A comparison of methods for experimentally determining prestress losses in pretensioned prestressed concrete girders

This paper presents a description and comparison of several experimental techniques used to determine the effective prestressing force in pretensioned prestressed concrete girders. The effective prestressing force was determined by three methods: (1) using vibrating wire strain gages that were embedded in the girders during fabrication; (2) load testing the girders to determine flexural cracking and crack reopening loads and then back calculating the losses; and (3) exposing a length of strand, attaching resistance strain gauges on the strands, and flame-cutting the instrumented strands. Several instruments were used to determine the flexural crack initiation and crack reopening loads. These included crack detection gages, concrete surface strain gauges, and LVDTs, as well as visual observations. Use of data from the strain gauges placed at the bottom surface of the girders was determined to be the most effective way of detecting flexural crack initiation and re-opening. Cracking loads determined from visual observation were significantly larger than those determined from the strain gauge data. The back-calculated prestress losses determined from the measured flexural crack initiation and re-opening loads of the girders were significantly larger than those determined from other experimental methods and those predicted by the PCI Committee and AASHTO LRFD methods. Losses determined by the other experimental methods and the predictions correlated more closely with those back-calculated using visually-observed cracking loads. These results indicate that prediction of losses based on the measured flexural cracking and crack re-opening loads using the basic theory of mechanics results in an overestimation of prestress losses. Consequently, girders may undergo flexural cracking and crack re-opening at lower loads than predicted using the basic theory of mechanics.