2.1 Geographic location

The Cuihongshan iron-polymetallic mine is situated at the northern foothills of Xiaoxinganling in Xunke County, Heilongjiang Province, Northeast China, with elevations between 426 and 668 m. The northern part of the deposit area features a river valley geomorphology, whereas the southern part comprises hilly mountainous terrain and flat areas along the Kuerbin River.

2.2 Engineering geological conditions

The mining area is located within a tectonic belt characterized by two sets of conjugate fractures, which exhibit a well-developed fracture network. The strata that host the ore body primarily belong to the lead mountain strata, which are characterized by a predominance of dolomitic carbonate rocks exhibiting significant chemical reactivity. Based on the borehole data, the strata prevalent in the exploration area consist of backfill crushed stone, loose sand, and gravel layers. Figure 1 shows that the underlying bedrock consists of limestone, granite, skarn, syenite, and magnetite layers within the lead mountain strata of the Lower Cambrian system.

2.3 Hydrogeological conditions

The largest river in the area, the Kuerbin River, measures 10–12 m in width and less than 2 m in water depth, crossing the northern part of the mine from east to west. It is the primary factor for water ingress in underground shafts. The hydrogeological conditions at the Cuihongshan iron-polymetallic mine are complex. The groundwater aquifers primarily comprise the loose layer aquifer of the Quaternary system and bedrock fissure aquifer, exhibiting favorable recharge conditions. The main sources of groundwater recharge in the mine are atmospheric precipitation and river water. There are mainly aquitards at the bottom of the Quaternary system and bedrock. However, gaps in these aquitards allow for groundwater recharge, and a close hydraulic connection exists with the Kuerbin River's water in the upper part of the area.

2.4 Cause of the accident

2.4.1 Water–sand mixture inrush process

In May 2019, during ore extraction at an elevation of 190 m coupled with engineering blasting activities, vibrations emanated from the goaf above an elevation of 310 m. These vibrations led to the dislodging of the rock layer on the northwestern side of the middle section, triggering the collapse of the ground surface. Subsequently, river water, laden with silt and sand, cascaded into the underground, flooding the remaining mine area through the 58-line filled roadway, ultimately leading to the accident. Within 22 h, the volume of the water-sand mixture reached 300 000 m3. A collapsed pit with a radius of 80 m and depth of 20 m manifested approximately 10 m east of the Kuerbin River directly. The collapsed pit was filled with accumulated water (Figure 2).

2.4.2 Mechanisms of the accident

The primary factors contributing to water-sand mixture inrush include material sources, pathways, energy, and storage space.

1. Material sources: Potential sources of the incident include the Kuerbin River, the collapsed pit, the loose layer aquifer, and the fractured aquifer. The chemical composition of the water was analyzed, focusing on factors such as sodium (Na+), calcium (Ca2+), bicarbonate (HCO3−), sulfate (SO42−), total hardness (TH), and total dissolved solids (TDS).

By comparing the hydrochemical factors, connections between aquifers can be established, and the water sources involved in the incident can be identified. Figure 3 shows that the water from the collapsed pit, originating from the river, closely resembles the six hydrochemical factors found in the middle section at a depth of 310 m. Core drilling at depths of 110–116 m revealed a significant quantity of loose sand and pebbles. A comparative analysis of particle size characteristics indicated that the grain size composition of the collapsed sand samples closely resembles that of the river sand sourced from the Kuerbin River. The accident material sources are attributed to the water from the Kuerbin River and the river sand located at the bottom of the river.

2. Pathways: A permeable geological structure zone, which is water-bearing, exists between the water from the Kuerbin River and the underground space. With excavation, the pathway allowed a mixture of surface water and sand to enter the underground space.

3. Energy and storage space: The water pressure of 1.17 MPa combined with the vibrations induced by engineering blasting at a depth of 190 m, propelled the water–sand mixture through the aquitards. The quarry located above 310 m served as a storage space for the water–sand mixture (Figure 4).

Given the complex distribution of water bodies and ore veins, treatment efforts became increasingly challenging. Therefore, the intricate mechanisms of grouting and filling require integrated treatment strategies.

2.5 Cement–sodium silicate slurry

Although single-cement slurry is inexpensive and easy to use, stratum fragmentation and high-water flow rates often cause it to spread before solidifying, rendering it ineffective for water plugging and stratum reinforcement. Some chemical slurries might address this issue, but they are costly and pose a high risk of environmental contamination (Cinar et al., 2020; Qin et al., 2023; Yu, 2023).

The two-liquid slurry meets treatment requirements while offering the advantages of low cost and a simpler process. Various factors, including temperature, cement strength, water–cement ratio, the volume ratio of the two liquids, and the Baume degree of sodium silicate, significantly affect the setting time and strength of the two-liquid slurry. The cementitious body exhibits accelerated setting and enhanced strength with increased temperature, a high Baume degree of sodium silicate, and higher cement strength. The water–cement ratio in the cement-sodium silicate slurry is suitably adjusted between 0.5:1.0 and 1.0:1.0 based on the site, with the volume of sodium silicate being one-tenth that of the cement slurry. The sodium silicate modulus was calibrated to range between 2.4 and 3.0, with a concentration range from 30 to 40 °Be, utilizing 32.5 MPa slag silicate cement.

3.1 On-site engineering methods for field investigations

3.2 Theoretical analysis

This study employed an orthogonal experimental design to construct an orthogonal table and an orthogonal array comprising four factors, each characterized by four distinct levels. The range analysis method was utilized to assess the data collected during the study. This methodology provides a comprehensive understanding of the relative importance of each factor influencing grouting efficacy.

3.3 Microscopic detection

A scanning electron microscope (SEM) was employed for the microscopic examination and characterization of the grouted rock samples. The SEM was a widely adopted technique that employs an electron beam to scan the sample's surface, thereby facilitating the acquisition of microstructural morphology, extraction of parameters, and determination of sample composition. Subsequently, the results of the cemented sample can be analyzed under a magnification of 500 times using the scanning electron microscope. Concurrent analysis of the energy spectrum at the interface can also be conducted. The objective of this analysis is to evaluate the influence of the slurry on the cementing and filling processes of fractured strata under flowing water conditions (Figure 5).

3.4 Experimental design

Figure 6 shows that the computer-based intelligent grouting system controls every step of the grouting process. Cement and water from the cement tank were blended into a slurry by the high-speed mixer using an air pump. After that, the slurry was piped to the low-speed mixer to keep it in a semi-solid state for grouting. To ensure the efficacy of grouting in water sealing and rock reinforcement, the slurry was ultimately injected into the base of the grouting boreholes using the grouting pump, where it was combined with sodium silicate.

Numerous scholars have analyzed the results of flowing water grouting through physical modeling tests. Four factors were determined: fissure width, water flow rate, grouting flow rate, and slurry solidification time play a dominant role in affecting the outcomes (Sui et al., 2015; Xu et al., 2023; Zhan et al., 2011). The measurable prototype parameters of these four factors in field engineering were the Lugeon value (A), which characterizes the stratigraphic fragmentation, water flow rate (B), grouting flow rate (C), and water–cement ratio (D).

Based on four typical strata, the stratigraphic fragmentation (A) will be revealed. Other parameters were essentially the same: water flow rates below 10 000 m3/day, a two-liquid slurry containing sodium silicate, and a 1:1 water–cement ratio where sodium silicate is one-tenth of the cement volume.

Based on on-site periods and flow rate conditions, the water flow rate (B) was divided into four levels: >10 000, 7000–10 000, 4000–7000, and <4000 m3/day. A two-liquid slurry with a water-cement ratio of 1:1 was utilized, and it is the fissure type.

The grouting pressure at the surface, which was modified by variations in the development of underground fractures, controls the grouting flow rate (C). There were four levels to the grouting pressure-provided range of flow rates: >110, 90–110, 70–110, and <70 L/min. The two-liquid slurry was also utilized in the fissure-type stratum with a water flow rate of 4000–7000 m3/day and a water–cement ratio of 1.0:1.0.

Four ratios were selected as the independent variable in the field to determine the water-cement ratio (D): 0.5:1.0, 0.8:1.0, and 1.0:1.0 for two-liquid slurry, and 1.0:1.0 without adding sodium silicate for single slurry 1.0:1.0(S). The water flow rate was 4000–7000 m3/day and the stratum belonged to the fissure type.

The experimental design was conducted based on actual field conditions, and the levels of the factors were categorized in Table 1.

4.1 Typical stratigraphic classification for grouting

The grouting pipe nozzle was situated 0.3 m above a pressure sensor. A positive grouting effect is indicated by a significant drop in grouting flow rate and an increase in pressure (Jia et al., 2022). The fissures, cavities, and conductivity of the strata significantly impact the design and efficacy of grouting. Figure 7 shows that the stratigraphic fragmentation could be divided into four typical strata: cavity, hidden, fissure, and complete types.

Cavity type (Figure 7a): Large, apparent hollows with a Lugeon value > 100 Lu were visible. The hollow volume ranged from 4 to 30 m3. When aggregate filling treatment is not used, the slurry spreads throughout the surrounding area as water flows through it. The slurry only adhered to the borehole wall partially, and the hollow could not be solidified and sealed.

Hidden type (Figure 7b): Exhibiting a Lugeon value ranging from 50 to 100 Lu, where no visible cavities were present. The fractured rock of the borehole wall has resulted in localized rockfalls. Despite an increase in the concentration of the water–cement ratio, achieving proper consolidation remains a significant challenge. Additionally, the presence of large fissures and concealed cavities necessitates the application of aggregate treatment to ensure structural stability.

Fissure type (Figure 7c): Fissures exhibiting a Lugeon value between 20 and 50 Lu were observed. The slurry grouting concentration was adjusted appropriately based on the magnitude of the Lugeon value and the degree of fissure development. The unit slurry consumption varied between 1 and 10 m3/m.

Complete type (Figure 7d): The stratum with a Lugeon value of less than 20 Lu remained intact. The grouting section length was set at 15–30 m, and the borehole wall appeared smooth. Direct grouting could be performed with unit slurry consumption of 0.5–5.0 m3/m and a water–cement ratio of 1.0:1.0.

The cavity type frequently experienced elevated water flow rates, which led to the washing away of the slurry and inhibited its solidification. To accurately assess the volume, morphology, and conductivity of the cavity, laser scanning utilizing the borehole laser 3D scanning system was indispensable. Subsequently, the cavity was filled with aggregates based on crushed stone before grouting. Figure 8 shows that the borehole laser 3D scanning facilitated the encapsulation of the calculated volume of 9.65 m3. Approximately 8.5 m3 of crushed stone was then utilized to fill the cavity, significantly enhancing the grouting effectiveness.

The hidden type was characterized by highly developed fissures that were loose and fragmented. High-water flow rates posed a significant challenge, as the grouting volume across the entire area might have been constrained during the initial phases of the project treatment. Consequently, this led to an inadequate water sealing effect in the region, as the application of crushed stone might have exacerbated the issue by obstructing the borehole. To address this challenge, finer particle sizes of sand were employed. Fine sand effectively percolated through narrow fissures, serving as a conduit to access hidden voids and larger fissures, facilitated by borehole packer tests and grouting procedures. It has been observed that filling these voids with fine sand yielded satisfactory grouting outcomes.

In the fissure type, when more fissures develop, the water–cement ratio of the two-liquid slurry was changed from 1.0:1.0 to 0.8:1.0 to enhance viscosity and mitigate dispersion caused by flowing water. This adjustment ensured effective treatment by facilitating the sealing of fissures without the necessity for aggregate placement. For the complete type, direct grouting sealing emerged as the optimal approach when utilizing a two-liquid slurry with a 1.0:1.0 water–cement ratio. Furthermore, the addition of sodium silicate was deemed unnecessary when levels were below 5 Lu.

4.2 Factors affecting the propagation filling of grouting with flowing water

4.2.1 Stratigraphic fragmentation

The larger the Lugeon value, the more fragmented the stratum. There were four types of grouting segments during the treatment process: >100, 50–100, 20–50, and <20 Lu. The correlation between grouting flow rate and grouting pressure was monitored throughout the grouting process.

For the cavity type (Figure 9a), the flow rate remained stable while the grouting pressure progressively increased. Once the crushed stone aggregate was added, the grouting effect had been significantly improved. Conversely, for the hidden type (Figure 9b), there was no increase in pressure, necessitating the artificial elevation of the surface grouting pressure. Although a rapid peak in grouting pressure was observed, it subsequently declined, leaving the sealing issue unaddressed. The grouting effect was visible after adding fine sand grouting. The grouting pressure of the fissure type (Figure 9c) increased noticeably and peaked at 237 min. A brief rise in flow rate was noted during a subsequent decline, which was attributed to rapid filling following the conduction of local fissures, as indicated by the observed minor wave peak. For the complete type (Figure 9d), the pressure rose quickly following grouting, peaking at 3.52 MPa after 70 min. The grouting flow rate then started to decrease drastically. Complete type and fissure type grouting had completion speeds of 10.27 and 2.32 m/h, respectively, and high grouting efficacy was demonstrated by their respective unit slurry consumption of 0.64 and 1.25 m3/m. Sealing cavities and hidden types require aggregate filling after grouting. Thus, the end unit slurry consumption was 107.65 and 13.62 m3/m, respectively. Increased development of fissures and cavities presented challenges in grouting, filling, and sealing, as the grouting pressure exhibited a slower rise in these instances. Therefore, it is advisable to avoid grouting lengthy sections at one time.

4.2.2 Water flow rate

It is essential to maintain downhole drainage and depressurization following the accident, as this allowed for the monitoring of treatment efficacy and recording of water flow rates.

During the pretreatment phase, the grouting pressure gradually increased as the water flow rate exceeded 10 000 m3/day (Figure 10a). After reaching its peak, the grouting flow rate gradually decreased, with a steep decline occurring only after approximately 440 min. The grouting pressure increased gradually when the water flow rate ranged between 7000 and 10 000 m3/day (Figure 10b). The grouting flow rate rapidly decreased following its peak. The grouting pressure increased rapidly when the water flow rate ranged between 4000 and 7000 m3/d (Figure 10c), with the grouting flow rate decreasing shortly after peaking. When the water flow rate was less than 4000 m3/day (Figure 10d), most of the nearby fractured rock and paths had been successfully blocked. The grouting pressure increased quickly, peaking at 4.25 MPa. Immediately after reaching peak pressure, a noticeable drop in the grouting flow rate was observed.

Grouting was completed at rates of 2.10, 2.79, 4.04, and 6.63 m/h under four different water flow conditions. The corresponding unit slurry consumptions were 4.99, 4.53, 2.91, and 1.22 m3/m. Higher water flow rates facilitated slurry dispersion, necessitating that a greater grouting flow rate was required to accomplish sealing. After the grouting pressure peaked, the slurry solidified near the pipe nozzle, and the grouting flow rate did not significantly drop due to substantial water flow at the edges. The flow rate decreased once the slurry thoroughly filled the fissures and edge areas. This observation suggested that grouting became more challenging and less efficient with higher water flow rates. However, grouting efficacy significantly increased when the water flow rate was low during the later stages.

4.2.3 Grouting flow rate

The grouting flow rate was adjusted according to variations in underground fracture development, controlled by surface grouting pressure. Different grouting flow rates were analyzed for boreholes with similar stratigraphic conditions during the same period.

A higher grouting flow rate was selected under conditions of high-water flow to counter the increased flow and achieve efficient solidification for blocking when the on-site grouting flow rate exceeded 110 L/min (Figure 11a). The grouting flow rate declined in a “stepped” manner as the grouting pressure gradually increased due to the solidification of liquid slurry beneath the grouting pipe's nozzle. Once the grouting pressure peaked, the grouting flow rate fell for flow rates between 90 and 110 L/min (Figure 11b). The grouting pressure increased significantly more quickly at a flow rate of 70–90 L/min (Figure 11c), peaking at 5 MPa after 100 min. To prevent surface bubbling caused by excessive grouting pressure and flow rate, grouting flow rates below 70 L/min (Figure 11d) were used when the strata integrity was particularly high. Grouting completion speeds of 2.70, 3.26, 6.65, and 7.42 m/h were corresponded to unit slurry consumptions of 6.01, 5.52, 0.66, and 0.53 m3/m, respectively. A higher grouting flow rate was more effective when stratigraphic conditions permitted, particularly in addressing fractured strata and high-water flow rates.

4.2.4 Water–cement ratio

Adding sodium silicate and increasing the slurry concentration significantly shorten the slurry solidification time, thereby preventing inefficient propagation filling under flowing water conditions.

After a few minutes, the grouting pressure in two-liquid grouting with a 0.5:1.0 water–cement ratio (Figure 12a) rose to 0.53 MPa and stabilized. An instantaneous decrease in the grouting flow rate indicated an excessively viscous slurry. Solidification occurred near the grouting pipe and the borehole, leaving empty space at the edges. This resulted in low grouting pressure, a small propagation radius, poor propagation effectiveness, and high grouting costs. Grouting pressure increased rapidly in a two-liquid slurry with a water–cement ratio of 0.8:1.0 (Figure 12b). A subsequent drop occurred as soon as the initial grouting flow rate increased, suggesting that the fissures in the strata had been effectively blocked. The typical water–cement ratio for a two-liquid slurry was 1.0:1.0 (Figure 12c), with the grouting flow rate decreasing as the grouting pressure peaked. The water–cement ratio for a single cement slurry in fissure-type strata was 1.0:1.0 S (Figure 12d), resulting in a longer process with a lower sealing success rate. The slurry consolidation time was slow, but the grouting pressure was low, and the grouting flow rate dropped quickly. The slurry's ability to fill voids and withstand high-water flow conditions increased with its viscosity, which accelerated consolidation and enhanced sealing capability but limited the propagation radius and treatment range.

4.3 Main effects

The stratum Lugeon value (A), water flow rate (B), grouting flow rate (C), and water–cement ratio (D) were identified as significant factors influencing the grouting effect under flowing water conditions in this study. Four factors, each with four levels, were used in an orthogonal experimental design to investigate systematically.

Table 3 summarized the results of the range analysis conducted to evaluate the GSR derived from each orthogonal design.

In Table 3, K1, K2, K3, and K4 represent the average GSR values associated with specific factors, illustrating the influence of different factor levels on grouting and water sealing effectiveness. R signifies the maximum difference in GSR across various factor levels, serving as a critical index that reflects the variability of orthogonal data. The larger the value of R, the greater the influence on GSR. The results indicate that RA > RB > RD > RC, elucidating the factors affecting the effects of grouting and water plugging in the order of stratum fragmentation > water flow rate > water–cement ratio > grouting flow rate.

Figure 13 shows that the GSR decreased as water flow rate and stratigraphic fragmentation increased. The characteristics of the strata should be considered when selecting the grouting flow rate and water–cement ratio.

4.4 Propagation filling modes

Inspection core boreholes were positioned 28 days after the completion of the grouting project to assess the efficacy of the treatment applied to the collapsed rock mass and to investigate the characteristics of slurry propagation and filling. Figure 14 shows core boreholes drilled in the three areas (a, b, and c) with the highest unit slurry consumption to analyze the grouting situation of the collapsed rock mass at the site. These three areas were selected for drilling and coring due to their high unit slurry consumption and notable degrees of stratigraphic fragmentation and conductivity. On the one hand, the slurry filling mode was more clearly revealed in these locations due to the drilling and coring activities. This additional information facilitated the identification of the propagation filling mode that would be utilized subsequently. On the other hand, these regions exhibited high unit slurry consumption, indicating greater treatment difficulty. Once treated, the effectiveness in more accessible regions could be evaluated, providing a comprehensive assessment of treatment efficacy.

The characteristics of grout propagation filling and the proportion of slurry within the cemented body under flowing water conditions were established through the observation and analysis of borehole coring data. Five modes of rock–slurry cemented body fillings were identified: pure slurry, big crack, small crack, small karst pore, and pore penetration. The interrelation and analysis of the four typical pre-grouting strata and the five diffusion filling modes were presented in Figure 15.

In the pure slurry mode, grouting pressure drove the slurry into each cavity, specifically intended to fill cavities or pathways. Subsequently, once the narrow connection was sealed, the slurry accumulated and gradually filled the void. Sections of core exposures ranged in length from 15 to 50 cm and exhibited a slurry content exceeding 90%. The coloration was predominantly gray, indicating high strength.

In the big crack mode, the slurry followed the direction of significant fissure development as it spread. Once the fissure width narrowed or after the material had been displaced a considerable distance, the hydration reaction became sufficient to complete the cementation sealing. The filling angle ranged from 30° to 60°. This type typically exhibited lengths between 5 and 30 cm, fissure widths exceeding 2 cm, and cement content ranging from 30% to 90%. The inclined rock–slurry interface was notably more susceptible to separation compared to other locations.

In the small crack mode, the primary filling angles of the rock slurry were near horizontal and vertical. The grouted fissures measured less than 2 cm in width, with the slurry propagating along the formed small fissures. The slurry content varied between 10% and 30%. The rock–slurry body exhibited resistance to separation and generally demonstrated high strength.

In the small karst pore mode, propagation filling occurred within the small pores created by karst erosion. These pores exhibited various forms and were typically closed. Under karst conditions, the slurry filled and consolidated rapidly. The pore diameter varied from 0.5 to 3.0 cm, containing less than 10% slurry. The robust rock–slurry body rendered separation nearly impossible.

The pore penetration mode was typically observed in shallow strata at depths of less than 70 m, where it was inferred that the slurry had infiltrated the upper loose layer that had collapsed. This type was infrequently exposed, and the slurry, with a content ranging from 10% to 30%, was squeezed into the pore spaces of the loose layer to enhance infiltration and filling. This mode exhibited greater overall strength with minimal potential for dispersion.

Figure 16 show figures of the rock–slurry cemented body that had been magnified 500 times. The pure slurry mode (Figure 16a) exhibited flocculated and agglomerated slurry particles with diameters ranging from 5 to 50 μm. Furthermore, pores with diameters of 10–50 μm were responsible for water vaporization due to cement hydration. The presence of C–S–H gels and CH constituted 94% of the slurry's composition, which was primarily composed of O, Ca, C, and Si. The overall compact appearance of the structure suggested effective internal cementation of the slurry.

In the big crack mode (Figure 16b), a distinct demarcation fissure was observed at the cementation interface between the slurry and the rock. The fissure width measured approximately 50 μm, with a development angle of nearly 45°. The magnetite fraction exhibited a smooth, flat structure with high density, while the slurry surface displayed flocculation and agglomeration. The principal components of the rock included O, Fe, Ca, Si, and C, which collectively constituted 96.83% of its overall composition.

The small crack mode (Figure 16c) exhibited a compact structure within the skarn body. A transition zone was present at the cementation interface between the rock and the slurry, facilitating a complete bond between the two materials. Fissures approximately 20 μm in width had developed on both sides of this transition zone. The majority of this mode's constituents comprised O, Ca, C, Si, and Mg, collectively accounting for 97.94% of the total composition.

The small karst pore mode (Figure 16d) was characterized by a pronounced contrast between the rock mass and the slurry within the transition zone. This mode exhibited thin sheets of laminated CH adhered to the rock mass alongside the development of numerous irregular micro-fractures. O, Fe, Ca, Si, and C elements made up the majority of this mode's constituents, accounting for 97.81% of the total composition.

The pore penetration mode (Figure 16e) exhibited flocculation, agglomeration, and early hydration, resulting in the outward radial growth of needle-like structures emanating from the cement particles. O, Ca, C, Si, and Mg made up 98.61% of this mode's composition. The development of C–S–H gel was markedly pronounced, with the gel's needle-like structures stacking upon one another, thereby forming a denser structure. This mode also exhibited the emergence of numerous intersecting micro-fissures.

Fissure filling and pore penetration were principally cementation methods under flowing water grouting. The filling effectiveness was deemed quite satisfactory, as indicated by the five filling modes. The dimensions of these micropores and micro-fissures were relatively small. The hydration products within the structure were widely dispersed. The subsurface fissures and voids had been successfully filled by the treatment measures developed and modified in response to the study. These measures encompassed adjustments to the parameters corresponding to the four influencing factors, in addition to treatment measures for the four strata.

It could also inform on-site treatment strategies based on the characteristics of the propagation filling mode. It became essential to consider the placement of aggregates in cavities larger than 50 cm, given that the exposure of the pure slurry section ranged from 15 to 50 cm; this could also serve as an indicator of the surrounding environment, as a high frequency of pure slurry modes suggested a greater frequency of nearby cavities. Consequently, it was advisable to promptly inspect the area for any untreated or inadequately treated voids. Numerous fissure modes suggested that fissures had formed and the local rock mass had been fractured. To determine whether the treatment met the requirements, it was imperative to evaluate the strength of the rock mass. The exposure of the small karst pore mode indicated that karst had developed in the area. Consequently, further research is necessary to ascertain whether the karst has received adequate treatment, along with a focus on downhole drainage. Based on the conditions of the permeable mode's penetration filling, the treatment of the loose layer will be enhanced and supplemented.

4.5 Field applications

The results of field engineering tests were directly applied to the field treatment. The grouting flow rate exceeded 110 L/min when treating all cavity and hidden types following the placement of aggregate, as well as fissure types with water flow rates ≤10 000 m3/day. The grouting flow rate ranged from 90 to 110 L/min for treating fissure types with water flow rates between 4000 and 10 000 m3/day. The grouting flow rates ranged from 70 to 90 L/min when treating fissure types with water flow rates less than 4000 m3/day and complete types with water flow rates exceeding 7000 m3/day. The grouting flow rate had been set at 70 L/min for treating all complete types with water flow rates <7000 m3/day. A two-liquid slurry with a water–cement ratio of 0.8:1.0 was utilized when the Lugeon value exceeded 15 Lu. Two-liquid grouting with a water–cement ratio of 1.0:1.0 was applied when the Lugeon value was between 5 and 15 Lu. A 1.0:1.0 water–cement single-liquid slurry without sodium silicate added had been utilized if the Lugeon value was less than 0.5 Lu.

During the application phase, the optimal horizontal combination (A4B4C3D2) had been validated. The stratum was the complete type with a water flow rate of less than 4000 m3/day and a grouting flow rate of between 70 and 90 L/min, and the 0.8:1.0 two-liquid ratio was chosen. Figure 17 shows that the grouting process took only 170 min, indicating high efficacy, using the 60–75 m grouting section in the No.CK14 grouting borehole, as an example, provided additional support for the accuracy and validity of the orthogonal experimental design.

Table 4 shows the statistical results for various stratigraphic conditions and water flow rates during the treatment application. GSR values exceeding 90% indicated a substantial level of efficacy for the implemented treatment. It was observed that as stratum fragmentation and water flow rates decreased, GSR values tended to increase, indicating enhanced grouting efficacy. Furthermore, strata with high-water flow rates had been identified as the primary environments where grouting failures occurred. Likewise, failures across varying water flow rates predominantly occurred in strata exhibiting higher degrees of fragmentation. Multiple grouting treatments had been conducted on the same grouting section, with consistent intervals maintained between each treatment, to address the grouting of all sites under varying conditions and achieve overall success in site treatments.

5.1 Stratigraphic fragmentation and water flow rate

The findings of the field experiment show that high-water flow rates and extensive stratigraphic fragmentation decrease grouting efficacy, while increasing the grouting flow rate may enhance sealing ability, which is consistent with previous physical tests (Hu et al., 2024; Zhan et al., 2011). Furthermore, as water flow rates increase, the unit slurry consumption of the field grout also increases. The field treatment places the aggregate into the grouting sections of hidden and cavity types as soon as possible.

It is essential to note that, even under identical background conditions, the relationship between grout flow rate and unit slurry consumption in the field differs from that observed in previous physical test results. In scenarios involving high-water flow rates and fractured strata, grouting at low flow rates may prove ineffective. High flow rate grouting (>110 L/min) is required in these circumstances, albeit at a higher grout consumption. Conversely, lower grout consumption is achieved by using low flow rate grouting (<70 L/min) in intact strata.

Some prior studies (Du et al., 2021; Jin et al., 2021; Li et al., 2020) indicated that fracture width has a lesser impact on grouting effectiveness compared to water flow rate. Field orthogonal results indicate that fissure development has a greater significance than water flow rate (RA = 97.778 > RB = 39.375). This disparity may lead to a variation in the order of importance. This is because the fissures observed in physical tests tend to be smaller than those present in real engineering contexts.

5.2 Water–cement ratio and grouting flow rate

Consistent with earlier research results (Jin et al., 2021; Liu et al., 2023; Sui et al., 2015), a more viscous water–cement ratio was not always necessary. Field experiment results indicate that slurries with a water-cement ratio exceeding 0.5:1.0 tend to condense rapidly and heavily, leading to poor diffusivity and mobility. While these slurries are easier to solidify and seal, they increase costs and limit the treatment range.

The selection of field grouting flow rates differs from those suggested by previous research. It is not necessarily true that a higher grouting flow rate yields better results. The choice of grouting flow rates should be based on the specific conditions of the field stratum and the water flow rate. A greater grouting flow rate can overcome the flowing water conditions at high-water flow rates, facilitating solidification and accumulation. However, using a high grouting flow rate (>110 L/min) in intact strata may lead to surface bubbling. A low grouting flow rate (<70 L/min) is more appropriate in such cases.

5.3 Models and applications in the field

This study identifies four distinct types of grouting strata and associated treatment approaches that can be directly implemented in the field, a capability not achievable through physical testing. Additionally, previous physical tests classified the degree of sealing into three simple categories: completely sealed, partially sealed, and unsealed (Jin et al., 2021; Sui et al., 2015). This field research systematically summarizes and rigorously analyzes five distinct propagation filling modes, which can influence and guide field grouting based on various types and characteristics.

5.4 Limitations

The primary focus of this study was on the mechanical characteristics of grout propagation filling. However, there is still more to be done; chemical concerns will be a subject of future research. Additionally, the results from field tests demonstrated that this treatment plan significantly reduced the hydraulic conductivity of the upper aquifer. Despite this success, some cracks remain hydraulically connected to the surrounding untreated areas. To achieve comprehensive waterproofing, a horizontal waterproof curtain engineering treatment will be implemented.