An Expert System for Qualitative XRD Analysis of Sedimentary Rocks

FANG, J. H., University of Alabama, Tuscaloosa, AL, DARRELL WRIGHT, Shelton State College, Tuscaloosa, AL, and H. C. CHEN and C. L. LIU, University of Alabama, Tuscaloosa, AL

Mineral identification using x-ray powder diffractometry (XRD) requires human judgment and heuristics. Thus, the task is admirably suited for an expert system approach. Expert systems are computer programs which emulate human expertise. The power of an expert system is derived from the knowledge the system embodies, rather than from search algorithms. An expert system helps solve problems for which well-defined algorithmic solutions are difficult to obtain.

We have coded an expert system, XRAYS, to identify minerals via x-ray diffractograms. The system emulates the well-known manual Hanawalt method, thus avoiding the black-box approach of some computer search/match programs. The mineral subfile of the JCPDS file is stored in a database file, from which the Hanawalt groups are created. The expert system then carries out a manual search following exactly the steps prescribed for the Hanawalt method. In the program, both peak positions and intensities are represented by fuzzy numbers. Fuzzy comparisons and fuzzy arithmetic operations are employed in searching for matches. A list of candidate minerals is output in decreasing order of confidence. Other information, such as chemistry, rock type, suspected minerals, etc. can be coded as product on rules, thereby further narrowing the list of candidate minerals. Examples composed of typical mineral suites in sedimentary rocks will be given.

AAPG Search and Discovery Article #91004 © 1991 AAPG Annual Convention Dallas, Texas, April 7-10, 1991 (2009)