Seismic Data Inversion Using Geologic Constraints

Robert Vest

Seismic data inversion has an inherent non-uniqueness problem due to the band-limited nature of the data. Using an iterative modeling process known as the conjugate gradient method, geologic constraints in the form of interval velocities can be used to improve the solution of the inversion process. By typing the seismic data to sonic logs, we can develop constraint windows, where a beginning bed velocity, an ending bed velocity, and the minimum and maximum interval velocities between these two beds can be input into the inversion process. Employing penalty functions ensures that the algorithm will honor all input constraints. The constraint windows are independent of each other except for the boundaries. Therefore, noise, diffractions, or multiples in one window will not ffect any other constraint window. The final inversion (output in interval velocities) will have the same character and feel as the sonic logs in the area.

AAPG Search and Discovery Article #91034©1988 AAPG Southwest Section, El Paso, Texas, 21-23 February 1988.