Figure 1 shows the application of the model to a simple case where the cross section of the bar is locally reduced by 10% over a length equal to one bar diameter。 The gauge length is taken as 20 bar diameters。 Plot A shows the assumed stress-strain relationship for the reference undamaged bar。 The plot shows an initial linear-elastic portion followed by a short yield plateau and a strain-hardening phase。 The ultimate tensile strength of the bar is taken to be 1。15 times the yield strength。 Plot B shows results for the damaged case。 Stress is based on the reference cross section。 A comparison of Plots A and B shows no change in the apparent yield strength of the bar as a result of local damage, but the ultimate tensile strength is reduced by 10%。 As code strength procedures are based on yield rather than ultimate tensile strength, however, this reduction would have little effect on calculated section strength。 The major change in mechanical characteristics is in ductility—a property of particular significance for plastic analysis and seismic resistance—which is reduced by approximately 50% in this illustration。

Despite the local damage, apparent yield strength is maintained because the ultimate tensile strength of the steel can be exploited where the cross section is locally reduced。 The corresponding ultimate strains at the locally reduced section develop only over the short length of the pit, and hence have little effect on overall elongation。 Strain capacity at the reduced section is exhausted before appreciable yielding can develop in the remainder of the bar and thus overall elongation at fracture is reduced。

MECHANICAL TESTS ON BARS WITH

SIMULATED PITTING DAMAGE

A series of tests to investigate the influence of local damage on mechanical characteristics of reinforcement were conducted in the Structural Engineering Laboratories at the University of Brescia。 In this part of the study, corrosion damage was simulated by removing a section of bar using a multifluted, hemispherical end mill with a cylindrical shank (Fig。 2)。 This enabled a realistic simulation of the pitting corrosion。 The tests were carried out on deformed B500B bars having diameters of 12, 16, 20, and 24 mm。 The steels used in the tests conformed to draft European regulation requirements。5 Several degrees of section reduction were created using mills of 4, 6, 8, and 10 mm radii。 The proportion of the cross section removed at the most damaged section (on the axis of the milled defect) were 5, 10, 20, 30, 40, and 50% of the nominal area of the bar section。

Reference undamaged and damaged bars were subjected to tensile testing。 The deformation of the bar in the damaged area was measured by two linear variable differential trans-formers (LVDTs) with gauge lengths equal to five times the diameter of the bar。 One transducer was placed in front of the machined defect, while the other was placed behind it on the intact side of the bar。 Elongations plotted herein were obtained from the average of the two measurements。 Load and displacement data were logged by a data acquisition system at a frequency of 1 Hz。

Sample results from the tensile tests are presented in Fig。 3, which shows load-displacement plots obtained from a 12 mm-diameter bar with various degrees of damage machined using an end mill of 8 mm radius。 The reduction of the maximum load is proportionate to the damaged area, while the reduction in the force at yield is slightly less-than-proportional to the section loss。 The main change in mechanical performance,

Fig。 1—Analysis of bar containing single defect。

Fig。 2—Schematic of machined defect geometry。

however, is the significant reduction in bar ductility caused by the absence of yielding out with the damaged area。

The marked reductions in ductility measured on 12 mm-diameter bars are summarized in Fig。 4, which shows the variation in displacement corresponding to a postpeak load equal to 99% of the maximum load as a function of residual cross section at the defect; ductility of bars with 5 and 50% of section loss at the damage is reduced by 30 to 40% and by approximately 80%, respectively。 The diameter of the artificially induced defect does not exert a significant influence on mechanical properties。 The other bar diameters tested show similar results。