Research and application of waveform clustering with variable time window in dolomite distribution prediction

Abstract

Dolomite is one of the rocks favorable for oil and gas accumulation in marine strata, but its distribution pattern is difficult to predict by geophysics. Many scholars have proposed research methods including attribute and lithological inversion to study the distribution of dolomite, and some scholars have proposed wave clustering methods. Among them, for waveform classification, the existing method usually uses equal-thickness time windows to select the target layer for waveform clustering analysis. For areas where the thickness of the reservoir varies greatly in vertical and horizontal directions, the corresponding relationship between the waveform and the reservoir is not equal to the thickness. This method cannot characterize isochronous sedimentation phenomena. From the perspective of seismic sedimentation, this paper transforms the time domain unequal-length seismic signal into the frequency domain with isochronous interface as the constraint, selects an appropriate constant bandwidth, carries out waveform clustering with variable time window, and analyzes seismic facies. This method has achieved good results in the prediction of dolomite distribution in complex areas such as Ordos.