FUZZY CLUSTER ALGORITHM FOR THE AUTOMATIC IDENTIFICATION OF JOINT SETS

The task of identifying and isolating joint sess or subgroups of disco ntinuities existing in data collected from joins surveys is not a triv ial issue and is fundamental to rock engineering design. Traditional m ethods for carrying put the task have been mostly based on the analysi s of plots of the discontinuity orientations or their clustering. Howe ver, they suffer from their inability to incorporate the extra data co lumns collected and also lack in objectivity. This paper proposes a fu zzy K-means algorithm, which has the capability of using the extra inf ormation on discontinuities, as well as their orientations in explorat ory data analysis. Apart from taking into account the hybrid nature of the information gathered on joints (orientation and non-orientation i nformation), the new algorithm also makes no a priori assumptions as t o the number of joint sets available. It provides validity indices (pe rformance measures) for assessing the optimal delineation of the data set into fracture subgroups. The proposed algorithm was tested on two simulated data sets in the paper. In the first example, the data set d emanded the analysis of discontinuity orientation only, and the algori thm identified both the number of joint sets present and their proper partitioning. In the second example, additional information on joint r oughness was necessary to recover the true structure of the data set. The algorithm was able to converge on the correct solution when the ex tra information was included in the analysis. (C) 1998 Elsevier Scienc e Ltd. All rights reserved.