A knowledge-based system approach for hydrocarbon prospect and play analysis

This dissertation describes the architecture and implementation of a knowledge based system for hydrocarbon exploration. The primary purpose of this research was to investigate the use of Artificial Intelligence and knowledge acquisition techniques in the design and construction of a knowledge based system which defines the salient characteristics of a target hydrocarbon play. The system then identifies analogous oil fields which are stored in the geological database. The knowledge base is implemented as a set of rules that model the way an experienced petroleum geologist works with incomplete and inexact information from sources such as seismic cross-sections and well logs to identify prospect and play characteristics. A pattern clustering technique is used to find a (group of) known field(s) similar to the target play. Clustering enables the geologist to better analyze the target prospect and/or play and to use data from similar fields to simulate the geometry of the target facies. A sample case study illustrates the interaction between the system and an exploration geologist. The results of applying the pattern clustering technique are also exemplified.