MECHANICAL TESTS ON BARS WITH

ACCELERATED CORROSION DAMAGE

Although simulated damage provides a convenient and controlled way of investigating the effects of section loss, it does not adequately represent the true nature of corrosion attack。 This section of the study reports tests on bars subjected to corrosion attack。 To obtain results within a reasonable timescale, corrosion was accelerated under an impressed current。 Two types of specimens were tested: one set at Heriot-Watt University, Edinburgh (designated as the HW series), and a second at the University of Brescia (designated as the UB series)。 The HW specimens used plain round bars of 16 mm diameter, while the UB specimens used 20 mm-diameter high-yield ribbed bars。 Test bars were cast in concrete before being subjected to corrosion accelerated by anodic polarization。 This produced a mixture of general

Fig。 3—Plots of load versus elongation from tensile tests on 12 mm-diameter bars with 8 mm-radius machined defects of various depths。

Fig。 4—Displacement at peak load for various damage levels and geometries, 12 mm-diameter bars。

and pitting attacks of varying intensities throughout the length of the bar, and although it cannot be claimed that the procedure produced damage that exactly simulated long-term field corrosion, it is more representative than single-machined defects。

Bars in Series HW specimens were cast in the corners of cubic beam end-type bond specimens。 Specimens were conditioned under a cyclic wetting/drying regime of 1 day of wetting by 3% salt solution followed by 6 days at a relative humidity of approximately 70%。 Stainless steel angle sections were placed on the test specimens and connected to the negative terminal of a controlled-current power supply。 Test bars were connected to the positive terminal of the supply。 A conductive foam was placed between cathodic plates and concrete to ensure good connectivity。 Current densities of between 0。01 and 0。05 mA/cm2 were applied during the no-spray part of the conditioning cycle。 A total of 25 bars were tested。

In the six specimens in Series UB, bars were cast in concrete cylinders, and after demolding, were conditioned under an impressed anodic potential while partly immersed in a 5% salt solution。 At the end of the conditioning period, bars were broken out of concrete and cleaned in an acid bath。 The bars were subjected to varying degrees of corrosion。 Some had only superficial corrosion with no significant loss of section, while others underwent severe corrosion with up to 72% of the section lost at the most severely damaged section。

Measurements of corrosion attack

Because of the different levels of corrosion attack induced in the two sets of specimens, different methods were used to measure section loss。 In the less heavily corroded HW series, bars were weighed to determine the average loss of the cross section and average corrosion penetration was calculated。 Bars were also examined in detail, and dimensions of the largest pits were measured by a graduated magnifier and micrometer。 The movable anvil of the micrometer was a conical shape with a point radius of 1 mm to allow it to reach to the bottom of the larger pits。 The area lost at the pit was calculated as π/4 times the depth and breadth of the pit。 Test bars had mean section loss (based on weight loss measurements) of up to 4%, with a maximum reduction at an inpidual pit of 8%。

Residual cross sections of the more heavily corroded UB series bars were measured using a liquid displacement technique。 The test bar was placed in a narrow calibrated measuring cylinder, and a known quantity of liquid was added to the cylinder。 The volume of the test bar could then be determined from the incremental rise in liquid level。 The procedure was repeated to obtain a profile of the cross section along the entire gauge length。