Evaluation of in-situ stress state and hydraulic failure in an unlined pressure tunnel: A case study from Norway

• Calibrated 3D stress model reproduces the Bjørnstokk tunnel failure conditions. • σ₃ locally drops to 1.6–2.2 MPa within weakness zone, enabling jacking & fracturing. • Sub-horizontal joints show σₙ < Pw, making them prone to opening. • Hydraulic failure arises from low σ₃, weak zones, and joint orientation. • Study shows empirical approach miss critical local stress perturbations. Accurate evaluation of in-situ rock stress is critical for assessing the risk of hydraulic failure in unlined pressure tunnels/shafts. This study assesses the 2016 hydraulic failure at the Bjørnstokk Hydropower Plant in Norway, where leakage occurred from an unlined pressure tunnel despite satisfying traditional empirical confinement criteria in the design. Expanding on the established role of three-dimensional (3D) numerical modeling, this study presents an applied, field-validated demonstration of how calibrated 3D stress analysis can reliably predict hydraulic jacking and hydraulic fracturing in geologically and topographically complex settings, thereby enhancing the design reliability of unlined pressure tunnels/shafts. The calibrated numerical model incorporating detailed terrain geometry, mapped weakness zone, and the regional stress reproduced the in-situ stress state scenario along the unlined pressure tunnel at the failure location. The analysis shows that minimum principal stress (σ 3 ), oriented predominantly in vertical direction, locally dropped to values as low as 1.6 to 2.2 MPa within the weakness zone, explaining the onset of hydraulic jacking and fracturing. Normal stress analysis on major joint sets further demonstrated that sub-horizontal joint sets were subjected to normal stresses less than 2.6 MPa, making them critically susceptible to opening under operating pressures. The findings illustrate that hydraulic jacking and fracturing in unlined pressure tunnels results from a combination of low σ 3 , unfavorably oriented discontinuities, and the geological weakness zones. The study highlights the need for integrating site-specific stress measurements, joint-normal analysis, and calibrated 3D modeling into unlined tunnel design, as empirical overburden criteria alone may be insufficient to capture local stress perturbations.