2.1 Slurry hydration temperature test

To investigate the intensity and phase changes of the slurry hydration reaction, the hydration temperature was monitored in three groups (A2/B2/C2) of slurries with variations in fineness and ratios (mf/mc = 0.4). A TDS-T303 temperature detector and a transparent acrylic container measuring 50 mm × 50 mm × 50 mm were used in the experiment. The slurry was prepared and placed in a holding tank, with an initial temperature of 20°C. The experiment lasted for 240 min, and the test data were taken every 1 min.

2.2 Slurry shear test

2.3 Slurry hardening course test

According to the provisions in a Chinese code (GB/T4509-84), the experiment was conducted using an LHZR-5B needle penetration tester. The test thermostat plate was set at a temperature of 20 ± 1°C. The specimen was a cylindrical block with a height of 60 ± 1 cm and a radius of 3 cm.

Two experiments were conducted. One focused on the homogeneity of slurries. The test pieces were all made of first-grade fly ash, with mf/mc = 0.4, which were cured for 18 h in the same environment. To ensure the accuracy of the experiment and reduce errors, four measurement points were taken for each section, taking the average value as the final result. Another experiment focused on the effect of fly ash dosage on the hardening time of slurries.

3.1 Mechanisms of hydration of fly ash particles

3.2 Analysis of experimental results on slurry hydration temperature

The hydration reaction between fly ash and cement generates significant heat. By monitoring these temperature changes, it is possible to investigate the various stages of slurry hydration and how they are affected by the fineness and content of fly ash. Table 4 shows the experimental data, and Figure 2 shows the temperature change curve during slurry hydration.

The monitoring results showed that the temperature profiles of all three groups of slurries had two peaks at different times. The first peak of each group appeared at 40, 60, and 80 min, respectively. The second peak appeared at 180, 200, and 220 min, respectively. In terms of time sequence, the peak temperature of the slurry in group A2 appeared earliest and that group C2 appeared the last. The reason is that the slurry in group A2 had the highest fineness of fly ash, which resulted in the largest contact area between its particles and a faster reaction rate with cement. This led to a higher rate of temperature increase in A2. The fluidity of the slurry is positively related to the temperature, and thus the fluidity of the slurry and the optimum grouting time can be controlled by increasing the fineness of fly ash.

The hydration reaction of fly ash consumed a significant amount of Ca(OH)2, resulting in the formation of a silicic acid gel and fibrous calcium sulfo aluminate crystals with greater toughness. Calcium sulfo aluminate crystals have better resistance than Ca(OH)2 crystals, and their tensile properties promote the strength enhancement of the filler material. Calcium silicate hydrate has high strength and low alkalinity, which increases the quantity of hydrated cementitious substances in cement (Blissett & Rowson, 2013; Lam et al., 2000) and greatly improves the quality of the resultant slurry. This conclusion needs to be combined with the results of the density experiments below, from which it is shown that fly ash allows for a more complete hydration reaction of the cement, increasing the density after hardening. However, the hydration of fly ash must withstand alkaline conditions, and ensuring the minimum value of Ca(OH)2 is therefore essential, that is, the amount of fly ash substitution is limited.

The content of the glass body in fly ash and the specific surface area of the glass body determine the reaction rate (Lee et al., 2003). In the reaction with Ca(OH)2, the glass body is easily corroded to release free SiO2 and Al2O3, after which the hydration reaction takes place. Therefore, the finer the fly ash, the better its effect as an admixture.

4.1 Analysis of the results of direct shear experiments

4.1.1 Shear strength of slurries after 2 h

The shear strength curve of slurries after 2 h is shown in Figure 3. Under the condition of mf/mc = 0.4 and vertical pressure of 400 kPa, the maximum shear stress of the slurry in Groups A2, B2, and C2 was 46, 57, and 64 kPa, respectively. Group A2 had the lowest strength and Group C2 had the highest strength. The slurry with fly ash with higher fineness showed lower shear strength in the early stage, indicating that increasing the fineness of fly ash in the slurry can reduce the early strength and increase the overall strength of slurries. Pumpability prevented the slurry from hardening prematurely and locking the shield shell. The setting time of slurries can be controlled by adjusting the fineness and ratio of fly ash.

4.1.2 Shear strength of slurries after 18 h

After the shield was advanced for 18 h, the shield passed the previous segment. At this time, the slurry in the shield gap must maintain high shear strength to reduce settlement. First, second, and third grades of fly ash were used in the slurries in Groups A2, B2, and C2, respectively, and the vertical pressure was 400 kPa.

The shear strength curve of the slurry in Group A is shown in Figure 4. The maximum shear stress of A0 (mf/mc = 0), A2 (mf/mc = 0.4), A4 (mf/mc = 1.0), and A6 (mf/mc = 2.5) slurry was 250, 430, 32.6, and 23 kPa, respectively. In summary, the unique physical structure and chemical properties of fly ash can significantly improve and enhance the structural strength, homogeneity and compactness, and shear strength of slurries. If the cement content in the slurry is insufficient, sufficient hydroxide ions cannot be provided, resulting in insufficient hydration and a decrease in the strength of slurries.

Figure 5 presents the shear strength curve of Group B, where the maximum shear stress was 170 kPa, which was 2.5 times lower than that of Group A2, with mf/mc = 0.4. The maximum shear stress of B4 (mf/mc = 1.0) and B6 (mf/mc = 2.5) was 32 and 21 kPa, respectively, indicating that the fineness of fly ash has a huge influence on the strength of the slurry. With an increase in fineness, the particle contact surface increased and thus the sufficiency of the hydration reaction increased, resulting in increased later strength. The overall strength of the slurries in Groups B4 and B6 was lower than that of the slurries in Groups A4 and A6, but the difference was small. This indicates that when mf/mc increases to a certain value, increasing or decreasing the fineness of the fly ash has a relatively small effect on the final strength of the slurry. Moreover, the fly ash itself has low strength after hydration, and therefore cannot completely replace cement.

The shear strength curve of Group C is presented in Figure 6. The maximum shear stress was 140 kPa, compared to that of 20 kPa for Group C4 (mf/mc = 1.0). The fineness of the fly ash in Group C2 was four times lower than that in Group B2, and the strength difference was about 1.2 times. Under the same conditions, the difference between the fly ash fineness of Groups A and B was two times. With the decrease of fineness of fly ash to a certain value, its influence on the strength of the slurry gradually weakened.

Figure 6c shows the slurry strength curve of Group C6 (mf/mc = 2.5). Under the vertical pressure of 200 kPa, the sample was damaged and overflowed from the shear box when the shear displacement reached 1.5 mm. The failure occurred after three repeated operations, indicating that the slurry in Group C6 could not withstand the vertical pressure of 200 kPa. The strength was low, and the reinforcing effect of the third-grade fly ash was poor at the same proportion.

4.2 Analysis of viscosity test results

Viscosity is an important parameter reflecting the characteristics of cementitious materials. After injection of slurries into the excavation gap, excess viscosity will cause problems such as uneven filling, accumulation, and pipe blockage (Zhang et al., 2021). The hydration process of fly ash slurries is generally similar to that of ordinary cement slurries, as shown in Figure 7. The hydration process of fly ash slurries mainly included the following stages: (1) Bulking period. This stage was mainly affected by the montmorillonite in the sodium soil, which increased the viscosity of slurries. (2) Induction period. This was mainly controlled by the cement. When the cement was initially hydrated, a weak alkaline environment was formed, releasing heat. (3) Neutralization period. Ca(OH)2 and other alkaline substances reacted chemically to form gelling substances such as calcium silicate hydrate and calcium aluminate hydrate. (4) Alkali increasing period. Due to the addition of water glass, the viscosity of slurries decreased in a short period. However, with the further increase of hydroxide ions, the reaction between the cement and fly ash became more intense, and the viscosity of the slurry gradually increased. (5) Strengthening period. The hydration reaction of slurries continued, and the viscosity continued to increase until the slurry hardened.

4.3 Analysis of hardening test results

The penetration results of the six sections of Groups A2 (with first-grade fly ash), B2 (with second-grade fly ash), and C2 (with third-grade fly ash) after 18 h are shown in Figure 8. The penetration of the six sections of Group A2 showed a small change concentrated in the range of 2.0–2.5, and the span was small, indicating the consolidation of slurries. The properties were relatively consistent. Group B2 performed slightly worse than Group A2 for a larger span. Group C2 showed the worst performance, and the penetration span was the largest. This indicates that the fineness of fly ash has a greater impact on the consistency of the strength of the slurry after coagulation. The higher the fineness, the higher the slurry strength and the higher the consistency.

Figure 9a shows the softness and hardness curves of the first-grade fly ash for mf/mc = 0.4. Within the 120 min before the slurry was allowed to stand, the penetration of the three groups of slurries was larger, ranging from 10 to 17 mm. In the 200-min interval, the needle penetration from greatest to least was Group A2 > Group B2 > Group C2, indicating that the slurries with high-fineness fly ash showed higher penetration in the early stage with greater softness and lower strength. Within the time interval of 10–100 min, the hydration reaction of the slurries in Groups B2 and C2 was faster than that of the slurries in Group A2, and the rate began to slow down after more than 100 min. To better explain the influence of fly ash fineness on the change of slurry strength, the concept of “mutation domain” was introduced to indicate the mutation range of slurry strength. At this time, the strength of Groups B and C began to be exceeded by that of Group A, indicating that the high-fineness fly ash slurry lasted for a long period of time, with low strength in the early stage and high strength in the later stage.

The softness and hardness curves of the first-grade fly ash for mf/mc = 1.0 are shown in Figure 9b. The difference is that the mutation domain appeared at different times. The mutation domain in the group with a higher fly ash ratio appeared later than that in the group with a lower fly ash ratio. The mutation domain in the mf/mc = 0.4 group appeared at about 220 min, while that in the mf/mc = 1.0 group appeared at about 300 min. This indicates that increasing the fly ash ratio can delay the mutation domain, that is, it can prolong the hydration time of slurries and avoid the early strength phenomenon of slurries.

Combined with the shear strength, it can be seen that increasing the fineness of fly ash can effectively increase the later strength of slurries and reduce their early strength. Note that this finding fits in with the early low shear strength properties.

4.4 Analysis of the results of the solidification rate experiment

Figure 10 shows the solidification rate of slurry homogeneity. According to the characteristics of fly ash, adding fly ash to a slurry will promote the hydration reaction and reduce its early strength. The solidification rate represents the ratio of the final volume of slurries to their initial volume (expressed as a percentage) and can reflect the filling effect of slurries after firming under the condition that the strength index is satisfied. The results showed that the three types of slurries with fly ash of different fineness all showed a high solidification rate for mf/mc smaller than 0.75, and the solidification rate gradually decreased with the increase of mf/mc. In the mf/mc range of 0.75–2.5, the finer the fly ash, the higher the solidification rate of slurries. After this interval, the fineness of fly ash had a greater influence on the solidification rate, while the effect of the content of slurries was small.

4.5 Analysis of density experimental results

From Figure 11, it can be observed that the slurry density of first-grade fly ash was higher than those of the other two groups, while the slurry density of control group A0 was lower than that of A1. This indicates that the addition of fly ash can effectively increase slurry density. Based on the density performance of the three groups, it can be concluded that increasing fly ash content beyond a certain range can decrease slurry density. Thus, to achieve the desired grouting effect, the proportion of fly ash and cement needs to be reasonably controlled within a specific range.

5.1 CT slice porosity

Two cross-sections of slurries after 48 h are shown in Figure 12. These sections showed numerous gray spots that were preliminarily hypothesized to be unreacted fly ash and cement granules within the slurry. To further analyze the strength performance of the solidified fly ash content, X-ray CT scanning was performed on the sample, and the slurry structure was measured using this method. The CT scanning method allows for visual observation of the pore structure distribution within the slurry, with the advantages of high efficiency, convenience, and excellent imaging effects (Henry et al., 2014; Villagrán-Zaccardi et al., 2018; Zhou et al., 2019).

Figure 13 shows the CT scanning results of A0 (mf/mc = 0), A2 (mf/mc = 0.4), and A6 (mf/mc = 2.5) groups including first-grade fly ash (Cui et al., 2019). The cylindrical samples had a radius of 15 mm and a height of 60 mm. The scanning results revealed that Group A0 (without the addition of fly ash) had a higher degree of porosity (black) and uneven distribution of pores. Additionally, there were several agglomerates with excessive density (white), indicating that the cement hydration was incomplete. For mf/mc = 0.4, the overall porosity of slurries reduced, and the pore distribution became uniform, with almost no high-density agglomerates. This demonstrated that the slurry was fully hydrated and possessed excellent integrity and high strength after consolidation. When the mf/mc ratio was increased to 2.5, the slurry showed large-diameter pores in the middle with an even distribution. Further, a small content of high-density agglomerates (hoop-shaped particles) appeared in the upper half of the slurries, after the addition of fly ash. Fly ash itself had a lower hydration strength than cement and cannot entirely replace cement.

In addition, combined with the results of the shear experiments, it can be found that when mf/mc = 0.4, the slurry had the lowest number of pores and the pore size was smaller than those of the other two groups of slurries. The shear strength was higher and the shear strength curve was smoother.

5.2 Number of microscopic pores and microchannels

A solidification slurry also has many tiny pores and channels, which will squeeze and connect with each other to form larger defects when subjected to external forces. These slow developments are often overlooked and lead to adverse consequences, and it is therefore necessary to perform statistical calculations and analyses of these small holes and channels (Suzuki et al., 2017). The results of the statistical analysis of pores (Table 5) demonstrate the following findings: (1) the number of pores with an equivalent radius greater than 10 μm in the slurry of Group A2 was the lowest and that in Group A6 was the highest and (2) the average radius of the pores in Group A2 was the smallest, and that in Group A6 was the largest. The statistical analysis of the pore channels reveal that (1) the number of pore channels with an equivalent radius greater than 10 μm in the slurry of Group A2 was the lowest and that of Group A0 was the highest and (2) the average pore channel radius of Group A2 was the smallest. Although the number of channels in Group A6 was less than that in Group A0, the average equivalent radius was about 2.5 times that of Group A0. It can be seen that the addition of fly ash reduced the number of internal pores in the slurry, and the pores were smaller (Zhao & Zhou, 2020). This helps to improve the strength of slurries after consolidation. Although fly ash can improve the uniformity of the pore distribution, a high content can easily lead to large-diameter pores, which is not conducive to the grouting of shield tunnel gaps.

5.3 CT 3D pore distribution

AVIZO image processing software was used to create 3D models of the CT data, allowing for an intuitive presentation of slurry porosities with three different contents of fly ash (Bernardes et al., 2015; Henry et al., 2014; Zhou et al., 2019). Figure 14 shows the CT 3D pore distribution results of the three groups. Group A0 showed poor performance, with continuous pores observed in the upper section of slurries and an uneven pore distribution. Group A2 performed better, with a smoother surface and fewer pores. However, Group A6 showed the worst performance, with continuously connected pores appearing in the middle and lower sections of slurries. Moreover, pores with diameters greater than 8 mm were present throughout the slurry. These results indicate that controlling the amount of fly ash can effectively promote cement hydration, reduce the number of pores after consolidation, optimize the slurry structure, and improve its strength.

5.4 Macroporous defects determined by CT scans

Defects with large diameters are often the direct cause of soil damage, and there are many such defects in fly ash slurry. Therefore, it is necessary to analyze and count the number of these large pore defects to help determine the optimal proportion of fly ash in the slurry (Skarżyński et al., 2019). To make the pore distribution more obvious, the Bounding Box in AVIZO software was first positioned on the scanned CT slice images, after which the pores were adjusted by Interactive Thresholding. Then, new slice data were obtained, and the postprocessing of volume rendering was performed to reveal the pore space. The geometric distributions are shown in Figures 15-17.

It can be seen from Figure 15 that there were continuous pores in the upper part of the slurry in Group A0, which had a pore diameter of 5–10 mm. Moreover, the middle and lower parts of slurries performed well, and the pore distribution was uneven. Figure 16 shows the slurry in Group A2, where the number of pores in the slurry was small, the distribution was uniform, and there were fewer large-diameter pores. Figure 17 shows the slurry in Group A6. Large-diameter pores appeared in the lower two parts of slurries, but the pores in the middle were smaller and evenly distributed. The results demonstrate that the addition of fly ash can optimize the hydration structure of the slurry and reduce the porosity. However, if the content of fly ash is too high, the slurry can easily segregate and the strength will be low. From this microscopic representation of pores, it can be found that the pores of slurries significantly reduce the shear strength of slurries. The addition of fly ash reduced the pores of slurries, optimized the internal structure of slurries, and improved their overall strength.

5.5 Slurry pore network model

From the model results, it can be seen that the pore channels in the slurries in Groups A0, A2, and A6 were all located in the horizontal direction, and there were fewer pore channels in the vertical direction. The reason is that due to the influence of gravity and earth pressure, the distance between the pores in the vertical interval was long and connected channels could not be formed (Zhao et al., 2019). Second, the number of pore channels in the slurry of Group A2 with mf/mc = 0.4 was the lowest, while that in the slurry of Group A6 with mf/mc = 2.5 was the highest (Figure 18). From the microscopic pore performance of the three groups of slurries, it can be seen that fly ash can optimize the slurry structure, increase the late shear strength of slurries, improve the filling efficiency, and help to reduce displacements.