Abstract: The hydraulic and mechanical properties of rock masses are affected significantly by the discontinuities, and thus it is vital to characterize the development of various discontinuities for the fundamental stability analysis of rock masses. The hydraulic and mechanical behaviors are different for the discontinuities with the same dip direction but different other parameters. Although the conventional method is widely used to classify the discontinuities by the dip direction, this approach is challenging to apply for those discontinuities having the same orientation but different other parameters. Therefore, we propose a new method based on quantum particle swarm optimization algorithm for partitioning multivariate discontinuities data. An objective function is established by the similarity measure of discontinuities. To search global optimal solution, quantum particle swarm optimization algorithm is employed. The newly proposed method classifies the discontinuities by their multivariate parameters, which is also validated by categorizing the simulated data into groups. Finally, the new method is applied to analysis the data of multivariate discontinuities collected from the Songta dam site on the NuJiang River, and there is good agreement with the in-situ measurement.

Key words: rock mechanics, discontinuities, data partitioning, quantum particle swarm optimization algorithm, clustering method