Testing, numerical modelling and design of corroded fixed parabolic steel arches under compression-bending loading combinations

Steel arches exposed to aggressive environments over prolonged periods are particularly susceptible to corrosion, which can markedly impair their structural integrity and reliability. Nevertheless, the mechanical response and residual strength of such corroded arches remain insufficiently understood in the existing literature. To address these gaps, this paper presents a comprehensive experimental and numerical investigation on welded I-section steel arches with local corrosion. Two fixed parabolic steel arch specimens, simulated through machining-induced defects, were fabricated and tested under half-span three-point concentrated loading. The influences of local corrosion and boundary conditions (fixed and pin-ended) on the in-plane failure modes, load-displacement responses, and longitudinal strain development were systematically examined. The test results revealed that local corrosion induced pronounced stress concentrations and triggered premature failure within the corroded region, thereby altering the buckling mode of the fixed arches. Following this, the nonlinear finite element (FE) models were developed and validated against the experimental observations. FE models were subsequently adopted to perform parametric studies, aiming to expand the database of ultimate strengths for corroded fixed arches across a broad spectrum of geometric parameters (cross-sectional dimensions, slenderness ratio, and rise-to-span ratio), corrosion characteristics (location, width, length, and depth), and loading combinations. Finally, the equivalent corrosion depth and bending moment amplification factor were defined and introduced to explicitly quantify the effects of local corrosion and second-order effects, and the controlling section-based design approach was proposed for predicting the in-plane ultimate strength of corroded fixed steel arches. Comparisons with both test and FE results revealed that the modified approach provides accurate and reliable predictions.