Application Of Acoustic Full-wave Logging In Effectiveness Evaluation Of Metamorphic Rock Reservoirs In Bohai Oilfield

As a kind of unconventional reservoir,metamorphic rock reservoir has been paid more and more attention in the wave of unconventional oil and gas exploration and development.BZ gas field is a large integrated condensate gas field discovered in Bohai oilfield in recent years,with proven reserves of up to 100 billion cubic meters,showing a good prospect of exploration and development.This area is a buried hill reservoir of Archaean metamorphic rock,mainly composed of gneiss.As a metamorphic rock reservoir,BZ gas field is characterized by low porosity and low permeability,and its reservoir space is dominated by fractures.It is difficult to identify effective reservoirs in this area.Reservoirs are usually divided into effective reservoirs and ineffective reservoirs according to whether the production of oil and gas in the reservoir has commercial exploitation value.For newly discovered reservoirs,effective reservoirs are generally identified according to the previous research experience by using dominant lithology,physical property parameters,fracture development and other indicators.The lack of a set of targeted technical methods for the effectiveness evaluation of metamorphic buriedhill reservoir.In view of the difficulty of reservoir effectiveness evaluation,this paper carries out the study of acoustic full wave train logging method,and evaluates the reservoir effectiveness quantitatively through various acoustic parameters.Firstly,the calculation methods of various acoustic parameters in full-wave train logging are introduced,and the close relationship between acoustic parameters and reservoir effectiveness is clarified in principle.Then,the logging data are processed by the array acoustic wave module in the logging interpretation platform.The calculation accuracy of each curve is greatly improved by subsection processing,parameter adjustment and time difference correction,and the time difference,amplitude,array attenuation,reflection coefficient,anisotropy coefficient and Stoneley wave permeability of the mode wave are obtained.In order to evaluate reservoir effectiveness better,the average attenuation calculation method is introduced,and the average attenuation calculation of mode wave is realized by connecting program in software.It is found that these acoustic full wave train logs show different response characteristics in the effective reservoir and the invalid reservoir.Therefore,the average values of various acoustic parameters in the effective reservoir and the invalid reservoir are calculated according to the comprehensive interpretation results table,and the sensitivity of these parameters to the variation of reservoir effectiveness is analyzed.The acoustic parameters sensitive to reservoir effectiveness were optimized based on the discriminating effect of the crossplot,and the evaluation chart and standard of reservoir effectiveness based on the sensitive acoustic parameters were established.The accuracy rate was 89.71% in three new Wells,and good application effect was achieved.In order to solve the problem that the two-dimensional mapping of some acoustic parameters is not effective enough to accurately identify effective reservoir,machine learning method is tried to reveal the implicit relationship between acoustic parameters and reservoir effectiveness.Three supervised machine learning methods,namely weighted KNN,BP neural network and random forest,were used to train and verify all data.The coincidence rates of the three methods were 91.8%,93.45% and 94.41%,respectively,which were verified in new Wells.The machine learning method does not need to calculate the average value of each reservoir segment,and the processing process is more convenient and fast.It makes up for the shortcomings of chart method in analyzing reflection coefficient and Stoneley wave permeability,and improves the effectiveness evaluation scheme of metamorphic rock buried-hill reservoir based on sensitive acoustic parameters.By means of map and machine learning,the effectiveness evaluation scheme of metamorphic rock reservoir for BZ gas field is established by using full wave train acoustic logging,which can accurately identify effective reservoirs,improve the accuracy of logging interpretation,meet the requirements of field application,and solve the problem of the lack of a set of targeted effectiveness evaluation methods of metamorphic rock buried-hill reservoir in this area.