Abstract:Identification of discontinuity parameters of rock mass through field investigation of joint sets or subgroups is fundamental to rock engineering design. A fuzzy C-means cluster analysis method based on genetic algorithm for automatic identification of joint sets is introduced. This method eliminates the local optimality disadvantages of fuzzy C-means cluster algorithm and the subjectivity of traditional methods such as pole and contour plots in determination of clustering demarcation. Based on field measured data of joint sets,analysis steps,parameter selection,cluster validity,and determination of dominant direction for identification of the joint sets by the method,are discussed.