Application of Rapid-Setting Belitic Calcium Sulfoaluminate Cements in Pretensioned, Prestressed Concrete

Abstract

In recent years, urgency for adopting sustainable materials has come to the forefront of the cement and concrete industries. Belitic calcium sulfoaluminate (BCSA) cement is an alternative cement which has a production energy demand of up to 60% less than portland cement and emits 30-50% less CO2. Although BCSA cement has been used in the U.S. for pavements, bridge decks, floors, and tanks, information on characteristics relevant to prestressed concrete is lacking. The purpose of this research was to establish strand bond, prestress loss, and time-dependent performance characteristics for BCSA cement concrete to increase production efficiency and member performance for precast, prestressed concrete. BCSA cement-based concrete can reach minimum strength for prestress transfer much sooner than portland cement concrete due to inherent rapid hardening properties. These time savings could allow precasters to clear stressing beds more quickly and streamline productivity, but the impact of early loading on long-term behavior is unknown. Mix designs for self-consolidating BCSA cement concrete were developed to provide sufficient workable time at various temperatures while reaching compressive strength required for prestress release in 2-3 hours. Material properties including flexural strength, elastic modulus, creep, and shrinkage were established. Untensioned large block pullout tests were used to characterize bond performance with age and fifteen 20 ft long prestress girders were fabricated and tested for strand transfer length, development length, and long-term prestress losses. Measured strand bond strengths were adequate and transfer and development lengths were shorter than code predictions. The measured material properties were used to model prestress losses which were compared with the values measured from the girder tests. Measured prestress losses were significantly lower than existing equations indicated, while models developed from material property data produced more accurate estimates.