Seismic Inversion is a very simple task on paper. Essentially, Seismic Inversion is the conversion of reflectivity data (which is normally obtained in any exploration task) into impedance data, which is obtained by nothing more than the velocity multiplied by the density. With something so simple, it really seems to beg belief that seismic inversion has achieved a status as one of the most ill-posed problems in geology but indeed, it has obtained this less than glowing reputation.
The problem with seismic inversion lies in the fact that the process is iterative – the process of extracting reliable velocity information from seismic stacking velocities and applying it to the sonic log will indeed give a trend, but the issue with most model driven methods is the fact that applying a model to another model would be the obvious – applying a model to your data will indicate points which may very well be inaccurate.
It falls to the interpreter to assess each point on this result, and establish well to seismic ties in order to fill in the gaps. This process can be long and arduous, as absolutely every single data reference point has to be interrogated in order to establish that it is likely that the reference point will reflect something which is there. These reference points are subject to interrogation and interrogation, and working out the average chance that a point of reference is where it is suggested is a very good reason for the fact that seismic inversion is still treated with faint suspicion despite advances such as simulated annealing inversion and coloured inversion in recent years.
Using either Kingdom or Petrel in order do this can be rather inaccurate, and problems can be overcome with a plugin called InSeis which is developed for both programs for the express purpose of answering the problems associated with seismic inversion. On Kingdom, the plugin is known as Kingdom Seismic Inversion, and on Schlumberger’s Petrel, it is known as InSeis.
Comments
Post a Comment