The Producing Method Research Of Synthetic Seismogram Of Thin Interbedding

With the improvement of the exploration degree, it has been changed fromlooking for constructed oil and gas reservoir into a search for hidden oil and gasreservoir in the old fields in Bohai Bay in eastern China. In these areas, there aremany layered reservoirs of multi-cycle and thin interbedding of sandstone andmudstone. The sand bodies with large quantity are small, mostly bar strip andpotato-like. The reservoir thickness uneven distribution is thinner, and the horizontalchange is very quick. The lithology of the upper and lower layers in thin interbeddingis different, which lead to the propagation speed of seismic waves is different, so thewave impedance difference exists between the two layers. Opposite reflectioncoefficient of the interface in the top and bottom of each stratum and the smallreflected signal time difference make the top and bottom interface reflection besimilar to reverse overlay, which weaken the reflected energy and make it difficult todistinguish the reflected wave of the interface in seismic profiles. In order to betterrecognize thin interbedding, ensure accuracy of seismic and geological horizoncalibration, the following research is done.According to the geological characteristics of thin interbedding in study area,synthetic seismogram is made to guide the recognition of thin interbedding on seismicprofile. In the normal process of making synthetic seismogram by forward simulation,Zoeppritz equation is used to derive the reflection coefficient, but the thininterbedding doesn’t meet the conditions of Zoeppritz equation, because thickness oftwo media on the interface’s sides are both small. Therefore, this article derivesequations of reflection and refraction coefficient of thin interbedding while simulatingit, and then studies them under the condition that the thin interbedding is abstractingmedium. Design a model and compare reflecting coefficients derived by this methodand Zoeppritz equation respectively. And design models with layers of differentnumber, different thickness and different petrophysical parameters. The reflection coefficient of models can be obtained with this method. Then producing syntheticseismic record and some guiding significance conclusions for seismic reflection waveidentification of thin interbedding can be obtained.Sometimes the lithology changes can’t be precisely indicated and the reservoirand surrounding rock are difficult to be distinguished according to the acoustic welldata, which lead to the synthetic seismograms accuracy is not high. On the basis ofcorrection and features analysis of a variety of well data, multi-well data which hasgood reservoir characteristics is used to reconstruct the acoustic curve. The acousticcurve after the reconstruction combines all the information of well logs involved inthe reconstruction. It not only can reflect the formation velocity and impedancecharacteristics, but also reflects the lithological differences, resulting in more accuratesynthetic seismograms.In order to obtain the seismic wave with high-accuracy, the model test is used toanalyze the relationship between the actual seismic wavelet frequency and thefrequency of synthetic seismograms to correct the seismic wavelet frequency in orderto get a better initial synthetic seismogram. Statistics scan analysis is used to selectseismic traces with a higher signal to noise ratio and better quality. The exact leastsquares solution considering the impact of truncation error and the DC component isapplied while extracting wavelet, and taking into account the absorption effect of theformation, several time windows are selected to extract the seismic wavelet.Curve reconstruction method is used to process acoustic logging data of theBohai Bay region, and the reconstructed curves highlight the characteristics of theformation velocity and acoustic impedance, which a better synthetic seismogram isobtained. Using the above seismic wavelet extraction method can get a better seismicwavelet, and the synthetic seismograms produced with extracted wavelet have ahigher math with the well seismic traces.