Page 6 - ShalePlay.indd
P. 6

See the geology in  the seismic by  classifying or mapping  complex geologies with supervised neural  networks...


     Seismic Facies at Wells                           Seismic Facies on Horizon                         Shale Capacity                                     Thinned Fault Likelihood



































     Identifying sands, silts and shales with          The classified seismic volume can now be          Shale capacities can be computed by                Secondary porosity can be as important as
     similar acoustic impedances is challenging.       interpreted and extracted from  seismic           combining individual supervised neural             shale capacity or lithology.  Thinned-fault
     Using supervised neural networks to               horizons, with the relative probability for       networks for brittleness, TOC and porosity,        likelihood, phase congruency, and GLCM
     classify the seismic lithologies or facies        each facies classified.                           or by training directly with effective shale       azimuthal texture measurements are non-
     improves risking and drilling decisions.                                                            capacity from well data. This can be done at       curvature based techniques to determine
                                                       The objective siltstone class is shown  on        sub-seismic sample rates to obtain higher          fracture orientation and density. Pore-
     While the acoustic impedance curves in            horizon by the orange colors. Areas without       resolution measurements, better matching           pressure can also be computed through
     each well track vary from well to well (blue      silt that are limestones are displayed in         log-scale observations.                            neural network techniques as a proxy for
     curve), the neural network  classification        white and gray colors. The clarity obtained                                                          fracture density.
     systematically determines the silt (orange)       through classifying neural networks is            Above, brighter colors indicate higher
     and shale (blue) through generalization.          straightforward.                                  capacity, and several high capacity lenses         In this example the normalized shale
                                                                                                         can be observed within the Spraberry and           capacity horizon is overlain by thin fault
                                                                                                         Wolfcamp intervals (between the blue and           likelihood. The faulting or fracturing
                                                                                                         pink horizons).                                    magnitude is illustrated by brighter colors.
   1   2   3   4   5   6   7   8