Research On The Research Fractures Prediction Methods Based On Random Fractal Theory

It is important to predict reservoir fractures in oil & gas exploration and development. The distribution of fractures has strong heterogeneity, randomicity and complexity. So the Geophysical responses such as seismic record or logging data which describe the reservoir character have non-liner trait like randomicity and fractal. The thesis based on random fractal theory, utilizing fractional Brownian motion model, using R/S analysis technique and wavelet transform method to estimate Hurst exponent which depicts the complexity of seismic signal. Hurst exponent can be used to predict reservoir fractures. Meanwhile, transform seismic amplitude to grey digital image, using edge detection algorithm based on the discrete fractional Brownian random field model to predict reservoir fractures is another method studied in the paper.Both R/S analysis technique and wavelet transform method can estimate Hurst exponent efficiently. Relatively, the wavelet transform method restrains noise better. In fractural area, because the elasticity is heterogeneity, seismic wave absorbed and scattering, the seismic record is more complex. Thus, the Hurst exponent in fractural area is less than in non-fractural one. Practical examples show that the methods have good effect for predicting reservoir fractures.The grey digital image transformed from seismic amplitude slice has fractal character. Its Hurst exponent change indicates the fracture's density, direction etc. Using edge detection algorithm based on the discrete fractional Brownian random field model to predict reservoir fractures has good effect, there are theoretical and practical examples placed in the paper.