Reservoir models integrating seismic data are generally obtained as a solution of a complex inverse problem involving geophysical models and geostatistical methods. We propose here a new methodology to obtain high-detailed reservoir models in terms of facies and rock properties as a solution of a linearized inverse problem given well log and seismic data. This methodology overcomes the assumption of Gaussian distribution of reservoir properties by using Gaussian mixture models to describe the multimodal behaviour of the data. In this method we combine geostatistics and linear inverse theory to sample from the posterior distribution of a Gaussian mixture linear inverse problem to obtain multiple realizations of reservoir properties characterized by multimodal behaviour.


Article metrics loading...

Loading full text...

Full text loading...

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error