Three-dimensional forward analysis and real-time design of deep tunneling based on digital in-situ testing

• In -situ digital test-based forward analysis method was developed to dynamically update the rock mass parameters and perform subsequent 3D numerical simulations. • GZZ-based 3D numerical results agree better with physical model tests compared with the Mohr–Coulomb, Hoek–Brown, and Drucker–Prager strength criteria. • Stress and deformation evolution mechanisms during deep tunneling are revealed. • Steel arch distortion of a deep tunnel is closely related to σ 2 and σ 3 , and the effect of σ 2 increases with increasing buried depth. • Longitudinal and overall stiffness of the support system should be enhanced in deep tunneling design. The mechanical behavior of a deep rock mass cannot be precisely evaluated with the traditional theory and model used for shallow tunneling. Moreover, the deep rock mass parameters are primarily determined from laboratory rock tests or field displacement-based back analysis. There is still a lack of a real-time, rapid, and dynamic forward analysis method for deep tunneling. In this study, three-dimensional (3D) numerical results based on the Mohr–Coulomb (MC), Drucker–Prager (DP), Hoek–Brown (HB), and generalized Zhang–Zhu (GZZ) strength criteria were compared and analyzed using physical tunnel models with different buried depths. Subsequently, practical models of shallow-, medium-, and deep-buried tunnels were established to investigate the stress and deformation evolution during tunneling, and a digital in-situ test-based forward analysis method was developed to dynamically update the rock mass parameters and perform subsequent 3D numerical simulations. (1) GZZ-based numerical results agreed well with the physical model tests; (2) large differences in the tunnel excavation responses (stress and deformation) between the 3D strength criterion (GZZ) and the 2D strength criteria (MC and HB) indicate the significant effect of the intermediate principal stress ( σ 2 ) on tunnel stability, and these differences are more significant in deep tunnels because of the widely distributed plastic zone; (3) steel arch distortion is closely related to the minimum principal stress ( σ 3 ) and σ 2 of the surrounding rock near the tunnel face, and the effect of σ 2 increases with the buried depth; however, steel arch distortion in the plane strain tunnel section can be equally attributed to both principal stresses ( σ 3 and σ 2 ). This indicates that complex engineering problems cannot be accurately analyzed with the plane strain model, particularly in relation to deep tunnels; (4) overemphasis on the radial stiffness of the support system seems inappropriate, and more attention should be paid to the longitudinal and overall stiffness of the deep tunnel support. The digital in-situ test-based 3D forward method and design theory developed in this study may contribute further to the dynamic construction and design of deep and ultra-deep rock tunnels.