- Facies Predictions

- Facies Predictions

Main | Mathematical Methods | CRAVA User GuideDownload


XML Model | Inversion | Estimation | Forward modelling | Rock Physics | Facies Predictions


CRAVA has the ability to create facies probabilities. This is set by the <facies-probabilities> keyword under <action> and with inversion mode,

<actions>
  <mode> inversion </mode>
 <inversion-settings>
    <facies-probabilities> yes </facies-probabilities>
  </inversion-settings>
</actions>

The facies probabilities are computed based on the inversion results and a distribution for inversion values given facies computed from filtered well logs, where the filter is defined by the inversion.

Probability volumes will be computed for all facies seen in wells, and if the command <facies-probabilities-with-undef> is used, an additional undefined probability cube is also generated, indicating areas where the inversion values are too far away from well data to make reliable predictions.

If <use-absolute-elastic-parameters> is set to ’yes’, facies probability is computed based on inverted parameters including background model. If the distribution for elastic parameters for each facies is constant over the inversion volume, using absolute values is more stable. However, if there are trends in the elastic parameters, the relative values are more robust. The default is to  use relative values.

In order to get reliable probabilities, we need good prior probabilities. By default, CRAVA computes the average fraction of each facies in the relevant interval of the wells. This can be overridden using the <prior-probabilities> command, which allows specification of these. Note that probabilities must be given for each facies. Probabilities can either be given globally, with <probability>, or as a full 3D trend, using <probability-cube>. The latter takes a grid file as argument, and the corresponding grid must cover the inversion volume. A prior value for undefined facies is set with the command <uncertainty-level>.

Example of facies probabilities file can be downloaded here.