Experimental investigation on long-term behavior of prestressed concrete beams under coupled effect of sustained load and corrosion

Abstract

Prestressed concrete beams are widely used in construction, while they may be attacked by the corrosion medium during the service. Past research works show that coupled effect of high stress and corrosion can significantly deteriorate the performance of prestressed concrete structures. This article presents an experimental investigation on long-term behavior of prestressed concrete beams under the coupled effect of sustained load and corrosion. During the accelerated corrosion process, six prestressed concrete beams were subjected to different levels of sustained loads, and time-dependent prestress loss and concrete stains were recorded and analyzed. It is observed that the local corrosion (i.e. pitting corrosion) of the steel strands was significantly affected by the sustained load level, and the coupled effect led to more serious damage on the beams than individual sustained load or corrosion. Bending loads were finally applied to investigate the influence of corrosion on flexural capacities of these beams. It was found that the flexural capacities and failure mode of the beams were significantly influenced by corrosion. Meanwhile, the corrosion has more significant effect on tension strength of strands rather than bond strength.

Keywords

corrosion coupled effect flexural capacity long-term behavior prestress loss prestressed concrete beam sustained load

Introduction

Corrosion is one major cause of long-term deterioration in pretensioned concrete (PSC) structures (Harries, 2009; Rinaldi et al., 2010) and could be more serious than that in reinforced concrete (RC) structures due to the high compressive stress applied through prestressed tendons (El Menoufy and Soudki, 2014). On one hand, corrosion of prestressed tendons leads to the cracking and spalling of concrete cover and reduces cross section of prestressed tendons (Jeon et al., 2019; Li et al., 2006; Zhao et al., 2012); on the other hand, the cracking and spalling of concrete cover provide paths for corrosion medium into steel rebars and promote the development of corrosion (Duffo et al., 2004). In addition, tendon corrosion weakens the bond between the tendon and concrete (Wang et al., 2017), and degraded the tensile strength of strands (Li and Yuan, 2013; Wang et al., 2018), resulting in time-dependent prestress loss. For PSC structures, prestress loss plays an important role in the change of structural performance (Caro et al., 2013), and the loss includes the elastic shorting loss during the construction stage and the long-term prestress loss induced by creep and shrinkage of concrete and relaxation of prestress tendon (Bymaster et al., 2015). The mechanism of long-term prestress losses is complicated, and to reveal the mechanism of long-term losses, a number of experimental and numerical studies have been conducted (Chai et al., 2019; Guo et al., 2011, 2012, 2018; Kim et al., 2012; Saiidi et al., 1998).

Past investigations on the ultimate capacity and serviceability of corroded RC beams (Ballim and Reid, 2003; Dong et al., 2017; Yang et al., 2015; Yoon et al., 2010) revealed that the sustained load level had a significant influence on corrosion initiation and could promote the development of corrosion propagation, and eventually changed the failure mode of corroded beam. Wang et al. (2018) further studied the flexural behavior of bonded post-tensioned concrete beams with corroded strands. As the corrosion level increased, the failure modes of beams changed from concrete crushing to strand fracture, which seriously decreased the ductility of the beams.

However, so far, no investigation has been made with regard to the long-term behavior of PSC beams under the coupled effect of sustained load and corrosion. In this study, six PSC beams were prefabricated in factory and loaded inside a laboratory for 550 days, and the effects of sustained load level and corrosion level of steel strand were compared and analyzed. During the long-term loading, ambient temperature, relative humidity, prestress loss, and strains were monitored to investigate the time-dependent prestress loss of pretensioned concrete beams. The flexural capacities of the beams were tested through incremental loading at the end of 550th day.

Description of experiments

Specimen design

Six PSC beams were prefabricated in factory, and Figure 1 shows the profile and cross section of the beams. Concrete casting was conducted twice for these beams. First, the beams were precast with the C40 concrete and the height of 350 mm. Eighteen days after, secondary concrete casting was conducted at beam top with the C45 concrete and the thickness of 100 mm so as to simulate the RC floor. Deformed HRB400 rebars with the diameter of 18 mm was used at the beam top, and deformed HRB335 rebars with a diameter of 8 and 10 mm were used for erection. Six HRB335 stirrups with a diameter of 8 mm were spaced at 50 mm near two beam ends, and at 100 mm in the other part. Three low-relaxation steel strands with a diameter of 12.7 mm were selected as the prestress tendons and placed at distance of 67 mm from the bottom of the beam. The steel strands with the nominal strength f_ptk = 1860 MPa were tensioned to 1398 MPa prior to first concrete casting, and 9 days after, the steel strands were relaxed to realize the prestress transfer. The six beams were designed with different corrosion conditions (i.e. with or without corrosion) and subjected to different sustained loads (being 0, 50, and 100 kN), as shown in Table 1; note that 50 and 100 kN represent about 30% and 60% of the design ultimate bearing capacity of the beams, respectively. Mechanical properties of concrete and steel were tested according to the Chinese material test standard GB/T 50152-2012 (2012) and are shown in Tables 2 and 3.

Figure 1.

Structural dimensions and layout of steel reinforcements of the beams (unit: mm): (a) layout of steel reinforcements and (b) section 1.

Table 1.

Parameters of sustained load and corrosion (unit: kN).

Beam	Sustained loading	Accelerated corrosion
B-N-0	0	No
B-Y-0	0	Yes
B-N-50	50	No
B-Y-50	50	Yes
B-N-100	100	No
B-Y-100	100	Yes

Table 2.

Mechanical properties of the concrete (unit: MPa).

Pouring part	Design strength	Number of specimens	Compressive strength, average value	Elastic modulus, average value
First casting	C40	3	51.6	33,100
Second casting	C45	3	55.8	34,700

Table 3.

Mechanical properties of steel bars (unit: mm or MPa).

Type of steel	Diameter, $D$	Yield strength, $f_{y}$	Ultimate strength, $f_{u}$	Elastic modulus, $E_{0}$
Steel bar (HRB335)	8	343	462	209,000
Steel bar (HRB335)	10	351	465	210,000
Steel bar (HRB400)	18	415	569	211,000
Steel strand	12.7	1623	1931	196,000

Sustained load and accelerated corrosion

The beams were loaded in a four-point bending configuration. The beams B-N-50, B-Y-50, B-N-100, and B-Y-100 were subjected to sustained loading using the bolted-anchorage system, as shown in Figure 2. Beams B-N-50 and B-Y-50 were combined as a group, and beams B-N-100 and B-Y-100 were combined as a group. During the loading, the bolt forces were monitored through load cells at beam ends and adjusted if needed.

Figure 2.

Coupled application of sustained load and corrosion: (a) schematic representation of the coupled application and (b) beams under loading.

Simultaneously, the six beams were subjected to wet and dry cycles using the direct current (DC) power, as shown in Figure 2(a). There were 7 days in a dry–wet cycle, “wet” for 4 days and “dry” for 3 days. During the “wet” process, the corrosion areas of the specimens were covered with sponge soaked in 5% sodium chloride solution, stainless steel mesh, and plastic film, respectively. Meanwhile, a constant current (150 μA/cm²) was applied between the steel strands in the beams and stainless steel mesh. During the “dry” process, the power supply was turned off and the sponge and stainless steel mesh were removed, allowing the concrete to become drier. The sustained load of all beams remained unchanged during the dry–wet cycles.

Measurement of tendon force

In this study, the time-dependent tendon force was measured through the elasto-magnetic (EM) sensor placed at mid-span of the beams, which is a nondestructive and noncontact test method based on the inherent magnetic-elastic effect in the ferromagnetic material. The change in tension force can be obtained by measuring variations in magnetic permeability of prestressed tendons (Guo et al., 2018). Before using the EM sensors to measure the prestress force, all the EM sensors were calibrated in laboratory. According to the research of Chen and Zhang (2018), the external media such as water, cement grouting, and plastic duct have no influence on the measurement results. The vibrating wire strain gauges (VWSGs) were used in the beams at mid-span to monitor concrete strain, and temperature and relative humidity were monitored through the thermo-hygrometer. The layout of measuring points and the measurement instrument can be seen in Figure 3.

Figure 3.

Measurement instrument and layout of measuring points.

Tests of flexural capacity

After 550 days of loading and corrosion, the anchors of the beams were removed, and the six PSC beams were monotonically loaded under bending, as shown in Figure 4. Three linear variable differential transformers (LVDTs) were used to obtain the displacement at mid-span of the beams. The beams B-Y-0, B-Y-50, and B-Y-100 were monotonically loaded to failure step by step. Considering the safety and range of the loading instrumentation, the beams B-N-0, B-N-50, and B-N-100 were monotonically loaded to 500 kN step by step, and the load increment was 10 kN per step for all the beams.

Figure 4.

Flexural load test (dimension unit: mm): (a) loading system and instrumentation and (b) specimen under test.

Results and discussion

Corrosion and cracking

After the flexural tests, 2-m-long steel strands in the mid-span of the corroded beams were taken out from concrete. All the steel strands were cleaned and measured according to the method outlined in Chinese standard GB/T 50082-2009 (2009). The average corrosion ratios $(η)$ of these steel strands were calculated according to the weight loss, as shown in equation (1)

η = \frac{g_{0} - g}{g_{0}} \times 100 %

(1)

where g₀ and g are the masses of the steel strands before and after corrosion, respectively.

When approaching the failure stage, the steel strands broke suddenly for the beams B-Y-0, B-Y-50, and B-Y-100, as shown in Figure 5. Note that local corrosion (pit) significantly affects the mechanical properties of steel strand (Darmawan and Stewart, 2007; Jeon et al., 2019). The minimum diameters of individual steel wires near the fracture of the steel strands were measured using the micrometer to further characterize the local corrosion, as shown in Table 4, and the minimum diameter can be regard as the representation of the local corrosion damage of steel strand. According to Table 4, the minimum diameter of wires near the fracture was found in beam B-Y-100, followed by beam B-Y-50 and beam B-Y-0. Due to the use of same accelerated corrosion method and the same test environment, the average corrosion ratios of each beam were close. However, the local corrosion damage was significantly affected by the sustained load level. The larger the sustained load was applied, the more serious the local corrosion damage was. For the beam B-Y-100, three wires of the steel strand were fractured during the coupled application of sustained load and corrosion, indicating that the sustained load can promote the local corrosion of steel strands.

Figure 5.

Fracture of steel strands.

Table 4.

Corrosion situation of steel strands (unit: % or mm).

Specimen	Average corrosion ratio	Average diameter of steel wire (edge wire/center wire)	Minimum diameters of individual wires near the fracture
B-Y-0	29.83	3.510/3.619	2.352–3.197
B-Y-50	30.07	3.504/3.613	1.934–3.232
B-Y-100	27.22	3.575/3.685	0–3.465

Although there was no direct contact between the stirrups and the steel strands, the steel corrosion of stirrups at the bottom of the beam was significantly larger than that of the steel strands. In some locations, the corrosion even resulted in the fracture of the stirrups, as shown in Figure 6. The possible reason is that compared with the steel strands, the concrete cover for the stirrup is thinner, and corrosion medium may invade from the corner in two directions; therefore, the corrosion of the stirrups was more serious.

Figure 6.

Corrosion fracture of stirrups.

The distribution of corrosion-induced cracks of the beams B-Y-0, B-Y-50, and B-Y-100 (cracks prior to the flexural loading) are shown in Figure 7. No surface-crack was found in beams B-N-0, B-N-50, and B-N-100. Although the stirrups were seriously corroded, no transversal crack due to sustained load and stirrup corrosion appeared in these corroded beams, indicating that prestress force can effectively restrict the appearance and development of transversal crack, even though serious corrosion (of about 30%) occurred in the steel strands. Since the corrosion of stirrups was more serious at beam corners, the confinement of concrete was weaker at the corners; therefore, the widest cracks in the beams appeared near the corners, as shown in Figure 8.

Figure 7.

Distribution of corrosion-induced cracks: (a) B-Y-0, (b) B-Y-50, and (c) B-Y-100.

Figure 8.

Corrosion-induced cracks of the beam B-Y-100: (a) top surface and (b) profile.

Time-dependent tendon force

Figure 9 presents the ambient temperature and relative humidity during the application of sustained loads, where they were measured manually almost every day except for the weekends and Chinese holidays. The monitored test period lasted for 550 days, including 59 dry–wet cycles. The highest and lowest temperatures in that period were 36.8°C and 2.4°C, respectively, with a mean value of 22.4°C, and the highest and lowest relative humidity were 99% and 44%, respectively, with a mean value of 72%.

Figure 9.

Measured ambient temperature and humidity: (a) ambient temperature prior to sustained loading, (b) ambient temperature during sustained loading, (c) relative humidity prior to sustained loading, and (d) relative humidity during sustained loading.

Figure 10 shows the change in tendon forces prior to the application of sustained loads, where the six beams show similar behavior. Prior to the prestress transfer, the change in tendon forces was relatively stable; while after the prestress transfer, the forces decreased rapidly during the first 37 days and then became relatively stable. At the end of 51st day, the sustained load and corrosion were applied to the beams. The change in tendon force under the coupled effect of load and corrosion can be seen in Figure 11. For the uncorroded beams B-N-0, B-N-50, and B-N-100, the changes of prestress force were similar in general, and the loss was influenced by season change, being faster in summer and slower in winter. For the corroded beams B-Y-0, B-Y-50, and B-Y-100, the tendon force seemed to be stable in the first 300 days. Then, the tendon forces began to decrease significantly. For the beam B-Y-100, the corrosion fracture of the three wires resulted in a sudden drop in tendon force. Meanwhile, there were significant corrosion cracks in B-Y-100, as shown in Figure 8. Note that the corrosion can decrease the bond between steel strands and concrete which promotes the loss of prestress force; besides, corrosion reduces the cross-sectional area of steel strands which increases the tendon force. Table 5 summarizes the total loss of tendon forces. For the uncorroded beams, the maximum total loss of tendon force was found in beam B-N-100, followed by the beam B-N-50 and the beam B-N-0. However, the differences were small, indicating that the sustained load can promote the loss of tendon force, but the effect is relatively insignificant, when no corrosion was combined. For the corroded beams, the maximum total loss of tendon force was found in beam B-Y-100, followed by beam B-Y-50 and beam B-Y-0, showing that the corrosion can promote the loss of tendon force. The total loss of tendon force of the beam B-Y-50 was larger than the sum of the total losses of prestress force of the beam B-N-50 and B-Y-0. Similarly, the total loss of tendon force of the beam B-Y-100 was larger than the sum of the total losses of tendon force of the beam B-N-100 and the beam B-Y-0. This indicated that the beams under coupled loads and corrosion have larger loss of tendon force than the sum of those under single effect of sustained load or corrosion. The difference is more significant for the beams under larger sustained load.

Figure 10.

Change in tendon force prior to the application of sustained load and corrosion.

Figure 11.

Change in tendon force after the application of sustained load and corrosion: (a) B-N-0, B-N-50, B-N-100; (b) B-Y-0, B-Y-50; and (c) B-Y-100.

Table 5.

Total loss of tendon forces at mid-span (unit: kN).

Specimen	Prior to the application of sustained load and corrosion		After the application of sustained load and corrosion
	Before prestress transfer	After prestress transfer
B-N-0	1.3	6.3	2.1
B-Y-0	0.9	7.2	4.7
B-N-50	1.2	7	2.3
B-Y-50	1	8.2	8.9
B-N-100	1.6	7.7	2.7
B-Y-100	0.7	6.3	59.7

Figure 12 presents the development in concrete strains obtained from the embedded strain gauges at mid-span prior to the load application, while Figure 13 shows the values during the application of sustained load and corrosion. Note that the VWSGs were embedded in concrete at the centroid of steel strands; therefore, based on deformation conforming condition, the strain increment of concrete equals the strain increment of steel strands. The strain increments measured by VWSGs only contained the shrinkage and creep strain of the concrete, excluding the relaxation loss of steel strands, so it satisfies the relationship shown in equation (2)

Δ σ_{pl} = Δ ε_{c} E_{p} + σ_{pr}

(2)

where $Δ σ_{pl}$ is the stress increment of steel strand; $Δ ε_{c}$ is the strain increment of concrete; E_p is the elastic modulus of steel strand; and $σ_{pr}$ is the loss of tendon stress due to relaxation.

Figure 12.

Development of strain in concrete prior to the application of sustained load and corrosion.

Figure 13.

Development of strain in concrete during the application of sustained load and corrosion: (a) B-N-0, B-N-50, B-N-100 and (b) B-Y-0, B-Y-50, B-Y-100.

According to Figures 10 –13, similar trends of the monitoring data measured by EM sensors and VWSGs can be found. Based on equation (2), the relaxation loss of prestress tendons was calculated, as shown in Table 6. After the change in tendon forces became relatively stable, the maximum loss of tendon strain due to relaxation was found in beam B-N-100, followed by beam B-N-50 and beam B-N-0. Therefore, it can be concluded that the sustained load can promote the relaxation loss of prestress tendon. Due to the corrosion-induced cracks in beams, the monitored data (measured using VWSGs) could not exactly reflect the strain increment of concrete at the centroid of steel strands. However, the change in strain can be used to evaluate the concrete damage. According to Table 7, the maximum change values of concrete strain were found in beam B-Y-100, followed by beam B-Y-50 and beam B-Y-0. Therefore, the coupled application of sustained load and corrosion can result in larger concrete damage than the sum of individual sustained load and corrosion.

Table 6.

Loss of tendon forces at mid-span due to relaxation (unit: kN).

Specimen	Prior to the application of sustained load and corrosion		Under the coupling effect
	Before prestress transfer	After prestress transfer
B-N-0	1.40	2.37	0.54
B-Y-0	1.10	2.93	–
B-N-50	1.25	3.34	1.09
B-Y-50	1.14	3.54	–
B-N-100	1.47	4.25	1.26
B-Y-100	0.77	2.91	–

Table 7.

Change in concrete strain at mid-span (unit: με).

Specimen	Prior to the application of sustained load and corrosion		After the application of sustained load and corrosion
	Before prestress transfer	After prestress transfer	After the application of sustained load and corrosion
B-N-0	5.44	−204.14	−81
B-Y-0	10.44	−221.68	59
B-N-50	2.78	−190.26	−63
B-Y-50	7.66	−242.24	113
B-N-100	−6.46	−179.14	−75
B-Y-100	3.88	−175.92	198

Flexural capacity

Figure 14 shows the flexural cracks and broken strands near the cracks, where more significant corrosion was found near cracks at pure bending region. Figure 15 shows the load–displacement curves of the six beams. Compared with the uncorroded beams, the corrosion resulted in a significant decrease in flexural capacities of beams B-Y-0, B-Y-50, and B-Y-100. Meanwhile, due to the facture of three wires of the steel strands, the failure of the beam B-Y-100 occurred very early with relatively small deflection. During the flexural capacity tests, the fracture of the steel strands caused the failure of the corroded beams B-Y-0, B-Y-50, and B-Y-100, showing unexpected brittle characteristic. It can be concluded that the corrosion in steel strands strongly affects the flexural capacities and failure mode of the beams; meanwhile, corrosion has more significant effect on tension strength than bond strength, since no significant bond-slip was observed prior to strand rupture.

Figure 14.

Flexural cracks and broken strands near cracks: (a) B-Y-0, (b) B-Y-50, and (c) B-Y-100.

Figure 15.

Load–displacement curves of the beams.

Conclusion

This study experimentally investigates the time-dependent prestress loss and flexural capacity of PSC beams under the coupled application of sustained load and corrosion. Based on the test results, the following conclusions can be drawn:

The local corrosion (i.e. pitting corrosion) of the steel strands was significantly affected by the sustained load level. The larger the sustained load was, the more serious the local corrosion damage was. Prestress force can effectively restrict the appearance and development of transverse cracks, even though serious corrosion (with the corrosion ratio of about 30%) occurred in the steel strands.

According to the measurements of tendon forces and concrete strains, the coupled application of sustained load and corrosion can result in larger total loss of prestress force and concrete damage than the sum of those due to individual sustained load and corrosion. The difference is more significant for the beams under larger sustained load.

The corrosion in steel strands strongly affects the flexural capacities and failure mode of the beams at the corrosion ratio of 30%. During the flexural capacity tests, the fracture of the steel strands caused the failure of the corroded beams B-Y-0, B-Y-50, and B-Y-100, all showing unexpected brittle characteristic. Meanwhile, it can found that the corrosion has more significant effect on tension strength than bond strength.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors thank the support from the Ministry of Science and Technology of the People’s Republic of China under grant no. 2018YFE0206100 and the Postgraduate Research & Practice Innovation Program of Jiangsu Province under grant no. KYCX18_0114.

ORCID iD

Tong Guo

References

Ballim

Reid

(2003) Reinforcement corrosion and the deflection of RC beams—an experimental critique of current test methods. Cement and Concrete Composites 25(6): 625–632.

Bymaster

Dang

Floyd

, et al. (2015) Prestress losses in pretensioned concrete beams cast with lightweight self-consolidating concrete. Structures 2: 50–57.

Caro

Marti-Vargas

Serna

(2013) Prestress losses evaluation in prestressed concrete prismatic specimens. Engineering Structures 48: 704–715.

Chai

Guo

Chen

, et al. (2019) Monitoring and simulation of long-term performance of precast concrete segmental box girders with dry joints. Journal of Bridge Engineering 24(6): 04019043.

Chen

Zhang

(2018) EM-based monitoring and probabilistic analysis of prestress loss of bonded tendons in PSC beams. Advances in Civil Engineering 2018: 4064362.

Darmawan

Stewart

(2007) Effect of pitting corrosion on capacity of prestressing wires. Magazine of Concrete Research 59(2): 131–139.

Dong

Zhao

Wang

, et al. (2017) Crack propagation and flexural behaviour of RC beams under simultaneous sustained loading and steel corrosion. Construction and Building Materials 151: 208–219.

Duffo

Morris

Raspini

, et al. (2004) A study of steel rebars embedded in concrete during 65 years. Corrosion Science 46(9): 2143–2157.

El Menoufy

Soudki

(2014) Flexural behaviour of corroded pretensioned girders repaired with CFRP sheets. PCI Journal 59(2): 129–143.

10.

GB/T 50082-2009 (2009) Standard for test methods of long-term performance and durability of ordinary concrete. Available at: http://www.lancarver.com/UpFiles/pdf/2014-04-08/040809434663.pdf

11.

GB/T 50152-2012 (2012) Standard for Test Method of Concrete Structures. Beijing: Architecture & Building Press.

12.

Guo

Chen

, et al. (2018) Monitoring and analysis of long-term prestress losses in post-tensioned concrete beams. Measurement 122: 573–581.

13.

Guo

Liu

(2012) Deflection reliability analysis of PSC box-girder bridge under high-speed railway loads. Advances in Structural Engineering 15(11): 2001–2011.

14.

Guo

Sause

Frangopol

, et al. (2011) Time-dependent reliability of PSC box-girder bridge considering creep, shrinkage and corrosion. Journal of Bridge Engineering 16(1): 29–43.

15.

Harries

(2009) Structural testing of prestressed concrete girders from the Lake View Drive Bridge. Journal of Bridge Engineering 14(2): 78–92.

16.

Jeon

Lee

Lon

, et al. (2019) Equivalent material model of corroded prestressing steel strand. Journal of Materials Research and Technology 8(2): 2450–2460.

17.

Kim

Park

, et al. (2012) FBG sensors encapsulated into 7-wire steel strand for tension monitoring of a prestressing tendon. Advances in Structural Engineering 15(6): 907–917.

18.

Melchers

Zheng

(2006) Analytical model for corrosion-induced crack width in reinforced concrete structures. ACI Structural Journal 103(4): 479–487.

19.

Yuan

(2013) Effects of corrosion on bond behaviour between steel strand and concrete. Construction and Building Materials 38: 413–422.

20.

Rinaldi

Imperatore

Valente

(2010) Experimental evaluation of the flexural behaviour of corroded P/C beams. Construction and Building Materials 24(11): 2267–2278.

21.

Saiidi

Hutchens

Gardella

(1998) Bridge prestress losses in dry climate. Journal of Bridge Engineering 3(3): 111–116.

22.

Wang

Zhang

, et al. (2018) Bond behaviour of corroded strand in pretensioned prestressed concrete beams. ACI Structural Journal 115(6): 1803–1812.

23.

Wang

Zhang

, et al. (2017) Simplified model for corrosion-induced bond degradation between steel strand and concrete. Journal of Materials in Civil Engineering 29(4): 04016257.

24.

Yang

Fang

Yuan

, et al. (2015) Effects of reinforcement corrosion and repeated loads on performance of reinforced concrete beam. Advances in Structural Engineering 18(8): 1257–1271.

25.

Yoon

Wang

Weiss

, et al. (2010) Interaction between loading, corrosion, and serviceability of reinforced concrete. ACI Materials Journal 97(6): 637–644.

26.

Zhao

, et al. (2012) Crack shape and rust distribution in corrosion-induced cracking concrete. Corrosion Science 55: 385–393.