Performance Study and Reliability Analysis of Desert Sand Concrete Under FTC

Abstract

Using desert sand (DS) to pour concrete is a feasible idea to solve the shortage of river sand. However, the frost resistance of desert sand concrete (DSC) is a key problem that must be solved for applying DSC in practical engineering. Therefore, this study on the performance and reliability analysis of DSC under a freeze–thaw cycle (FTC) was carried out. The DSC specimens were subjected to the FTC test with desert sand replacement ratios (DSRRs) of 0%, 20%, 40%, 60%, 80%, and 100%. Then, the appearance, mass, relative dynamic elastic modulus (RDEM), and compressive strength of DSC were analyzed and discussed. Moreover, the damage mechanism of DSC was discussed via microstructural analysis. The results indicated that as the FTCs increased, the mass loss rate of concrete increased, while the RDEM and compressive strength decreased. Among the samples, DSC-40 showed the best resistance to the FTC. After 250 cycles, the changes in mass, RDEM, and compressive strength of DSC-40 were 2.73%, 15.19%, and 27.2% lower than those of DSC0. Finally, the DSC reliability model was established by using the Weibull probability distribution method. Among all groups, DSC-40 showed the best reliability, and the failure life was 287 FTCs, which was approximately 1.55-times longer than DSC0. This model could provide a theoretical basis for the durability evaluation and life prediction of DSC structures in cold regions.

1. Introduction

The shortage of natural fine aggregate resources is a serious problem in current engineering construction. Therefore, seeking a fine aggregate to replace river sand is a hot topic [1,2]. The application of desert sand (DS) in concrete has attracted the attention of many scholars in recent years [3,4,5]. The deserts of China are mostly located in cold areas. The concrete structure in those areas is subject to freeze–thaw cycles (FTCs). The FTC significantly negatively affects the performance and reliability of concrete structures. Thus, a study on the performance and reliability of desert sand concrete (DSC) under the FTC is necessary.

Over the past decade, numerous studies have been carried out on the freeze–thaw damage of DSC. They provided valuable insights into the performance and durability of DSC under the FTC. Bai et al. [6] analyzed the flexural fatigue behavior and damage evolution of aeolian sand concrete (ASC) under FTCs. Bai et al. [7] established a damage degradation model of DSC from a macro-microscopic perspective. Bai et al. [8] conducted a cross-scale study on the mechanical properties and frost resistance durability of ASC. Li et al. [9] conducted a multi-scale study on the durability degradation mechanism of ASC under FTCs. Li et al. [10] explored its durability deterioration mechanism under carbonization and FTCs. They provided comprehensive insights into the performance and lives of ASC in complex environments. Liu et al. [11] studied the FTC resistance of DSC and offered valuable information for the application of and improvements in DSC. Dong et al. [12,13] studied the capillary water absorption characteristics of ASC under sulfate FTCs based on fractal theory, conducted an FTC test, and established a damage model for ASC. They provided comprehensive insights into the properties and durability of ASC under different conditions. Qiao et al. [14] developed an interpretable machine learning model to predict the freeze–thaw damage of DS and fiber-reinforced concrete and offered insights for understanding and preventing such damage in the concrete. Gong et al. [15] revealed the FTC resistance of DSC and provided a life prediction method for the durability assessment of DSC. Wang et al. [16] used DS to prepare 3D-printed concrete, studied the effects of sand and cement ratio on printing properties and mechanical properties, and revealed the relationship between interface porosity and mechanical anisotropy. Ji et al. [17] studied the mechanical properties of reclaimed rubber DSC and analyzed the CO₂ emission, microoptical structure, and solid freeze–thaw resistance mechanical properties of the concrete. Wang et al. [18] studied the effect of adding graphene oxide on cement-based composites of dune sand. Hou et al. [19] studied the effects of polypropylene fiber and glass fiber on the frost resistance of DSC. Wei et al. [20] studied the mechanical properties of ASC prepared using the alkali treatment of aeolian sand and zeolite powder.

So far, the damage mechanisms of DSC under FTC have been explored from various aspects. Meanwhile, reliability was crucial for the practical application of DSC. However, the study of the reliability of DSC under FTC was limited. Therefore, this study aims to investigate the performance and reliability of DSC under FTC. The mass loss, dynamic elastic modulus, and compressive strength of DSC under FTC were analyzed. The damage evolution and reliability of DSC were analyzed based on the Weibull probabilistic approach. The relationship between the reliability of DSC and the number of FTCs was revealed. It provided a theoretical basis for the promotion and application of DSC in a cold area.

2. Materials and Methods

2.1. Experimental Material Properties

River sand came from the Manas River in Xinjiang, and the desert sand was collected from the Gurbantunggut Desert. Figure 1 shows the apparent morphology of fine aggregate. The physical properties of fine aggregates are shown in

Table 1. Physical properties of fine aggregates.

Types of Sand	Apparent Density (kg/m3)	Stacking Density (kg/m3)	Fineness Modulus (mm)	Mud Content (%)	Water Absorption (%)
River sand	2038	1350	2.58	2.2	0.8
Desert sand	2630	1615	0.198	1.9	2.1

[2]. DS is finer and more absorbent than river sand. The cement was PO42.5 Portland cement (Tianye Cement Plant, Shihezi, China). Coarse aggregates were obtained from quarries near Shihezi city, with particle sizes ranging from 5 to 26.5 mm. The polycarboxylate superplasticizer was HSC-brand (Sobute company, Jiangsu, China). Testing water was drinking water (Shihezi, China). Detailed material properties are provided in the authors’ previous studies [1,2,3].

2.2. Mixed Proportion Design

The mixed proportion design of DSC is shown in

Table 2. Mixed proportion design of DSC.

Specimen Number	Water–Binder Ratio	Sand Rate	Water (kg/m3)	Cement (kg/m3)	Water Reducer (kg/m3)	Coarse Aggregates (kg/m3)	River Sand (kg/m3)	Desert Sand (kg/m3)
DSC-0	0.4	0.3	160	400	1.6	1288	552	0
DSC-20	0.4	0.3	160	400	1.6	1288	441.6	110.4
DSC-40	0.4	0.3	160	400	1.6	1288	331.2	220.8
DSC-60	0.4	0.3	160	400	1.6	1288	220.8	331.2
DSC-80	0.4	0.3	160	400	1.6	1288	110.4	441.6
DSC-100	0.4	0.3	160	400	1.6	1288	0	552

Note: DSC indicates desert sand concrete, followed by number indicating the DSRR.

. The strength grade was C40 with a fixed water–binder ratio and sand rate. The desert sand replacement rate (DSRR) of DSC was 0%, 20%, 40%, 60%, 80%, and 100%. Detailed mixed proportion design studies are provided in the authors’ previous studies [1,2,3]. No less than three sets of specimens are used in each group for a single test.

2.3. Test Method

2.3.1. FTC Test

The FTC test was conducted according to GB/T50082-2009 [21]. Each cycle lasted 4 h. The maximum number of cycles set in this experiment was 250 cycles, and the interval gradient was 25 cycles. Two kinds of specimens were used in the FTC test, which were 100 mm × 100 mm × 400 mm prism specimens and 100 mm × 100 mm × 100 mm cube specimens. Before the FTC, DSC was cured in the standard curing environment for 24 days, then immersed in water for 4 days. Then, the specimens were placed in the rubber containers of the test instrument, keeping the water 5 mm above the top surface of the test specimens. The test temperature of the specimens was kept within the range of (−18 ± 2) °C and (5 ± 2) °C. The FTC test setup is shown in Figure 2.

2.3.2. Mass and Relative Dynamic Elastic Modulus (RDEM) Test

The mass and RDEM test was conducted according to GB/T50082-2009 [21]. The mass of DSC was measured via a TD50001C electronic scale (Tianma Company, Shenzhen, China), and the RDEM was detected using a DT-20 concrete RDEM meter (Zhongke Luda, Yinchuan, China). The measurement interval was 25 cycles. These test specimens were 100 × 100 × 400 prisms. Each group was no less than three.

2.3.3. Compressive Strength Test

The compressive strength test was conducted according to GB/T50081-2019 [22]. The compressive strength was measured using a universal testing machine. These test specimens were 100 × 100 × 100 cubes. Every 50 cycles were used as uniaxial compression tests. The force-controlled loading method was used, and the loading rate was 0.5 Mpa/s.

2.3.4. Scanning Electron Microscope Tests

The deterioration of concrete microstructure has a significant negative effect on its macro performance. The sample size of 1 cm³ was extracted from the specimen after FTCs. These samples were immersed in anhydrous ethanol, followed by grinding, polishing, and gold coating for preparation. Subsequently, the microscopic morphological changes and the interface transition zone (ITZ) of the DS mortar were observed using ZEISS scanning electron microscope (Sigma Company, Ronkonkoma, NY, USA). The SEM scale is 10,000:1.

3. Experimental Results and Analysis

3.1. Appearance Damage

Figure 3 shows the appearance damage of DSC during the FTC. The appearance damage of DSC became more and more serious as the FTCs increased. Before the FTC, some original surface defects of DSC were caused by the production and curing process, but its appearance was smooth and complete. After 100 cycles, the surface mortar of the specimen began to peel off, and scattered small pits appeared on the surface of the specimen. After 150 cycles, a large area of the specimen surface mortar was peeled off, and coarse aggregate was exposed. After 250 cycles, the surface mortar of the specimen was almost completely peeled off, and a large amount of coarse aggregate was exposed.

The appearance damage of DSC during the FTC was manifested by cracking, spalling, and slagging of the concrete surface. The coarse aggregate was exposed in the most damaged area. This damage were caused by the repeated freeze–thawing of the solution next to the surface of the DSC during the FTC. Figure 4 shows a diagram of the freeze–thaw damage of DSC [23]. The surface structure of DSC and the ice formed by the frozen solution form the ice–concrete composite interface (Figure 4a). As the number of FTCs increased, the surface temperature of DSC changed rapidly, while the internal temperature of DSC changed slowly. Therefore, the temperature gradient was formed between the inner and outer layers. The tensile stress on the concrete surface occurred due to the temperature gradient [24]. The specimen surface became loose when the tensile stress exceeded the ultimate tensile stresses (Figure 4b). At the same time, the expansion stress was caused by the freezing of water inside the specimen surface as the FTC progressed. The hydrostatic pressure was caused by the compression of unfrozen water on the pore wall under the influence of expansion and contraction [25]. The growth of pores in the concrete caused by these stresses and pressure encouraged the migration of water molecules to the interior (Figure 4c). The bonding between mortar and aggregate was weakened as these stresses and pressure increased. Finally, the specimens were destroyed because of the crack that extended in specimens (Figure 4d).

3.2. Freeze–Thaw Damage of DSC

3.2.1. Mass Loss Rate (MLR)

Figure 5 shows the MLR of DSC during the FTC.

Table 3. MLR values of DSC during FTCs.

FTCs/n		0	25	50	75	100	125	150	175	200	225	250
MLR/%	DSC-0	0	0.314	0.671	1.287	2.012	2.892	3.570	4.303	5.375	6.652	7.741
	DSC-20	0	0.224	0.762	1.240	2.264	3.125	3.859	4.681	5.764	6.968	8.365
	DSC-40	0	0.102	0.212	0.422	0.690	1.067	1.432	1.879	2.715	3.637	5.016
	DSC-60	0	0.216	0.362	0.461	0.832	1.171	1.540	2.118	3.046	4.049	5.607
	DSC-80	0	0.236	0.627	0.993	1.387	1.861	2.399	3.241	4.086	5.289	6.675
	DSC-100	0	0.243	0.647	1.173	1.743	2.282	2.964	3.635	4.786	5.931	7.180

shows the MLR values of DSC during the FTC. The damage degree of DSC was assessed by measuring the mass change in DSC. The MLR of DSC decreases with the increase in FTCs. At 25 cycles, the MLR of DSC increased slowly. This result indicated that the damage of DSC caused by FTCs was minor during this stage. After 75 cycles, the MLR of DSC increased rapidly, and the differences between the groups became apparent. The mass loss of DSC was due to the spalling of cement mortar and the loss of fine aggregate from the specimen surface.

The DSRR influenced the MLR of DSC. As the DSRR increased, the MLR of DSC showed a pattern of increasing and then decreasing. The order of MLRs for DSC after 250 cycles was as follows: DSC-40, DSC-60, DSC-80, DSC-100, DSC-0, and DSC-20. The lowest MLR loss rate in DSC-40 was 2.73% lower than DSC0. This is because DSC-40 has the best particle size gradation. The best particle size gradation of DSC40 resulted in a higher density and a reduction in pore space [26,27]. This result showed that the appropriate addition of DS increases the resistance of concrete to FTC and reduces the mass loss caused by FTC.

3.2.2. Relative Dynamic Elastic Modulus (RDEM)

Figure 6 shows the RDEM of DSC during the FTC.

Table 4. RDEM values of DSC during FTCs.

FTCs/n		0	25	50	75	100	125	150	175	200	225	250
RDEM/%	DSC-0	0.00	3.29	8.39	16.25	20.88	26.72	30.35	35.20	41.32	48.48	56.39
	DSC-20	0.00	3.07	10.42	16.42	23.70	29.41	34.62	38.64	45.14	52.31	58.62
	DSC-40	0.00	2.42	4.32	7.51	11.44	14.38	18.20	21.62	25.63	32.03	41.20
	DSC-60	0.00	3.27	5.32	9.41	12.77	17.10	20.14	24.17	27.59	32.95	41.98
	DSC-80	0.00	3.90	7.11	12.89	16.96	21.65	24.69	29.90	34.52	42.15	50.31
	DSC-100	0.00	3.80	8.06	14.70	18.14	23.35	27.20	31.92	36.96	46.31	53.87

shows the RDEM values of DSC during the FTC. The internal damage of concrete was evaluated by measuring the RDEM. The RDEM of the specimens decreased as the number of FTCs increased. In the initial stage of FTC, the RDEM of DSC increased slowly, and the difference between groups was small. After 100 cycles, the RDEM of DSC increased rapidly, and the differences between the groups became apparent. This result indicated that the internal structure of DSC deteriorated during the FTC. The expansion stress and the hydrostatic pressure were generated inside the DSC and acted on its pores and micro-cracks. This led to the expansion of micro-cracks, damage to the pore structure, and gradual loosening of the concrete structure [28]. The expansion stress was caused by the water in the pores expanding due to freezing. The DSC was cracked when the expansion tension exceeded the ultimate tensile stress. Therefore, the internal structural deterioration and damage accumulation of DSC led to freeze–thaw damage.

As the DSRR increased, the RDEM of DSC exhibited a trend of initially increasing and then decreasing. The order of RDEMs for DSC after 250 cycles was as follows: DSC-40, DSC-60, DSC-80, DSC-100, DSC-0, and DSC-20. The highest RDEM in DSC-40 was 15.19% higher than DSC-0. This result was because of the properties and the admixture amount of the DS. DS was an ultra-fine sand with a smooth surface, low friction, and high water absorption [29]. When the DSRR was 20%, DS led to earlier micro-cracking in ITZ and weakening of the cement mortar bond strength. When the DSRR was 40–60%, DS could effectively absorb the water spilled from the cement slurry. This led to an increase in the compactness of the concrete and an improvement in the bond strength between the cement mortar and aggregates. When the DSRR was 100%, the excess desert sand dehydrated the cement mortar. This resulted in a weaker bonding effect between the mortar and aggregate [30].

3.2.3. Compressive Strength

Figure 7 shows the compressive strength loss rate (CSLR) of DSC during the FTC. The CSLR was a key indicator of concrete resistance to the FTC [31]. Under loading, micro-cracks were systematically generated at stress-concentrated zones within the matrix of DSC, progressing through three distinct growth phases: nucleation, propagation, and ultimate coalescence into through-thickness fractures. The failure mechanism exhibited four characteristic stages:

• Stage I (Elastic Domain): At stress levels below 30% of peak capacity, initial crack formation occurred preferentially at pre-existing microdefects. No discernible surface deformation was observed, with the material maintaining linear elastic behavior.
• Stage II (Elastoplastic Transition): Between 30 and 85% ultimate load, the micro-crack grew. The surface crack emerged.
• Stage III (Stress Localization): At 90–100% peak load, vertical macrocracks aligned with principal stress vectors became visible, accompanied by localized crushing of surface aggregates.
• Stage IV (Structural Failure): Post-peak strength degradation occurred through progressive shear band formation. Dominant diagonal cracks propagated, culminating in a complete loss of structural integrity.

From Figure 7, it can be observed that the CSLR of DSC gradually increased during the FTC. In the initial stage of FTC, the CSLR of DSC increased slowly, and the difference between groups was small. After 100 cycles, the CSLR of DSC increased rapidly, and the differences between the groups became apparent. The bond strength between the aggregates and the cement mortar weakened during FTC, which was caused by the increased internal damage of DSC. Under stress, the cracks in DSC occurred along the surface of the coarse aggregates, especially the bonding surface of the coarse aggregate and cement mortar [31].

As the DSRR increased, the CSLRs of DSC exhibited a trend of initially increasing and then decreasing. The order of CSLRs for DSC after 250 cycles was as follows: DSC40, DSC-60, DSC-80, DSC-100, DSC-0, and DSC-20. The specific values were 53.7%, 57.2%, 71.1%, 77.2%, 80.9%, and 84.4%. The lowest CSLR in DSC-40 was 27.2% lower than DSC-0. The addition of DS had a two-sided effect on the compressive strength of the concrete. The addition of DS could absorb the water overflowing from the cement mortar and improve the compactness of concrete, but smooth, low-strength desert sand reduced the strength of concrete. When the DSRR was 40%, the advantages of DS far outweighed the disadvantages. DSC-40 had a higher density and friction in the cement mortar and aggregates. When the DSRR reached 20%, DS destroyed the advantage of the friction of the polygonal particles of aggregates. When the DSRR reached 100%, the overall strength of the concrete was reduced due to the lower strength of the desert sand [32,33].

3.3. Microscopic Analysis

Figure 8 and Figure 9 illustrate the microstructure of the DSC before and after FTCs. The change in the concrete microstructure affected the macroscopic properties. Microscopic changes in the cement matrix and ITZ of DSC before and after FTC were analyzed to clarify the mechanism of the freeze–thaw deterioration mechanism of DSC [34,35].

From Figure 8, no obvious initial cracks were detected on the cement mortar of DSC before the FTC. The ITZ structure of DSC was relatively complete. The cement mortar and ITZ structure of DSC-40 were the densest. A few micro-cracks and micro-pores were visible in the DSC-0 cement mortar. The cement mortar and ITZ structure of DSC-100 were loose. Connected micro-pores and fine micro-cracks existed in the cement mortar of DSC-100, which was related to the high content of DS.

From Figure 9, the cement mortar of DSC changed from dense to sparse after FTCs. The ITZ structure of DSC deteriorated after the FTCs, and holes and cracks were observed. The order of the degree of microscopic damage for DSC after FTCs was as follows: DSC-40, DSC-60, DSC-80, DSC-100, DSC-0, and DSC-20. For DSC-0, the initial cement mortar cracks were noticeably wider, and the internal structure was extremely flimsy. In particular, the ITZ structure was semi-exfoliated, and cracks in the cement mortar at the bonding surface were evident. For DSC-40, the cement mortar appeared porous and exhibited an overall flimsy appearance. The cracks and holes developed and expanded through the interior of the cement mortar. The damage to DSC-40 was significantly less than that of DSC-0. When the DSRR was 40%, the concrete density increased, the porosity decreased, and the adhesive strength of the cement mortar and ITZ structure was improved [36]. The advantage came from giving full play to the mutual filling effect of desert sand and river sand and the lubricating effect of desert sand [7,37]. In addition, the excessive incorporation of desert sand reduced the fluidity and affected the generation of cement hydration products and the degree of hydration reaction. For DSC-60, DSC-80, DSC-100, and DSC-20, the cement mortar separated from the aggregate, with obvious pore cracks and severe damage to ITZ occurred. The microscopic damage to these groups was less than that of DSC-0. This result corresponded to the macroscopic damage.

4. Reliability Analysis Model

Reliability is the ability of a product to complete the specified function under the specified conditions and within the specified time. The probabilistic measure of reliability is reliability. The failure life of concrete can be predicted by analyzing its reliability. The Weibull distribution function is proposed based on brittle failure statistical theory, statistics, and probability theory. The Weibull distribution function has been widely used in material reliability analysis and life prediction. The reliability analysis model of concrete under FTC can be described using the Weibull distribution [38,39]. Therefore, the reliability analysis model of DSC under FTC was established based on Weibull probability distribution.

A probability density function based on a two-parameter Weibull probability distribution was proposed to study the failure probability of DSC under the FTC [40]. The function is shown in Equation (1).

$$f\left(n\right)={\displaystyle \frac{a}{b}}{\left({\displaystyle \frac{n}{b}}\right)}^a\exp \left[-{\left({\displaystyle \frac{n}{b}}\right)}^a\right]$$

(1)

where a represents the shape parameter of the probability density function curve, which is the change rate of the instantaneous failure rate following FTCs and is related to the failure mode of concrete; b means the size parameter, which affects the growth rate of the failure rate and is related to the ability of concrete to resist freeze–thaw damage; n denotes the number of FTCs.

The RDEM can effectively characterize the freeze–thaw damage degree of DSC [41]. Equation (2) describes the cumulative damage of DSC following FTCs:

$$D\left(n\right)=1-{\displaystyle \frac{E_n}{E_0}}=1-P_n$$

(2)

where n represents the number of FTCs, D_n means the accumulation degree of freeze–thaw damage of DSC, E₀ and E_n denote the RDEM of DSC before and after FTCs.

The failure probability and the damage degree of DSC were accumulated simultaneously with FTCs, and they had the same logical relationship and changing characteristics. Thus, the cumulative damage of DSC can be equivalent to the failure probability, and the freeze–thaw random damage evolution equation of DSC was obtained in Equation (3):

$$D\left(n\right)=1-\exp \left[-{\left({\displaystyle \frac{n}{b}}\right)}^a\right]$$

(3)

Furthermore, a reliability function (4) was established that described the relationship between the reliability of the specimen and the number of FTCs [42].

$$R\left(n\right)=\exp \left[-{\left({\displaystyle \frac{n}{b}}\right)}^a\right]$$

(4)

The coefficients of the Weibull probability distribution function are shown in Table 3.

From

Table 5. Coefficients of Weibull probability density function.

Specimen Number	A	b
DSC-0	1.33568	305.5357
DSC-20	1.37716	276.1621
DSC-40	1.31762	484.6548
DSC-60	1.201	495.1015
DSC-80	1.22052	387.9058
DSC-100	1.25195	349.4554

, the shape parameters of DSC were consistent with those of DSC-0. The a was greater than 1. a represents the time dependence of the material failure mode and reflects the rate change of damage accumulation in the FTC. The effects of a on failure probability were as follows (Table 5).

(1)

When a < 1, the failure probability decreased with the increase in FTCs, corresponding to the initial stage of freeze–thaw micro-crack initiation. At this time, the damage caused by the ice expansion of pore water in the material had not formed a dominant damage path. The failure mode was a random failure.

(2)

When a = 1, the failure probability did not change when the FTCs increased, corresponding to the stable damage stage. The micro-crack growth and self-healing effects (such as partial pore closure) in the material reached a dynamic balance. The failure mode was an early failure.

(3)

When a > 1, the failure probability increased with FTCs, corresponding to the damage acceleration stage in the late freeze–thaw period. At this time, micro-cracks are connected to form macroscopic cracks. The failure mode was a loss failure.

b represents the number of FTCs required for a material to reach a critical damage state and is directly related to the material’s ability to withstand phase transition stresses.

The larger the b value, the longer the material characteristic life, indicating that its internal structure can effectively delay the expansion of freeze–thaw damage. The decrease in the b value reflects the deterioration of frost resistance. The repeated phase change of pore water in the FTC leads to the accumulation of internal defects.

The relationship between the failure probability growth rates of each group of DSC was DSC-60 > DSC-40 > DSC-80 > DSC-100 > DSC-0 > DSC-20. This result showed that the freeze–thaw failure of DSC was a wear failure, which was mainly caused by the accumulation of damage and fatigue of DSC, and the failure probability of DSC increases with FTCs.

The curves of probability density function and reliability of DSC in each group are shown in Figure 10 and Figure 11. From Figure 10, the curve peak of DSC-20 shifted to the left and was narrower overall compared with DSC-0. The life of DSC-20 was less than that of DSC-0, and the dispersion decreased. The curve peaks of the DSC-40, DSC-60, DSC-80 and DSC-100 curves were all after the DSC-0 baseline. When DSRR exceeded 20%, the peak of the curve shifted to the right, and the freeze–thaw damage life and dispersion of DSC were improved compared with DSC-0. This result indicated that adding desert sand enhanced the frost resistance of DSC.

From Figure 11, all groups of curves showed a decreasing trend. The curve of DSC-20 had the steepest slope and the fastest rate of destruction. The DSC-40 curve had the smallest slope and the slowest speed of destruction. Under the same FTCs, DSC-40 had the maximum reliability, followed by DSC-60. Based on the dichotomous method, the freeze–thaw failures of DSC-0, DSC-20, DSC-40, and DSC-60 were 185, 170, 287, and 281 times. The failure reliability of DSC-40 and DSC-60 was significantly better than that of DSC-0.

5. Conclusions

In this experiment, the FTC test was conducted on DSC with different replacement rates. The performance and reliability analysis of DSC after FTCs were examined. From this investigation, the following conclusions can be inferred:

(1)

As the FTCs increased, the MLR of concrete increased, while the RDEM and compressive strength decreased. Among the samples, DSC-40 showed the best resistance to the FTC. After 200 FTCs, the changes in mass, RDEM, and compressive strength of DSC-40 were 2.73%, 15.19%, and 27.2% lower than those of DSC0. The results showed that incorporating DS can improve the internal structure of concrete and effectively increase the resistance of concrete to CSE.

(2)

Damage to DSC was caused by the diminished adhesion effect of ITZ and the growth of cracks in cement mortar structures. According to the microscopic analysis, the addition of desert sand can optimize the concrete particle gradation and improve the initial defects of cement mortar and ITZ.

(3)

A reliability analysis model of DSC under FTC was developed based on the Weibull probability distribution. According to the model, the damage probability and damage degree of DSC accumulated simultaneously during the FTC. According to the results of the model, DSC-40 had the highest reliability, and its failure life was 287 FTCs. The results showed that the addition of DS can effectively increase the failure life of concrete under CSE.