A digital twin-based multiscale framework for predicting full-scale TBM rock cutting performance from miniature point load tests

• A digital twin-based multiscale framework is developed to bridge miniature point load tests with full-scale TBM rock cutting force prediction. • An automated rock parameter calibration procedure is introduced to iteratively optimizes material properties via Newton convergence of experimental and simulated point load tests. • A coupled DDA&4D-LSM is adopted to resolve tool-rock interaction mechanics, enabling accurate prediction of full-scale TBM rock cutting forces. Tunnel Boring Machine (TBM) cutter head performance prediction relies heavily on costly and labor-intensive full-scale rock cutting tests, whereas existing numerical models struggle with limitations in scale resolution, dimensionality, and fracture dynamics. This study introduces a digital twin framework that integrates automated rock parameter calibration—based on small-scale physical tests—with advanced numerical simulations to enhance large-scale TBM cutting predictions. Using point load tests on miniature rock specimens, an inverse analysis algorithm iteratively optimizes material properties, including elastic modulus, tensile strength, and cohesion, by calibrating them against experimental load-displacement curves. These optimized parameters drive a digital twin–based coupled Four-Dimensional Lattice Spring Model (4D-LSM) and Discontinuous Deformation Analysis (DDA) framework to predict full-scale TBM rock cutting forces. Validation against linear cutting machine (LCM) experiments confirms the model’s ability to replicate cutting forces (i.e., normal and rolling) and fracture propagation in jointed rock masses across varying penetration depths. Further application to conical cutter head tests demonstrates the framework’s predictive accuracy, with rolling forces falling within the range of experimental data. By bridging laboratory-scale testing with field-scale TBM operations, this methodology offers a cost-effective digital twin approach for optimizing cutter head design and rock fragmentation strategies, advancing geotechnical engineering through data-driven, high-fidelity simulations.