Effects of sustained loading and corrosion on the performance of reinforced concrete beams

This paper presents an experimental investigation on the behaviour of reinforced concrete (RC) beams under simultaneous loading and reinforcement corrosion. Corrosion of reinforcements within beams were created by an accelerated method using a 5% NaCl solution combined with a constant impressed current. Three different corrosion durations at 5, 10 and 20 days and four levels of sustained loading at 0, 15%, 30% and 60% of ultimate loading capacity were applied to the beams. Totally 13 RC beams were tested to examine the corrosion of reinforcements, cracking of beams, and structural behaviour of the corroded beams. Test results indicate that corrosion of reinforcements increases with the sustained loading but undergoes an initially increasing rate followed by a decreasing rate. Higher loading level and longer corrosion period are prone to cause the brittle failure of RC beams. Increasing the sustained loading extends the longitudinal crack but not the crack width. The joint effects of sustained loading and corrosion duration, compared to the single effect of either one factor, are more significant on the performance of RC beams in terms of corrosion of reinforcements, failure mode, ultimate loading capacity, and deformation ability. At a higher sustained loading level, beams’ ultimate loading capacity and deformation ability decrease more significantly with the corrosion periods. It is also found that a lower loading increases the flexural stiffness of RC beams, but a higher loading level instead decreases it.


Introduction
Corrosion of reinforcements has been recognized as one major cause of structural degradation in reinforced concrete (RC) structures [1]- [3], especially for those exposed to the marine environment. This is mainly attributed to the reduction of reinforcement section area as well as cracking and/or spalling of concrete induced by the expanded corrosion products [4].
Probabilistic modes considering corrosion of reinforcements have also been proposed to predict the service life of RC structures [5]- [7]. Corrosion variability of reinforcements in the structures is properly considered in their modes, leading to a more accurate predictions for RC structures. Wang et al. [8] reported that the width, density and tortuosity of cracks are the main parameters affecting the chloride diffusion of concrete. Moreover, corrosion of reinforcements causes the bonding deterioration between the reinforcements and the concrete, which significantly affects the safety and service life of the infrastructures [9]. Therefore, it is necessary to estimate the performance of RC structural members subject to different levels of corrosion.
A large number of studies have been focusing on investigating the behaviour of RC beams in the presence of corrosion and loading [10]- [12]. Longitudinal tensile strains on the tensile surface of corroded beams under the service loadings would increase monotonically with the corrosion periods at a decreasing rate [10]. It has been known that the joint action of corrosion and loading would increase the deflection of corroded RC beams [11]. Nevertheless, the joint effect of corrosion and loading on the performance of RC beams could be further studied, such as flexural stiffness. It was previously found that a low-level corrosion could enhance the bond between reinforcements and concrete [13] as well as increase the flexural stiffness of the beam [14]. As the further increase of corrosion, the flexural stiffness of the corroded beams under a constant loading was found to become constant. This is mainly attributed to the secondary longitudinal strains of reinforcements from the expansive corrosion products, which supersedes the increase in bond between reinforcements and concrete [15]. However, it remains unclear how would the flexural stiffness of RC beams be affected when the loading is further increased.
Corrosion rate of reinforcements within the RC beams would significantly affect the service life of RC structural members. Currently, there are controversial findings regarding the corrosion rate of beam reinforcements subject to various levels of sustained loading. Yoon et al. [16] stated that the corrosion of RC beams would increase at an increasing rate under high levels of sustained loads. While other researchers (e.g., Liu and Weyers [17]; Weyers [18]) found that the corrosion rate of reinforcements would decrease as the corrosion level increased.
Differently, Zhe et al. [19] reported that the corrosion of reinforcements within RC beams developed in a stochastic manner. Therefore, the effect of sustained loading on the corrosion rate of reinforcements within RC beams needs further study.
Existing studies demonstrated that crack width on corroded beams under sustained loading was wider than that without loading. However, most studies (e.g., Zhu et al., [12]; Du et al., [20]; Yin et al., [21]) focused on studying the cracking behaviour of RC beams after corrosion process. Cracks would propagate on the corroded beams subject to a constant loading as the corrosion period increased. There has been limited research investigating the development of cracks for the RC beams with different corrosion levels. The effect of sustained loading on the cracking behaviour of corroded RC beams remains unclear.
The objectives of this research lie in that: (1) to investigate the effect of sustained loading on the corrosion rate of reinforcements within RC beams; (2) to characterize the cracking behaviour of corroded RC beams under multiple sustained loadings and corrosion periods; and (3) to analyse the joint effects of sustained loading and corrosion on the flexural behaviour of RC beams, specifically, to study the flexural loading capacities and load-displacement response of the corroded beams. The research findings would provide insights on the joint effects of sustained loading and corrosion levels on corroded RC beam's behaviour, especially at high loading levels (i.e., 30% to 60% of ultimate loading capacity).

Specimens
Totally 13 RC beams with the same reinforcement details and concrete strength were prepared for experimental tests. Each beam had the cross-section of 120 mm × 200 mm and the length of 1,700 mm. The beam section was reinforced with two T12 in the tension zone and two R8 in the compression zone. T12 is the HRB335 high strength steel bars with the nominal diameter of 12 mm while R8 is the HPB235 round steel bars with the diameter of 8 mm. The longitudinal reinforcements in the tension zone extended beyond both ends of the RC beam, and the extended portion of reinforcements were polished, bonded with copper wires, and sealed by epoxy resin. Stirrups R6.5 were provided along the beam at spacing of 80 mm except the constant moment zone in the middle of beam. R6.5 is the HPB235 round steel bars with the diameter of 6.5 mm. The clear span of each beam was 1,500 mm. The dimension and reinforcement details of the beam are shown in Figure 1. The C40 concrete was used to cast the RC beams. In the concrete mix, the ordinary Portland cement 42.5 was used. The natural gravel with the maximum size at 20 mm was adopted as the coarse aggregate and the river sand with the fineness modulus at 2.6 was used as the fine aggregate. To guarantee the same concrete strength, all beams were cast by one batch of readymix concrete and were cured under the same condition. The HRB335 high strength steel bars were used as the longitudinal reinforcements while mild steel bars were adopted as the stirrups.
The specimens are exposed to different sustained loadings and corrosion conditions. Four levels of sustained loadings at 0, 15%, 30% and 60% of ultimate loading capacity of the beam were adopted. Under each loading level, three levels of corrosion durations at 5, 10 and 20 days were applied to the beams. In addition, one RC beam without sustained loading and corrosion exposure was included as the control specimen. Further detailed experimental programme can be found in Table 1.

Corrosion setup
The test setup for the beam consists of a sustained loading system and an accelerated corrosion system as illustrated in Figure 2. The load was applied through a mechanical jack installed between the specimen and the reaction frame. A load cell was installed to monitor the loading level for each beam. After reaching the loading level for each beam, the sustained loading level was maintained through adjusting the jack during the corrosion process. The accelerated corrosion system comprises of a direct current (DC) source and a salt water spray cycling. The DC source supplied the maximum voltage and current of 30 V and 3.0 A, respectively. The current applied on each RC beam was controlled by a separate DC source. The bottom reinforcements were connected to the positive electrode of the DC supply.
A stainless steel plate with the length of 1,450 mm was set at the bottom of the RC beam and connected to the negative electrode. The corrosion current was controlled by the steady flow.
The current density of the RC beam reinforcements was calculated at 0.01 mA/mm 2 according to the Faraday's law. The 5% NaCl solution was sprayed to the RC beam to simulate wet-dry cycling, which was designed to provide the RC beams with corrosion environment. Two PVC pipes were installed at both sides of the beam along the longitudinal direction. Sprinklers were placed at 100 mm spacing along the PVC pipes, which were connected to the pump through hoses. As shown in Figure 2(b), the accelerated corrosion system could form the water circulation to provide controllable spraying. The wet-dry alternate condition was set in the 24hour wetting followed by 24-hour drying cycle.

Measurement of reinforcement corrosion
Two measurement methods were employed to determine the corrosion levels of reinforcements, including cross-section area loss and mass loss of reinforcements. According to JTJ270-98 Testing Code of Concrete for Port and Waterway Engineering [22], the corroded reinforcements were first immersed in acid solution for 30 mins before being sent to alkali solution for another 10 mins. Subsequently, reinforcements were wetted followed by the drying process. The samples were weighted at the precision level of 0.01g. The mass loss of reinforcements is calculated by Equation (1).
where W and Wc are the mass of reinforcements before and after the corrosion, respectively.
For cross-section area method, the area of cross-section was calculated based on the averaged diameter of six measurements for the corroded reinforcements. Cross-section area loss of reinforcements can be calculated by Equation (2).
where F and Fmin are the cross-section areas of reinforcements before and after the corrosion, respectively.

Figure 3. Flexural testing setup
Upon the completion of the corrosion process, the corroded RC beams were washed before being placed in the flexural testing setup shown in Figure 3. The four-point load was applied to the beam. The maximum loading capacity of the frame was 5,000 kN. The applied load and deflection at the mid of the beam were collected for analysis.

Corrosion of reinforcements
The corrosion of reinforcements within the beams subject to different sustained loadings and corrosion periods was examined. Figure 4 shows the typically corroded reinforcements taken from different beams. It was observed that the outer surface of the reinforcement (close to concrete cover) exhibited more severe corrosion as compared to the inner surface of reinforcement. This would promote the formation of corrosion caused by the oxygen concentration between the outer and the inner surfaces of reinforcements, and also accelerate the corrosion process [23]. Expansion of reinforcements due to the accumulation of corrosion products would subsequently induce cracks on the beams. For beams under a constant sustained loading (e.g. 60% of ultimate loading capacity) as seen in Figure 4, corrosion levels of reinforcements increased with the corrosion period. At the advanced stage of corrosion, the pitting corrosion could be found at a certain location along the reinforcement. When the corrosion level further increased, the significant reduction in the cross-section area of reinforcements due to the pitting corrosion was found (e.g. specimen BL-60-3). The corrosion level of reinforcements was quantitatively assessed in terms of mass loss and section area loss as shown in Table 2. Two types of corrosion were identified for the reinforcements, including the surface corrosion and the pitting corrosion. For the specimens corroded on surface only (e.g. specimens in series BL-0 and BL-15), ratio of mass loss to section area loss was approximated to be 1.0, which indicated that both mass loss and section area loss of reinforcements could reflect the corrosion level of reinforcements. When reinforcements were corroded in a uniform manner, both methods were effective to measure the corrosion states. For the specimens corroded with pits, however, the ratio of mass loss to section area loss was much smaller than 1.0. Thus, mass loss of reinforcement could not appropriately represent the corrosion levels as the occurrence of pitting corrosion significantly affected mechanical property of reinforcements due to the reduction of section area. Therefore, it is recommended to examine the corrosion level of reinforcements through section area loss especially for reinforcements with pitting corrosion.  Corrosion rate of reinforcement varied slightly when increasing the loading level from 0 to 15% of ultimate loading capacity of the beam. However, a sharp increase in section area loss was observed when increasing loading level from 15% to 30% of ultimate loading capacity. The corrosion rate of reinforcements increased with the sustained loading if the load was below 30% of ultimate loading capacity. However, a further increase of sustained load to 60% of ultimate loading capacity would not continuously increase the corrosion rate of reinforcements. This was probably attributed to the constant width of cracks when increasing the sustained load to 60% of ultimate loading capacity, which limited the ingress of moisture and oxygen to the reinforcements [24]. It indicated that the RC beams subject to a medium level of loadings (e.g. 15% or 30% of ultimate loading capacity) experienced the fastest corrosion of reinforcements as indicated by the largest slopes of corrosion rate lines between 15% and 30% of ultimate loading capacity in Figure 5. It means that the corrosion rate of reinforcements of the beams under loading level from 15% to 30% was higher than other levels. The highest corrosion rate of reinforcements initiated at loading level at 15% of ultimate loading capacity. This loading level could be identified as the critical loading for accelerating the corrosion of reinforcements in beams. In general, corrosion rate of reinforcements within the RC beams would first increase with sustained loading to a medium level, but then decrease with further increase of the sustained loading.

Cracking behaviour
Cracking behaviour of the RC beams under various loadings and corrosion periods was characterized into three stages, including corrosion without cracking, corrosion of longitudinal reinforcements and leakage of corrosion products. Figure 6 shows a typical example of the beam under 15% of ultimate loading capacity and different corrosion periods. The first stage started from corrosion till the occurrence of cracks on the beams. During this stage, the stress caused by the corrosion of reinforcements was smaller than the tensile strength of concrete, which in turns to no cracking on the beam as shown in Figure 6 Fig.7 are the crack width, with the unit of mm.

General behaviour and failure modes
Flexural test was performed to examine the structural performance of RC beams after exposing to different levels of corrosion. Three types of failure mode for the corroded beams were identified. In the first type, the corroded beams failed upon the fracture of tensile reinforcements. It occurred to the beams under severe corrosions only (e.g. 29.01% section area loss in reinforcements). After the yielding of the corroded longitudinal reinforcements, the development of flexural cracks in the constant moment zone was negligible as the applied load increased. Once the applied load reached around 90% of ultimate loading capacity, several main cracks formed quickly followed by the sudden fracture of longitudinal reinforcements.
The specimen failed in a brittle manner and possessed the lowest loading capacity, e.g. for specimen BL-60-3 as shown in Figure 8(a). In the second type, the RC beam failed with concrete crushing in the compression zone after the yielding of longitudinal reinforcements, which commonly happened to the beams with limited corrosion. The main cracks were observed in the constant moment zone of the beam after the yielding of longitudinal reinforcements. As the applied load increased, cracks developed in both length and width followed by the crushing of concrete in compression zone as shown in Figure 8

Load-displacement behaviour
The load-displacement relationships at the mid-span of the corroded beams are shown in Figure   9. The control beam without the sustained loading and the corrosion exposure (i.e. specimen BL-0-0) is also included in Figure 9 for comparison. Compared to the control beam, structural behaviour of the corroded beams degraded in terms of ultimate loading capacity and deformation ability. This depends on the corrosion level of reinforcements. Increasing sustained loading and/or corrosion period would decrease ultimate loading capacity and deformation of beams. The beams subject to a corrosion only performed similar to the control beam as seen in Figure 9(a), particularly for the beam experienced a short corrosion period (e.g. specimen BL-0-1). As seen in Figures 9 (b)-(d), the ultimate loading capacity and the deformation ability of beams would be further reduced as the sustained loading and corrosion period increased. Apart from the decrease in ultimate loading capacity, deformation ability decreased gradually for the beams under each loading level. As the RC beam subject to 60% of ultimate loading capacity and 20 days' corrosion period underwent the brittle failure, the failure load and deformation were only 25% and 13.2% as that of the control beam, respectively.
There was a joint effect of sustained loading and corrosion period on the ultimate loading capacity and the deformation ability of RC beams.   to the cracking of the concrete cover [13], which reduced the moment of inertia of the beam section. To maintain the flexural stiffness of the corroded beams, the critical corrosion level for reinforcements was identified at 22.86% section area loss. Further increasing corrosion level would decrease the flexural stiffness of the beam. This critical corrosion level was slightly higher than 11.7% in the study of Almusallam et al. [14].

Conclusions
An experimental study was conducted to investigate the behaviour of RC beams under simultaneous sustained loading and reinforcement corrosion. Afterwards, flexural behaviour of the corroded RC beams was studied. The joint effects of sustained loading and corrosion periods on beam's ultimate loading capacity and deformation were studied. The test results and discussion lead to the following conclusions: (1) Corrosion level of reinforcements should be estimated based on the section area loss rather than the mass loss, especially for those with rust pit. Increasing corrosion period and loading level aggregates the corrosion level of reinforcements within beams. Particularly, loading level higher than 15% of ultimate loading capacity significantly accelerates the corrosion of reinforcements. Reinforcements in the beams under loading level of 15% to 30% exhibit the highest corrosion rate.
(2) Propagation of cracks on the corroded beams are not proportional to the corrosion levels caused by the sustained loading and the corrosion period. Increasing sustained loading would not increase the crack width caused by the reinforcement corrosion, but increase the length of longitudinal cracks.