224 Geophysics in Geothermal Exploration Figure 7.1 Seismic inversion versus direct modelling. The objectives of a seismic inversion are: • To optimize impedance values. They depend on the intrinsic properties of the rock, while the seismic signal depends on their contrast. Seismic characterization, sometimes called quantitative interpretation, is a discipline that links elastic properties (or other signal derivatives) with key reservoir properties (lithology, porosity, fluid...) or fractures. Learned directly from raw results (unsupervised) or guided by well data (supervised), the discipline allows the estimation of key properties through the application of machine learning techniques. • To reduce random noise, depending on the inversion technique used, and therefore improve legacy seismic data and its subsequent seismic attributes, revealing better the faults and fractures, or facilitating seismic interpretation. After describing key concepts related to these disciplines, we will describe the methodology using InterWell, the software solution from Beicip-Franlab, part of IFPen group, for seismic inversion, seismic characterization and time-depth conversion. This will be followed by a practical case study. 7.1 Technical background 7.1.1 Seismic gathers and partial stacking Following acquisition and a sequence of data processing and migration, preserving amplitudes, seismic data are gathered such that traces correspond to the same CMP (Common Middle Point) or CDP (Common Depth Point) depending on the
RkJQdWJsaXNoZXIy NjA3NzQ=