D-tect sequence stratigraphic interpretation system (ssis)

Development of a seismic sequence stratigraphic interpretation system on top of the innovative d-tect seismic attribute processing and pattern recognition system. Decomposition and analysis of seismic data in a transformed, geological time domain.

Sequence stratigraphy is the study of rock relationships within a chronostratigraphic framework, wherein the succession of rocks is cyclic and is composed of genetically related stratal units (Posamentier, 1988, SEPM Special Publication 42). The concepts of sequence stratigraphy have revolutionised geologic interpretation and yielded valuable new exploration and production tools for oil, gas and coal. Sequence stratigraphy is used amongst others for well log correlations, geologic modelling and predictions of reservoir quality in un- and poorly explored basins. Sequence stratigraphy was directly derived from seismic stratigraphy, which in its turn was introduced in the 1970's (AAPG memoir 26). Seismic stratigraphy is composed of two components: 1) Study of seismic sequence boundaries (the study of seismic reflection patterns and reflector relations) 2) The study of seismic facies. To our knowledge there are no commercially available seismic systems that support sequence stratigraphic interpretations. Such interpretations are still largely done by manually picking unconformities and by describing seismic facies in qualitative terms. The patterns are sometimes described through attribute clustering techniques, but none of the commercial systems (including the current version of D-TECT) produce results that honour sequence boundaries. We intend to develop a system that will perform a semi-automated clustering of seismic data into sequence stratigraphic units and new visualisation tools that allow seismic data to be studied in geologic time. A two-step approach will be chosen. First of all, the most important sequence boundaries will be defined. Both the seismic volume and existing interpretations of major sequence boundaries in wells will be used and integrated. On well logs sequence boundaries can be detected by studying the cyclic nature of the log response (this is the basis of CycloLog, a commercial log analysis tool). We will investigate whether the output of such a well log analysis can be used to identify major seismic boundaries. We will also exploit the unique spatial characteristics of seismic data to detect sequence boundaries (using reflector terminations). This requires developing new attributes and tracking algorithms to map boundaries that separate sequences of a different order. The second step in the process, once separate seismic sequences (which can thus be considered as sub-volumes of the seismic cube) have been identified, is seismic facies mapping within the individual sequence. The multi-attribute neural network techniques already existing in D-TECT will be further developed for this specific purpose. Traditionally seismic events are mapped one-by-one using some kind of seismic horizon tracker. In D-TECT a new seismic interpretation module is currently under development. In this module horizons and faults are tracked simultaneously. This allows geometry constraints to be incorporated, which in turn avoids extensive manual editing that is common in conventional systems where horizons and faults are mapped separately and sequentially. In the sequence stratigraphic module we want to go one step further. We do not want to limit ourselves to mapping the main events only. Instead we want to map all seismic events within a chronostratigraphic framework. Additional events marking higher order boundaries will be filled through some form of automated tracking, yet to be developed. The clustering algorithm that needs to be developed shall be constrained to the boundaries found in this way. Finally we intend to develop algorithms that allow us to follow all seismic events on a sample-by-sample basis and thus create a set of horizons that could be referred to as a 'reflector cube'. Flattening horizons is an established way of viewing data that are spatially correlated in a stratigraphic sense. If we are able to map all horizons it should also be possible to flatten these and stack them to create a new volume that allows us to study seismic data in the stratigraphic domain. Keywords: sequence stratigraphy, seismic clustering, horizon flattening.
Project ID: 
3 058
Start date: 
Project Duration: 
Project costs: 
1 700 000.00€
Technological Area: 
Tectonics, Seismology
Market Area: 

Raising the productivity and competitiveness of European businesses through technology. Boosting national economies on the international market, and strengthening the basis for sustainable prosperity and employment.