The clustering of discontinuity sets is not always a trivial task, especially when only the pole density diagram is used, the classical method. This process is extremely subjective since the size of the counting circle, the pole overlapping, and the presence of outliers between families make it difficult to define their characteristics. In these cases, it is useful to apply numerical and classical methods together. For that, this article proposes an algorithm based on the Fuzzy K-means method that allows the clustering of the discontinuities without the influence of these factors. The algorithm had its results compared to two fracture sets studied in literature and it has proved its efficiency.
Discontinuity families; Fuzzy K-means; Clustering analysis
