Variation on thermal damage rate of granite specimen with thermal cycle treatment

Abstract Temperature significantly affects the physical and mechanical properties of granite. To have a comprehensive understanding of the thermal cycle effect on uniaxial compressive strength (UCS) and thermal damage rate, a series of thermal cycle experiments on granite specimens were carried out with five types of designed temperatures and five types of cycle number of thermal treatments. The experimental results indicate that UCS decreases and thermal damage rate increases as temperature and thermal cycle increase. UCS of specimens cooled in water condition after thermal damage treatment are lower than those cooled in air condition. In addition, two new phenomena related to thermal damage rate were observed. Firstly, previous studies have shown that a rapid value reduction of UCS of specimens with one thermal cycle treatment under air cooling condition can be observed at 400∘C. While the temperature threshold for the specimens treated with more than one thermal cycle under water cooling condition increases to 550∘C. Secondly, a thoroughly antipodal evolution law of the thermal damage rate for the specimens with multiple thermal cycle treatments is also observed as compared to those treated by only one thermal cycle. These differences might be induced by the different microcrack initial time and their development speed. The new findings are important to understand the failure mechanism and variation process of physical and mechanical properties of granite specimens subjected to thermal cycles.


Introduction
Granite is a type of rock widely used for constructions (i.e., buildings, tunnels, bridges, etc.) owing to its good physical and mechanical properties, such as high uniaxial compressive strength (UCS) and modulus of elasticity [1]. However, during the life cycle, it is inevitable to encounter sudden external environmental factors which will reduce the physical and mechanical properties. Temperature is one of such key factors. Great changes in temperature, especially subjected to high temperature, different thermal expansion of different granite components will lead to a creation of inter-granular compressive and tensile forces. Once these thermal stresses exceed the strength, the original cracks propagate and new cracks initiate [2,3]. Consequently, rapid degradations of the physical and mechanical properties of the granite occur [4][5][6], which results in reducing the service life in constructions or affecting the stability of the constructions. It is therefore important to have a comprehensive understanding of the thermal effect on physical and mechanical properties of the granite.
Extensive laboratory studies have been carried out to investigate the thermal effect on physical and mechanical properties of granite [7][8][9]. Jansen et al. [10] studied the thermal induced microcracks of Lac du Bonnet granite using ultrasonic imaging and acoustic emission monitoring. Chaki et al. [11] evaluated the influence of thermal damage on porosity, permeability and ultrasonic wave evolutions. Shao et al. [12] investigated the cool rate effect on physical and mechanical properties of granite. Chen et al. [13] assessed the thermal heating rate effect on physical and mechanical properties of Beishan granite. A series of uniaxial compression tests were carried out by Yang et al. [14] to evaluate the effect of high temperature treatments (200, 300, 400, 500, 600, 700 and 800 ∘ C) on the crack damage, strength and deformation failure behavior of a type of granite. Hu et al. [15] [16] presented an investigation of thermal effect on micro-properties of granite through X-ray CT technique. These studies were all based on one time thermal treatment in term of heating the specimens to a designed temperature with a specific heating rate.
Tiskatine et al. [17] performed a series of thermal cycle experiments on two granite specimens. The charge cycle was carried out in an oven for one hour at 650 ∘ C with a heating rate of 25 ∘ C/min, and a discharge cycle was carried out in ambient air for thirty minutes at 20 ∘ C with a cooling rate of 20 ∘ C/min. The grain specimens disintegrated into small pieces after only ten thermal cycles. This study indicates that thermal cycle significantly affects the microcracks and physical and mechanical properties. So far, corresponding studies are rarely reported and comprehensive understanding of the thermal cycle effect on granite physical and mechanical properties are still insufficient. However, it is important for us to have an accuracy stability assessment of the buildings or projects constructed by granite subjected to several time fires or to several cycles of high temperatures in a fire.
In this study, a series of thermal cycle experiments on 93 granite specimens were carried out with five designed temperatures and five thermal cycles as well as related uniaxial compression tests were conducted for aforementioned aim. Based on the experimental results, the relationships among the UCS, the elevated temperature and thermal cycles were analyzed. What's more, to better evaluate the degree of the degeneration of the UCS induced by the thermal damage, a thermal damage index, Dσc, was defined. The temperature and thermal cycle effects on Dσc and the rate of Dσc changes were also discussed.

Specimen preparation
The granite specimens to be tested were cored from a granite block which was collected from Zhangzhou, Fujian  Table 1. Triplicate specimens were prepared for each group test in standard size recommended by International Society for Rock Mechanics standards [18]. The diameter was 50 mm and length was 100 mm. The end surfaces were flat with an accuracy of 0.02 mm.

Thermal cycle treatment
The charge cycle was carried out in a furnace (type CYM-100) to reach a designed temperature with a heating rate of 10 ∘ C/min, and a discharge cycle was carried out in water to have a quick cooling down to the water temperature (20 ∘ C). In order to investigate the temperature and cy-  Table 2, five group tests for each temperature were conducted with different thermal cycles, such as for 250 ∘ C, thermal treatment were carried out by 1, 5, 10, 15 and 20 times. Therefore, a total of 25 group tests for water cooling condition were carried out. Simultaneously, for comparison, 5 group tests under air cooling (AG) condition (cooled in air to room temperature at 25 ∘ C) were also conducted with one time thermal treatment (250 ∘ C, 350 ∘ C, 450 ∘ C, 550 ∘ C and 650 ∘ C). What's more, 1 group of specimens were set to measure the original UCS (UCS 0 ) of the granite without thermal treatment at room temperature (25 ∘ C). In other words, totally 31 group tests were carried out. Each group test was repeated three times based on the prepared triplicate specimens.

Uniaxial compression test
Uniaxial compression test was conducted on an MTS816 servo-controlled rock mechanics experimental system (as shown in Figure 1). Prior to the application of a constant axial strain loading rate (1.0×10 −5 /s), silicon grease was greased to the top and bottom surface of the specimen to mitigate the end friction effect between the specimen and testing machine.

Results
The mean UCS of each group test and the thermal damage index, Dσc, which was defined to evaluate the degree of the degeneration of the UCS induced by the thermal damage were shown in Table 2. UCS of each test was directly  Table 2 shows that the mean value of UCS 0 is 125MPa. It can be found in Figure 2a that for specimens treated with the same thermal cycles, UCS decreases while Dσc increases as temperature increases. Especially between 550 and 650 ∘ C, there is a significant reduction of UCS of specimens cooled in water, i.e., the USC of specimen with one thermal cycle decreases from 59 to 39 MPa (corresponding Dσc increases from 0.528 to 0.712, see the black line with square). The comparison between the specimens (1 thermal cycle) cooled in air (AG) (marked with gray line with rhombus) and cooled in water (marked with black line with square) indicates that the cooling type does not affect the UCS of the specimens at 250 ∘ C much, while it affects the UCS more significantly as the temperatures increase to 350 ∘ C and 450 ∘ C, the UCS of specimens cooled in water are smaller than those cooled in air. The same phenomenon also can be observed at 650 ∘ C. It might be due to more microcracks were formed by cooling in water at those temperatures. Figure 2b indicates that thermal cycle significantly affects the UCS, UCS decreases as the thermal cycle increases, especially at low temperature. One possible reason is that when the specimens are subjected to low temperatures, very few microcracks are formed inducing a small Dσc (see Figure 2b). The specimens thus have a high original UCS (one thermal cycle). The increment of the thermal cycles forms more microcracks and reduces the UCS significantly with a greater acceleration of the increment of Dσc. While for the specimens are subjected to high temperature, much more microcracks are formed at the first  thermal cycle resulting in small original UCS. It is difficult to quickly form more microcracks in the following thermal cycles. Therefore, the reduction rate of UCS in high temperature is smaller.

Discussions
It can be observed from Figure 2a that the temperature thresholds of having a rapid value reduction of UCS of specimens under different cooling conditions are different. In order to improve the reliability of the observed phenomenon, some test results of UCS of specimens subjected to various temperatures and cooled in AG condition from previous studies [2,[19][20][21][22][23] are shown in Figure 3a with the same type test results obtained from this study. These results indicate that the UCS of granite specimens cooled in AG condition have a rapid reduction at 400 ∘ C which is lower than those cooled in water condition. The acceleration of the value of the Dσc also can be observed at the same temperature in Figure 3b. The studies of Etienne and Poupert [24], Nasseri et al. [25,26], Chaki et al. [11], Chen et al. [22] and Inserra et al. [27] suggested that the rapid reduction of UCS at 400 ∘ C is primarily due to the microcrack initiates at such a temperature and results in an acceleration of linear expansion α (a ratio of increment in specimen length direction to the specimen length, see the red and brown dotted lines in Figures 3a and 3b). However, in this study, the temperature threshold of significant reduction of UCS of granite specimens cooled in water condition delays to 550 ∘ C, especially in specimens with large thermal cycles (see Figure 2a). This phenomenon might be caused by that microcracks initiated earlier after specimens were treated by larger thermal cycles and cooled in water which results in having a low UCS at low temperature (i.e., the UCS at 250 ∘ C is 36 MPa after 20 thermal cycles). As temper-  ature increases to 350 ∘ C and 450 ∘ C, the temperature does not have the ability to generate more microcracks. Since the temperature exceeds 550 ∘ C, more newer microcracks are generated quickly, causing rapid reduction of UCS at 550 ∘ C.
The comparison between Figures 2a and 3b shows that the evolution of the rate of the Dσc change for specimens cooled in air or water conditions are different, especially for the temperature thresholds of the acceleration of the increment of Dσc which are 400 ∘ C and 550 ∘ C, respectively. To better understand the details of the evolution of the rate of Dσc change, two types of the rate of Dσc change including dD σcT and dlogD σcN were defined in Eqs. (2) and (3), respectively.
where D σcT is defined as the Dσc related to the specimen subjected to a T( ∘ C) temperature; D σcN is defined as the Dσc related to the specimen treated after N thermal cycles. The comparison between the specimens (1 thermal cycle) cooled in AG condition (marked with gray solid line shown in Figure 4a) and cooled in water (marked with black solid line shown in Figure 4a) shows that the cooling type affects the evolution of the dD σcT . As shown in Figure 4a, dD σcT of AG condition is in an ascending order before 400 ∘ C, then decreases as temperature increases. This result is in consist with previous studies shown in . Figure 5 shows that the maximum value of dD σcT in different studies are among 400-550 ∘ C. In the ascending order stage of dD σcT , corresponding rate of linear expansion change (dα/dT) also increases which also can be observed in Figure 5. Figure 6 indicates that the acceleration of the increment of dα/dT results in an increment of linear expansion (α). The increasing of α represents the amount of the formed microcracks increasing.
While for the water cooling condition (black solid line in Figure 4a), there is a high dD σcT at the initial temperature (300 ∘ C) induced by more formed microcracks because of quick cooling process in water. During 350-500 ∘ C, temperature effect on dD σcT is small until the temperature reaches to 550 ∘ C. What's more, Figure 4a also shows that the specimens treated by more numbers of thermal cycles (i.e., 5, 10, 15 and 20) also have a high dD σcT at the initial temperature (300 ∘ C) but a thoroughly antipodal evolution of dD σcT as that of AG condition. dD σcT decreases when the values of temperature between 300 ∘ C and 450 ∘ C, and changes to increase as temperatures higher than 450 ∘ C. This indicates that thermal cycle influences the evolution process of the dD σcT . More microcracks are produced at 250 ∘ C when more of thermal cycles are carried out. The following temperatures ranging from 300 ∘ C to 450 ∘ C damage the specimens slightly further, causing smaller dD σcT with more times of thermal cycle treatment at 300 ∘ C and dD σcT decreases between 300 ∘ C and 450 ∘ C. When the temperature exceeds 500 ∘ C, thermal cycle treatment damages the specimens more quickly and dD σcT starts to increase. Figure 4b indicates that dlogD σcN is in descending order as thermal cycle increases, especially under the thermal cycle number among 1-10. What's more, low temperatures relates to a high and rapid reduction of the dlogD σcN .

Conclusion
According to the experimental results obtained from a series of thermal cycle experiments on 93 with five types of designed temperatures and five types of cycle number of thermal treatments, the following conclusions can be drawn: 1. Uniaxial compressive strength (UCS) of granite specimens decreases as temperature and thermal cycle increase. 2. In general, UCS of granite specimens cooled in water condition after thermal damage treatment are lower than those cooled in air condition. 3. Previous studies have shown that a rapid reduction of UCS of specimens with one thermal cycle treatment and under air cooling condition can be observed at 400 ∘ C. While for the specimens treated with more times of thermal cycles and under water cooling condition, a new temperature threshold of significant reduction of UCS is observed in this study which is 550 ∘ C. 4. The same phenomenon can also be observed from the thermal damage index, Dσc, which increases as temperature and thermal cycle increase. 5. What's more, a thoroughly antipodal evolution of the rate of Dσc changes for the specimens treated by several times of thermal cycles and under water cooling condition during heating process is also observed in this study as compared to that for the specimens with one time of thermal treatment and under and under air cooling condition.