Study On Seismic Random Inversion Method And Its Application In Thin Layer Prediction Of Chengdao East Slope

Reservoir prediction is the core of oil and gas exploration.In recent years,with the deepening of oil and gas exploration and development in China,the focus of work has gradually shifted from structural reservoir to lithologic reservoir.The corresponding reservoir thickness also gradually becomes thinner.This puts forward new requirements for the research and application of exploration methods that can adapt to the new situation.It not only requires high accuracy of the method,but also requires the results to meet the actual geological conditions of the work area.This paper studies and applies the stochastic inversion method,and solves the above problems through the application of geological facies control modeling,regularization operator and characteristic waveform.The Dongdong slope area of Chengdao studied in this paper is located in the shallow sea area with a water depth of 5-10 m in the south of Bohai Bay.The main target layer is the lower half of Dongying Formation.The overall oil-bearing sand body is deeply buried,the reservoir is greatly affected by the structure in the transverse direction,the thin interbedding of sand and mudstone is relatively distributed,and the main frequency of seismic data is low.To solve the above problems,this paper first studies and analyzes the seismic phase controlled nonlinear random inversion method and seismic waveform indication inversion method,and analyzes the advantages and disadvantages of random inversion method in the inversion process.Combined with logging,geological data,seismic characteristic waveform and previous research results,two random inversion methods are used to apply the actual data in the work area.The practice shows that the inversion results correspond well with the logging data,and the thin sand body of about 2-10 M can be identified,and the transverse distribution conforms to the geological law.Under the control of seismic facies controlled initial modeling,regularization operator and seismic characteristic waveform,the prediction results are in line with the geological law and effectively enhance the ability to identify thin layers.It provides a reference for the application of stochastic inversion method in thin layer prediction.