TL;DR
This paper presents a high-fidelity approach to modeling temperature variations in concrete bridge box girders using Finite Element Method (FEM) and environmental simulation software. The study reveals that standard international design codes (AASHTO, Chinese, and New Zealand) can deviate from actual measured temperature gradients by 20% to 40%, particularly during the sensitive construction phase.
What Problem Does This Paper Address?
Thermal effects—driven by solar radiation and diurnal temperature swings—are a leading cause of non-linear stress and cracking in reinforced concrete bridges. While engineers rely on design "norms" or specifications to predict these gradients, these standards are often generalized and may not account for local microclimates or the specific geometry of a hollow box girder.
The challenge is especially acute during construction, where the internal air temperature of the girder's "hollow" section fluctuates in ways that are difficult to measure or predict. This paper seeks to bridge the gap between theoretical design codes and real-world thermal behavior by providing a more accurate simulation framework.
Methods
The researchers employed a multi-stage simulation and validation workflow:
- Environmental Modeling: They used Ecotect software to simulate the internal air environment of the box girder based on local meteorological data (wind speed, solar path, and ambient temperature).
- Finite Element Analysis (FEA): A 3D transient heat transfer model was built using DC3D8 elements. The model solved the Fourier heat conduction equation, accounting for boundary conditions like convection, thermal radiation, and solar absorption.
- Field Validation: To verify the model, the team installed 36 thermometers across three different girder sections (A1, A4, and A6) of a real-world high-speed railway bridge. Measurements were taken between March and June 2018.
- Mathematical Fitting: The resulting temperature-depth data were fitted to power exponential functions to determine the "index $a$," a critical parameter used in design formulas to describe the shape of the temperature gradient.
Datasets
- Field Measurements: A custom dataset of internal and external temperatures from a 32m span reinforced concrete bridge project.
- Meteorological Data: Local weather records including monthly temperature extremes (ranging from below -20°C to above 30°C) and wind speeds (5–15 km/h).
- Design Specifications: The study used the AASHTO LRFD (USA), New Zealand Code, and Chinese Code (JTG D60-2015 and Tb/10002.3-2005) as benchmarks for comparison.
Key Insights and Ideas
- The "Indoor" Air Buffer: The air trapped inside the hollow box girder doesn't just stay at ambient temperature; it acts as a thermal buffer that hits its peak roughly at 2:00 PM, often staying 5°C–6°C higher than the outside air in the morning.
- Code Inaccuracy: The most surprising finding was that existing design specifications are least accurate when external temperatures are high and the difference between internal and external temperatures is small. In these cases, the "fitting line" of the code deviates significantly from reality.
- Web Complexity: While top and bottom slabs follow a standard power exponential curve, the vertical "webs" of the girder are more complex. The authors found that for thin-walled sections, a single exponential function is insufficient, and a two-power exponential function provides a much better fit.
- Sectional Sensitivity: Thinner girder sections (like A6) are more sensitive to environmental changes than thicker sections (like A1), leading to higher temperature gradients.
Key Results
- Simulation Accuracy: The numerical results showed "good agreement" with field measurements, validating the use of Ecotect for predicting internal cavity temperatures.
- Gradient Errors: The study found that the standard "index $a$" used in the Chinese code (typically $a=5$ for vertical gradients) actually varied between 4.17 and 11.21 in real-world conditions.
- Quantified Deviation: The error between measured data and standard design norms was found to be between 20% and 40%. Specifically, the simulated "index $a$" for the top slab of section A1 was 19.4% higher than measured on a spring day, but nearly 68% higher on a summer day, highlighting the seasonal volatility of thermal gradients.
Limitations and Future Work
- Scope of Geometry: The study focused on a specific box girder design for high-speed rail; results might vary for different bridge types (e.g., truss or cable-stayed).
- Material Properties: The model assumes standard concrete properties (C50). Variations in concrete mix or the presence of heavy reinforcement could alter thermal conductivity.
- Future Directions: The authors suggest that this simulation method could be integrated into real-time structural health monitoring systems to provide early warnings for thermal stress-induced cracking during the curing process.
Reference: Liu, H., Wu, S., Zhang, Y., & Hu, T. (2023). Finite Element Simulation of Temperature Variations in Concrete Bridge Girders. Fluid Dynamics & Materials Processing, 19(6), 1551-1572.
Summary updated