Stratigraphic Correlation of Well Logs Using Relational Tree

J. H. Fang, F. Tsay, P. F. Lai

A heuristic waveform correlation scheme of well logs is based on a relational tree matching. Waveforms (well logs) are represented in a data structure known as a relational tree. A relational tree provides a complete description of the contextural relationships, as defined by peaks and valleys of the waveforms.

The correlation scheme consists of a distance-measuring process using all possible peak attributes. First, a distance function is defined for any two nodes in terms of peak attributes. To find the best match for a given node of a given waveform, we measure the distance between the given node and each node from a predefined subset of the second waveform. The closest one is considered to be the matched node. The process is repeated for every node in the waveform.

This quantitative correlation method has been implemented and tested with well logs from the Black Warrior basin, north Alabama. Results showed that the procedure has the capability of handling the thickening and thinning of strata, as well as missing intervals.

AAPG Search and Discovery Article #91043©1986 AAPG Annual Convention, Atlanta, Georgia, June 15-18, 1986.