Grip-Specimen Interaction in Uniaxial Restrained Test

Abstract

Synopsis: Restrained tests are used to evaluate the risk of early age cracking and the cracking sensitivity of concrete mixtures. One test that has become common in recent years is the active uniaxial restrained test in which the length change due to shrinkage is recovered by applying external load to maintain the concrete sample at constant length. The length change is measured by linear variable differential transformer (LVDT), which is used as the control signal in this test. In such tests, the dog-bone geometry is used to grip the ends. To ensure a fully restrained test, the LVDT response to the loads and to shrinkage should reflect the deformation in the concrete sample. Therefore, the grip-specimen interaction should not interfere with the measurement of deformation, and this depends on the instrumentation and how the LVDT is attached to the concrete specimen. Some experiments in the literature have the LVDT attached to the steel grips, a practice vulnerable to possible error due to the interaction between the grip and the concrete. This study considered two methods of attaching the LVDT. First, the LVDT is attached to the steel grips; second, the LVDT is attached to the concrete within the zone of reduced cross-section. The results indicate that attaching the LVDT to the grips results in errant measurement of the shrinkage stress, creep, and elastic strains due to the grip-specimen interaction. The consequences will be false interpretation of fully restrained shrinkage and creep characteristics because the grip-specimen interaction leads to a partially restrained test. The study suggests mounting the LVDT to the concrete sample away from the grips to achieve a fully restrained test. Results for two concrete mixtures with w/c ratio of 0.51 and 0.56 are discussed for both methods of attaching LVDTs.