A novel fractal-order elasto-visco-plastic creep damage model and its engineering-scale creep parameter determination method

The effective prediction and evaluation of the long-term stability of deep-buried tunnels are crucial for tunnel design, construction, and operation. The creep model is key to predicting time-dependent behavior, and the accuracy of time-dependent deformation predictions is determined by the creep parameters. This paper introduces a novel fractal-order elasto-visco-plastic creep damage (FEVPD) model that incorporates long-term strength into the damage evolution equation within the framework of continuum damage mechanics. The model effectively captures the three-stage creep behavior of various rock types and predicts their creep lifespans under different stress levels. The FEVPD model was implemented in FLAC3D using C++. Additionally, in determining the creep parameters of rock at the engineering scale to address the high computational cost of parameter inversion, an improved genetic algorithm was developed with adaptive perturbation, elitism, and dynamic mutation mechanisms. Application to field monitoring data from the Jinping II hydropower station tunnel demonstrated that the FEVPD model improved the prediction accuracy of time-dependent deformation by 32.68% compared to the classical Burgers–Mohr (CVISC) model. The enhanced inversion method also reduced the final error by 26.0% and 22.7% for the FEVPD and CVISC models, respectively, compared with the standard algorithm. Finally, this model was used to predict the long-term stability of the tunnel. The results provide a reliable and efficient framework for modeling and predicting creep behavior in deep rock engineering.