Application Research Of Separate Frequency Seismic Attributes In Fractured Reservoir Prediction

Conventional seismic attributes are usually extracted from the whole frequency band seismic data,causing the loss of effective information in different frequency bands.Compared to conventional seismic interpretation methods,the separate frequency seismic interpretation technology can implement analysis and interpretation using seismic data with different frequency bands obtained by time-frequency analysis method.Additionally,employing low-frequency and high-frequency information separately can help to achieve interpretation results with higher resolution and accuracy.According to conventional seismic attribute extraction methods,this paper studies the separate frequency seismic attribute extraction method based on wavelet transform.First of all,a series of data with different frequency bands can be yielded by using the continuous wavelet transform method to decompose the original seismic data.Then,the seismic attributes are extracted separately from the separate frequency seismic data,and the horizon attributes are extracted according to the interpreted horizons.In the study of seismic attribute optimization methods,the principal component analysis(PCA)and the locally linear embedding(LLE)are implemented,and their optimization effects are compared.Finally,the separate frequency seismic attributes and the LLE optimization method are applied to the fractured reservoir prediction.The applications of field data show that separate frequency seismic attributes can effectively predict the development of fractures from different frequency bands and different scales,and the prediction results are more reasonable than those of the conventional seismic attributes technology.