1. Introduction
Degradation of stones used as building and construction materials under severe environmental conditions is very important for many applications such as railroads, roads, and rock slopes, especially in cold areas [
1,
2,
3,
4,
5,
6,
7]. Natural rocks are subject to at least one freezing–thawing cycle annually. These repeated cycles lead to rapid changes in engineering characteristics and hence limit rock durability in the course of time [
8]. Rock damage in cold regions is influenced by a number of frost weathering cycles, rock types, temperatures, applied pressures, and rock parameters including mineral content, porosity, and pore size distribution [
9,
10]. When water turns into ice, it expands in volume by up to 9%, thus giving rise to an increase in the stress inside the pores. When the stress reaches the tensile strength of the rock, new micro-cracks are formed, and the present ones are deepened and widened; therefore, these changes can influence the rock mechanical properties and its durability [
11,
12]. The physical, mechanical, and petrographical properties of rocks are very important in the selection of natural stones used as construction and building materials subject to freezing–thawing weathering [
13]. Characterizations of petrographical properties are significant tools to evaluate the durability of rock materials against accelerated weathering agents and provide information about the mineralogical composition and textural properties and provenance of the rock materials’ origins [
6].
Fort et al. [
14] used Schmidt hardness and ultrasound velocity for in situ testing of rock characteristics of stone monuments and their quarries. They concluded that the results of these two nondestructive tests are reliable indicators of the degree of rock deterioration. Takarli et al. [
10] studied the influence of freezing–thawing cycles on microstructure in saturated granitic rocks by testing compression wave velocity and permeability. They founded that a decrease in compressional wave velocity and permeability measurements for these rocks. However, Martínez-Martínez et al. [
15,
16] and Hashemi et al. [
17] reported that the spatial attenuation is more sensitive ultrasonic property for detecting damage processes in carbonate building stones. Yavuz et al. [
18] established a model to evaluate the index parameters (compressional wave velocity, Schmidt rebound and mechanical unconfined compressive strength) of deteriorated carbonate rocks from the initial index properties after 20 freezing–thawing cycles. Saad et al. [
19] studied the effect of water flows into porous network on freezing–thawing cycles of rocks. Karaca et al. [
20] experimentally studied the Böhme abrasion and wide-wheel abrasion parameters before and after frost weathering tests. They proposed the statistical models for abrasion values before and after frost cycles. Tan et al. [
21] investigated the deterioration in the mechanical characteristics of granite as a function of freezing–thawing processes by uniaxial and triaxial compression tests. Mutluturk et al. [
22] postulated a mathematical model to predict the decrease of integrity process of rocks due to freezing–thawing processes and provided different meaningful properties for rock decay or rock durability. Sousa et al. [
23] derived an empirical formula between the rock strength and porosity for a sample of granitic rocks. Bayram [
24] developed a statistical model to predict the rate of decrease values in uniaxial compressive strength. In this model, impact strength, elastic modulus, and absorption of water were used as independent variables. Baud et al. [
24] proposed an analytical model in which unconfined compressive strength, initial porosity, and crack density in a rock were related. Many studies investigated the decay of physico-mechanical properties of rocks due to freezing–thawing action [
10,
16,
18,
21,
25,
26,
27,
28,
29,
30,
31].
In the previous works, estimation of damaged stones used as building and construction material due to freezing–thawing conditions were classified into two groups—physical and mechanical. The physical parameters mainly include compressional wave velocity, porosity, and mass, which can be noninvasively measured. The destructive mechanical parameters include compressive strength, modulus of elasticity, tensile strength, and impact strength. The destructive mechanical test of rock materials due to freezing–thawing action is the most direct and plain method for estimating the uniaxial compressive strength decrease for stones used for building and constructions. Unconfined compressive strength is the most important mechanical method to determine the strength of materials. However, it is time-consuming, expensive, and sample preparation is difficult. The strength loss of rock materials against frost conditions is an important indicator to select stones for building and construction purposes and assess the stability of rock engineering in cold regions. Therefore, predicting the strength parameters by the other physical and mechanical nondestructive test methods is popular and preferable. However, most of the previous studies are focused on destructive mechanical tests and some different numerical methods. Therefore, these tests are insufficient to predict the unconfined compressive strength of carbonate building materials due to frost weathering process. Moreover, this test is generally applied with natural environmental conditions.
Our current research aims to predict the unconfined compressive strength loss from nondestructive physical tests of some carbonate building materials such as limestone against frost weathering. Specifically, it aims to obtain significant or reliable relationships between unconfined compressive strength decrease and nondestructive parameters variations including compressional wave velocity, spatial attenuation, and porosity of studied materials. It also, aims to provide useful information regarding their durability in order to ensure the long-term viability or sustainability of these materials used for constructions against frost weathering conditions.
The remainder of this paper is organized as follows. In
Section 2, materials description, and laboratory experiments setup details such as petrographic analysis, frost attack process, unconfined compression test, porosity, water absorption, and ultrasonic detection system are presented.
Section 3 presents the results and discussions in addition to evaluation of the relationships between
RDσc and some nondestructive properties of studied limestone samples. Finally, conclusions are drawn in
Section 4.
3. Results and Discussions
The mean values of the initial physical and mechanical properties of studied fresh limestone specimens (unweathered) are summarized in
Table 2.
The unconfined compressive strength, porosity, compressional wave velocity, and spatial attenuation were obtained after the end of each 10 cycles. Thirty samples for each type of rock (Six samples for each 10 cycles), and their mean values and standard deviation are given in
Table 3,
Table 4,
Table 5 and
Table 6, respectively. As can be seen, the greater the number of cycles, the greater the porosity and spatial attenuation and the lower the unconfined compressive strength and compressional wave velocity of all studied limestone samples. The relation between unconfined compressive strength (
σc), compression wave velocity (
Vp), and the number of cycles for studied samples is presented in
Figure 1 and
Figure 2.
To indicate the variation in critical parameters, the change ratios of unconfined compressive strength, porosity, compressional wave velocity, and spatial attenuation were obtained. The rate of increase and decrease determines the degree of rock material breakage and it changed between 0 and 1. A rating value of 1 was assigned to rocks that were fully damaged after the end of 50 frost cycles. The increase in ratios of an effective porosity and spatial attenuation after the end of every 10 frost cycles was determined using Formula (2) outlined by Khanlari et al. [
8]:
where
RI is the increase ratio of a property (effective porosity, and spatial attenuation),
rn is the property after the end of each 10 frost cycles, and
ro is the property value of fresh specimens. The relationship between the increase ratio values of effective porosity (
RIpe) and the number of frost cycles of studied samples (
N) are indicated in
Figure 3. As can be observed in this Figure, after 50 cycles of frost action, ZK01-H2 and ZK01-H1 limestone samples indicate the highest increase ratio in porosity (
RIpe), with 1.1795 and 1.0983, respectively, in comparison with the other specimens. Limestone samples of ZK02-H1 and ZK02-H2 show slightly higher increased ratio in porosity with 0.9385 and 0.8718, respectively, whereas CK-H1 limestone samples exhibit the lowest increase ratio with 0.6923 in comparison with the other specimens. These variations are considered to be related to the mineralogical composition of these specimens, especially calcium carbonate (calcite) mineral, which exhibits a higher expansion–contraction property in the frost weathering action, and these changes lead to rock breakage or deterioration [
42]. According to the degree of variation due to the frost attack process, new cracks or fractures have been observed in limestone samples (namely ZK01-H2). As shown in
Figure 1, there were slightly higher changes in the ZK01-H2 specimens after the 10th to 40th cycle, while in other samples,
RIpe increased with the increased number of frost cycles.
On the other hand, the relationship between
RIαs and number of frost cycles of specimens (
N) are illustrated in
Figure 4. As can be seen, after 50 cycles of frost, ZK01-H2 samples have the highest increase ratio in spatial attenuation (
RIαs), with 0.0630 compared to other samples, while ZK02-H2, ZK01-H1, and ZK02-H1 samples have increased ratios of spatial attenuation with 0.0490, 0.0370, and 0.0340, respectively. Finally, CK-H1 samples have the lowest increase ratio, with 0.0270 compared to other samples. Artificial weathering tests led to internal disintegration of rock materials, which scatters the ultrasound waves and results in increase of spatial attenuation [
14]. Therefore, the variation degree of these parameters depends on internal deterioration degree of the rock material [
43], which determines the durability of materials. Depending on these results, it can be said that the ZK01-H2 samples have the lowest durability due to having the highest values of
RIαs, while CK-H1 samples have the highest durability because
RIαs has the lowest values for these samples due to frost process. As shown in
Figure 4,
RIαs increases with an increased number of frost weathering cycles of studied limestone samples.
The high absorption of water, breakage, or disintegration and the weakening and expansion of natural rocks lad to the development and formation of microfractures and cracks due to frost action, and hence, the reduction in the strength of the stones is higher. The decrease in ratios of the unconfined compressive strength and the compressional wave velocity after the end of each 10 cycles of frost were determined using Equation (3),
where
RD is the decrease ratio of the property (unconfined compressive strength, and compressional wave velocity),
rn is the property after the end of each 10 frost cycles, and
ro is the initial property of fresh sample. The relationship between the decrease ratios in unconfined compressive strength (
RDσc), compressional wave velocity (
RDvp) values, and the number of frost cycles of specimens (
N) are given in
Figure 5 and
Figure 6, respectively. It is observed that the progression of the frost cycles caused reduction in unconfined compressive strength and compressional wave velocity. According to Mokhtari and Behnia [
44], there are some critical parameters affected on the rock materials strength and elasticity including rock type, mineralogical composition, porosity and water absorption capacity. As can be seen in
Figure 5, after 50 cycles of frost action, ZK01-H2 specimens have the highest decrease in unconfined compressive strength value (
RDσc) with 0.3903 and hence the highest deterioration because of their higher porosity and water absorption in addition the noticeable presence of talc mineral. Therefore, ZK01-H2 is the most sensitive to frost process in comparison with the other samples. Ruedrich et al. [
45] indicated that the water and moisture content of natural rock materials are significant factors during frost action. Wet freezing–thawing process proceedings are considered limited because natural rock materials have to be moistened before frost action. The presence of water in the rocks is an important parameter for the damage action [
46].
On the other hand, limestone samples of CK-H1 exhibit the highest resistance to deterioration because these samples have the lowest value of
RDσc with 0.0909 due to having the lowest moisture content and porosity under frost weathering action in comparison with the other samples. Erguvan [
47] showed that the percentage reduction in uniaxial compressive strength must be less than 20% for rocks used as building materials. If the reduction value of these materials is more than 20%, they should not be used in cold areas. Depending on these results, most studied samples can be used in cold areas as building and construction materials except limestone samples namely ZK01-H2 should not be used in cold areas due to having high percentage reduction in unconfined compressive strength.
Figure 6 presents the relationship between the decrease ratio values of compressional wave velocity (
RDvp) and the number of frost cycles of specimens (
N). After 50 cycles of frost, limestone samples of ZK01-H2 have the highest decrease in compressional wave velocity value (
RDvp) with 0.0386, which resulted in the highest internal deterioration in comparison with the other samples while, samples of CK-H1 have the highest resistance to damage due to having the lowest
RDvp value with 0.0104 compared to the other samples. The reduction in index properties of rock materials after frost cycles are found to be closely related initial index parameters including porosity, absorption of water, and mineralogical composition. Accelerated weathering processes lead to rock material breakage or degradation and result in variation in the index characteristics of the rocks [
42]. According to the present results, it is observed that high porosity, water absorption, and mineralogical composition are critical characteristics for determining the strength loss of studied limestone samples against frost attack processes.
Relationtionships Evolution between RDσc and Limestone Properties Variations
The correlation diagrams were used to investigate the relationships between decrease rate of unconfined compressive strength (RDσc) and other nondestructive physical characteristics variations of studied samples, including the increase ratio in porosity (RIpe) and spatial attenuation (RIαs) and the decrease ratio in compressional wave velocity (RDvp) realized due to frost attack cycles. Moreover, we attempt to find the higher correlation between these parameters. The dataset was analyzed using linear regression analysis and the coefficient of determination (R2) was determined for each regression. Correlation coefficient was determined to assess the strength and direction of the correlations the correlation between RDσc and the other disintegration parameters.
Figure 7,
Figure 8,
Figure 9,
Figure 10 and
Figure 11 present the relationships between
RDσc and
RIpe &
RIαs at 10, 20, 30, 40, and 50 cycles, respectively, and the relationships between
RDσc and
RDvp are illustrated in
Figure 12,
Figure 13,
Figure 14,
Figure 15 and
Figure 16.
As can be observed in
Figure 7,
Figure 8,
Figure 9,
Figure 10 and
Figure 11, there are no significant correlations between
RDσc and
RIpe. As these Figures show, the low values of their coefficients of determination (
R2) and hence the correlation between them is poor. As shown in these diagrams, depending on the number of cycles, a moderately correlation was observed between
RDσc and
RIpe after the end of the 10th and 20th cycles of frost. The moderate values of their coefficients of determination with
R2 of 0.58 and 0.61, respectively. In addition, their scatter plots show that some points are close to the regression lines and the others are far and thus the correlation between them are moderate. The progression of frost cycles caused a decrease in relationship between
RDσc and
RIpe. The greater the number of cycles, the lower the correlation coefficients.
Figure 10 and
Figure 11 indicate that there is a strong correlation between
RDσc and
RIαs with large values of
R2 of 0.8584 and 0.8344, respectively, after the end of 40th and 50th frost cycles, and their scatter plots indicate that the stronger the relationship, the closer the data points fall to the regression lines also, which corresponds to the best fitting linear relationship for obtaining the decrease of unconfined compressive strength. It is observed that the progression of frost cycles led to increase in relationship between
RDσc and
RIαs. These higher relationships indicate that the increase ratio of spatial attenuation (
RIαs) is a more sensitive ultrasound parameter for detecting any degradation processes and formation of new microcracks in materials during frost processes (depending on the number of frost cycles). Therefore, the highest relationship between
RDσc and
RIαs can be used for pre-estimation of unconfined compressive strength loss (
RDσc) for studied limestone types.
As can be seen in
Figure 12,
Figure 13,
Figure 14,
Figure 15 and
Figure 16, the further the progression of frost cycles, the higher the relationship between
RDσc and
RDvp.
Figure 13 and
Figure 16 indicate that there is a strong correlation between
RDσc and
RDvp with
R2 of 0.9089 and 0.9025, respectively, at the end of 20th and 50th frost cycles. Consequently, these relations are a strong indicator for pre-determining the unconfined compression strength loss of studied samples.
These correlations also represent a positive and linear relationship. According to the number of cycles, it can be observed that the increase rate of spatial attenuation (RIαs) and the decrease rate of compressional wave velocity (RDvp) are very sensitive for analyzing the fabric deterioration of limestone samples against frost weathering processes.
Therefore, the ultrasonic detection system is a powerful technique for investigating the durability of studied samples during frost attack processes. Overall, it is concluded that the RIαs factor related to spatial attenuation obtained at the end of 40th frost attack cycles can be used to estimate the unconfined compressive strength (RDσc) reduction. Moreover, RDvp parameter related to compressional wave velocity decrease determined at the end of 20th and 50th cycles of frost can be used for pre-estimation of unconfined compressive strength decrease of selected studied samples without tests. These relations may be useful at the preliminary stage of design but should be used with caution and only for the selected samples. The results also support the use of nondestructive tests and provide valuable data when choosing limestone samples.
4. Conclusions
The employment of nondestructive methods to determine the mechanical strength and durability of carbonate building and construction materials under frost conditions is gaining great significance. For this purpose, the present study aimed to determine the relationships between unconfined compressive strength decrease and some nondestructive physical parameters variations including the increase ratios of porosity and spatial attenuation and the decrease ratios of compressional wave velocity of some limestone types used as building and construction materials and subjected to 50 frost attack cycles. This provided useful information for the assessment of their long-term durability in order to ensure the long-term viability or sustainability of these materials used for constructions under frost action.
The main conclusions drawn from the present study can be summarized as follow:
(1) Most of studied limestone samples can be used as building and construction materials for different outdoor applications in severe cold areas, except ZK01-H2.
(2) The high porosity, water absorption capacity, and presence of some accessory minerals such as talc are critical parameters for determining the strength losses of samples from nondestructive physical parameter changes against frost attack processes.
(3) There is a high linear relationship between RDσc and RIαs at 40th frost cycles with R2 of 0.8584. Moreover, there is also a strong linear correlation between RDσc and RDvp at the end of 20th and 50th frost cycles with R2 of 0.9089 and 0.9025. Consequently, the ultrasound measurements including compressional wave velocity decrease and spatial attenuation increase factors can be successfully used to pre-determine the unconfined compression strength loss of only the studied limestone types against frost conditions without testing. It also may be useful at the preliminary stage of design but should be used with caution and only for limestone types in many engineering applications in cold regions.
This study also creates a paradigm for future studies for pre-estimation of unconfined compressive strength loss of studied limestone used as building materials subject to frost weathering. However, to evaluate the relations between materials characteristics clearly, future studies need to use different conditions and construction materials in order to obtain more important results.