Stress Relaxation of Steel Tendons Used in Prestressed Concrete Under Conditions of Changing Applied Stress

Abstract

Several methods have been proposed for predicting the total prestress loss in prestressed concrete structures. Inherent in all of these methods is the concept that the applied elastic strain on the steel tendons changes with time due to creep and shrinkage of concrete. A procedure most often used to determine the steel tendon's stress relaxation contribution to total prestress loss is a time iteration technique. The stress relaxation that occurs in each time interval is predicted from an equation describing stress relaxation as a function of time and initial stress level. This equation is normally derived from data obtained from constant-strain tests as per ASTM E328. During the time iteration program, however, the initial stress level is different for each time interval because of changing elastic strains in the steel tendons due to concrete creep and shrinkage. In some procedures under consideration by the prestressed concrete industry, the actual stress on the tendon at the start of each time interval is assumed to be the correct value for initial stress. Total relaxation is then determined by a summation of the relaxation during each time interval. The object of the present research is to investigate the effects of variable applied elastic strain on stress relaxation results during a continuous test and to evaluate procedures for assessing these effects. To accomplish this, stress relaxation tests were carried out on 6.35-mm (0.250-in.) diameter prestressing wire. Results of variable elastic strain application during the test were compared with constant-strain stress relaxation data. Based on these results, a new procedure is proposed for predicting the effects of variable elastic strain on the stress relaxation-time relationship. This new procedure incorporates the effect of stress relaxation in prior time intervals on predicted stress relaxation in a given interval. This is accomplished analytically by defining an “effective initial stress” and an “effective time” to be used in the mathematical expression for stress relaxation during each time interval. This new procedure is shown to be more accurate in predicting actual stress relaxation results under conditions of changing applied elastic strain.