In modern numerical modelling software, built-in rockbolt models are available to practitioners for predicting rockbolt performance in underground excavation design. The accuracy of rockbolt performance assessment using load, strain, or displacement limits relies on the ability of the numerical rockbolt model to accurately simulate in-situ rockbolt load development and distribution. The determination of the onset of rockbolt yield, the transition to post-yield behaviour, and the impact of rockbolts on rockmass stability are all dependent on appropriately modelled pre-yield rockbolt behaviour. This research investigates the accuracy of two conventional built-in rockbolt models in the industry-common 2D Finite Element Method (FEM) program RS2 for the prediction of pre-yield axial load development and distribution in fully grouted rebar (FGR) rockbolts. Conventional built-in rockbolt models using input parameters from manufacturer specifications and literature review are compared to high-resolution fiber-optic sensing data from laboratory axial pull testing, and model results are seen to poorly represent rockbolt load development and distribution. This research thus proposes and validates a novel, built-in FGR rockbolt model calibration approach using the bond shear stiffness (B.S.S.) parameter in the Elasto-Plastic Interface (EPI) rockbolt model. The calibrated EPI rockbolt model is verified at laboratory-scale and tunnel-scale in a hard, jointed rockmass, and is seen to demonstrate realistic load development and distribution compared to fiber-optic sensing data. Overall, this research delivers practical recommendations for selecting and calibrating a representative, practical built-in rockbolt model for FGR rockbolts, improving upon the demonstrated underperformance of conventional built-in rockbolt models.
Fischer et al. (Wed,) studied this question.