As in tunnels or underground infrastructures, walls and inverts of metro stations can be used as geothermal collectors. The thermo-mechanical behaviour of concrete used in such walls was studied at early age using experimental and numerical approaches. A laboratory-scale reinforced thermo-active concrete block, designed to reproduce diaphragm-wall-type restraint and realistic thermal boundary conditions, was instrumented with embedded vibrating-string extensometers to monitor internal temperatures and strains from casting prior to any operational thermal activation. Semi-adiabatic calorimetry was used to derive the adiabatic temperature-rise curve for calibrating the thermal–hydration model. Simulations performed with CESAR-LCPC using the coupled TEXO–MEXO framework reproduced the measured temperature and strain evolutions at several locations, demonstrating the robustness of the proposed workflow when combined with an effective-modulus approach based on fib Model Code 2010 for early-age stress relaxation. The validated model was then used to investigate scale effects and boundary heat exchange, and to compare CEM III/B and CEM I concretes. Cracking risk was screened using a stress-to-strength indicator and quantified by a volumetric cracking index. A complementary spatial analysis was introduced to characterise the location of tensile overstress zones and the influence of reinforcement. Peak CI values at 24 h under laboratory boundary conditions were about 2.20% for CEM III/B and 2.12% for CEM I, while in-situ boundary conditions reduced CI to about 1.78% and 0.77%, respectively; at later ages, CI remained limited, typically below 1% for the investigated configurations. The results highlight the role of base restraint in controlling the position of critical zones and the influence of reinforcement, including bar diameter, on the intensity of local stress concentrations. Finally, operational thermal cycling representative of geothermal activation produced very small measured deformations, confirming that the structural integrity of the block was maintained at early age and after activation. • Thermo-mechanical behaviour of concrete block was modelled with CESAR-LCPC. • The concrete block is representative of walls and inverts of metro stations. • The model is consistent with strains vibrating-string extensometers data. • Strains depend on elastic, thermal, autogenous shrinkage and creep effects. • The risk of cracking at early age remains low and localised in the block.
Mesbah et al. (Thu,) studied this question.