Compressive Strength Gain Behavior and Prediction of Cement-Stabilized Macadam at Low Temperature Curing

For cement-based materials, the curing temperature determines the strength gain rate and the value of compressive strength. In this paper, the 5% cement-stabilized macadam mixture is used. -ree indoor controlled temperature curing and one outdoor natural curing scenarios are designed and implemented to study the strength development scenario law of compressive strength, and they are standard temperature curing (20°C), constant low temperature curing (10°C), day interaction temperature curing (varying from 6°C to 16°C), and one outdoor natural temperature curing (in which the air temperature ranges from 4°C to 20°C). Finally, based on the maturity method, the maturity-strength estimation model is obtained by using and analyzing the data collected from the indoor tests. -e model is proved with high accuracy based on the validated results obtained from the data of outdoor tests. -is research provides technical support for the construction of cement-stabilized macadam in regions with low temperature, which is beneficial in the construction process and quality control.


Introduction
Cement-stabilized macadam is low-dose cement-base-stabilized mixture and its cement dosage is 5% or so; it is generally used for base layer of pavement construction in China [1]. Whether or not the compressive strength of cement-based materials largely depends upon the curing process is well known, in which both the curing temperature and time are particularly important [2,3]. For conventional laboratory test of compressive strength, the curing is typically carried in constant temperature environment at 20°C in many national specifications [4][5][6]. But for the pavement construction project, the actual outdoor curing temperature varies with weather. Specification requires that more than 5°C temperature can be carried out for construction [4]. However, in the northern seasonal frozen areas such as China's Heilongjiang Province, although the temperature is more than 5°C in the month of April, the temperature changes a lot and is very unstable. Due to the large temperature difference between day and night and the fact that usually does not reach 20°C during curing time, the compressive strength sometimes cannot meet the requirements, causing the coring to be loose. Since the strength cannot be confirmed, one cannot reasonably arrange the next process [7]. Based on this special temperature condition, there is a strong need to study the compressive strength gain laws under such different low temperature curing conditions. In that regard, several indoor and outdoor experiments are designed accordingly in this paper to conduct such study.
ere are many research efforts made to study the impact of curing temperature on cement-based materials, such as Portland cement-stabilized soil, light weight cemented soil, sand, coal fly ash, and lime blends [8][9][10]. As for the curing temperature, many research studies have been reported for the high temperature, and most results showed that high temperature curing can increase the initial compressive strength [11,12]. e compressive strength and tensile strength of cement-stabilized marine soils which were used as road construction materials were studied under cured temperatures varying from 40°C to 60°C in Wang's research work [13]. Escalante-Garcia et al. [14] tested the compressive strength of hydration at five temperatures ranging from 10°C to 60°C, and the results showed that high temperature can improve the initial compressive strength, but it can actually decrease the strength in the long term. Wang et al. [15] conducted tests of calcium sulphoaluminate cement at different curing temperatures (i.e., 0°C-80°C) in order to study the influence of hydration evolution on compressive strength.
e results indicated the early-age compressive strength increases with increasing temperature but decreases at the temperature ranging from 40°C to 80°C, and the compressive strength is mainly affected by hydration degree.
About low temperature curing, several studies have been reported in the literature. Price [16] showed that the strength of concrete mixed developed significantly slower at low temperature than that at room temperature. Husem et al. [17] tested the compressive strength of ordinary and high performance concrete under standard curing (at 23 ± 2°C) and other low temperature curing (at 10, 5, 0, and −5°C, resp.). e results indicated that the strength at 10°C and less than 10°C was lower than that in the standard curing. Kim et al. [18] investigated the strength development for curing histories with 5°C, 20°C, and 40°C temperature, which indicated that the concrete strength at low temperature was less than that at standard temperature initially but was almost the same with time. Marzouk et al. [19] performed the tests at five temperatures ranging from −10°C to 20°C over 3 months and found that there was a proportional relationship between compressive strength and temperature.
In addition, in terms of strength prediction, many literatures have shown that maturity theory is appropriate and better in the strength prediction than some other methods [20,21]. In 1951, Saul et al. [22] first proposed the "maturity" concept, which was defined as the product of the curing time and temperature. It was pointed out in the famous Nurse-Saul maturity function that when the maturity is the same, the strength will also be approximately the same. It is well noticed that the Nurse-Saul maturity model has been constantly improved and modified later, and different mathematical models have been adopted to predict the strength. For example, in Chitambira's model, the equivalent age as an index was proposed which combined both the curing age and curing temperature [23].
ere was a linear relationship between double logarithmic strength and logarithmic maturity under different curing temperatures. Jeong et al. [24] calibrated the relationship of relative strength and maturity by the moisture factor.
A review of the existing literature revealed that although there have been many studies on other cement-based materials, less research efforts were made to the 5% cementstabilized macadam. Many studies were devoted to the effect of curing temperature on strength. However, most of them focused on high temperatures, and, furthermore, almost all the curing (be they under either high temperature or low temperature) was made under varied constant controllable temperature at the laboratory chamber. It is important to note that such curing failed to account for the alternate changes in the temperatures during actual days and nights (like the construction project), and there have been no tests conducted under outdoor natural conditions. As such, this purpose of this study is to focus on the strength gain law of 5% cement-stabilized macadam mixture at low temperature, which meets the actual temperature of the construction project. e maturity theory will be employed to predict the compressive strength. Appropriate function will be selected and relevant parameters will be calibrated and obtained by using and analyzing the experimental data.
e research results will provide technical support for the construction of cement-stabilized macadam in regions of low temperature, which is beneficial to the construction quality and process control.

Descriptive Analysis of Temperatures in Harbin Area
Harbin City, Heilongjiang Province, China, is located in the north latitude 44°04′∼ 46°40′, mainly plain, belonging to the north temperate zone continental monsoon climate, and the temperature changes fast in the spring and autumn. Annual precipitation reaches 400 mm-600 mm, moisture coefficient is in the range of 0.25-1.25, and the average maximum permafrost is 120 cm-240 cm. e temperature distribution from the 15 th to 30 th of April from 2012 to 2014 at Harbin City is shown in Figure 1.
e trend of the high temperature and the low temperature during the construction period is basically similar. Most of the high temperatures are distributed in the range from 15°C to 20°C, and most of the low temperatures fall in the range from 5°C to 10°C. e average high temperature is 16°C and the average low temperature is 6°C. Figure 2 showed the daily temperature data from the 15 th to 30 th of April, 2014, at Harbin City. e data of other years follow similar pattern. At about 2 am-4 am the temperatures were the lowest, from 5 am the temperatures began to rise consistently for 9 hours at a high rate, at 12 pm-14 pm the temperatures reached the highest, and then the temperatures started to decrease continuously for 15 hours at a relatively low rate.

Indoor and Outdoor Testing Plans
According to the law of temperature variation, three indoor testing cases and one outdoor test were designed. e temperatures of three indoor tests were determined according to nearly 3 years of data at Harbin, as shown in Figure 3, and the outdoor tests started at the 17 th of April, 2015.
e cylinder specimens of 150 mm × 150 mm size with 5% cement-stabilized macadam were prepared according to the stabilized macadam mix design. Under three different curing temperatures, unconfined compressive strength tests were carried out on a daily basis.
Case 1 (standard temperature curing): the standard curing was in full accordance with the requirements of the specification operation in which the temperature was 20°C. e unconfined compressive strength test was carried out from the 3 rd day to the 7 th day. e compressive strength of the 7 th day (i.e., the 7 th standard strength) was used as a standard for reference.
Case 2 (constant low temperature curing): the curing temperature was 10°C that was determined according to the average high and average low temperatures weighted by time during nearly three years. e compressive strength was tested and the tests would not stop until the compressive strength became more than the 7 th standard strength.
Case 3 (day interaction temperature curing): the temperatures were changed in the test chamber to simulate the large variations in the day and night temperatures. As shown in Figure 3, the high temperature was kept at 16°C from 7 am to 15 pm for 8 hours, and the low temperature was 6°C from 16 pm to 6 am for 14 hours. From 6 am to 7 am, the temperatures were increased from 6°C to 16°C, and from 15 pm to 16 pm, the temperature was reduced from 16°C to 6°C. Also, the compressive strength would continue to be tested beyond the 7 th day until the strength was more than the 7 th standard strength.
Case 4 (outdoor natural temperature curing): according to the weather forecast data, the test started on April 17, 2015. e specimens were placed in test pit. e pavement base layer and curing methods were simulated and the compressive strength was tested from the 7 th day until the strength was more than the 7 th standard strength. e specific operating procedure and temperature measuring method are discussed as follows. 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Hour  First, a depth of 15 cm pit was dug and the bottom was flattened. e specimens were then neatly placed into pit, and the gap was filled with fine aggregate and compact. e upper was covered with a white geotextile to keep moisture, and the water was sprayed on the surface at noon every day. Specimen placement pictures were shown in Figure 4. ree specimens were used for temperature measurement. On each specimen, four temperature sensors were embedded in the upper, the middle external, the bottom, and the central parts of the body, which were used to measure the temperature of different parts of each specimen. Figure 5 is the schematic diagram showing the location of temperature sensors, among which the central sensor was embedded in the specimen production process and the external three sensors were fixed on the surface later. e pictures showing the central sensors and middle external sensors were given in Figure 6. During the outdoor curing period, the hand-held thermometer was used to measure the temperatures and the measurement frequency was 1 reading/hour.

Cement Performance.
e cement used in the experiment was Harbin TIANE 425 #. e technical indexes of cement are shown in Table 1. Note that the cement dosage is 5% of the aggregate weight.

Aggregate Grade.
e aggregates used were of four sizes: 2 cm-3 cm, 1 cm-2 cm, 0.5 cm-1 cm, and 0 cm-0.5 cm. e gravel used was in line with the "Road Pavement Construction Technical Specifications (JTJ034-2000)" requirements. e aggregate composite grade is shown in Table 2.

Compaction Test.
To prepare for the specimen making, the maximum dry density and the optimum water content of the mixture were determined through compaction tests. According to the procedures described in "Test Procedure for Stabilized Materials for Highway Engineering Inorganic Binder (JTG E51-2009)", the optimum water content was 6.8% and the maximum dry density was 2.144 g/cm 3 .

Unconfined Compressive Strength Test.
e specimens were made and kept in the curing chamber. In accordance with the requirements, the curing temperatures in three cases were controlled at 20°C and 10°C and in the range from 6°C to 16°C. e specimens were subjected to the unconfined compressive strength tests according to the designed testing plan. Figure 7 is the compressive strength gain law of three-case indoor tests. As for the standard curing temperature of 20°C (Case 1), the strength increases as the curing time increases, and the gain rate is initially high but gradually decreases until the 7 th day. e strength is 3.5 MPa which can meet the standard requirements. Under the constant low temperature of 10°C conditions (Case 2), the compressive strength increases continuously with the increase of the curing time, but the gain rate is less than that under the standard curing condition. e compressive strength is 2.2 Mpa at the 7 th day, accounting for only 62.9% of the 7 th standard strength. e compressive strength does not achieve the 7 th standard strength until the 14 th day. In the day interaction temperature curing of 6°C-16°C (Case 3), the compressive strength also increases with the increase of the curing time, but the gain rate is less than that for the standard curing and is also slightly less than that under the constant low temperature curing condition. e compressive strength is 2.1 Mpa at the 7 th day, which is only 60% of the 7 th standard strength under the standard curing condition. e compressive strength does not reach the 7 th standard strength until the 14 th day.

Temperature Transfer Law of the Specimens in Outdoor
Natural Environments. Figure 8 shows the day temperature curve at each position of the specimens on April 20, 2015. It can be seen that the temperature variation in the specimens was similar to that of the air temperature, and the fluctuation range in the upper part was larger than that in the middle and the lower parts. e difference between the central and the middle external was small, which indicated that the temperature transfer was small in the horizontal direction. Temperature transfer law of the specimens in outdoor natural environments is presented as follows.
(1) From 6 am, the temperature started to increase, and the temperature difference between the upper, the middle, and the bottom parts also gradually increased. (2) At 11 am∼14 pm, the temperature difference between the upper and bottom part reached the maximum of 8°C, while the upper and middle temperature difference was about 6°C and the middle and bottom temperature difference was about 2°C. is clearly indicated that the temperature demonstrated a nonlinear pattern in the depth direction. In other words, the heat received by the surface was the most significant; then the heat reduced noticeably when it was transferred to the middle and was almost nonexisting until the bottom. e "Temperature × Time" data was used as an index to analyze the curing status at each position of the specimens. e "Temperature × Time" cumulative sum at each position of the specimens in outdoor natural environments was calculated from the 7 th day and was shown in Table 3. e "Temperature × Time" of the 7 th day standard curing was calculated to be 3360°C·h.
As can be seen from Table 3, when the curing continued to the 12 th day, the "Temperature × Time" value at the upper position reached 3569°C·h, which exceeded the standard curing at the 7 th day of 3360°C·h. However, it was only 2498°C·h at the bottom position and 2979°C·h at the central position. Based on maturity theory, it could be considered that the compressive strength at the upper position has achieved the 7 th standard strength, while that at the middle and bottom positions did not reach the 7 th standard strength.
is can also be a good explanation as to why coring in the construction site can sometimes fail in which only the upper part is solid and the bottom part is fairly loose, as shown in Figure 9.

Strength Gain Law under Outdoor Natural Tempera-
ture Curing. Figure 10 shows the strength gain law under outdoor nature temperature curing. e compressive strength gained with increasing curing days. e strength at day 7 was 2.2 MPa accounting for only 62.9% of standard curing and achieved the 7 th standard strength when the number of days reached 13.

Comparisons of Strength Gain Law and
Establishing the Maturity-Strength Model

Comparison of Strength Gain Law under Four Curing
Conditions. Figure 11 provides the comparisons of compressive strength gain curves under different curing conditions. e following conclusions can be made.
(1) In all four cases, the compressive strength increased with the increasing curing time. e gain rate of low temperature curing was lower than that under the standard curing temperature curing. e gain rates can be sorted in descending (from high to low) order: standard temperature curing > outdoor natural temperature curing > constant low temperature curing > day interactive temperature curing, in which the difference between the last two was insignificant.
(2) e strength gain curves under four cases were in accordance with the logarithmic curve with the function form being f � a ln(M) − b. After the model calibration, it was found that the average gain rate for the standard temperature was      (5) Among three low temperature curing cases, the curves of the constant low temperature and the outdoor natural temperature were the same until the 11 th day, both of which were also very close to the day interaction temperature case although the day interactive gain was the slowest among these three cases. e maturity theory will be used to explain this result in the next section.

Maturity-Strength Model Estimation and
Forecasting. e cement-stabilized macadam mixture consists of mainly cement, graded macadam, and water. e composition is similar to that of the cement concrete. e only difference lies in the cement dosage. Maturity theory has been widely used to predict the strength of cement concrete. As such, from the material composition point of view, the prediction function can be established based on maturity theory to forecast the compressive strength of 5% cement-stabilized macadam mixture. Because cementstabilized macadam can be seen as cement concrete with low-dose cement, there are four functions that can be used based on the existing studies of cement concrete, including   Central  7  2057  1727  1427  1690  8  2360  1987  1641  1946  9  2660  2247  1853  2200  10  2965  2515  2068  2462  11  3265  2779  2280  2719  12  3569  3045  2498  2979  13  3877  3315  2720  Journal of Advanced Transportation the power function, logarithmic function, exponential function, and hyperbolic function [25]. e maturity of three indoor experiments was calculated and shown in Tables 4 and 5. e relationship between the maturity and strength in three cases is shown in Figure 12. It seems that the logarithmic functions f � a ln(M) − b are the best predictive curves under all three cases, and, therefore, it was used as the preferred function for cement-stabilized macadam mixture. In addition, by combining the data under all three cases and developing a single predictive model, the parameters a � 1.9358 and b � 12.183 were obtained by fitting the compressive strength and maturity data, and the correlation coefficient was R 2 � 0.9907. In short, the Maturity-Strength prediction model of 5% cement-stabilized macadam mixture was f � 1.9358 ln(M) − 12.183.
For the outdoor natural curing cases, the central position data was used for maturity calculation, It should be noted that one hour was used as the temperature range, was then accumulated into one day, and was again accumulated across days to get maturity value. Using the obtained function f � 1.9358 ln(M) − 12.183 to predict the compressive strength under the outdoor curing case, the results were shown in Table 6. Note that these results were very close to the tested strength and the correlation coefficient was as high as 99.865%, which clearly indicated the high accuracy of the model. According to the model, the compressive strength of low temperature curing can be predicted with maturity, which provides a reference to calculate the strength and determine the schedule of the construction project for engineering applications.

Conclusion
e present study discusses the compressive strength gain law of 5% cement-stabilized macadam at low temperature curing, with a particular focus on the varied temperature curing which is similar to the varied air temperature in the real world.
In this paper, experiments under three indoor temperature curing cases and one outdoor natural curing were   Table 5: e maturity data in Case 2 and Case 3 (°C·h). four cases and that the gain rate at low temperature was smaller than that at standard temperature. e gain rates can be sorted in descending order: standard temperature curing > outdoor natural temperature curing > constant low temperature curing > day interactive temperature curing. e standard strength reached 3.5 MPa on the 7 th day, while the others accounted for only 62% or so. Numerical results also indicated that to achieve the same strength of 3.5 MPa, the number of days required under each low temperature case was 14 days for both the constant low and day interaction temperatures and 13 days for the outdoor temperature. According to the temperature data and the strength information collected by several indoor tests, an estimated model f � 1.9358 ln(M) − 12.183 was established to predict the strength based on the maturity theory. e model is proved to have the ability to predict with high accuracy based on the validated results obtained from the data of outdoor tests.
As the line of research matures in the future, the characteristics associated with the compressive strength in the long term can also be investigated with more data collected over time.

Data Availability
e 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.