Stochastic simulation of karstic systems: data conditioning.

@inproceedings{BardyGM2011,
 abstract = { By their strong spatial heterogeneities, karstic networks considerably modify underground flows, and make their modeling important for both petroleum and aqueous reservoir engineering. But the common lack of data and the complexity of the phenomena involved in their formation increase the difficulties to use a universal genetic approach. This work aims at building a stochastic simulator of karstic systems, taking into account the most of the data. It is based on a previously developed static approach, which computes the skeleton of the network(low-cost paths between sources and sinks). The karst envelope is then determined with the ODSIM method which stochastically simulates the cave dimensions around the karst skeleton.
To be more consistent with karst genesis processes, we propose to take integrate surfaces such as bedding planes or paleo-water-level which promote karstic network development. We also use tracer data obtained by hydrologists to constrain the karst connectivities. Finally, well data, by imposing the presence or absence of karst, are used to guide the simulation of the karst skeleton. In the case of paleokarsts, the size of the collapsed zone determined from logs constrains the simulation of the paleocave during the ODSIM step. The first results obtained on a synthetic dataset show the potentiality of the proposed method. },
 author = { Bardy, Gaetan AND Collon, Pauline AND Pellerin, Jeanne },
 booktitle = { Proc. 31st Gocad Meeting, Nancy },
 title = { Stochastic simulation of karstic systems: data conditioning. },
 year = { 2011 }
}