Damage tolerance and failure analysis of tie-down cables after long service life in a cable-stayed bridge

Highlights

• Protection removal and long-time environment action determined the strands failure.

• Assisted damaged was promoted by the localized dissolution of Zn-coating.

• Stress corrosion tests allowed assessing the damage tolerance of strand-wires.

Abstract

The paper examines the damage tolerance of prestressing steel wire-strands failed after being exposed without protection to environmental action for 30 years as tie-down cables of a cable-stayed bridge. The influence of corrosion pitting, cross-section losses and possible assisted cracking on the tensile behavior of the Zn-coated prestressing steel strand-wires was experimentally determined, as well as their susceptibility to aggressive environments. A damage tolerance assessment was completed by performing fracture tests of as-received wire-samples exposed to an aqueous solution of ammonium thiocyanate while tensile loaded by applying a constant load followed by an increasing load at low strain rate. The fractographic analysis of the wires broken in-service and in-laboratory proved the induced damage to be very similar and provided evidence of environmental damage in some of the service ruptures. The damage tolerance assessment is in agreement with the two types of wire ruptures found in the failed wire-strands. These sequentially occurred over the 30 years of service, with the brittle ruptures of the mostly exposed peripheral wires being due to environmentally assisted cracking and preceding the ductile ones of the remaining wires by mechanical overloading.

Introduction

The main objectives of present research were to determine the damage causes and fracture micro- and macro-mechanisms that induced failure after only 30 years of service of two seven-wire prestressing steel strands in distinct tie-down cables of an urban cable-stayed bridge over one of the most important rivers in the Iberian Peninsula. The influence of damage tolerance of the prestressing steel on the failures was addressed as the nucleus of the research regarding the subsequent structural assessment of the bridge tie-down cables. For the purpose of this research, damage is understood as any externally induced alteration of the geometrical configuration and/or the component materials of a structural component able to weaken the loading capacity.

The high-strength seven-wire strands used in structural engineering for stay-cables and prestressing tendons are generally manufactured from six cold-drawn eutectoid steel wires that are helically wound around a central one. The design tensile loads of these strands are typically limited to 45% of their guaranteed bearing capacity when forming cables for cable-stayed structures [1], [2]. Distinct corrosion protection systems have been applied to the strands for decades, one of them often used consisting of galvanizing the steel wires and covering each strand with a polyethylene sheath filled with grease [3], [4]. However, an increase in partial or catastrophic bridge failures involving cables [5], [6], [7], post/prestressed tendons [8] and stay-cables [9] registered in the last decade shows that these protection measurements are not perennial, which entails the necessity of assessing the damage tolerance of prestressing steel after a prolonged service time, especially when the occurrence of environmentally assisted damage represents a significant risk. Previous research concerning the in-service failure of strand-tendons in various Spanish bridges [5], [6] showed that their rupture occurred in locally unprotected zones due to construction limitations or deficiencies, related with the exposure to aggressive environments that favored stress corrosion, or corrosion-assisted fatigue [5], [6], [7], [8], [9], [10].

The broken tie-down strands were the source of samples for testing and experimental research, which is focused on the damage effects on the macroscopic tensile behavior of the strands after 30 years of service, damage tolerance to corrosion and stress corrosion phenomena, and micro- and macro- mechanisms of failure. The experimental program is aimed at determining the environmentally induced damage mechanisms, as well as the mechanical strength and stress corrosion resistance of the steel wires. To this end, tensile and stress corrosion tests under constant load (SCC-CT) and under slow strain rate (SCC-SSRT) are combined and correlated with fractographic analysis using scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDX) techniques. These concern the fracture and lateral surfaces of the wires broken both in service and in laboratory testing.

As schematically shown in Fig. 1a, the cable-stayed bridge in which the two strand failures occurred structurally consists of a continuous reinforced concrete girder supported by a number of inclined stay–cables attached to two porticoed pylons and to four load-bearing porticoed piers. According to the bridge design, the intermediate caisson pier boxes work as sliding bilateral supports of the girder concerning the vertical displacements. The bilateral support condition was materialized through tie-down cables housed in the piers and able to assure the vertical anchoring of the girder (Fig. 1b). Each tie-down cable, of approximately 40 m length, was a bundle of 12 parallel strands of 15.7 mm nominal diameter. Each strand was made of seven galvanized high-strength strength eutectoid steel wires of 5.3 mm diameter and inserted with grease into a polyethylene sheath of 2 mm thickness. The strands were exposed to the environmental action inside the pile chambers with this protection, but the lower ends emerged from the anchor heads about 0.5 m with the protection being reduced to a superficial grease layer and the Zn coating of the wires, since the polyethylene sheath had been removed for constructive needs related to the anchoring process.

The failures were detected at only two days of difference inside the two intermediate piles, with both of them consisting of the rupture of one strand (Fig. 1a) and being located close to the pile bottom. As shown in the picture taken on-site (Fig. 1b), these areas were covered over time with accumulations of detritus coming from of decaying and dead leaves, birds, and small rodents whose decomposition locally increased the environment aggressiveness. The failures occurred at the lower end of the strands, where the polyethylene sheath had been removed. In one of them, the sheath began at about 30 cm from the rupture and in the other at fewer than 10 cm. Fig. 2a and Fig. 2c show the side of the ruptures belonging to the strand piece that remained attached to the upper anchor head. According to the macroscopic detail given in Fig. 2b the service-rupture patterns of the wires indicate a sequential strand collapse: three of the wires failed first in a non-ductile manner, which suggests some environmental assisted damage process, whereas the other four wires subsequently suffered a quasi-ductile fracture by mechanical overloading. This preliminary interpretation is further verified throughout the experimental program.