Reduction in the Constrained Shrinkage and Crack Risk of Different-Strength-Level Concretes with Saturated Ceramsite

Abstract

To reduce the shrinkage and cracking risk of concrete, internal curing technology was applied to modify concretes with different strength levels (C30 and C60) by incorporating saturated ceramsite. Three kinds of tests were carried out to study the effects of the incorporation of saturated ceramsite on the compressive strength, hydration temperature rise, and shrinkage behavior of the concretes, respectively. It was found that the internal hydration temperature rise of the concrete could be delayed and reduced due to the internal curing effect of ceramsite. Moreover, both the early-stage and long-term constrained shrinkage behavior of the internal cured concretes were monitored with embedded strain sensors and compared with free shrinkage behavior. For the C30 concrete system, with the incorporation of saturated ceramsite, its constrained shrinkage at 96 h was reduced greatly from 181 µε to 36 µε. Furthermore, the surface-attached-sensor method was also used to study the shrinkage behavior of the concrete beams and it was found that the sensor location affected the measured shrinkage values greatly. The model beams of both the C30 and C60 concrete systems shrunk significantly in the first month, and the highest cracking risk occurred in this period as well. The internal curing effect of saturated ceramite could significantly reduce the constrained shrinkage of the concrete beam, evidently diminishing the cracking risk. More importantly, compared to the ordinary concrete C30 system, it was revealed that such an internal curing effect was more effective in promoting the performance of the higher-strength concrete (C60). With this effect, the cracking risk of C60 concrete was reduced from 0.69 to 0.37 at 300 days and changed little from then on.

1. Introduction

It is well known that the quality of concrete is difficult to control. Due to various practical problems, including the accumulation of air bubbles during construction, the resulting concrete is prone to shrinkage and cracking. The shrinkage of concrete could lead to many problems, such as volume stability, steel corrosion, structural leakage and damage, and even catastrophic collapse; therefore, it has a great influence on the building’s structural health and service life [1]. For instance, the bridge guardrail is one of the important structural elements of a bridge and its cracking could be observed occasionally.

As the maintenance conditions are often limited at the construction site, an effective way to overcome the easy-cracking problem of guardrail concrete has been developed from the point of view of materials. It was suggested that internal curing techniques were efficient to reduce the self-shrinkage of concrete and diminish the concrete cracking risk that originated from insufficient curing [2]. During the long-term service life of concrete, the pre-incorporated internal curing component could provide internal water to continuously cure the hardened concrete [1,3,4]. Such internal curing materials can be categorized into two types, i.e., organic and inorganic materials. The organic materials are usually highly absorbent polymers or resins [5,6,7,8,9], such as super-absorbing polymers (SAPs). In the concrete, SAPs could release the pre-absorbed water for internal curing of concrete; however, this likely left more defective pores and led to the reduced strength of concrete generally [1]. The inorganic materials are mainly lightweight aggregates (LWA) [4,10,11,12,13,14]. Among these, ceramsite is one of the most important LWA for concretes [13,14,15,16,17,18,19,20,21,22], as it has its own advantages over those SAPs. Firstly, as a typical inorganic ceramic material, it has more compatibility with the inorganic cementitious matrix materials of concrete. Secondly, it can replace part of the coarse aggregates and thus reduce the structural deadweight without bringing evidently negative effects on the mechanical properties of concretes [23]. As revealed recently [9], the compressive strength of concretes incorporated with ceramsite was decreased much less than those incorporated with the SAP. Thirdly, the thermal conductivity coefficient of the ceramsite concrete is usually small and thus has a good thermal insulation performance [24]. It was found that the moisture content of LWAs had a great effect on the concrete shrinkage [11], and the concrete incorporated with saturated LWAs displayed evidently reduced self-shrinkage [12]. Ceramsite could be used singly or combined with other fillers to improve the performance of concretes [20,21,22], such as polymer fibers and rubber particles. Despite lots of research on the self-shrinkage of ceramsite concrete, more work needs to be carried out on this important internal curing technology, such as on the influence of the location of sensors on monitoring the long-term shrinkage behavior and the relative predicted cracking risk of the internally cured concrete system with different strength levels.

In this study, saturated ceramsite was used for internally curing the concrete with different strength levels. The effect of the internal curing effect on the mechanical properties of the resulting concretes and their hydration temperature rise will be studied. Moreover, both the early-stage and long-term shrinkage behavior of the internally cured concrete will be monitored and analyzed by model beam tests, as well as the different behavior between the constrained shrinkage and free shrinkage. Based on these analyses, the incorporation of saturated ceramsite on the cracking risk of concrete will be revealed. In the meanwhile, the internal curing effect of saturated ceramsite on the performance of the concrete with different strength levels will also be clarified.

2. Materials and Methods

2.1. Raw Materials

P.O 42.5 ordinary silicate cement was produced by Bozhou Conch Cement Company (Bozhou, China), and its main physical parameters are shown in

Table 1. The physical properties of cement.

Raw Material	Fineness (cm2/g)	Time of Initial Condensation (min)	Time of Final Condensation (min)	Standardized Consistency (%)	Heat Loss (%)
C30	3400	160	220	27	0.5
C60	3790	122	160	29	0.61

. The mineral admixtures included the grade II fly ash produced by Huainan Changhua Electric Industrial Corporation (Huainan, China); S95 grade mineral powder with an alkali content of 0.4%, loss on burning of 0.06%, water demand of 99%, and an activity index of 84% for 7 days and 107% for 28 days, respectively. The detailed properties of fly ash and mineral powder are shown in

Table 2. The physical properties of fly ash and slag.

Raw Material	Density (g/cm3)	Specific Surface Area (cm2/g)	Cl− (%)	SO3 (%)
fly ash	2.1	350	0.005	0.9
mineral powder	2.9	461	0.006	1

. Two concretes with different compressive strength levels, i.e., C30 (compressive strength of ~30 MPa) and C60 (compressive strength of ~60 MPa), were designed and utilized to study the internal curing effect of saturated ceramsite on their various properties.

The fine aggregate was medium-coarse sand with a fineness modulus of 2.54, which was produced in Ganzhou, China. And the coarse aggregate was a mixture of the gravels with sizes of 5–16 mm and 16–31.5 mm. Their detailed properties are listed in

Table 3. The physical properties of aggregate.

Raw Material	Apparent Density (g/cm3)	Bulk Density (g/cm3)	Clay Content (%)
Fine aggregate	2.62	1.62	0.1
Coarse aggregate	2.68	1.5	0.8

.

For the preparation of C60 concrete, silica fume was added and it had an average particle size of 0.31 μm, a density of 2.2 g/cm³, and a specific surface area of 143,100 cm²/g. The water-reducing agent was a polycarboxylic acid product with a water reduction rate of 29% and solid content of 25% and it was provided by Sika (Jiangsu, China) Building Materials Co.

For the internal curing materials, ceramsite clay particles with a size of less than 10 mm were used. They had a stacking density of 2.8 g/cm³ and barrel compression strength of 2.1 MPa with a saturated water absorption rate of 22.54%. The mixing water was just tap water.

2.2. Proportioning

Two types of concrete systems with different strength levels, i.e., C30 and C60, were chosen and utilized in this work. Different contents (0, 10%, 20%, and 30%) of saturated ceramsite were used to replace the coarse aggregate in the C30 concrete formulation, and the resulting samples were labeled as C30-0, C30-1, C30-2, and C30-3, respectively. A similar operation was also applied for the C60 concrete formulation. The detailed mix ratios of these two concrete component systems are given in

Table 4. The mix ratio of C30 and C60 concrete component system.

Sample	Cement	Aggregate	Gravel	Fly Ash	Mineral Powder	Silicon Powder	Water	Water Reducing Agent	Saturated Ceramsite
C30-0	216	1117	809	108	36	0	140	9	0
C30-1	216	1005	809	108	36	0	140	9	34
C30-2	216	894	809	108	36	0	140	9	67
C30-3	216	782	809	108	36	0	140	9	101
C60-0	312	998	722	104	52	52	130	12.5	0
C60-1	312	898	722	104	52	52	130	12.5	30
C60-2	312	898	722	104	52	52	130	12.5	60
C60-3	312	789	722	104	52	52	130	12.5	90

Note: C30 concrete: total cementitious amount 360 kg/m³, gravel content 42%, mineral admixture 40%, fly ash:mineral powder = 3:1; C60 concrete: total cementitious amount 520 kg/m³, gravel content 42%, mineral admixture 40%, fly ash:silicon powder:mineral powder = 2:1:1.

.

2.3. Test Methods

To simulate the constraint conditions of the actual concrete structure on the guardrail site, model concrete beams were fabricated for testing. Early shrinkage tests were carried out according to the code for “ordinary concrete long-term performance and durability test method standard” (GB/T50082-2009 [25]). Referring to GB/T50081-2002 [26] the code for “Standard for Test Methods of Mechanical Properties of Ordinary Concrete”, the compressive strength tests were performed and the cubic specimens of 100 × 100 × 100 mm were manufactured and tested on a WAW-600D-type electrohydraulic servo system (Beijng Zhongsheng Huaye Technology, Co. LTD., Beijing, China). According to the model beam test used by Bazant P. et al. [27,28], rectangular beams with a design span of 2000 mm and dimensions of 2000 mm × 400 mm × 200 mm were fabricated and the beam ends were secured with anchor bolts and steel plates, as shown in Figure 1.

The 215HAT-type high-precision steel string sensors(Kingmach Measurement & Monitoring Technology Co., Ltd., Changsha, China), resistance strain gauges, temperature and humidity sensors (Beijing Power Bausheng Technology Development Co., Ltd., Beijing, China), and the corresponding data acquisition devices (Figure 2) were utilized to monitor the temperature change and the shrinkage behavior of the concrete beams with time. The high-precision steel string sensor is capable of responding to a small load (<1 N); therefore, it works well in monitoring the shrinkage of concrete. All the collected data were analyzed with Origin software (version 8.0). For the internal temperature test of the concrete sample, an SCTH 2002 (Sonbest Company of Shanghai, Shanghai, China) data acquisition device was used to collect the data.

As shown in Figure 3, the temperature sensors were embedded at the end of the girder prior to the concrete placement to avoid the influence of early cracking of concrete on them. The data were recorded hourly within 48 h after the concrete placement, and daily thereafter. Shrinkage measurements were achieved through either the sensor embedded in the concrete or the ones patched externally on the beam surface, as shown in Figure 3. The strain gauges were placed before the concrete placement and fixed at the lower reinforcement area of the full girder to record the early shrinkage of the concrete in time. However, the external patched ones should be mounted immediately after 24 h once the concrete beam was demolded. For the sensor location on the side and bottom surface of the beam, a group of strain gauges was distributed every 500 mm, and each group of strain gauges included one at the upper edge and another at the lower edge. Additionally, a set of strain gauges was mounted in the middle of the beam. Long-term data were measured up to 365 days.

3. Results and Discussion

3.1. Compressive Strength of the Concrete Specimens

The effect of incorporating different contents of saturated ceramsite on the compressive strength of both normal and high-strength concrete systems was studied and at least three specimens were tested for every sample. For the ordinary low-strength concrete system (C30), the compressive strength of the internally cured concrete at 7 days, 28 days, and 90 days all decreased gradually with the increasing content of ceramsite, as can be seen in

Table 5. Compressive strength of the different concrete samples.

Sample	Compressive Strength at 7 Days (MPa)	Compressive Strength at 28 Days (MPa)	Compressive Strength at 90 Days (MPa)
C30-0	31.11 ± 1.08	42.22 ± 0.46	46.27 ± 1.61
C30-1	28.13 ± 2.16	37.97 ± 1.71	42.77 ± 2.47
C30-2	25.11 ± 1.88	33.72 ± 1.42	39.27 ± 1.63
C30-3	22.12 ± 1.35	29.47 ± 3.15	35.77 ± 1.87
C60-0	69.57 ± 3.53	82.62 ± 2.14	87.63 ± 3.04
C60-1	68.37 ± 5.18	80.57 ± 3.32	84.83 ± 4.67
C60-2	67.17 ± 2.81	78.52 ± 2.97	82.03 ± 3.52
C60-3	65.97 ± 4.07	76.47 ± 3.33	79.23 ± 4.73

. For the high-strength concrete system (C60), the same phenomenon occurred as well. However, its decrease was much less than that of its C30 counterpart. For instance, by incorporating 30% ceramsite, the compressive strength of the resulting C60-3 at 28 days was reduced by 6.15 MPa from the reference sample C60-0 (82.62 MPa), which was even less than half of that for the reduction in the C30 concrete system (12.75 MPa). This indicates the better applicability of the saturated ceramsite in the high-strength concrete system. It is noted that introducing 20% content of saturated ceramsite has a smaller negative effect on the mechanical properties of concrete. Taking the strength at 28 days as the example, the compressive strength value for C30-2 was 8.5 MPa less than that of C30-0, implying a reduction of 19.5%, while C60-2 was just 4.1 MPa lower than C60-0 and its reduction was only 4.9%. Therefore, the content of saturated ceramsite in the concrete was set to be 20% in the following research.

3.2. Hydration Temperature Rise Inside the Model Concrete Beams

The hydration temperature inside the model concrete beam was monitored in situ with the embedded temperature sensor over a 96 h period, and the corresponding curve of the temperature vs the time is shown in Figure 4.

As shown in Figure 4, the hydration temperature rise of concrete was not significant within the first 18 h after the concrete placement. After 18 h, the internal temperature of all the concrete samples began to increase. Throughout the entire test time, the internal temperature of the C30 concrete group increased maximally by ~6 °C. In comparison, the maximum temperature rise of the C60 concrete group increased by about 11 °C, which was significantly higher than that of the C30 concrete system. This implies that the larger thermal stress accumulated in the high-strength concretes leads to their probable larger autogenous shrinkage and higher cracking risk, which will be studied in the following research. Moreover, it was found that with the incorporation of saturated ceramsite in the concrete, the rapid rise in the internal temperature of the specimens was somewhat delayed by about 1–2 h for both concrete systems. The hydration temperature curves of their concrete samples were always below those of the ones without saturated ceramsite, despite that the differences were not so significant. This suggested that the internal curing effect of the saturated ceramsite was able to delay and reduce the internal hydration temperature rise during the hydration and curing of concrete. As suggested recently, the delayed and reduced hydration temperature rise was in favor of reducing the cracking risk of concretes [29]. Theoretically, the internal temperature rise will lead to the uneven curing of mass concrete and the accumulation of excessive thermal stress, which likely results in its autogenous shrinkage and cracking. The internal curing effect of the saturated ceramsite can effectively reduce the heat release rate of cement, regulate its curing and hydration process, delay and reduce its internal hydration temperature rise, and, therefore, reduce the shrinkage of the concrete beams.

3.3. Early Shrinkage Behavior of Concrete Beams

Early shrinkage values of concrete were measured continuously using internally embedded steel chord sensors in their model beams. The test results of the 4 samples within 96 h were shown in Figure 5, as well as those of the free shrinkage counterparts.

As can be seen from Figure 5, the internal shrinkage behavior of the concrete model beam was basically consistent with that of free shrinkage. All the shrinkage of the concrete increased generally with time. From 0 to 96 h, the internal concrete shrinkage values of the C30-0 and C30-2 model beams were 211 µε and 148 µε, respectively, while the corresponding free shrinkage values of the concrete material were 392 µε and 184 µε, respectively (Figure 5a), confirming that the early internal shrinkage of concrete was constrained and much less than the corresponding free shrinkage. This could be attributed to the presence of the steel columns and rivets at the ends of the beams, as well as the reinforcement rebars. It is well known that such constraints may lead to internal stresses within the concrete materials and even cause significant adverse consequences to the building structure. At 96 h, the constrained shrinkage was calculated as 392 µε − 211 µε = 181 µε. After the incorporation of saturated ceramsite into the concrete, the constrained shrinkage of the resulting C30-2 was calculated as 184 µε − 148 µε = 36 µε. These demonstrated that the internal curing effect of the ceramsite can significantly reduce the confined shrinkage and the internal stresses triggered by early shrinkage of concrete, and, therefore, diminish the early cracking risk. As shown in Figure 5b, similar shrinkage behavior was also found for the C60 concrete system, indicating that the internal curing technique also works well for high-strength concrete. Interestingly, the internal shrinkage curve of C60-2 (C60-2 internal) is almost overlapped with that of the free shrinkage curve of C60-2 (C60-2 free), confirming an even greater internal curing effect for the high-strength concrete compared to the ordinary C-30 concrete system.

3.4. Long-Term Shrinkage of Concrete Beams

The long-term internal shrinkage behavior of concrete beams was monitored with the internally embedded steel string sensors, and the data were recorded once a month over a period of 12 months. As shown in Figure 6, both the internal and free shrinkage of all four concrete beams increased abruptly within the first month and then changed little during the following time.

As shown in Figure 6a, the C30 concrete series showed shrinkage values of 300–600 µε in the first month, while their maximal shrinkage change in the following 11 months was only about 100 µε. For the C60 concrete series, they had shrinkage values of 700–900 µε in the first month (Figure 6b), with a similar maximal shrinkage change to that of C30 concrete in the following time. This means that the shrinkage of the concrete model beams could be almost completed in the early stage of curing and their highest cracking risk should also occur in this time.

For all the concrete beams, it can be seen that all the internal shrinkage curves were significantly lower than those for the free shrinkage values. Specifically, the internal shrinkage value was found to be even decreased in some concrete groups (for instance, C30-0 and C30-2), which could be caused by the restraints of the model beams and the creep relaxation of the concrete materials due to the extended stresses. As the concrete inside the model beam is restrained by the steel reinforcement, their contraction is also restrained, which leads to the accumulation of internal stress. Although such internal stress may not seriously affect the deformation of concrete beams in the early stage, the creep relaxation of concrete occurs under the extended action of internal stress and results in shrinkage reduction. Moreover, it is noted that the shrinkage values of both C30 and C60 concrete incorporated with saturated ceramsite are always lower than those of the reference counterpart during the long-term shrinkage tests, indicating that the internal curing effect of the saturated ceramsite can significantly reduce the long-term shrinkage of concrete in model beams as well.

3.5. Long-Term Shrinkage Behavior of Concrete Beams Monitored by Surface-Attached Strain Sensors

Besides the long-term shrinkage measurements of concrete beams by the internally embedded strain sensors, this shrinkage behavior could also be monitored with the surface-attached strain gauges, which should be more convenient from the point of view of practical applications on-site. The strain gauges were attached at three characteristic locations: (1) two strain gauges were located at the center of the beam; (2) four strain gauges were located 1/4 length from the end of the beam; and (3) four strain gauges at the girder end of the beam. The average data obtained from all the sensors at the same location were taken as the final strain value. As can be seen in Figure 7, for certain concrete beams, the maximum and the minimum shrinkage occurred at the center and the ends of the beam, respectively, while the shrinkage values at 1/4 length to the beam ends were always the medium ones. This could be related to the different restrain at different beam locations. As the model beam is restrained by fixed supports at the beam ends, early contraction of the concrete materials causes the tensile stresses within the beam and significant volume deformation as well. Since the support constraints at the beam ends are realized by steel columns and rivets, the shrinkage of concrete is more constrained at the beam ends; therefore, the corresponding shrinkage values are lower. In contrast, due to the fewer constraints in the middle of the beam, this part bears most of the early contraction of the entire beam, which results in its larger shrinkage values. Also, for this reason, the early cracks in the model concrete beams usually occur in the middle.

Moreover, it was found that the shrinkage of C30-2 specimens (Figure 7b) was always less than that of C30-0 (Figure 7a), no matter the data obtained in the middle of the beam or at the end of it. A similar phenomenon was also observed for C60-0 (Figure 7c) and C60-2 specimens (Figure 7d). These experimental results indicated again that the internal curing effect of the saturated ceramsite can significantly reduce the long-term shrinkage of concrete beams, which is consistent with the aforementioned results measured by the embedded strain sensors (Figure 6).

In fact, it is found that this surface-attached-sensor method does not seem to be a good option to monitor the long-term shrinkage of concrete. The main reasons for this are as follows: (1) The crack size of the concrete beams. The length of most strain gauges is about 10 cm, and their attached area are vulnerable to being penetrated by cracks, causing the failure of strain gauges and difficulties in obtaining long-term shrinkage data. (2) The glue used for fixing the sensors on the beam surface has certain aging problems during the long-term service. They usually become hard and brittle, which makes it hard to ensure the coordination between sensors and the concrete deformation process. In some cases, this could lead to the buckling of strain gauges or them even falling off. (3) Multiple strain gauges must be applied to minimize the test error as their locations have a great influence on the results.

3.6. Assessment of the Cracking Risk of Model Beams

The cracking risk of the concrete beam could be estimated by calculating the cracking stresses in the beams. Studies have shown that a greater risk of cracking occurred when the concrete shrinkage stress was ≥70% of its tensile strength [30,31,32]. The cracking risk of concrete can be defined with Equation (1) [33]. When the cracking risk reaches 0.7 or more, the concrete possesses a high probability of cracking.

$$\zeta =\frac{{\sigma }_s\left(t\right)}{f_t\left(t\right)}=\frac{{\epsilon }_s\left(t\right)E_c}{f_t\left(t\right)}$$

(1)

where ${\sigma }_s\left(t\right)$—maximum tensile stress at moment t; ${\epsilon }_s\left(t\right)$—shrinkage value of concrete at moment t; E_c—elastic modulus of concrete; and $f_t\left(t\right)$—tensile strength of concrete at moment t.

The elastic modulus of concrete was measured to be 3.4 × 10⁴ MPa, 3.25 × 10⁴ MPa, 3.92 × 10⁴ MPa, and 3.77 × 10⁴ MPa for C30-0, C30-2, C60-0, and C60-2, respectively. The tensile strength of concrete was generally approximated to be 0.1 times its compressive strength. According to the preliminary data, the variations in the compressive strength of the concrete samples with time were fitted and their fitted equations are shown in

Table 6. The fitted relationship between compressive strength and time.

Sample	The Fitted Equation
C30-0	F(x) = (48.88x + 0.7329)/(x + 3.516)
C30-2	F(x) = (42.38x − 1.84)/(x + 5.133)
C60-0	F(x) = (91.3x − 2.462)/(x + 2.371)
C60-2	F(x) = (86.92x − 3.41)/(x + 2.581)

. Therefore, the tensile strength at moment t could be obtained, and the cracking risk was calculated by combining it with the shrinkage data in Equation (1), as shown in Figure 8.

As can be seen from Figure 8, concrete generally has a larger cracking risk in the early stage than in the later stage. Particularly, the largest cracking risk of all the four kinds of concretes appeared within the first three days, indicating its great susceptibility to cracking in this period. It is also noted that all the cracking risks of the four concretes are less than 0.7, which was suggested to be the critical cracking risk value of concrete previously [30,32], as concrete has a great possibility of cracking once the value is ≥0.7. These results imply the lower cracking risk of our well-designed concrete systems. The curve for C60-0 is always above that of C30-0, indicating that, for the concretes without the incorporation of saturated ceramsite, the higher the strength level of concrete the greater its cracking risk. However, with the incorporation of saturated ceramsite, the cracking risk values of both the two concretes with different strength levels were reduced significantly, especially for the higher-strength concrete. For instance, the cracking risk of C60-2 after 300 days is about 0.37, which is reduced by more than 46% in comparison with that of C60-0 at the same time. More interestingly, it can be seen in Figure 8 that, besides the shift downwards of the curves for both C30-2 and C60-2 compared to those for C30-0 and C60-0, the curve of C60-2 is almost overlaying that of C30-2, indicating the remarkable internal curing effect of saturated ceramsite on reducing the cracking risk of the high-strength concrete sample (C60-2), which reaches almost the same low cracking risk level of the ordinary concrete sample (C30-2). Furthermore, the valuable internal curing effect of saturated ceramsite could work well for a very long time (at least 1 year), as shown in Figure 8. In fact, it can be seen that the cracking risk value of the internally cured high-strength concrete (C60-2) changes little after 150 days, indicating its long-term stability.

4. Conclusions

The shrinkage and internal temperature development of high-performance internally cured concrete were investigated. The conclusions are given as follows:

(1)

The inhibitory effect of saturated ceramsite on the hydration temperature rise of concrete was verified by tests on the model beam. With the incorporation of saturated ceramsite, the hydration temperature rise of concrete was reduced.

(2)

The constrained shrinkage behavior of internally cured concrete beams was revealed. Constraints may lead to internal stresses within the concrete beams as well as early-stage cracking. The internal curing effect of saturated ceramsite significantly reduced the confined shrinkage of concrete.

(3)

Both early-stage and long-term shrinkage development of the internally cured concrete beams were characterized. Most shrinkage of concrete beams over a one-year period was accomplished within the first month, after which it changed little. The internal curing effect significantly reduced the long-term shrinkage of the concrete beams.

(4)

Besides the method of embedded strain sensors, the surface-attached-sensor technique could also be used to monitor the shrinkage of the concrete beam. In this technique, the location of the strain sensor affected the measured shrinkage behavior greatly, and the largest shrinkage values were obtained in the middle of the concrete beam.

(5)

The maximum cracking risk of all the concrete beams within 12 months occurred in their early curing stage. For ordinary concretes, the higher the strength level, the greater the cracking risk. The incorporation of saturated ceramsite could significantly reduce the cracking risk of concretes, especially for higher-strength concrete.

In the bridge structure elements, long-term shrinkage of concrete may result in significant pre-stress loss and stiffness reduction. The internal curing technique via the incorporation of saturated ceramsite is effective in reducing the long-term shrinkage of concretes, leading to significant improvement in their volumetric stability and reduction in their cracking risk.