Deformation control of surrounding rock in underground caverns with high side wall is always a huge challenge in underground engineering construction, and the deformation and failure of surrounding rock are prominent in the excavation process. Aiming at the deformation and failure of high side wall during excavation of extremely broken surrounding rockmass of a hydropower station, a method of controlling surrounding rock deformation with steel support is proposed. Based on numerical simulation and field monitoring, the effects of different types of steel support structures and surrounding rock support are quantitatively evaluated. The results show that: The circular steel pipe support structure is obviously superior to the composite Ⅰ-steel support structure. The deformation of the surrounding rock of the side wall is effectively controlled after the steel tube support is adopted, the range and depth of the plastic zone are obviously reduced, the stress concentration phenomenon is improved, and the stability of the surrounding rock is significantly improved. The results of field monitoring and numerical simulation agree with each other and verify the effect of steel support. The relevant results can provide reference for the deformation control of the surrounding rock of the underground cavern with high side wall under the condition of extremely fractured rockmass.

Highly accurate microseismic (MS) localization is the basis for rock damage assessment and disaster warning. The engineering background noise mixed in the MS signal (s(ε)) seriously affects the subsequent analysis of the MS signal. A noise reduction method of singular spectral analysis–complementary ensemble empirical mode decomposition–wavelet threshold (SSA–CEEMD–WT) is proposed. The CEEMD, CEEMD–WT, and proposed methods are used for denoising the noisy Ricker wavelet. The signal-to-noise ratio (SNR) of the denoised signal (x_de(ε)) by the proposed method is 56.77% and 37.88% higher than those of CEEMD and CEEMD–WT methods, respectively. Moreover, an adaptive artificial bee colony (ABC) algorithm is applied for MS source (O(h₀, y₀, z₀)) location. The time to quantile difference is introduced as the objective function. The blast positioning test results prove that the proposed method improves the positioning accuracy of particle swarm optimization (PSO) algorithm and simulated annealing PSO (SA-PSO) algorithm by 44.12% and 47.64%, respectively. The MS positions of underground caverns reveal that the calculated clusters of MS events using the adaptive ABC algorithm are more concentrated at the structural plane and appearance deformation failure location and in good agreement with field survey and routine monitoring data.