Both the discrete and continuous particle packing models are used to design UHPC, but the influences of a water film covering the particle surfaces on the compactness of the particle system were not considered in these models. In fact, the water film results in a certain distance between solid particles (DSP), which affects the compactness of the particle system, especially for cementitious materials with small particle sizes. In the present study, the mixture design method for UHPC was proposed based on the Fuller distribution model modified using the DSP. Then, the components of cementitious materials and aggregates were optimized, and the UHPC matrices with high solid concentrations were obtained. The results showed that the solid concentration, slump flow, and compressive strength of the UHPC matrix reached 77.1 vol.%, 810 mm, and 162.0 MPa, respectively. By replacing granulated blast furnace slag (GBFS) with quartz powder (QP), the flexural strength of the UHPC matrix was increased without reducing its compressive strength. When the steel fiber with a volume fraction of 1.5% was used, the slump flow, compressive strength, tensile strength, and flexural strength of the UHPC reached 740 mm, 175.6 MPa, 9.7 MPa, and 22.8 MPa, respectively. After 500 freeze–thaw cycles or 60 dry–wet cycles under sulfate erosion, the mechanical properties did not deteriorate. The chloride diffusion coefficients in UHPCs were lower than 3.0 × 10⁻¹⁴ m²/s, and the carbonation depth of each UHPC was 0 mm after carbonization for 28 days. The UHPCs presented ideal workability, mechanical properties, and durability, demonstrating the validity of the method proposed for UHPC design.

Large deformation of surrounding rock is one of the major disasters that cannot be prevented in deep underground engineering. Several arch changes or even landslides may occur if the management is ineffective. In this study, an indoor model experiment is conducted using a similar geotechnical multifunction load simulation device and a surrounding rock strain and displacement monitoring system. The failure evolution process of surrounding rocks and the variation law of displacement and stress of a deep underground roadway under different burial depths are studied. The critical components for preventing large deformations in underground engineering have been identified. Model test results show that the designed deep roadway model loading system can meet the requirements of uniform loading in simulating different deep roadway loads and can satisfactorily realize cascade loading of various surrounding rocks in a deep roadway. The stress and displacement in the surrounding rock can also be efficiently measured with by the designed measuring device. The similarity test of physical simulation indicates that the instability risk of a deep roadway, especially the collapse disaster in the sidewall, is greater with the increase in depth when no support exists. The collapsed area of the surrounding rock gradually increases with the increase in the burial depth of the roadway. In the model test, no evident convergence deformation is observed in the surrounding rock of the roadway when the simulated depth is less than 800 m. The deformation of the sidewall and the arch waist increases with the increase in burial depth.

To address the practical problem of decreased anchoring effect due to the detachment of the exterior bolt head from the cave wall under dynamic loads, this paper developed a partially recoverable displacement energy absorption (RDEA) bolt. On the basis of preliminary static loading performance tests, a field comparative test was conducted on the blast resistance performance of the cavern reinforced by RDEA bolt. The overall damage and dynamic response of the RDEA-bolt-reinforced test section and the conventional steel-bolt-reinforced test section were compared. The research found that under the same test conditions, the macroscopic damage to the conventional steel-bolt-reinforced test section was more severe than that of the RDEA-bolt-reinforced test section. When the scaled distance was 0.93 m/kg^1/3, the ratio of the rebound displacement to the displacement peak at the arch top of the conventional steel-bolt-reinforced test section was 9.09%, while that of the RDEA-bolt-reinforced test section was 31.1%. The energy consumption characteristics of the RDEA bolt were described by the pressure peak value at the third working condition of the arch top. The pressure peak value of the RDEA-bolt-reinforced test section was 76.4%, lower than that of the conventional steel-bolt-reinforced test section. The arch top acceleration of the conventional steel-bolt-reinforced test section was about 1.35 times that of the RDEA-bolt-reinforced test section. The good blast resistance performance of the cavern reinforced by RDEA bolt was reflected from various aspects, such as macroscopic damage and the wall displacement, indicating that the RDEA bolt can not only weaken the effect of explosion load at the cavern location but also enable the reinforced cavern to have a good ability to resist deformation recovery after explosion, thereby having good application prospects.