Impact Resistance Behavior of Reinforced Concrete Beams Deteriorated due to Repeated Freezing and Thawing

Many existing reinforced concrete (RC) structures constructedmore than 50 years ago now require maintenance.+is is especially true in cold, snowy regions where significant frost damage deterioration of RC structures becomes a severe problem. In this study, falling-weight impact tests were performed to investigate the impact resistance behavior of RC beams degraded by frost damage. An RC beamwas subjected to approximately 900 freeze-thaw cycles to emulate the frost damage before the execution of the impact test. +e surface of the beam was remarkably scaled, and its coarse aggregate was exposed. +e degree of deterioration was evaluated by the distribution of ultrasonic propagation velocity. +e following conclusions were drawn. (1) +e ultrasonic propagation velocity of RC beams was significantly reduced following 872 freeze-thaw cycles. At the upper edge of the RC beam, the ultrasonic wave propagation velocity decreased from 4,000m/s to 1,500m/s in some parts. +is corresponds to a relative dynamic elastic modulus of approximately 14%. (2) +e residual deflection of RC beams with frost damage increased at most by 20% compared with beams without frost damage. +e increase in residual deflection was primarily related to the peeling of concrete at the collision site and the opening of multiple bending cracks. (3) According to the existing residual deflection calculation formula, an increase of 20% in the residual deflection corresponds to a decrease of about 17% in the bending capacity of the RC beam. When the relationship between the degree of frost damage deterioration and the impact resistance of RC structures is defined, existing structures subjected to accidental impact force from rockfalls are safer and can be maintained more efficiently.


Introduction
Many existing reinforced concrete (RC) structures constructed more than 50 years ago require considerable maintenance. In particular, in cold, snowy regions, significant frost damage deterioration of RC structures has become a severe problem. Both the appearance and the structural performance of these structures have degraded (see Figure 1). erefore, some research institutes have been promoting studies on the effects of frost damage on the material performance of concrete [1][2][3][4][5][6][7] and bond behavior with steel rebars [8][9][10][11][12]. Additionally, the load-carrying performance of RC members deteriorated by frost-damage has been examined [8,[13][14][15][16][17][18][19][20]. According to these studies, the relationship between the degree of frost damage deterioration of concrete and the load-bearing performance of RC structures has been identified. However, the number of the experimental studies that have been conducted on the impact-resistant behavior of RC structures has been limited.
In a previous study, the effects of frost damage degradation on the impact resistance of structures were investigated experimentally, such as wall railings and rockfall protection structures, subjected to collisions [21,22]. As a result, it was clarified that the impact resistant capacity of the RC beam decreased, and the failure mode of RC beams likely shifted from bending to shearing owing to the frost damage of concrete (see Figure 2). However, these research results are for small RC beams with cross-sectional dimensions of 60 mm × 100 mm and no shear reinforcement. In contrast, structural members, such as the existing RC beams and columns, are usually provided with shear-reinforcing bars. ey also have a broad cross section, so the fact that the degree of frost damage deterioration differs significantly between the surface and the inner parts of the member is considered. To evaluate the relationship between the degree of frost damage and structural performance, it is necessary to investigate RC members in conditions that resemble those of actual structures.
In this study, a drop-weight impact test of an RC beam having a cross-sectional dimension of 200 mm × 250 mm and a total length of 2 m was conducted to examine the impact resistance behavior of the RC beam that has been degraded by frost damage. Impact tests were performed after repeated freezing and thawing cycles that resulted in significant surface scaling and frost damage to the extent that the coarse aggregates were exposed. e degree of deterioration was evaluated based on an evaluation survey and the distribution of ultrasonic propagation velocity. e determination of the relationship between the degree of frost damage deterioration and the impact resistance of the RC structures will enable safe and efficient maintenance of existing structures subjected to accidental impact force from rockfall. Table 1 shows the proportional mix of concrete used for the RC beams. In this study, the water/cement ratio (W/C) was set at 55% to promote freeze-thaw degradation. e minimum amount of admixture required for concrete casting was added to reduce the amount of air. To minimize the increase in strength from age during the test, early-strength cement was used, and the RC beams were cured for six weeks. Figure 3 shows the dimensions of the test specimen. e specimen used in this experiment was a double-reinforced rectangular RC beam with a cross-sectional dimension (width × height) of 200 mm × 250 mm. e net span length was set to 1.4 m because both ends of the beam significantly deteriorated, as described later in the study. Two D13 (SD345) bars were arranged on the top and bottom as axial reinforcement, and a D6 (SD295A) bar was used as shear reinforcement. Table 2 shows a list of the calculated capacity of the beams. e flexural and shear load-carrying capacities in the table were calculated based on standard specifications of the Japanese Society of Civil Engineers (JCSE) [23]. e compressive strength of concrete without frost damage deterioration was 39.0 MPa. e yield strengths of the rebar were 353 MPa and 370 MPa at D13 and D6, respectively. e table shows that the calculated shear capacity divided by the calculated flexural capacity is 2.61, thus indicating that the design failure ultimately results from bending failure.

Overview of Accelerated Deterioration owing to Successive
Freeze-aw Cycles. Accelerated deterioration owing to freezing and thawing was performed by aerial freezing and thawing in water based on JIS A 1148 [24]. As shown in Figure 4, the minimum and maximum temperatures of the concrete during the freeze-thaw cycle were −18°C and 5°C, respectively. e temperature was controlled by a sensor inserted in the center section of the beam span. e beam used for temperature control was not used for loading tests. e time required for one freeze-thaw cycle was approximately 7 h. A total of 872 cycles were executed. It is known that the concrete subjected to frost damage leads to decreased static elastic modulus and ultrasonic propagation velocity values [1,25]. In this study, the ultrasonic propagation velocity in the beam width direction was measured by the transmission method at the positions shown in Figure 5 to evaluate the degree of frost damage deterioration of RC beams. e beam width was uniform and equal to 200 mm. Table 3 shows a list of test specimens. Specimens classified by N indicate no frost damage deterioration, whereas D indicates frost damage deterioration. H defines the falling height of the weight (m). e height of the weight was set at two levels, namely, H � 0.5 and 1.0 m. In the case of D-H1.0, these items were then classified by serial numbers. In the impact loading experiment, a 300 kg steel weight with a tip diameter of 200 mm was dropped freely from the set falling height H to the center of the span of the RC beam.

Outline of Impact Loading Test.
In this experiment, we adopted a single loading method wherein the free-fall test was performed once. Additionally, both fulcrums of the test specimen had a structure similar to a pin support that allowed for rotation and restrain lifting. e impact velocity of the weight was measured using a laser sensor. e impact loading test setup is shown in Figure 6. e measurement items were the weight impact force and reaction force, impact velocity, and loading point deflection. e weight impact force (impact load) and the reaction force were measured with a strain-gauge load-cell type. e load point deflection was measured with a laser noncontact displacement meter. Cracking progress was recorded using a high-speed camera (progressive scan method, sensor resolution � 1024 × 1024 pixels). e frame rate was 2,000 frames per second. Figure 7 shows the D-H0.5/H1.0 before the impact loading test as an example of the frost damage deterioration state. e photographs show that the concrete on the outer edge of the top surface of the beam exfoliated owing to frost damage, and coarse aggregate was exposed. In particular, the deterioration was remarkable at both ends of the beam because the frost damage deterioration gradually progressed by scaling of the surface of the beam.

Water Absorption Ratio of Beam.
In this study, the 24 h water absorption of RC beam was measured, and the degree of frost damage on the beam was estimated. e measurement method is as follows. (1) First, the mass of the RC beam in the air-dried state was measured. (2) e RC beam was then placed in a water tank filled with tap water and soaked for 24 h. (3) Finally, the beam was lifted from the water tank, and the water on the surface was gently wiped off.
e mass was measured again. e water absorption ratio R a was evaluated based on the following equation: where W w and W d are wet and dry masses of the beam. Figure 8 shows the dry and wet masses and water absorption of each RC beam. From this figure, it is observed that the dry masses of N-beams are approximately 245 kg. Its density can be estimated to be equal to 24.0 kN/m 3 . Dry  Shock and Vibration 3 masses of D-beams are smaller than those of N-beams. Specifically, in the case of D-H1.0-1, the mass of the beam is the smallest because the edge of the beam is significantly exfoliated. e magnitude relation of the wet mass is almost the same as that of dry mass. e water absorption ratio R a of N-beams was approximately equal to 0.6%. is value is much smaller than the typical moisture content of concrete (approximately 3%) [26]. is is because the absorption ratio in this study was evaluated by the total mass of the RC beam, as shown in equation (1). erefore, it can be observed that water is absorbed slightly inside the beam. e absorption ratios of D-beams are more extensive than those of N-beams. In particular, the absorption ratio of D-H1.0-2 is the largest    (approximately 2.3%), which is more than four times that of N-beams. Subsequent measurements of ultrasonic propagation velocity and impact loading experiments were conducted after specimen withdrawal from water and storage in air for approximately five days. Figure 9 shows the measurement results of the ultrasonic propagation velocity. From the figure, the propagation velocity of the N-beam is approximately 4,000 m/s and is equivalent to nondegraded concrete [27]. Conversely, for the D-beam, the propagation velocity in the upper part of the beam is slow. In particular, in the case of the D-H1.0-2 specimen, it can be observed that there is a broad distribution and increased deterioration response for parts that have a low propagation velocity. In the upper part of D-H1.0-2, the ultrasonic propagation velocity is the lowest and is approximately equal to 1,500 m/s. is corresponds to a decrease of 14% in terms of the relative dynamic elastic modulus (RDM) obtained using the following equation [1,25]:  Shock and Vibration 5

Ultrasonic Propagation Velocity Distribution.
where v n and v 0 mean the ultrasonic propagation velocity of concrete at the nth cycle and the beginning, respectively. is indicates that the evaluation of the degree of deterioration based on the propagation velocity of ultrasonic waves corresponds to the evaluation based on the water absorption of the beam. Figure 10 shows the weight impact force, reaction force, and load-point  deflection waveforms of all specimens at each weight falling height H. It is observed that the first wave response generated by the weight impact force waveform has duration of approximately 1 ms and amplitudes in the range of 500-700 kN, regardless of the presence or absence of frost damage deterioration. e amplitude tends to increase as the falling height of the weight H increases. Additionally, the maximum impact forces of the D-beams deteriorated by frost damage tend to be smaller than those of the N-beams. e reaction force waveform shows a waveform characteristic in which the high-frequency component is combined with a half-sine wave. e figure shows that the maximum amplitude and primary wave duration increase as the falling height of the weight increases. e duration can be an indicator of the stiffness of the RC beam when subjected to the same impact loading condition. For H � 0.5 m, the duration of the reaction force wave for D-H0.5 is longer than that for N-H0. 5.

Various Response Waveforms.
is indicates that the rigidity of D-H0.5 is smaller than that of N-H0.5. For H � 1.0 m, the duration of D-H1.0-1 is slightly longer than that of N-H1.0, and that of D-H1.0-2 is approximately 4 ms longer.
is tendency corresponds to the degree of deterioration of D-H1.0-2 being greater than D-H1.0-1, as shown in Figure 9(a). e maximum amplitude of the loading point deflection waveform increases as the set weight falling height H increases. Moreover, for all the specimens, it can be observed that the deflection remains after the first half-sine wave reaches its maximum value. Subsequently, it shifts to a damped free vibration. For H � 0.5 m, the maximum deflection is almost the same, regardless of the presence or absence of deterioration. is is because the degree of frost damage deterioration of D-H0.5 is relatively low, and the input energy is small. However, in the case of the specimen D-H0.5, there is minor deflection restoration at a slower rate than that of N-H0.5.
For H � 1.0 m, the deflection increases as the degree of deterioration increases.
is is attributed to decreased concrete strength, peeling of the upper concrete, and development of diagonal crack openings, as described in the next section. Figure 11 shows the distribution of cracks and ultrasonic propagation velocity in each specimen. All specimens generally exhibit symmetric bending deformation. As shown in Figure 11(a), when H � 0.5 m, bending cracks are observed at the center of the beam in both specimens. However, in the case of D-H0.5 with frost damage deterioration, horizontal cracks are observed in the upper side of the beam. e occurrence of horizontal cracks is observed because the upper degradation is significant, as shown by the ultrasonic propagation velocity distribution. In the experiment, high-speed camera images confirmed that split cracking occurred at the interface between the standard part and the deteriorated part during restoration. However, these cracks did not significantly affect the maximum deflection of D-H0.5, as shown in Figure 10  Shock and Vibration cracks occur diagonally in the downward direction at both sides from the loading point, and bending cracks are extensively distributed. Conversely, for beams D-H1.0-1/2, the damage is primarily at the center of both specimens. e damage is localized because a decrease in concrete strength became apparent at the center of the span, wherein a sizeable bending moment and shear force were applied. e cover concrete around the mid-span of the specimen D-H1.0-2, which has a high degree of deterioration, was peeled off. ese properties correspond to the ultrasonic velocity distribution. It is probable that, owing to these damages, the maximum deflection of D-H1.0-2 was more extensive than that of N-H1.0, as shown in Figure 10(b). Figure 12 shows the cracking behavior of the upper edge concrete at the center of the span after the experiment when the falling height H � 1.0 m. From the figure, the exfoliation of the cover concrete occurred at the weight collision part in each specimen. In the case of the nondegraded N-H1.0, delamination occurred along the beam width direction area. In contrast, in the degraded D-H1.0-1/2, the upper surface of the beam was peeled off. A peeled region is also observed in the axial direction.

Cracking Properties.
is tendency is remarkable in the D-H1.0-2, where the response deflection was large. Figure 13 shows high-speed camera images of the progress of cracking at a falling height H � 1.0 m. Results show that in the case of the N-H1.0 without deterioration, bending and diagonal cracks occurred at the elapsed time t � 2.5 ms. e initial bending rigidity of N-H1.0 was higher than other deteriorated beams, so that diagonal cracks such as punching shear cracks occurred. At t � 5 ms, in addition to the bending cracks at the center of the span, bending shear cracks at both ends were observed, and these cracks were opened at t � 10 ms. At t � 20 ms, the width of the cracks became more extensive, with concrete compressive failure at the upper edge.
In the cases of the deteriorated D-H1.0-1, there are no diagonal cracks at the elapsed time t � 2.5 ms. is is due to the appearance of increased damage at the weight collision part ( Figure 12). Subsequently, in the case of the D-H1.0-1, a bending crack opened at t � 5 ms, and a diagonal crack occurred on the right side at t � 10 ms. A bending crack appeared on the left side. At t � 20 ms, the diagonal crack on the right side opened further. is result is probably attributed to the fact that the degree of deterioration on the right side was slightly large. As shown in Figure 10(b), the effect of these cracks on the maximum deflection of the RC beam was not so large.
In the case of D-H1.0-2, three bending cracks occurred at t � 5 ms. ese cracks opened at t � 10 and 20 ms. e crack on the left side had an unusually large opening that corresponds to the ultrasonic propagation velocity distribution, wherein the damage was concentrated on the highly degraded part. It is considered that the maximum deflection of D-H1.0-2 increased owing to the occurrence of these multiple bending crack openings. erefore, it became clear that in the case of RC beams deteriorated owing to frost damage, the damage tends to concentrate on the highly deteriorated parts. It is possible to evaluate the impact resistance of RC beams that have been

Relationship between Residual Deflection and Input
Energy. Figure 14 shows the relationship between the residual deflection and input energy. e figure also shows the relational expression between the input energy E k and the residual deflection δ rs evaluated using equation (3) [28]. is equation can estimate the residual deflection δ rs by estimating the input energy E k owing to the falling of the weight and the calculated bending strength P us of the RC beam.
Herein, E k is the input energy (kJ), and P us is the static bending capacity (kN) of the beam. e static bending capacity P us (56.8 kN) of the beam was calculated according to the JSCE concrete specifications [23] using the material properties of concrete and rebar without degradation. For the N-beams without deterioration, the residual deflection proportionally increased as a function of the input energy similar to equation (3). Conversely, the residual deflections of the D-beams are larger than those of the N-beams. In particular, the residual deflection of D-H1.0-2 is approximately 1.2 times larger than that of N-H1.0. is may be due to the microcracks and scaling caused by frost damage and the decrease in the bending capacity of the RC beam due to decreased compressive strength.
erefore, it is observed that when the ultrasonic propagation velocity of the upper cover concrete is approximately 1,500 m/s (14% relative dynamic elastic modulus), the residual deflection δ rs of the RC beam is increased by approximately 20%. Additionally, according to the existing residual deflection calculation formula (equation (3)), an increase of 20% in the residual deflection δ rs corresponds to a decrease of approximately 17% in the bending capacity of the RC beam P us .

Conclusions
In this study, RC beams with shear reinforcement were fabricated, and weight falling impact tests were performed to investigate the impact resistance behavior of RC beams with frost damage deterioration. e RC beams suffered frost damage induced by successive freezing and thawing tests. e surface was scaled considerably, and the coarse aggregate was exposed. e degree of deterioration was evaluated by the rate of change of the ultrasonic propagation velocity. As a result, the following conclusions were inferred.
(1) e ultrasonic propagation velocity of RC beams was reduced considerably by the freeze-thaw action of approximately 900 cycles. At the upper edge of the RC beam, the ultrasonic wave propagation velocity decreased from 4,000 m/s to 1,500 m/s in some parts. is corresponds to a relative dynamic elastic modulus of approximately 14%. At this time, the water absorption ratio of the RC beam was approximately 2.3%.
(2) In the case of RC beams without deterioration, diagonal cracks extended downward from the loading point to both sides, and bending cracks were extensively distributed owing to the impact loading. Conversely, in the case of RC beams with frost damage, the damage was concentrated near the loading point. e concrete on the upper edge was peeled off, and multiple bending cracks were generated. (3) e residual deflection of RC beams with frost damage increased up to 20% compared with beams without frost damage. e increase in residual deflection was primarily related to the peeling of concrete at the collision site and the opening of multiple bending cracks. (4) According to the existing residual deflection calculation formula, an increase of 20% in the residual deflection corresponds to a decrease of approximately 17% in the bending capacity of the RC beam.
In the future, we will investigate the actual compressive strength and frost damage depth of concrete after frost damage deterioration of the RC beams. e bending strength would be calculated based on survey results, and the impact resistance would then be evaluated.

Data Availability
All the data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest regarding the publication of this paper.