Mixture Design Considerations

Shrinkage-Reduction of UHPC for Bridge Deck Overlays

---

Abstract

Bonded concrete overlays are frequently used to rehabilitate concrete bridges. This necessitates that the overlay material develops high bond to the substrate, high cracking resistance,and adequate durability. Typical thicknesses of bonded bridge overlays prepared with conventional concrete (CC) can vary between 50 and 125 mm. However, CC overlays can exhibit limited service life and can significantly increase dead load on the bridge structure. Ultra-high performance concrete (UHPC) is regarded as a promising for use in thin UHPC overlays of 25 to 50 mm. Such novel material can extend the service life of bridge decks and minimize additional dead load. However, UHPC is prone to high autogenous shrinkage and restrained shrinkage from the substrate, which can lead to cracking and delamination of the overlay material. Shrinkage mitigation strategies include the use of pre-saturated lightweight sand (LWS) for internal curing and CaO-based expansive agent (EXC). The LWS was used at 60% volume replacement of the rivers and in the investigated UHPC; the UHPC was proportioned with 70% river sand and 30% masonry sand, by volume. The EXC was used at 0, 5%, and 10%, by mass of binder. The UHPC was prepared with 0.2w/b and 20% and 25% fly ash and silica fume replacements, respectively. Figure1shows the effectiveness of LWS and EXC on total shrinkage, which is considered as the autogenous shrinkage at one day plus the drying shrinkage after 7 days of moist curing. The use of 60% LWS led to approximately 50% enhancement in compressive strength compared to the reference mixture made without any LWS and EXC. However, the incorporation of EXC in UHPC with LWS reduced strength given the reduction in water available for cement hydration, which was consumed by the hydration the CaO. Four UHPC mixtures made with EXC and LWS as well as CC overlay mixture were used to cast thin overlays of 25, 38, and 50 mm onto substrate specimens measuring 2×1×0.15m. The CC substrate was cast 6 months before the overlay placement. Strain gauges, relative humidity sensors,and thermocouples were installed at various locations at the bottom of the overlay materials near the interface with the substrate. In the initial six months,the slab specimens were kept indoor and were then moved outdoors (Rolla, MO). The variations in temperature are shown in Fig.2. Figure3showsthe variations of strain determined at edge locations of the composite slabs for samples cast with 38-mm thick overlays. The strain in the reference UHPC was approximately±100 m/m. The incorporation of 60% LWS resulted in approximately strain 0 to 200 μm/m. Furthermore, the addition of 5% and 10% EXC along with 60% LWS led to strain approximate values of 100 to 300 and 500 to 700 μm/m, respectively, thus indicating that the overlay material was always in compression. Pull-out tests were conducted after approximately 185, 350, and 500days.As shown in Fig. 4,the failure for the CC overlay specimens occurred at the interface, whereas that for UHPC specimens was in the substrate. The bond of the latter system varied between approximately 2 and 2.8 MPa compared to 1.5 MPa for the CC overlay. Similar observation was noted for the 25-and 50-mm thick UHPC overlays. Further work is underway to compare bond strength between the various UHPC mixtures with the CC substrate using specially designed test specimens.

How to Cite:

Teng, L. & Khayat, K. H., (2019) “Shrinkage-Reduction of UHPC for Bridge Deck Overlays”, International Interactive Symposium on Ultra-High Performance Concrete 2(1). doi: https://doi.org/10.21838/uhpc.9738