Brought to you by EarthWeb
IT Library Logo

Click Here!
Click Here!

Search the site:
 
EXPERT SEARCH -----
Programming Languages
Databases
Security
Web Services
Network Services
Middleware
Components
Operating Systems
User Interfaces
Groupware & Collaboration
Content Management
Productivity Applications
Hardware
Fun & Games

EarthWeb Direct EarthWeb Direct Fatbrain Auctions Support Source Answers

EarthWeb sites
Crossnodes
Datamation
Developer.com
DICE
EarthWeb.com
EarthWeb Direct
ERP Hub
Gamelan
GoCertify.com
HTMLGoodies
Intranet Journal
IT Knowledge
IT Library
JavaGoodies
JARS
JavaScripts.com
open source IT
RoadCoders
Y2K Info

Previous Table of Contents Next


2. GEOLOGICAL EXPLORATION AND MINERAL ANALYSIS

2.1. SCOPE AND HISTORICAL DEVELOPMENT

The pioneering effort behind MYCIN and PROSPECTOR (Gaschnig 81) encouraged similar work in many areas. Several laboratories replicated the PROSPECTOR effort. These early systems also spurred the interests of the oil industry. This development had significant effects on both R&D funding as well as on commercialization. The oil industry was a wealthy and technology-oriented industry, well prepared to recognize the potentials of expert systems. In particular, exploration departments and exploration contractors embraced the advent of this technology. Despite a long tradition with logging and seismic shooting that yielded hard facts about a geological area, exploration still relied on expert judgment, intelligent guessing, and many qualitative elements in the decision process. It took a long time for a person to learn. Expert system technology promised a formalization of the heuristics that resided with very experienced geologists. It pledged a future of effective knowledge sharing and improved data analysis, blending many types of data with apt knowledge of a particular geological field. Slight improvements could yield potential rewards in the multibillion dollar range.

Almost as important as PROSPECTOR was the DIPMETER ADVISOR (Davis 81). This was a system initiated by Schlumberger-Doll Research Centre (SDRC) and developed in part together with MIT and Fairchild Lab for AI Research. DIPMETER-ADVISOR proved to be an efficient aid in interpreting log data produced by dipmeter sensors dropped into bore holes. Data of this kind indicate how different types of energy interact with a geological formation. In this way, subsurface geological structures can be inferred. The interpretation of well logs can be very difficult. One of the reasons was (and still is) the rapid development of the well logging discipline and technology itself. In 1984, Schlumberger-Doll commercialized the system.

Due to the success of the system and because Schlumberger-Doll was a significant operator in the exploration business, DIPMETER ADVISOR drew a lot of attention from the industry. Many companies initiated their own R&D activities, both on their own as well as together with laboratories with a standing record of expert systems development. In Europe, Elf Aquitaine became a driving force. Amoco and Schlumberger-Doll locomoted the developments in the U.S. The work on MUD ( Kahn 84 ) carried out at Carnegie-Mellon University was also significant. By the end of the 1980s, the petroleum industry became a major R&D partner not only in terms of expert systems, but also within AI in general. AI-related research and system development covered many disciplines and eventually many types of knowledge-oriented approaches. The petroleum domain became a test bench for all kinds of systems. Because of its early initiatives and great spending the petroleum related work was significant in the development of the whole field of AI and thus also enabling developments in other areas of natural resources. At the World Conference on Expert Systems in Portugal in 1994, a panel session was held that addressed the history and standings of petroleum-related AI. The projection of the work within the petroleum industry could well be a mirror of AI itself. In particular, the exploration and resource management efforts reverberate the change in technology emphasis from a simple rule-based system (RBS) to a blend of case-based reasoning (CBR), fuzzy retrieval, and neural networks. The exploration-oriented work gave important incentives to other types of mineral exploration and in turn to other types of resource management. Today the resource oriented systems within the petroleum industry represent only a fraction of the total. The expansion in this particular area has been less than in other areas. This could be related to the fact that RBS for interpretation and analysis have reached a certain maturity, making it less interesting for further publication. Another reason is of course the fact that many petroleum companies invested heavily very early on and evaluated their returns prematurely. Many disappointments accompanied the real successes. This led to budget cuts and a redirection of attention on the account of the R&D departments.

If we look at the records of exploration systems that pertain to the group of systems that we discuss in this discourse, we have the following list:

  • Source rock evaluation
  • Fossil identification
  • Mineral identification
  • Mineral prospecting
  • Well log analysis
    • Quality control of well logs
    • Well log interpretation
    • Well log correlation
  • Dipmetering
  • Sedimentary environment
  • Classification
  • Basin analysis
  • Play analysis
  • Structural style identification
  • Gridding advice
  • Risk analysis
  • Overall resource assessment
  • Seismics
    • Geophone array design
    • Vibrosis parameter
    • Selection
    • VSP advice
    • Tape format identification
    • Velocity analysis
    • Processing sequence
    • Design
    • Interpretation
  • Remote sensing
    • Sedimentary environment
    • Image processing

Much of the work published on AI applications within the petroleum field has been channeled through the CAIPEP conference organized in Texas and later the EUROCAIPEP conferences in Europe. In 1995, these conferences were merged to form AI Petro. A fair cross-section of both R&D and commercialization can be found in the work of Braunschweig et al. (Braunschweig 95; Braunschweig 96).

The early systems like PROSPECTOR and DIPMETER ADVISOR were true representatives of the rule-based paradigm within the field. Many new systems followed these. Development trends are roughly depicted in Table 1.

The systems discussed represents a selection of systems reported over past years. Log and mineral analysis are recurrent subjects for several initiatives. Proper log analysis is still an issue more than 15 years after the first work on the DIPMETER ADVISOR to the system that Baldwin and colleagues built at the turn of the decade (Baldwin 90)

The continued interest can be explained in two different ways. The first work based on rule-based systems produced grossly inadequate results, so that new techniques had to be tried out. The debut based on rule-based systems was so encouraging that it gave rise to greater ambitions and thus the need to investigate new techniques. The truth probably lies somewhere between. The first RBS was a significant leap forward. Yet, it became clear that it had shortcomings. This gave rise to new AI-related initiatives. With renewed interests in neural networks at the end of the 1980s, this paradigm also became a research candidate for the log analyst and other geologists. In fact, today we see a tendency toward using a hybrid set of techniques to capture and represent different types of knowledge.

In the following paragraphs, we will discuss some of these systems in more detail.


Previous Table of Contents Next

footer nav
Use of this site is subject certain Terms & Conditions.
Copyright (c) 1996-1999 EarthWeb, Inc.. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of EarthWeb is prohibited. Please read our privacy policy for details.