Abstract
This paper presents a sine series solution for horizontally composite multi-span curved beams accounting
for both tangential and radial interface slip. The proposed method neglects the effects of shear deformation, sectional
warping, vertical uplift at the interface, while satisfying the assumption of linear elasticity. Based on the principle of
minimum potential energy, the governing equations and boundary conditions are derived with vertical displacement,
torsional angle, and interface slips in two directions as fundamental unknowns. The governing equations are further
transformed into a formulation involving two types of variables, namely vertical displacement, torsional angle, axial
force, and in-plane bending moment. Based on this formulation, the solution for simply supported single-span
horizontally composite curved beams is obtained. Using the superposition principle, the sine series solution for multi
span horizontally composite curved beams is then established. Two numerical examples are provided to validate the
accuracy of the proposed solution and to investigate its convergence behavior. The results demonstrate that the
proposed solution is computationally efficient, requires low computational effort, and yields high accuracy, thereby
offering a fast and effective tool for the preliminary design of multi-span horizontally composite curved beams
considering interface slip.
Key Words
multi-span curved composite beam; parameter analysis; semi-analytical solution; sine series;
slip; superposition principle
Address
Xuxi Qin:College of Traffic, Jilin University, Changchun 130025, P.R. China
Ting Zhang:College of Traffic, Jilin University, Changchun 130025, P.R. China
Guojin Tan:College of Traffic, Jilin University, Changchun 130025, P.R. China
Ruiyu Gao:College of Traffic, Jilin University, Changchun 130025, P.R. China
Weida Wang:College of Traffic, Jilin University, Changchun 130025, P.R. China
Abstract
Soil Steel Bridges combine the advantages of arch action, corrugations and soil's load carrying ability
and design is often constrained to empirical methods due to inadequate quantification of their behaviour. This study
identifies the effect of geometric parameters through experimental investigation and 3D finite element study. The
specimen embedded in sandy soil was tested to failure and both steel structural response and soil pressure were
obtained during both backfilling and loading stages. The strain readings demonstrated their ability to transfer loads
without undergoing failure of soil backfill. Soil pressure variations revealed that stresses transfer efficiently through
the confined soil above the structure corrugation. Parametric studies investigating soil cover depth, boundary
proximity and corrugation geometry were conducted. The studies demonstrated that increasing the soil cover depth
can enhance the ultimate load capacity simultaneously reducing vertical deflections. The optimum cover depth range
was obtained at the cover to span ratio range of 0.3 to 0.35. Boundary effects were found to diminish beyond a
critical width-to-span ratio, beyond which load capacity stabilized toward free-field conditions. Comparative analysis
of different corrugation shapes revealed that sinusoidal profiles provided more uniform stress distribution and lower
peak stresses than trapezoidal sheets. Comparison of bending moment and thrust variations highlight the inaccuracies
of using 2D large strain analysis for these structures or plane arch equations. The conclusions derive suggestive
ranges for geometric parameters and illustrate their structural application as load transferring underground structures.
Key Words
corrugated steel; finite element analysis; parametric study; soil-steel bridge; soil-structure
interaction
Address
Aswin Chembrakuzhi Perikathra:Department of Civil Engineering, National Institute of Technology Tiruchirappalli, Tamil Nadu, India - 620015
Raghavan Ramalingam:Department of Civil Engineering, National Institute of Technology Tiruchirappalli, Tamil Nadu, India - 620015
Abstract
Concrete-filled steel tubular (CFST) members are widely used in civil and marine engineering owing to
their high structural efficiency. However, accurate prediction of their ultimate strength remains challenging due to the
complex interactions among geometric properties, material characteristics, and loading conditions. In this study,
machine learning (ML) techniques are employed to predict the ultimate strength of circular CFST members under
both concentric and eccentric loading. An extensive database is compiled from the literature, comprising 2186
concentrically loaded and 545 eccentrically loaded test results. Six ML algorithms - SVR, KNN, ANN, DT, RF, and
XGBoost - are developed using both basic input variables and code-based intermediate parameters. The results
indicate that the ML models consistently outperform conventional design code formulae, with substantial reductions
in prediction error. Among the models, XGBoost exhibits the best overall performance, achieving R2 values generally
above 0.92 for both concentric and eccentric loading conditions. The corresponding reduction in RMSE reaches
approximately 30-70% compared with code-based predictions. The effectiveness of input features is found to depend
on both the algorithm and loading condition, with code-based intermediate parameters further improving model
performance in some cases. SHAP analysis shows that geometric parameters dominate the prediction process, with
the outer diameter, D, and composite section area, Asc, identified as the most influential features. Overall, the
proposed ML models provide an accurate and flexible alternative to traditional design approaches, with strong
potential for practical application in CFST strength prediction.
Key Words
CFST; input features; machine learning; ultimate strength; XGBoost
Address
Dongqi Jiang:Department of Protective Engineering, School of Safety Science and Engineering, Nanjing University of
Science and Technology, 200 Xiaolingwei St., Nanjing 210094, China
Jiaxin Fan:Department of Protective Engineering, School of Safety Science and Engineering, Nanjing University of
Science and Technology, 200 Xiaolingwei St., Nanjing 210094, China
Hongjian Du:Department of Civil and Environmental Engineering, National University of Singapore, 117576, Singapore
Yuexin Zhong:Department of Protective Engineering, School of Safety Science and Engineering, Nanjing University of
Science and Technology, 200 Xiaolingwei St., Nanjing 210094, China
Xiao Hu:Department of Protective Engineering, School of Safety Science and Engineering, Nanjing University of
Science and Technology, 200 Xiaolingwei St., Nanjing 210094, China
Abstract
The seismic performance of steel-reinforced concrete (SRC) columns subjected to reinforcement
corrosion remains insufficiently quantified, particularly with respect to system-level degradation mechanisms. This
study presents an experimental investigation on four SRC column specimens with corrosion levels ranging from 0%
to 15%, including one uncorroded reference specimen. Accelerated electrolytic corrosion was applied to the
longitudinal reinforcement and stirrups, followed by quasi-static cyclic loading tests under a constant axial load ratio
of 0.8. The test results indicate that progressive corrosion leads to pronounced nonlinear degradation of seismic
performance. When the average reinforcement mass loss reached approximately 15%, the peak lateral load capacity
decreased by about 34%, accompanied by a substantial reduction in energy dissipation capacity. Despite severe
deterioration of the reinforced concrete shell, all specimens exhibited flexure-dominated failure modes. The
embedded steel section effectively preserved post-peak deformation capacity and residual stiffness, resulting in a
clear decoupling between strength degradation and ductility retention.To quantitatively assess corrosion-induced
capacity loss, a preliminary multi-parameter assessment framework was developed by integrating reinforcement
corrosion ratio, crack length density, and average crack width using a PCA–MLR approach. Although based on a
limited experimental dataset, the proposed framework demonstrates the feasibility of combining corrosion and
cracking indicators for system-level seismic performance assessment of SRC columns.
Key Words
damage assessment; rebar corrosion; seismic performance; src column
Address
Juan Zhao:School of Urban Construction and Ecological Technology, Shanghai Institute of Technology,
Shanghai 201418, China
Shiao Wu:School of Urban Construction and Ecological Technology, Shanghai Institute of Technology,
Shanghai 201418, China
Dazhu Hu:School of Urban Construction and Ecological Technology, Shanghai Institute of Technology,
Shanghai 201418, China
Jinsheng Li:School of Urban Construction and Ecological Technology, Shanghai Institute of Technology,
Shanghai 201418, China
Caiyan Zhang:School of Urban Construction and Ecological Technology, Shanghai Institute of Technology,
Shanghai 201418, China
Zhiping Wu:School of Urban Construction and Ecological Technology, Shanghai Institute of Technology,
Shanghai 201418, China
Abstract
The massive accumulation of phosphogypsum (PG) poses significant environmental challenges.
However, the inherent brittleness and low elastic modulus of PG limit its direct application in construction field. This
study proposes a reinforced concrete-encased phosphogypsum-filled GFRP tube (CPGGT). This design leverages
the confinement effect of GFRP tubes to enhance the mechanical performance of the PG core, allowing it to partially
replace core concrete of RC conventional columns. An experimental program was conducted to investigate the axial
compressive behavior of CPGGT specimens, considering the number of longitudinal bars, GFRP tube thickness, and
core diameter. Experimental results indicate that the stiffness mismatch causes a stage-wise load transfer: the RC shell
governs the response initially, whereas GFRP confinement is significantly mobilized after the RC shell starts to
degrade (concrete crushing and steel yielding), leading to a distinct failure mechanism. Instead of brittle failure, the
CPGGT specimens exhibited significant ductility capacities. A distinct strain-hardening behavior was observed after
the yield point, particularly in specimens with larger core diameters, indicating a positive size effect. The interaction
between the GFRP tube and the low-modulus PG core effectively delayed the activation of full confinement,
allowing the column to sustain loads through large deformations. A theoretical model was also developed to predict
the stress-strain relationship and calculate the load-bearing capacity. The findings demonstrate that GFRP
confinement effectively transforms the low elastic modulus of PG from a defect into a source of ductility, changing
the failure process and enhancing the structural resilience of the composite columns.
Key Words
axial compression; composite column; GFRP tube; Phosphogypsum; theoretical model
Address
Shu-jie Yao:1)Research Center of Space Structures, Guizhou University, Guiyang, 550025, China
2)Guizhou Provincial Key Laboratory of Green Building and Intelligent Construction, Guizhou Province,
Guiyang, 550025, China
Zhuo-qun Liu:1)Research Center of Space Structures, Guizhou University, Guiyang, 550025, China
2)Guizhou Provincial Key Laboratory of Green Building and Intelligent Construction, Guizhou Province,
Guiyang, 550025, China
Ze-hua Wang:1)Research Center of Space Structures, Guizhou University, Guiyang, 550025, China
2)Guizhou Provincial Key Laboratory of Green Building and Intelligent Construction, Guizhou Province,
Guiyang, 550025, China
Qiang Wang:School of Civil and Transportation Engineering, Hebei University of Technology, Tianjin 300401, China
Feng Chen:School of Civil Engineering, Central South University, Changsha 410000, China
Ke-jian Ma:1)Research Center of Space Structures, Guizhou University, Guiyang, 550025, China
2)Guizhou Provincial Key Laboratory of Green Building and Intelligent Construction, Guizhou Province,
Guiyang, 550025, China
Abstract
The aim of this study is to experimentally investigate the behavior of cold formed steel (CFS) U and C
profiles under axial load. Axial load tests were applied to samples with different web hole ratios (0%, 20%, 40% and
60%) under normal environment and high-temperature effects. A total of 16 CFS columns were produced in the
experimental program. Another parameter of the study is different sample lengths (300 mm and 500 mm).
Experimental results show that increasing the web hole ratio and sample length negatively affects the axial load
carrying capacity. Due to high temperature, the decrease in elasticity modulus and yield strength limited the
deformation capacity of the material and caused a decrease in the axial load carrying capacity by 34.9%. At the high
temperature, the increase in stress concentration around the hole caused the deformation to be localized and the strain
to decrease by 55.95% in 300 mm long samples and 85.01% in 500 mm long samples. Based on this research, it can
be concluded that the formulas introduced in current design standards tend to overlook the impact of temperature and
hole ratio on the performance of CFS member.
Key Words
Cold-Formed Steel (CFS); global buckling; high temperature; local buckling; web holes
Address
Furkan Kara: Civil Engineering Department, Erzurum Technical University, Erzurum, 25030, Türkiye
Barış Bayrak: Civil Engineering Department, Kafkas University, Kars, 36000, Türkiye
Mahyar Maali:
1) Civil Engineering Department, Erzurum Technical University, Erzurum, 25030, Türkiye
2) Maali Çelik Ar-Ge Danışmanlık, Müh. İnş. Taah. Tarım ve Hayvancılık Company, Atateknokent, 25030 Erzurum, Türkiye
Abstract
The initial geometric imperfections critically influence the load-bearing behavior of Q460 cold-formed
thin-walled channel columns, consequently, the effects of imperfection modes and amplitudes require quantitative
and systematic investigations. This study uses laser scanning technology to capture comprehensive initial geometric
imperfection distributions across 16 specimens, summarizing the imperfections' morphology and amplitude laws. A
finite element method is developed and validated using axial compression test data. Based on functional modeling, a
simulation method for imperfections is employed to accurately determine global, distortional, local, and coupled
imperfections. The imperfection sensitivity of members with varied dimensions is then analyzed using a nonlinear
finite element method. The results indicate that slender members demonstrate a heightened sensitivity to global
imperfections, while stubbier members show greater vulnerability to local imperfections. The reduction in load
bearing capacity due to coupled imperfections typically surpasses that caused by individual imperfections. Members
with smaller cross-sections and higher slenderness ratios are found to be more sensitive to initial geometric
imperfections, whereas those with larger cross-sections and lower slenderness ratios are comparatively less sensitive
to such imperfections. Furthermore, the recommended imperfection modes and corresponding amplitudes are
summarized, and a prediction formula for the simplified imperfection reduction factor is proposed.
Abstract
In the present research an optimization algorithm is introduced which builds upon the classical non
directed random search by integrating quasi-Monte Carlo sampling, and its application in structural design
optimization and handling constraints are discussed. The Enhanced Random Search technique preserves the
robustness of conventional random search, while significantly enhancing convergence by approximating local
extrema through quadratic interpolation. At each step, quasi-random samples are generated to uniformly explore the
search space, while promising regions are identified using local interpolation. Additionally, a second-layer search is
introduced which refines the sampling domain around the current best solution and is particularly useful for highly
nonlinear functions. The performance of this algorithm is evaluated through benchmark numerical experiments and
structural design of a steel spring and a spatial steel truss structure under different load cases. These designs are
subjected to geometry, deflection, and stress constraints, and a penalty method for handling these constraints is
discussed. The results are compared and contrasted with those of other well-established algorithms and methods,
demonstrating the superiority of Enhanced Random Search and its accuracy and efficiency for design optimization of
steel structures. The advantages and limitations of the proposed method are also discussed in detail.
Key Words
global optimization; quadratic interpolation; quasi-Monte Carlo simulation; random search
algorithm; structural design optimization
Address
Mohammad Mahdi Javidan:Global Frontiers of Resilient EcoSmart City, Sungkyunkwan University, Suwon, Republic of Korea
Jinkoo Kim:Department of Civil and Architectural Engineering, Sungkyunkwan University, Suwon, Republic of Korea
Seunghee Park:AI plus K-Construction Resilience Research Center, Sungkyunkwan University, Suwon, Republic of Korea
