Frost Resistance Differences of Concrete in Frequent Natural Freeze–Thaw versus Standard Rapid Method

Abstract

In order to find the anti-freezing durability differences between concrete in the frequent natural freeze–thaw conditions in the northwest of Sichuan Province, China, and concrete in the rapid freeze–thaw conditions of the standard rapid method, the typical temperature and humidity of the northwest of Sichuan Province were simulated. The results showed that the average number of freeze–thaw cycles in the northwest of this province can reach up to 150 per year. The relative dynamic modulus of C30 ordinary concrete, which is 100% pre-saturated, still remained above 90% after 450 cycles in simulated environments. However, during the rapid freeze–thaw test, even the C30 air-entrained concrete failed after 425 cycles. Compared to the saturation degree of concrete itself, the continuous replenishment of external moisture during freeze–thaw cycles is a key factor affecting the frost resistance of concrete. Rapid freeze–thaw reduces the number of the most probable pore sizes in ordinary concrete, and the pore size distribution curve tends to flatten. The reduction rate of the surface porosity of air-entrained concrete before and after rapid freeze–thaw is only about one third of that of ordinary concrete.

1. Introduction

The Jiuma Expressway is located in the northwest of Sichuan Province, China, with an average altitude of 3000–3800 m. The climate and environment are characterized by a large temperature difference between day and night, with frequent alternation between positive and negative temperatures and repeated freeze–thaw cycles. At present, the standard method for evaluating the frost resistance of concrete is the rapid freezing method, but many studies have shown significant differences in the frost resistance of concrete under natural service conditions compared to those obtained using this method. Yuan Bin et al. studied the relationship between rapid freeze–thaw conditions and natural freeze–thaw conditions in hydraulic concrete indoor environments. Their investigation indicated that the deterioration phenomenon worsened as the lowest temperature of the cycle decreased [1]. Li Haoyu et al. established a freeze–thaw relationship between laboratory concrete and natural hydraulic structural concrete in the Jilin region of China based on the parameters of the dynamic modulus of C25 specimens and the ultrasonic speed index [2]. Studies have found that existing concrete freeze–thaw life prediction models are mostly based on the regression of experimental data to obtain the relationship between a certain indicator and the number of freeze–thaw cycles or modeling methods that use theoretical derivation to obtain evolution equations and then fit the parameters of the evolution equations with experimental data [3,4,5,6,7]. However, due to the assumptions of the model and the fact that projections are unverified by real service life, the applicability of frost resistance evaluation models is relatively limited. Currently, most engineering construction projects do not use the estimated frost resistance service life of models.

At present, the common methods used to improve the frost resistance of concrete include air entrainment and the internal mixing of admixtures. There seem to be inconsistent conclusions in the literature regarding the use of the air entrainment method to improve the frost resistance of plateau concrete. Li et al. compiled meteorological data from 20 meteorological stations in the Qinghai Tibet Plateau region over the past 40 years and obtained a formula for calculating the average annual freeze–thaw cycles through statistical regression. Experiments showed that air entrainment in concrete decreases linearly with a decrease in air pressure. Therefore, it is recommended to increase the air content of ready-mixed concrete in high-altitude areas compared to that used for plain areas [8]. Zhang et al. used nuclear magnetic resonance to analyze the pore structure of concrete in the Qinghai Tibet Plateau region before and after freeze–thaw cycles. They believed that concrete with a gas content of 3.2–3.8% had the best frost resistance. Nuclear magnetic resonance showed that the gas content in this range effectively enhanced the internal pore structure of hardened concrete [9]. Dai et al. showed that negative temperature curing can reduce the frost resistance of high-strength concrete, making it unable to achieve the same frost resistance and service life as positive temperature curing. The optimal initial air content for the frost resistance of high-strength concrete is 4–5% [10].

Some studies have also improved the frost resistance of concrete by adding admixtures. Qu et al. showed that the addition of basalt fibers and silica fumes can improve the frost resistance of concrete by 25% [11]. Sumanta Das et al. studied a microencapsulated phase-change material, which was incorporated into concrete layers to reduce the freeze–thaw effects on bridge decks. The effectiveness of the microencapsulated phase change for the protective layer of concrete was mainly demonstrated using finite element analysis and multi-scale simulations [12]. Dong et al. used n-tetradecane and expanded perlite to produce the composite phase-change material EPC14, and the study also showed that concrete with 20% addition of this material experienced the smallest amount of damage after 200 freeze–thaw cycles [13]. Sun et al. showed that the addition of 1–2 mm rubber fine aggregate, being raised from 0% to 5.6%, can increase the F100 of concrete from 76.6% to 86.5% [14]. Arasteh-Khoshbin et al.’s research showed that adding up to 6% nano-SiO₂ improved the compressive strength before and after freeze–thaw cycles [15]. Mohammed et al. studied the beneficial effects of anti-freeze–thaw measures by mixing fly ash, blast furnace slag, silica fume, and metakaolin with concrete to enhance the concrete’s performance [16]. Lee et al.’s research showed that concrete in which 10% of the ordinary Portland cement was replaced with air-cooled slag and water-cooled slag exhibited a similar durability against the freeze–thaw cycle as that of concrete in which 40% of the ordinary Portland cement was replaced with water-cooled slag only [17].

Chen et al. showed that saturation and water migration have a significant impact on the freeze–thaw damage of cement-based materials. In water environments, unsaturated specimens are more susceptible to frost damage before a freeze–thaw event [18].

In the literature, there is relatively little experimental research on the differential damage to concrete materials caused by the natural environment and rapid freezing environment under high-frequency freeze–thaw cycles. Model research is generally biased toward theory, and its guiding significance is limited for practical engineering. In this study, in order to explore the actual service status of concrete along the entire Jiuma Expressway in small regional environments and adopt targeted measures to improve the frost resistance and durability of concrete, we conducted freeze–thaw tests, simulating the natural temperature and humidity environment of the region. To compare the difference in influence on concretes under the rapid freezing method and the simulated natural freeze–thaw environment of concrete, the weight loss and the relative dynamic modulus of elasticity were tested every 25 freeze–thaw cycles. Moreover, a computerized tomography (CT) scan and a mercury injection test (MIP) were carried out. The reasons for the differences in the changes to the micro-pore structure of concrete before and after exposure to different freeze–thaw environments and the rapid natural freezing method were explained from a microscopic perspective. Finally, according to the research rules, the environment where the bridges were located—along the Jiuma Expressway—was analyzed. Measures to improve frost resistance and durability are provided for every kind of concrete facility along the Jiuma Expressway.

2. Materials and Methods

2.1. Materials

Commercial cement (grade 42.5R) was used in the experiments. The cement properties are shown in

Table 1. Properties of cements.

Requirement of Normal Consistency	Setting Time (min)		Density (g/cm3)	Specific Surface Area (m2/kg)	Loss (%)	Compressive Strength (MPa)		Bending Strength (MPa)
	Initial	Final				3 d	28 d	3 d	28 d
29.3	165	205	3.07	376	3.33	31.5	53.7	5.2	7.5

. Fly ash was used as a mineral admixture and the performance is shown in

Table 2. Performance of fly ash.

Fineness (45 μm Square Hole Sieve Residue) (%)	Activity Index (%)	Water Demand Ratio (%)	Loss (%)	Sulfur Dioxide (%)	Stability (Boil)
15	85	97	6.2	1.8	qualified

. Crushed sand was used as fine aggregate and the measured value of the fineness modulus is 2.78. Crushed stones with particle sizes of 4.75–19 mm and 19–31.5 mm were used as normal coarse aggregates. The sieving results of crushed stone of 4.75–19 mm are shown in

Table 3. Sieving results of 4.75–19 mm crushed stone.

Sieve Size (mm)	26.5	19	16	9.5	4.75	2.36
5–10 mm pass rate (%)	100	100	100	95.3	12.6	0.6
10–20 mm pass rate (%)	100	99.1	75.3	0.5	0.1	0.1

. A polycarboxylate-based superplasticizer was used to guarantee that fresh concrete had a considerable slump. The mixing content was 1% of the cement content, which kept the concrete slump value in the range of 160 mm to 220 mm. The test method was in accordance with the ASTM C143/C143M-20 [19]. A triterpenoid saponin air-entraining agent was used to guarantee that fresh concrete had a 4–5% air content and a mixing content of 1–1.5 ten-thousandths of the cement content. The dosage of the air-entraining agent was certified using the air content of freshly mixed concrete and the test method was in accordance with ASTM C 231/C231M-24 [20]. The properties of the cement were tested based on the Chinese standard GB 175-2020 [21]. The properties of the fly ash were tested based on the Chinese standard GB/T 1596-2017 [22]. The cumulative screening percentage of the natural gravel, with a size of 4.75–19 mm, was tested based on the Chinese standard GB/T 14685-2022 [23].

2.2. Mix Proportions

The mix proportions of the concrete are shown in

Table 4. Mix proportions of concrete (kg/m³).

Concrete ID	Cement	Fly Ash	Water	Natural Sand	Crushed Stone (4.75–19) mm	Crushed Stone (19–31.5) mm	Water Reducing Agent	Air Entraining Agent
C30-normal	322	56	165	807	642	428	3.8	0
C30-air	322	56	165	807	642	428	3.8	0.047
C50-normal	460	0	154	740	1109	0	4.79	0
C50-air	460	0	154	740	1109	0	4.79	0.064

. C30 and C50 concretes with or without air content were made to test the anti-frozen properties. The C30 concrete without air entrainment was classified as C30-normal. The C30 concrete with air content was classified as C30-air. The same categorization applies to C50.

2.3. Experimental Design

2.3.1. Rapid Freeze–Thaw Test

We prepared 3 blocks of C30 and C50 concretes with or without being air-entrained separately. The block sizes were all 100 mm × 100 mm × 400 mm. We placed the frost resistance test blocks of C30 normal concrete, C30 air-entrained concrete, C50 normal concrete, and C50 air-entrained concrete that had been cured for 28 days into the freeze–thaw test box in the rapid freeze–thaw test machine. The rapid freeze–thaw method was based on ASTM C666/C666M-15, Procedure A [24]. Figure 1 shows the procedure of the rapid freeze–thaw method. The relative dynamic modulus is calculated in Formula (1). It was used to characterize the degree of internal damage and durability of concrete after undergoing physical erosion, such as freeze–thaw cycles.

$$P_c={\displaystyle \frac{n_1^2}{n^2}}\times 100$$

(1)

• $P_c$ = relative dynamic modulus of elasticity, after c cycles of freeze–thaw, percent.
• n = fundamental transverse frequency at 0 cycles of freeze–thaw.
• $n_1$ = fundamental transverse frequency after c cycles of freeze–thaw.

2.3.2. Simulated Natural Environment Freeze–Thaw Test

A temperature survey was conducted before the simulation. The temperature data were obtained from the National Weather Service of China. Daily changes in high and low temperatures along the Jiuma Expressway are shown in Figure 2 and Figure 3. From October to April of the following year, the daily high and low temperatures crossed 0 °C every day. The Jiuzhi and Hongyuan are two typical plateau counties along the Jiuma Expressway.

In theory, not all positive and negative temperature changes will lead to freeze–thaw cycles. For example, a high temperature of only 0.1 °C for 2 h cannot cause the ice in concrete to melt into water, and an average high temperature of 5.5 for only 0.5 h cannot effectively melt the ice in the concrete. Based on existing experimental experience data, it is possible to complete a freeze–thaw cycle when the duration of high and low temperatures exceeds 4 h.

To understand possible freeze–thaw cycles in the region, hourly temperature data analysis was conducted on the coldest months (December and January) in Jiuzhi County. The data were also obtained from the National Weather Service of China. The average temperature above freezing point and below freezing point was calculated for dates spanning 0 °C, as shown in

Table 5. Average value of positive and negative temperatures during the coldest month in Jiuzhi District.

Classification	December	January
Positive temperature	5.5	3.7
Negative temperature	−14.8	−13.1

. Further detailed temperature surveys were conducted in December and January to observe the duration of positive temperatures throughout the day in the coldest month, as shown in

Table 6. Variation in temperature by hour during the coldest month in Jiuzhi District.

Positive Temperature Duration (h)	Frequency (d)
	December	January
0	10	6
1	1	0
2	1	0
3	0	3
4	1	3
5	3	2
6	4	5
7	4	6
8	5	3
9	1	1
10	1	2
Accumulated freeze–thaw cycles	21	25
Accumulated completely frozen	10	6

.

From Table 5, it can be seen that, in December in Jiuzhi County, during the period of crossing 0 °C, the average high temperature was 5.5 °C and the average low temperature was −14.8 °C. In January in Jiuzhi County, during the period of crossing 0 °C, the average high temperature was 3.7 °C and the average low temperature was −13.1 °C. From Table 6, it can be seen that the average high temperature ranges from 3.7 °C to 5.5 °C. The positive temperature can last for 7 h in a day. The average low temperature ranges from −13.1 °C to −14.8 °C and can last for 17 h in a day. According to the temperature requirements of the rapid freeze–thaw method in ASTM C666/C666M-15, the freezing temperature is 18 ± 2 °C. The melting temperature is 5 ± 2 °C and the melting time is not less than 1/4 of the entire freeze–thaw cycle time. It can be seen that the environmental temperature change in this area is very close to the freeze–thaw temperature of the rapid freeze–thaw method. The temperature difference and duration of high and low temperatures can complete one freeze–thaw cycle per day in this area. As shown in Table 6, even in the coldest month, there are positive and negative temperature changes that meet the freeze–thaw cycle conditions more than 20 days.

A core sampling was taken at the edge of the beam slab of a water-crossing small arch bridge on a national expressway near the Jiuma Expressway. During a period of no snowfall, the measured moisture content of the concrete was about 60%, which was close to the annual average atmospheric humidity of 65% in the surveyed area. Within 24 h after snowfall, the measured moisture content of the core test concrete reached 100% saturation. A test was conducted to prepare natural freeze–thaw test blocks based on 100% saturated water.

After 28 days of curing, we placed the test block in distilled water and then removed it after 4 days. We wiped it until the surface of the test block was dry and then weighed it. We wrapped and sealed it tightly with more than 10 layers of cling film. We placed the sealed test block in a high and low temperature alternating test box and set it to melt at 5.5 °C for 7 h and freeze at −14.8 °C for 17 h. We simulated the freeze–thaw behavior in a natural environment so it was consistent with the rapid freeze–thaw method. We removed the test piece every 25 cycles, along with the cling film, weighed the mass loss, and tested the fundamental transverse frequency according to the China Standard GB/T 50082-2009 [25] and calculated the relative dynamic modulus. If it was found that the moisture content of the test block had decreased during weighing, the test block was placed in water again to absorb water until saturation. We then resealed and placed it in a freeze–thaw test box with a temperature that alternated between high and low. If there was no significant change in moisture content loss during weighing, it would be directly sealed. Figure 4 and Figure 5 show the sealing situation between the test device and the test block. Figure 6 shows the procedure of the simulated freeze–thaw test.

2.3.3. Pore Structure Change Test

Using industrial CT to detect the pore changes of C30 concrete before and after freeze–thaw tests, only data within a depth range of 5mm from the concrete surface were selected for statistical analysis. The block test numbers are shown in

Table 7. Specimen code for CT.

Type	Test Conditions	Specimen ID	Cycle Number
Ordinary concrete	Quick freezing	N-ac	25
	Simulation	N-st	300
Air-entrained concrete	Quick freezing	A-ac	25
	Simulation	A-st	300

. The surface pore situation should be tested with a complete concrete surface without obvious cracking. Therefore, the CT test was performed before the concrete was destroyed by the freeze–thaw test. Ordinary concrete shows weakness in resisting the rapid freeze–thaw cycle, while 25 cycles are safe according to test experience. Although air-entrained concrete is much better at bearing the rapid freeze–thaw cycle, the contrast is considered with the same cycle number with ordinary concrete. Thus, for the rapid freeze–thaw test, the cycle number is 25 for all kinds of concrete. For the simulated freeze–thaw test, the cycle number is 300 for all kinds of concrete.

To avoid the influence of the forming process on the pore structure, the test blocks used for mercury intrusion testing were the same batch of formed test blocks. At 28 days of curing and after the freeze–thaw test, one group of C30 normal concrete and one group of C30 air-entrained concrete test blocks were all individually broken. The particles required for mercury intrusion testing were prepared by the samples. The specimen code for MIP is shown in

Table 8. Specimen code for MIP.

Types	Test Conditions	Specimen ID	Cycle Number
Ordinary concrete	Quick freezing	MIP-N-ac	50
	Simulation	MIP-N-st	425
Air-entrained concrete	Quick freezing	MIP-A-ac	425
	Simulation	MIP-A-st	425

.

3. Analysis of Experimental Data

3.1. Analysis of Rapid Freeze–Thaw Test Results

Figure 7 shows the variation in the measured relative dynamic modulus of ordinary concrete and air-entrained concrete with the number of freeze–thaw cycles. Figure 8 shows the appearance of the specimen when it is damaged. As shown in Figure 8, the C30 ordinary concrete was destroyed after 50 freeze–thaw cycles in the rapid freeze–thaw test. The relative dynamic modulus decreased to 60%. Although the mass loss did not exceed 5%, it had already reached 4.98%. The C30 air-entrained concrete reached the failure condition when the measured mass loss was 5.73% and the relative dynamic modulus was 94.6% and the number of freeze–thaw cycles reached 425. Based on the appearance, when the C30 ordinary concrete is damaged, there is obvious looseness and slag dropping when touched by hand. Stable fundamental frequency can no longer be measured. However, when C30 air-entrained concrete is damaged, the surface mortar falls off and the aggregates are exposed. There is no looseness after cleaning the surface of the blocks. The overall integrity of the test block is good. The stable fundamental frequency can still be measured. From Figure 7, it can also be seen that, for C50 concrete, the relative dynamic modulus did not decrease and there was no significant mass loss after 725 rapid freeze–thaw tests.

3.2. Analysis of Simulated Freeze–Thaw Test Results

Figure 9 shows the relative dynamic elastic modulus of ordinary concrete and air-entrained concrete in the simulated freeze–thaw environment. Figure 10 shows the appearance of the specimen after 450 cycles in a simulated environment. It can be seen that the C30 and C50 concrete exhibit good performance in terms of freeze–thaw resistance whether they are air-entrained or not. After 450 freeze–thaw cycles in a natural simulated environment, the relative dynamic modulus is still above 90% and there is basically no mass loss.

The C30 normal concrete failed after 50 cycles in the rapid freeze–thaw test, while it remained intact after 450 cycles in the simulated environment test. In the rapid freeze–thaw test, the test block is completely immersed in the test box and water is added to the box to completely submerge the test block. During the freeze–thaw test, the block is completely exposed to water. During the melting process, the volume of water in the concrete decreases and the water around the test piece is saturated. During replenishment before the freezing process, the ice volume expands. Theoretically, free water should be squeezed out. However, the water between the box and the test block also undergoes freeze–thaw. The icing pressure affects the block. The water in the concrete pores cannot be effectively squeezed out. It can only freeze inside the concrete pores, thereby exacerbating the damage to the internal pores of the concrete due to ice compression. In the simulated natural temperature and humidity environment test, although the test block was sealed with water in advance during the freeze–thaw process, the area around the test block was covered with cling film and the outside of the cling film was in air. There was no rigid constraint around the test block. The overall external constraint was weaker than that of the rapid freeze–thaw method. The pore water in the concrete expanded during the freezing process, and some free water could seep out of the concrete.

In theory, there is a “frost resistance durability conversion factor” that can be calculated by using the ratio of the number of freeze–thaw cycles at which the same concrete achieves the same mass loss or relative dynamic modulus under rapid freeze–thaw cycles and natural freeze–thaw cycles. Moreover, it can be predicted that different types of concrete have different conversion factors for frost resistance durability, which is related to the different pore structures of different types of concrete. This can be supported by the functional relationship between concrete pore parameters and concrete dynamic modulus loss under different degradation factors established by Yu et al.’s research result, which shows that, when predicting macroscopic dynamic modulus loss based on changes in micro-pore structure parameters, the weight coefficients of models established for different types of concrete (ordinary concrete, concrete with added fly ash, and concrete with added fly ash and air-entraining agent) are different [26].

3.3. Analysis of Changes in Concrete Pore Structure

3.3.1. Surface Structure

Table 9. Porosity variation of the shallow concretes before and after the freeze–thaw.

Specimen ID	Before Freeze–Thaw (%)	After Freeze–Thaw (%)	Change Rate (%)
N-ac	1.10	1.42	29.1
N-st	1.13	1.06	−6.2
A-ac	4.35	4.77	9.7
A-st	4.28	4.13	−3.5

presents the porosity data of different types of concrete within 5 mm of the surface before and after exposure to different freeze–thaw environments. It can be seen that the porosity of the surface of ordinary concrete changes significantly before and after the rapid freezing test. The porosity of air-entrained concrete increases before and after the rapid freezing test. The change rate is only about one third of that of ordinary concrete. After simulating natural environmental freeze–thaw, as shown in Table 9, the porosity of concrete decreased. This may be due to the lack of external water exchange in the concrete. The insufficient frost heave of water in its own pores causes surface pore damage. During the melting process, due to the sealing and curing effect of the cling film, the concrete further hydrated, resulting in a denser surface. This is consistent with the experimental data, showing a slight increase in the relative dynamic modulus of C30 concrete after freeze–thaw in the simulated environment.

Table 10. Bubble spacing coefficient variation of the concrete before and after the freeze–thaw.

Specimen ID	Before Freeze–Thaw (μm)	After Freeze–Thaw (μm)
N-ac	527	673
N-st	536	522
A-ac	165	185
A-st	542	534

shows the variation in the bubble spacing coefficient before and after freeze–thaw cycles. The smaller the bubble spacing coefficient, the better the frost resistance of concrete. It can be seen that the bubble spacing coefficient of both ordinary concrete and air-entrained concrete significantly increases before and after rapid freeze–thaw and the micro-pores deteriorate. The deterioration of ordinary concrete is more severe. Before and after simulating the freeze–thaw cycles in natural environments, there is little change in the micro-pores on all the surfaces of air-entrained concrete and ordinary concrete, which is basically consistent with the change law of total porosity shown in Table 9.

3.3.2. Internal Pore Structure

The particles used in the mercury intrusion test are sampled from the entire specimen after breaking. The pore structure of the sample is tested after all particles are mixed. The results represent the overall pore structure changes of the concrete specimen before and after freeze–thaw. Figure 11 shows the changes in pore size distribution of concrete before and after rapid freeze–thaw and simulated natural environment freeze–thaw tests. It can be seen that, between the two freeze–thaw test methods, the maximum pore size of ordinary concrete does not change much but the content decreases and the pore size distribution curve tends to be uniform. Meanwhile, in the air-entrained concrete, the maximum pore size and content of air-entrained concrete do not change significantly. The proportion of the most probable pore sizes in air-entrained concrete also changes relatively little, indicating that the pore structure of air-entrained concrete is less susceptible to freeze–thaw cycles [27,28]. From Figure 11b, it can also be seen that, after simulating the natural freeze–thaw environment, the pore size distribution of air-entrained concrete at 100–1000 nm shows a trend towards small pores and the number increases. This may be because there is no water exchange between the concrete and the outside world in the simulated freeze–thaw test. The internal water cannot evaporate so it supports a self-curing effect [26,29,30]. This microstructural change is also consistent with the slight increase in the relative dynamic modulus of the macroscopic frost resistance performance index of the simulated natural freeze–thaw environment, as shown in Figure 11. The formula of the relative dynamic modulus is shown in Formula (1).

4. Discussion

As shown in Figure 9, to obtain the cycle number for achieving the same damage for the rapid freeze–thaw cycle as the simulated freeze–thaw cycle, the cycle test needs to be performed year by year. However, the duration and funding of this study are both limited. The freeze–thaw cycle number cannot be detected for achieving the same damage level for the rapid freeze–thaw as the simulated environment. Therefore, this study was unable to establish a proportional conversion relationship between the rapid freezing method and the simulated natural freeze–thaw method. In the future, it is recommended to conduct in situ tests directly on components of the same size and environmental degradation and regularly cut the required size specimens for relative dynamic modulus testing.

5. Conclusions

In this study, we simulated the representative temperature and humid environment in high-altitude and cold areas in the northwest district of Sichuan Province, China. We compared natural environment freeze–thaw with the rapid freeze–thaw test method. We used the relative dynamic modulus and quality loss of concrete with the number of freeze–thaw cycles as evaluation indicators. Using CT scanning and MIP testing, the surface pore structure and internal pore structure were tested before and after freeze–thaw. The surface porosity, bubble spacing coefficient, and overall pore size distribution from a microscopic perspective were used as evaluation indicators to explore the differences in the frost resistance ability of concrete under the rapid freeze–thaw method and the high-frequency natural freeze–thaw environment. The main research findings are as follows:

(1)

The rapid freeze–thaw method has a more severe destructive effect on concrete than simulating natural freeze–thaw cycles. The rapid freezing method of C30 air-entrained concrete was damaged 425 cycles during rapid freezing. In the simulated natural environment, the relative dynamic thaw modulus of C30 and C50 concrete remained above 90% after 450 freeze–thaw cycles and there is basically no mass loss.

(2)

Air entrainment has a positive effect on improving the surface resistance to the freeze–thaw erosion of concrete. Although the surface porosity of C30 air-entrained concrete increased before and after the rapid freezing test, the change rate was only about one third of that of C30 ordinary concrete. The surface porosity of air-entrained concrete remained basically unchanged before and after simulating natural environment freeze–thaw cycles. Both rapid freeze–thaw and simulated environmental freeze–thaw can reduce the number of the most probable pore sizes in normal concrete. The pore size distribution curve tends to flatten and homogenize the pores. However, the proportion of the most probable pore sizes in air-entrained concrete changes relatively little. The simulation of natural environment freeze–thaw has little effect on the internal pore structure of both ordinary concrete and air-entrained ones.

(3)

During the freeze–thaw cycle, the continuous replenishment of external moisture into the concrete is an important factor affecting its frost resistance. With the same number of freeze–thaw cycles, the morphology of the test blocks in simulating high-frequency freeze–thaw natural environments is better than that of the rapid freeze–thaw method after freeze–thaw cycles.

(4)

For the high-frequency freeze–thaw in the natural environment in the northwest of Sichuan Province, China, an average number of over 100 freeze–thaw cycles happen per year. It is recommended to introduce high-quality bubbles with a bubble spacing coefficient of 4–5% between 200 μm and 250 μm into the premixed concrete components of C30 and below for the culvert wall body, the eight-shaped retaining wall, the section below the water level line of the cross-river pier and abutment, and the back of the concrete retaining wall on abundant water content in the mountains. In addition to introducing high-quality bubbles with a bubble spacing coefficient of 200 μm to 250 μm into the premixed concrete during the construction process, it is also recommended to seal the components after construction to significantly improve their frost resistance, such as covering them with epoxy mortar, a polyurethane coat, and a silane impregnated layer.