ABSTRACT
Successful in-service performance of modular bridge joint systems (MBJS) is critical for the lifetime maintenance and management of bridge structures. These joint systems, often used to accommodate large expansion movements in modern bridges, are assemblies of steel and elastomeric components and exhibit complex response under ambient loading. In particular, the elastomeric components exhibit non-linear hyperelastic andhysteretic response, which contribute to the energy dissipation of the system and significantly affect the fatigue stress cycles experienced by the connection details under vehicular live load. Understanding the dynamic characteristics of the system, including the energy dissipation behavior, is essential for effective service design. Suitable rate dependent material constitutive relationships for the elastomeric components are necessary for accurate simulation of the MBJS response.
Research was performed to quantify the non-linear elastic and hysteretic behavior of sliding springs within MBJS. Material models incorporating nonlinear and time-dependent material behavior were evaluated based on accuracy and ease of use. For consideration of non-linear large-strain elasticity and time and strain-rate dependency, the Bergstrom-Boyce (1997) hysteresis model, or BB model, was considered. This model decomposes the response of the material into two networks acting in parallel: (1) network A, an equilibrium response represented by a hyperelastic spring; and (2) network B, a time-dependent deviation from equilibrium represented by a similar hyperelastic spring in series with a dashpot. The dashpot within the BB model considers the nonlinearity associated with natural reptation motion, i.e. interspersed uncoiling, of polymer chains within the material over time, which leads to the viscoelastic behavior. A mechanistic-phenomenological hybrid strain energy density function developed by Kilian (1981), also called the Van der Waals (VDW) model, was considered for the hyperelastic springs within the BB model.
The best possible set of parameters for the VDW and BB models for the sliding spring material was determined using a code written in MATLAB (2013), a numerical computing environment, which used a genetic algorithm to optimize the material parameters. The code also used commercial FEA software packages ABAQUS (2015) as an analysis engine for solving the BB model, which was readily available.
The resulting simulation compared with the experimental test data is shown in Figure 1. The results indicate that the non-linear, time dependent behavior of the sliding springs within MBJS were typical of reinforced elastomers and can be sufficiently simulated using the BB model. This model is readily available in commercial FEA software such as ABAQUS. Attempts to improve the response of the model near small-strain by considering alternate optimization processes and hyperelastic models are ongoing. Experimental and analytical stress-strain curves for uniaxial compression of the sliding spring. https://s3-euw1-ap-pe-df-pch-content-public-p.s3.eu-west-1.amazonaws.com/9781315207681/cd556cd4-4dcf-4efe-8e29-56fc67b8bfbd/content/fig173_1.tif"/>
