Résultat de la recherche
Conditioning Facies Simulations with Connectivity Data
When characterizing and simulating underground reservoirs for flow simulations, one of the key characteristics that needs to be reproduced accurately is its connectivity. More precisely, field observations frequently allow the identification of specific points in space that are connected. For example, in hydrogeology, tracer tests are frequently conducted that show which springs are connected to which sink-hole. Similarly well tests often allow connectivity information in a petroleum reservoir to be provided. To account for this type of information, we propose a new algorithm to condition stochastic simulations of lithofacies to connectivity information. The algorithm is based on the multiple-point philosophy but does not imply necessarily the use of multiple-point simulation. However, the challenge lies in generating realizations, for example of a binary medium, such that the connectivity information is honored as well as any prior structural information (e.g. as modeled through a training image). The algorithm consists of using a training image to build a set of replicates of connected paths that are consistent with the prior model. This is done by scanning the training image to find point locations that satisfy the constraints. Any path (a string of connected cells) between these points is therefore consistent with the prior model. For each simulation, one sample from this set of connected paths is sampled to generate hard conditioning data prior to running the simulation algorithm. The paper presents in detail the algorithm and some examples of two-dimensional and three-dimensional applications with multiple-point simulations.
An Improved Parallel Multiple-Point Algorithm Using a List Approach
Among the techniques used to simulate categorical variables, multiple-point statistics is becoming very popular because it allows the user to provide an explicit conceptual model via a training image. In classic implementations, the multiple-point statistics are inferred from the training image by storing all the observed patterns of a certain size in a tree structure. This type of algorithm has the advantage of being fast to apply, but it presents some critical limitations. In particular, a tree is extremely RAM demanding. For three-dimensional problems with numerous facies, large templates cannot be used. Complex structures are then difficult to simulate. In this paper, we propose to replace the tree by a list. This structure requires much less RAM. It has three main advantages. First, it allows for the use of larger templates. Second, the list structure being parsimonious, it can be extended to include additional information. Here, we show how this can be used to develop a new approach for dealing with non-stationary training images. Finally, an interesting aspect of the list is that it allows one to parallelize the part of the algorithm in which the conditional probability density function is computed. This is especially important for large problems that can be solved on clusters of PCs with distributed memory or on multicore machines with shared memory.