Rock Blastability Evaluation based on K-Means Clustering and Entropy Weight TOPSIS Method

The distribution of blasting fragmentation in open pit mines has a direct impact on subsequent excavation, transportation, and crushing operations. To effectively control the fragmentation distribution of blasted rocks in different regions of graphite mines, a new model for evaluating rock blastability was developed using the K-means unsupervised cluster learning method and entropy weight TOPSIS evaluation method. Evaluation indexes including rock density, dynamic energy dissipation rate, dynamic compressive strength, average strain rate, and brittleness index were selected. Through entropy weight calculation, it was determined that the degree of rock breakage is most influenced by the brittleness index and least influenced by the average strain rate. The model was then applied to an actual graphite mine to assess its effectiveness. The rock blastability was divided into 10 grades based on this evaluation model. The average particle size of rocks under different grades was calculated and it was observed that as blastability grade increased, so did the average particle size. This finding demonstrates clear classification characteristics and validates the efficacy of our model. From the perspective of rock mass type of graphite ore, the rock explosibility is ranked from easy to difficult: schist, gneiss, granodiorite, mixed rock. Combined with the analysis of microscopic observation results of graphite ore, it can be seen that the lithology changes from schist to mixed rock, and the graphite crystalline content in the rock decreases, and the graphite ore explosibility grade is also higher and higher. Additionally, there exists a linear positive relationship between density/energy dissipation rate/dynamic compressive strength with rock blastability while negative correlation is observed with respect to average strain rate/brittleness index.