This is an old revision of the document!
Temperature evolution due to hydration in insulated column
First, affinity hydration model was calibrated to released heat from isothermal calorimetry. Cement CEM I 42.5R from Radotín, Czech Republic was used for this purpose, see the Figure below.
The same cement was used in concrete production thus parameters of the affinity model remained the same. A column 0.5 x 0.5 x 1.5 m was cast in two steps with approximately 3 days time lag between casting. Thermal insulation 0.25 m thick was present on all vertical faces of the column. Bottom and top parts had equal 0.75 m length. Top parts were covered by a foil to prevent water evaporation and excessive heat loss. Figure below shows experimental setup.
Simulations reflected external fluctuation of ambient temperature and casting sequence. During the first simulation, so called blind stage, no experimental results were provided and other parameters of the model were estimated. The second stage calibrated especially heat transfer coefficients and extra induction time to experimental data.
Two computational scales are defined here, see Figure on the right
- Level of cement paste. CEMHYD3D material model predicts the evolution of discrete microstructure on the scale of micrometers and returns liberated heat.
- Structural level. The heat balance equation is solved with finite elements. Several finite elements are grouped together and mapped to one CEMHYD3D instance. Nine OOFEM's instances are used in the simulation.
The Figures below show temperature evolution during concrete hardening and induced out-of-plane stress when considering B3 model for concrete creep. The simulation runs on a left symmetric part of the arch cross-section. Optimal position of cooling pipes is apparent. Note that the cooling turns off after several hours which detaches natural Dirichlet's boundary conditions and changes number of equations. The flat bottom subfigure shows the 2D triangular mesh and the assignment of hydration models to groups of finite elements on the cross-section. The right Figure validates the multiscale simulation with the temperature in the core of the cross-section. Temperature remained below 65oC during summer casting, which was found acceptable.
Animation of temperature field in the column.
Created 11/2012 by Vít Šmilauer. Acknowledgements belong to Jan L. Vítek and K. Chmelíková who conducted experimental part.