S U M M A R Y


The behavior of structural concrete members strongly depends on the bond between reinforcement and surrounding concrete. In pretensioned prestressed concrete members, concreting is done around the prestressing reinforcement. At release, the prestressing force is transferred by bond to the concrete along the transfer length.

Transfer length is an important design parameter. However, the assumptions included at the standards lead to expressions of calculation yield values very heterogeneous and do not consider the increases of the transfer length through time.

The hypothesis that is supported in this thesis assumes that the prestressed concrete member modifies its stress-strain state with time because of the phenomena of shrinkage and creep of concrete and relaxation of strand, what produces unavoidable prestress losses. The aim, therefore, is to analyze the influence of these phenomena in the bond behavior of prestressing reinforcement.

It has performed a review of the literature about parameters that influence bond, testing techniques to determine the transfer length and prestress losses, and expressions that quantify this length and models for predicting such losses.

An experimental program has been developed and a test method (ECADA+) conceived which consists of a revised and improved version of the method ECADA, in order to measure both the initial and long-term transfer length from the control force supported by the strand. Slips strands have been also measured and longitudinal strains of concrete obtained. A comparative analysis of the long-term transfer lengths obtained from the three techniques has been performed, and an additional factor in expressions for long-term transfer length calculation has been proposed.

Additionally, the prestressing force measured by the test ECADA+ and the obtained from strain of concrete were compared and an overestimation of the prestressing force to the method ECADA+ was observed. Therefore, it has been determined an underestimation coefficient of prestress losses which depends on the prestressing of the cross section of the specimen, both for the initial response and as deferred. Finally, prestress losses from the measured parameters using different calculation methods have been estimated, and such losses have been compared with experimental results.