Recieved:

19/10/2026

Accepted:

13/11/2026

Page: 

–

doi:

http://dx.doi.org/10.17515/resm2026-2495st1019rs

Views:

6

Curvature-guarded two-sided Newton and secant bounds for elastic sway buckling of plane frames

Alyaa Abdulrazzaq Azeez1, Furat Y. Al-Ghalibi2

1Ministry of Education, Najaf, Iraq
2Structures and Water Resources Department, College of Engineering, University of Kufa, Najaf, Iraq

Abstract

Trial-and-error determination of the elastic critical buckling load of symmetric plane frames requires repeated stiffness analyses. Al-Sarraf (1979) proposed estimating this load from one or two trial loads using two criteria—a shear-including criterion (Φ) and a shear-free joint-rotation criterion (Φ_θ)—and interpreting the resulting Newton-type and secant estimates as lower and upper bounds. This study re-examines the method across three benchmark structures from the original paper: a two-story portal frame, a five-story single-bay frame, and a truss. Member stiffnesses are derived from the exact solution of the beam-column equation, and the structures are analyzed using a general exact stiffness method. In the verification analyses, the published stability factors, one-trial bounds, and exact critical load of the two-story frame are reproduced to four significant figures. For the five-story frame, the exact critical load (λ_c=0.8099) and multi-trial bounds are confirmed within 0.1% to 0.3% of the published figures. For the truss, the exact load is found to be 1.8793 (a 0.2% difference from the published 1.876), though the original bounds could not be reproduced due to ambiguous half-strut stability-function conventions in the source paper. The findings demonstrate that the two-sided bounding property is not universally valid. A parametric study of 72 single-bay portal frame cases—varying storey counts, stiffness ratios, and trial loads—reveals that the upper estimate can fall below the exact load in multi-story frames, and both estimates can fail near the pole of the no-shear stiffness at the member Euler load. A convexity lemma is presented to explain these behaviors. A proposed curvature test evaluated between the trial load and the Newton estimate proves conservative, ensuring no accepted estimate falls on the wrong side of the exact load across the parametric dataset. Additionally, a Halley-corrected estimator reduces the median error of the lower criterion from 0.53% to 0.07%, though without guaranteeing a strict bound. Ultimately, the results support applying this method to low-rise symmetric frames with moderate stiffness ratios, provided the curvature guard is applied before accepting the estimates as bounds.

Keywords

Elastic critical load; Sway buckling; Stability functions; Two-sided bounds; Newton’s method; Secant method; Halley’s method; Portal frame; Truss

Cite this article as: 

Share This Article
LinkedIn
X
Facebook
journal cover
News & Upcoming Events