--> --> Abstract: Rule-Based Expert System for Seismic Data Interpretation, by Neelu J. Ahuja and Dr. Parag Diwan; #90081 (2008)

Datapages, Inc.Print this page

Rule-Based Expert System for Seismic Data Interpretation

Neelu J. Ahuja and Dr. Parag Diwan
College of Engineering Studies University of Petroleum & Energy Studies

Seismic exploration is one of the major tools, extensively used by petroleum geologists, to identify geological structures, suitable for oil reservoirs. The process of Seismic Data Interpretation facilitates accurate and meaningful inference which can be tied to the subsurface geology. This interpretation is a complicated process and depends not only on sophisticated technology, but also, to a large extent, on the experience and interpretive powers of field geophysicists. Seismologists routinely carry out interpretation, but there is fair amount of uncertainty in terms of geological structure, which is partly because of extreme geological complexities and partly because there are no formal rules, and each ‘expert’ uses his individualistic knowledge-base of unwritten thumb rules, that he has developed over years. Responding to this need, development of an expert system prototype for seismic data interpretation, is undertaken which is facilitated by the knowledge engineering process.

In order to develop the expert system prototype, a general purpose expert system shell has been identified, and suitably tailored to this specific task domain. Knowledge solicited, analyzed and converted into set of rules (Rule Base) written in a logical sequence in a declarative if…then format, inferring the existence of the geological structure of potential hydrocarbon bearing entities, has been incorporated into the knowledge-base of the shell. The knowledge-base embodies, knowledge both factual, which is widely shared and commonly agreed upon, as well as heuristic. The latter is less rigorous, more experimental and largely individualistic.

Another vital component of the shell is the inference engine which is a reasoning structure, which manipulates and uses the knowledge in the knowledge-base appropriately, to forward chain if…then rules to form a line of reasoning. It employs a data driven approach and derives recommendations or conclusions from the knowledge-base, by matching facts against patterns and determining which rules are applicable.

The processed seismic data available in an internationally accepted “SEG-Y” format are ported in the shell, and a marker of certain amplitude is picked and followed across the successive traces, on time axis. A sudden shift of this marker would get inferred as a fault. The knowledge-base holds the fact, that the presence of a discontinuity in reflected wave amplitude is indicative of existence of a fault plane. The inference engine matches this fact, with the observed discontinuity and interprets a fault in its conclusion. Thus, the knowledge-base and inference engine co-operate to simulate the reasoning process that human expert pursues in analyzing a problem and arriving at a conclusion.

The power and competency of the expert system lies in its accumulation and assimilation of high quality domain knowledge, just as human mind develops the expertise as more and more experience is gained, in a given field.

The process of knowledge accumulation is continuing with the hope of building a prototype capable of interpreting a reasonably complex seismic section in terms of its geological structure.

Presentation GEO India Expo XXI, Noida, New Delhi, India 2008©AAPG Search and Discovery