Fracture mechanism and dynamic compressive strength prediction of microwave-assisted stress wave magnetite blasting based on GA-LSTM

The mechanical properties of metal rock mass are important factors affecting the wide application of microwave-assisted stress wave blasting in geological exploration and mining, as well as blasting efficiency and safety. In this paper, the genetic algorithm optimization of long short-term memory (GA-LSTM) model is applied to predict the dynamic compressive strength (DCS) of magnetite. With heating time and heating power as the test variables, the simulated blasting test of microwave-treated magnetite is carried out by using separated Hopkinson pressure bar. The fracture characteristics of magnetite under microwave-assisted stress wave blasting under laboratory conditions are revealed. Results show that under different microwave heating conditions, the decrease rate of DCS of the specimen is 19.95% to 56.67% and the P-wave velocity is reduced by 2.21% to 13.22%. The black crystallized spots near the fracture zone dominated by iron minerals are caused by microwave heating, which becomes the main reason to enhance the fracture dissociation of magnetite. Under the synergistic action of microwave and dynamic load, irregular section and rough zone are produced inside the magnetite, which intensifies the impact and extrusion between particles and promotes the deep breaking of the magnetite.