Tensile viscous response of Strain Hardening UHPFRC under high restraint and isothermal conditions

Strain Hardening Ultra High Performance Fibre Reinforced Concretes (SH-UHPFRC) have a high elastic limit (around 10 MPa) and high tensile strength (up to 20 MPa) and exhibit significant strain hardening (1 to 4 ‰) under tensile loads. Because of these appealing mechanical properties associated to their outstanding protective properties, SH-UHPFRC are extremely effective materials for strengthening and improving the durability of existing structures with minimized dead weight. However, their cast on site application induces significant eigenstresses due to deformations restrained by existing structural members. Depending on the stiffness of the existing structure, up to more than 90% of the free autogenous shrinkage deformations can be hindered. These eigenstresses can cross the elastic limit, leading to complex couplings at serviceability between SH-UHPFRC viscous response and their damage properties in the strain hardening domain. On the other hand, the optimization of their use at serviceability requires maximising their tensile stress level, even entering the hardening domain. No experimental data is currently available on the tensile response of SH-UHPFRC under full restraint conditions, with loading rates as low as that corresponding to the rates imposed by restrained shrinkage deformations. Furthermore, the limit of applicability of linear viscoelastic models with respect to the apparent elastic limit is unclear. In the present paper, the effects of various restraint conditions on the tensile viscous response of SH-UHPFRC was determined experimentally and modelled using the finite element software DIANA. The necessary data required for the modelling were collected from experimental tests using a Temperature Stress Testing Machine starting at very early age directly after casting, under isothermal temperature conditions at 20°C. Two tests were conducted at an intermediate restraint of 60% under passive stroke control, representative of composite bridge decks and two tests were conducted at 100% restraint under closed loop deformation control, representative of overlays on kerbs for instance, over a period of one month each. Under full restraint conditions, the eigenstresses clearly entered the strain hardening domain of the UHPFRC (up to 15 MPa) for both tests, with a slow evolution over a period of more than one month, only driven by restrained shrinkage. The tests were also accompanied by Vibration Resonance Frequency (VRF) measurements on companion cylindrical specimens to determine the evolution of the elastic modulus with age, and fracture tests on companion dumbbell specimens at different ages. The ageing linear viscoelastic response of the material was represented by an ageing generalised Maxwell model, fitted on the experimental data for moderate eigenstresses (below 25% of the tensile strength). The influence of the scatter of free shrinkage measurements on the fitting of the linear viscous model was studied by means of a sensitivity analysis.