--> --> Adaptive Eigen-structure Classification and Stochastic Decorrelation Filters for Coherent Interference Suppression in the Acoustic Zoom Method

International Conference & Exhibition

Datapages, Inc.Print this page

Adaptive Eigen-structure Classification and Stochastic Decorrelation Filters for Coherent Interference Suppression in the Acoustic Zoom Method

Abstract

The Acoustic Zoom® (AZ) Method is an unconventional oil and gas exploration technology for 3D/4D seismic imaging that offers unique resolution and direct focusing of non-specular backscatter energy returns in land and marine environments. The receiver array has sixteen spokes (eight lines through a hub) at ∼22.5°increments. This array was deployed to intersect GGS's Wrangler 3D survey in Wilson County, Texas, over the prolific Eagle Ford, Austin Chalk, and Buda (limestone) formations. The AZ purpose-designed array—covering 12.5 km2, encompassing over 4000 receivers, and with recorded frequencies up to 170 Hz—acquired over two terabytes of data. Five vibroseis locations were established in a cross configuration at a quarter of wavelength separation. At each of the five vibroseis locations, 512 sweeps were generated and vertically stacked for a total of 2560 sweeps. By design, AZ accentuates the rich content of non-specular backscatter energy by directly probing underlying geophysical properties of the earth where conventional migration accentuates specular reflections from ambiguous impedance changes in the subsurface. The totality of energy backscattered in the direction and range of a corresponding beam forms each AZ image. AZ adds value to existing 3D surveys by reconstructing complementary components of recorded energy that 3D seismic rejects as incoherent noise. Imaging of non-specular returns requires significant attenuation of coherent background interference (e.g. ground roll, specular reverberation) achieved by combining the narrow beam-width of the receiver array with adaptive classification and filtering of specular energy using Singular Value Decomposition, Stochastic Spectral Decorrelation, and advanced Eigen-structure methods. This approach replaces conventional filters that could introduce artifacts greater than the non-specular signals being sought.