Discrete element framework for modeling tertiary creep of concrete in tension and compression. (September 2018)

• Record Type: Journal Article
• Authors: Boumakis, Ioannis; Di Luzio, Giovanni; Marcon, Marco; Vorel, Jan; Wan-Wendner, Roman
• Is Part Of: Engineering fracture mechanics. Volume 200(2018)
• Page Start: 263
• Page End: 282
• Publication Date: 2018-09
• Subjects: Time to failure -- Rate effect -- Lattice discrete particle model -- Micro-prestress -- Solidification; Fracture mechanics -- Periodicals; Rupture, Mécanique de la -- Périodiques; Fracture mechanics; Periodicals; 620.112605
• DOI: 10.1016/j.engfracmech.2018.07.006
• Languages: English
• ISSNs: 0013-7944

Abstract

Highlights: Modeling tertiary creep requires rate-dependent damage and linear viscoelasticity. Linear viscoelastic model without distinction between tension and compression. Failure times for both compressive and bending tests are numerically reproduced. Rate effect dominates the time to failure over concrete creep. Time to failure curves may follow a sigmoid function. Abstract: In this contribution, a computational framework for the analysis of tertiary concrete creep is presented, combining a discrete element framework with linear visco-elasticity and rate-dependency of damage. The Lattice Discrete Particle Model (LDPM) serves as constitutive model. Aging visco-elasticity is implemented based on the Micro-Prestress-Solidification (MPS) theory, linking the mechanical response to the underlying physical and chemical processes of hydration, heat transfer and moisture transport through a multi-physics approach. The numerical framework is calibrated on literature data, which include tensile and compressive creep tests, and tests at various loading rates. Afterwards, the framework is validated on time-to-failure tests, both for flexure and compression. It is shown that the numerical framework is capable of predicting the time-dependent evolution of concrete creep deformations in the primary, secondary but also tertiary domains, including very accurate estimates of times to failure. Finally, a predictive numerical study on the time-to-failure response is presented for load levels that are difficult to test experimentally, showing a deviation from the simple linear trend that is commonly assumed. Ultimately, two alternative functions for time-to-failure curves are proposed that are mechanically justified and in good agreement with both, experimental data and numerical simulations.