the Mallat wavelet algorithm is used to process the original seismic data.Seismic data is processed by frequency division,and the instantaneous amplitude and wave impedance properties of the study area are obtained respectively by using the improved wavelet algorithm and sparse pulse wave impedance inversion.Based on the relationship between the interpretation of sedimentary microfacies in the well and the reflection characteristics of seismic profiles,a sedimentary microfacies sample set is established,and the attribute characteristics of different sedimentary microfacies are characterized by the appropriate attribute specification method,and the set of seismic attribute feature samples is constructed.On this basis,the seismic attribute combination of channel fill microfacies,natural dike microfacies and flood plain microfacies is optimized by using the random forest MDI feature importance evaluation algorithm.Then,the stochastic forest machine learning algorithm is used to model the sedimentary microfacies probabilistic prediction,and the sedimentary microfacies of the target strata is fuzzy discriminant based on the probabilistic prediction model.On the plane,the predicted results in the vicinity of the large river channel are in perfect agreement with the micro-phase interpretation conclusion of each well,and the microfacies boundary is clear,and the plane contact relationship is in accordance with the theoretical fluvial facies deposition law.In vertical direction,the predicted results are also in accordance with the interpretation results of well logging sedimentary microfacies.In addition,the intelligent prediction of the river boundary is clearer in contrast with the existing results of the wave impedance inversion slice and Shengli oilfield.The reliability of intelligent predictive results is proved by review.