Long-term extreme dynamic response of suspension bridges: Part II - Case study and discussion of results

The long-term extreme response framework presented in Part I is applied to the Halsafjorden Bridge, a 2000-m main-span suspension bridge planned in Norway’s E39 project. A finite-element model of the structure is combined with measured static coefficients and aerodynamic derivatives for various angles of attack, enabling multimodal coupled buffeting and flutter analyses including parametric effects of turbulence. Three wind scenarios are examined: a baseline climate (100-year return-period mean wind speed of about 46 m/s), a higher-turbulence climate with the same mean wind speed distribution, and a more severe climate with both higher turbulence intensity and 100-year return-period mean wind speed (about 52 m/s). The comparison with classical Davenport’s approach estimates shows that the long-term response associated with a 100-year return period is generally higher, although explicitly including the static response in the long-term formulation reduces the gap. The average parametric effect systematically amplifies torsional response and reduces lateral response, and its influence increases with wind severity: for the dynamic response, variations reach about −4%/+8% (lateral/torsional) in the baseline scenario, about −10%/+16% in the high-turbulence scenario, and about −12%/+23% in the most severe wind scenario. Considering turbulence intensity as an additional environmental random variable further increases the predicted extremes, particularly in torsion. Flutter annual probability remains low (up to 3.6×10⁻⁴), yet it can significantly contribute to the long-term response associated with very high return periods. The study clarifies when classical simplified design assumptions remain acceptable and when more advanced approaches are necessary.