A system of coupled transient differential equations governing heat, mass transfer, and pore pressure built up in porous media ( concrete) , subjected to intensive heating, is derived. Water vapor and liquid water are considered separately in the mass transfer formulation. The primary unknowns are temperature, water vapor content, and pore pressure of the gaseous mixture. A nite element formulation and corresponding owchart of computat ions of all required data are presented. The numerical example solved represent s a cross section of a concrete column exposed to re. The domain and time distributions of temperature, pore pressure, water vapor, and liquid water content are presented. Computed pore pressure is higher than those usually reported by ot her analytical studies. The in uence of some initial parameters ( permeability, initial water content, and porosity) on maximum pore pressure is investigated.

A characteristic feature of concrete under uniaxial compression is the development of cracks parallel to the loading direction. A damage constitutive model proposed by Ortiz [Ortiz, M., 1985. A constitutive theory for the inelastic behaviour of concrete. Mech. Mater. 4, can predict the transverse tensile stress responsible for these cracks by considering the interaction between the aggregate and the mortar and the development of damage in the latter. When concrete is exposed to high temperature, as is the case during fire, the failure mode is thermal spalling. In order to improve the prediction of the stresses involved in this failure OrtizÕs model is extended to account for the effects of high temperature. Published experimental results for uniaxial and biaxial compression at high temperatures are used to calibrate the temperature dependence of some of the material properties. The transient creep strain is accounted for by modifying the constrained thermal strain. The stress analysis is coupled with hygro-thermal analysis of heat, mass transfer and pore pressure build-up. The effect of pore pressure on the damage evolution is modeled by applying a body force in the stress analysis module proportional to the pressure gradient. A numerical example of concrete under fire is solved and the computed results are discussed. Spalling is predicted when the damage variable reaches its maximum value of unity. The predicted depth and time of spalling for a range of variation of permeability and initial liquid water content are presented. They are in good agreement with published experimental results.

This paper presents a comprehensive model for simulating coupled heat and mass transfer in concrete when it is subjected to a fire. The model considers not only the transport of dry air, water vapour and liquid water but also the evaporation of liquid water and the dehydration of bound water. It can be applied to multidimensional problems. The governing equations describing this coupled heat and mass transfer phenomenon are solved using finite element methods with a standard Galerkin approximation. Results related to temperature, pore pressure, moisture content, the level of evaporation and dehydration of water within concrete for a standard fire curve are given. The influence of the liquid water transport on the spatial distributions of temperature and pore pressure is investigated through the use of parametric studies. The present results can be used to predict the potential of concrete spalling.

SynopsisThis paper is concerned with matching a fully three-dimensional ͑3D͒ viscoelastic numerical simulation with experimental results obtained using a multi-pass rheometer for both an entry-exit slit flow and a cross-slot geometry. The 3D code simulates the time evolution of steady flows using a multi-mode Rolie-Poly constitutive equation. A test polydisperse polystyrene was characterized for both its linear and non-linear viscoelastic response and the rheological parameters were used for the simulation with matching boundary conditions for the flow. Both overall pressure difference and flow birefringence were compared for the entry-exit slit flow and good matching between simulation and experiment was found for the three different depth geometries tested. The 10 mm depth results ͑depth to width aspect ratio of 6.7:1͒ also showed that a 2D simulation gave a close match to both 3D simulation and experimental results. The flow birefringence fit between experiment and simulation for the cross-slot case, while reasonable, did not match as well as the slit and the results demonstrate that the cross-slot geometry is very sensitive to the extensional behavior of the melt. In addition, examples of the application of the 3D code are given for a monodisperse polystyrene, where the match to experiment proved as good as that of the test polydisperse polystyrene.

No abstract