Reliability analysis of steel frame structural systems via the generalized plastic hinge method and SVM approach
Da Lian Bai,Yan Xue Liao,Zhi Yi Wu,Jia Liang Wang,Cun Peng Liu
Abstract
Steel frames are widely used because of their light weight, high ductility, and construction convenience. However, existing approaches for ultimate load-carrying capacity analysis struggle to balance accuracy and computational efficiency. The plastic-zone method provides high accuracy but low efficiency, whereas the conventional first-order plastic-hinge method is computationally efficient but neglects the influence of axial force on plastic hinge development, overestimating the structural ultimate load-carrying capacity. To address these issues, this study proposes a generalized plastic-hinge method based on a generalized yield criterion. An equivalent homogeneous yield function is introduced to construct an element-level load-carrying ratio, which captures the coupled effects of the axial force and bending moment on plastic hinge development while preserving proportional scaling with the external load. This formulation enables efficient and accurate evaluation of the ultimate load-carrying capacity. Subsequently, a small set of ultimate load-carrying capacity samples generated by the proposed method is used to reconstruct an explicit limit-state function via a support vector machine optimized by the particle swarm optimization for the reliability analysis. The numerical results demonstrate that the proposed method can accurately and reliably evaluate the structural ultimate load-carrying capacity. The reliability results obtained from the reconstructed limit-state function are in excellent agreement with those obtained from the Monte Carlo Simulation.
