Predicting the distribution of reservoirs by combining variable wavelet model of seismograms with wavelet edge analysis and modeling

Xie, Yujiang and Liu, Gao (2014) Predicting the distribution of reservoirs by combining variable wavelet model of seismograms with wavelet edge analysis and modeling. Journal of Applied Geophysics, 101, 116-123. (doi:10.1016/j.jappgeo.2013.12.005).

Abstract

Reservoir prediction with its unique role in oil and gas fields is constantly facing new challenges, such as high-resolution seismic data and fast-accurate impedance inversion are needed. Generally, conventional methods used to enhance the resolution of seismic data, for example the spectral whitening, sometimes called balancing or broadening, is hard to yield valuable results as the seismic wavelets change during traveling subsurface. Besides, impedance inversion used in reservoir such as acoustic impedance inversion (AII) also confronts problem—low computational efficiency when more geological and geophysical parameters are taken into consideration in the modeling inversion. Based on these questions, in this study, a joint approach is presented. The first approach is the variable wavelet model of seismograms (VWMS), which is carried out by a series of processes such as time partition and frequency domain processing, to enhance the resolution of the seismic traces. Another approach that can improve the computational efficiency of the AII is the acoustic impedance inversion based wavelet edge analysis and modeling (AII-WEAM). In this approach, the algorithms of the AII were replaced by the modified very fast simulated annealing (MVFSA), to improve the inversed speed. By using a gas reservoir predicting example, we show that the joint approaches produce results that are feasible and reliable after comparing with the well data. Hence, these joint approaches have great potential to be the next-generation tools for reservoir description and prediction.

This record has no associated files available for download.

Contributors

Download statistics