Abstract
This article uses a four-span continuous spatially slanting heterotypic stay cable arch bridge as the background, and establishes finite element models of the main girder using both a single-beam model and a beam-grid model. The calculation results of the two models show that there are certain differences in the deformation and stress of the main girder and arch ribs, but these differences are not significant. Similar projects can refer to the analytical results described in this paper and, considering the actual conditions of bridge engineering, choose and establish appropriate models.
Introduction
In modern structural engineering, simulating the behavior of a bridge under load is a critical step for both construction and long-term structural health monitoring. For these simulations, engineers frequently employ the single-beam method, which is computationally efficient and straightforward to implement. However, this method simplifies the structural properties of the bridge deck into a single line, potentially ignoring how loads distribute across wide web plates and overlooking "shear lag"—a phenomenon where stress is unevenly distributed across a wide cross-section.
As modern bridges feature wider spans and more complex, irregular (heterotypic) geometries, a practical question arises: Does the simpler single-beam method remain sufficient, or do its inherent simplifications significantly skew the safety and deformation calculations of a complex arch bridge? This study addresses this by comparing the single-beam method with the more granular beam-grid method.
Methodology
The researchers conducted a comparative Finite Element Analysis (FEA) using Midas Civil software, focusing on the Echeng City Bridge, a 460-meter continuous skewed-space irregular tied-arch bridge. Two distinct models were constructed under identical load conditions, including self-weight, anchor box weight, permanent loads, and cable forces:
- Single-Beam Model: A streamlined model consisting of 1,298 elements. The bending and torsional stiffness of the main girder are concentrated into a single equivalent beam.
- Beam-Grid Model: A more granular model consisting of 2,373 elements. The main girder's cross-sections are divided into an equivalent spatial framework of longitudinal and transverse grids. This explicitly accounts for shear lag and the proportional distribution of permanent loads across different sections of the deck.
Key Insights and Comparative Results
Macro-level Agreement
For high-level structural checks, the single-beam method proved highly reliable. The overall trends in stress and deformation across the bridge's arch ribs and main girders were remarkably similar between both models. Specifically, the deformation difference in the arch ribs was negligible, averaging only 1% across all measured points.
Localized Discrepancies
The most significant structural insight relates to where the models diverge. The single-beam model assumes the entire width of the bridge deck acts uniformly. However, the grid model reveals that the outer edges of the deck (which are rigidly connected to the suspender cables) behave differently than the inner sections.
- Outer Longitudinal Beams: The difference in calculated deformation was relatively small, ranging from 6% to 9%.
- Inner Longitudinal Beams: The single-beam method significantly underestimated deformation here. The grid model showed deformations 13% to 17% higher than the single-beam model.
This occurs because the inner longitudinal beams rely on the limited stiffness of the transverse connecting beams to transfer their load to the suspenders. The single-beam method lacks the resolution to capture this localized inner-deck deformation.
Conclusion
The choice of modeling technique ultimately depends on the specific requirements of a project. While the single-beam method is sufficient for assessing macro-level stability and arch rib performance, the beam-grid method is necessary for designs featuring very wide decks where inner-beam deformation could be a critical failure point. Designers can use these findings as a benchmark to balance computational efficiency with the required simulation accuracy for complex heterotypic structures.
Summary updated