Neural network seismic prediction of sand and mudstone lithology of Pinghu Formation in Xihu Sag

The traditional seismic P-wave impedance inversion method has the problems of low lithologic resolution and multi-solution,and it is hence difficult for the inversion results to meet the requirements of finely characterizing the lithologic distribution.In this paper,by constructing a normalized pseudo-gamma curve containing lithology and P-wave impedance information as a lithology index indicator curve,the neural network method is used to convert seismic data into a gamma data volume which is more closely related to lithology.Through the neural network seismic inversion,the sand and mudstone lithologic inversion data volume is obtained.This method was used to invert the sand and mudstone lithology of the Pinghu Formation in the Xihu Sag.Compared with traditional methods,the prediction accuracy of the mudstone thickness is up to 93%,which more accurately characterizes the distribution of underground sand and mudstone,and provides a basis for later oil and gas exploration.