Simulating Sedimentary Successions Using Syntactic Pattern Recognition Techniques
- PDF / 351,982 Bytes
- 17 Pages / 441 x 666 pts Page_size
- 54 Downloads / 241 Views
Simulating Sedimentary Successions Using Syntactic Pattern Recognition Techniques1 E. June Hill2 and Cedric Griffiths2 Sections from a sedimentary succession can be simulated using a process which includes both probabilistic and deterministic information. The inclusion of both of these types of information allows the production of geologically realistic simulations which contain the required level of heterogeneity. The process uses syntactic pattern recognition techniques and is based on the formal description of a geological model using a grammar. The simulations can be conditioned on well data. KEY WORDS: reservoir, stochastic, syntactic, pattern, grammar.
INTRODUCTION In many regions of economic interest (such as oil fields), knowledge of a particular package of sedimentary rocks is restricted to the information provided by a few widely spaced drilled wells. The wells provide a very narrow, almost continuous, vertical succession of data and it is necessary to interpolate the data into the space between wells. When manually reconstructing geology from limited data, geologists will typically produce the simplest model that honours the data. However, it has been found that oil reservoir predictions have failed because these models are overly simplistic and so stochastic techniques are widely used for reservoir modelling (e.g. Haldorsen and Damsleth, 1990; Damsleth and others, 1992; Srivastava, 1994). Stochastic models can control the size and density of sedimentary bodies and can therefore more accurately reflect the heterogeneity present in a sedimentary succession. This process of stochastic prediction is called conditional simulation. However, it does not produce geologically realistic models of sediment distribution and, for this reason, it is desirable to include deterministic information, based on geological observations, in the stochastic models (Damsleth and others, 1992). 1Received
5 August 2005; accepted 22 October 2006; Published online: 12 April 2007. Scientific and Industrial Research Organisation, Petroleum Resources, P.O. Box 1130, Bentley, WA 6102, Australia; e-mail: [email protected]
2Commonwealth
141 C 2007 International Association for Mathematical Geology 0882-8121/07/0200-0141/1
142
Hill and Griffiths
analog model conditional data
grammars (rules, symbols and attributes)
parser
string of symbols
static geology model
conditional symbols
Figure 1. An analog model is described in the format of a grammar. The grammar and conditioning data are used by the parser program to generate stochastic geological models which are made up of vertically stacked strings of symbols.
A number of methods which can produce discrete object-based geological models have been developed, these all use marked point processes (e.g. Hatløy, 1994; Tyler, Henriquez, and Svanes, 1994; Deutsch and Tran, 2002). A completely different method is introduced here, based on syntactic pattern recognition. The syntactic method is not only object-based but it is also object-oriented. Objectoriented means that a geologi
Data Loading...
