Multi-stage reinforcing construction method for granite stratum high-filling roadbed

Abstract

The invention discloses a multi-stage reinforcing construction method for a granite stratum high-filling roadbed, and belongs to the technical field of roadbed construction in road engineering. The method aims at solving the problems that the granite stratum and roadbed filler interface bonding is weak, water drainage in the high-fill body is not smooth, and stability control in the construction process is difficult. The method comprises the following steps: scabbling the surface of the granite foundation, and laying a cement bentonite slurry transition layer to enhance interface bonding; a vertical drainage system is arranged after the first layer of filler is laid; then layered filling and reinforcing are conducted, geogrids are laid on each layer, door-shaped anchor rods are used for anchoring, then a water-permeable gravel layer is covered, and circulation is conducted till the designed elevation is reached; meanwhile, monitoring points are arranged on the center line and the slope toe of the roadbed, and construction parameters such as filling pause, reinforcement layer addition or compaction process adjustment are dynamically adjusted according to horizontal displacement data. The method is mainly used for building the high-fill roadbed on the granite stratum, and the overall stability and construction safety controllability of the roadbed can be effectively improved.

Description

Technical Field

[0001] This invention relates to the field of roadbed construction technology in civil engineering. More specifically, this invention relates to a multi-stage reinforcement construction method for high-fill roadbeds in granite strata. Background Technology

[0002] Several technical challenges are typically encountered when constructing high-fill roadbeds on granite strata. The surface of granite is usually dense and smooth, exhibiting significant differences in physical and mechanical properties compared to the overlying roadbed fill. This difference results in weak interfacial bonding, making the interface prone to relative slippage, creating potential weak points that affect the overall stability of the roadbed. Traditional cleaning methods or simple leveling treatments are often insufficient to effectively improve the bonding between these rigid and flexible materials.

[0003] High-fill roadbeds bear significant loads, generating substantial additional stresses during construction and service life. This places higher demands not only on the bearing capacity of the underlying foundation but also on the compaction and drainage of the roadbed fill material itself. Conventional layered filling and compaction methods sometimes fail to ensure uniform density within the high-fill structure, and if pore water pressure cannot dissipate promptly, it can delay foundation consolidation and potentially lead to uneven settlement. Furthermore, traditional construction quality control relies primarily on post-construction compaction testing, lacking real-time monitoring of the stress and deformation state within the roadbed during the filling process.

[0004] For reinforcing fill, the laying of geogrids and other reinforcing materials is commonly used. However, under high fill loads and potential lateral deformation, the interfacial friction between the reinforcing material and the fill material may be insufficient, leading to the reinforcement being pulled out or failing to fully utilize its tensile strength, thus weakening the reinforcement effect. Simultaneously, dynamic changes during construction, such as fluctuations in fill material properties and gradual increases in load, cause the safety state of the roadbed to constantly change. Traditional design based on fixed parameters and phased acceptance monitoring models struggle to capture the gradual changes in stability in a timely manner and issue warnings before critical states. Remedial measures are often only taken when obvious signs of deformation appear, at which point the difficulty and cost of adjustments have increased significantly. Summary of the Invention

[0005] One object of the present invention is to solve at least the above-mentioned problems and to provide at least the advantages that will be described later.

[0006] To achieve these objectives and other advantages according to the present invention, a method for multi-stage reinforcement of high-fill roadbeds in granite strata is provided, comprising the following steps: S1. Foundation treatment: The surface of the granite strata is cleaned and roughened to a depth of 8mm to 15mm to form a rough bonding surface; S2. Transition layer construction: A layer of cement-bentonite slurry transition layer is laid on the roughened rock surface. Before the slurry initially sets, a layer of manufactured sand with a particle size of 5mm to 10mm is evenly spread to form a reinforced bonding interface; S3. First Layer Filling and Drainage System Setup: After laying and compacting the first layer of fill material, a vertical drainage system is set up; the fill material layer is 0.3m to 0.5m thick crushed stone soil, which is compacted 6 to 8 times using a vibratory roller at a speed of 2km / h to 4km / h, achieving a compaction degree of 95% to 97%; the vertical drainage system consists of multiple plastic drainage strips with filter membranes arranged in a square grid, with the insertion depth penetrating the fill material layer and contacting the granite strata; S4. Layered Filling and Reinforcement: Starting from the first fill layer, a reinforcement layer, a permeable crushed stone layer, and the fill layer are laid and compacted, repeated multiple times until the design elevation is reached. The reinforcement layer includes geogrid and portal anchors. The tensile strength of the geogrid is 80kN / m to 120kN / m. It is tensioned and fixed during laying, with an overlap width of 0.4m to 0.5m. Immediately after laying, portal anchors are used for anchoring. The spacing of the portal anchors is 1.5m to 2.0m, with their two legs driven into the fill layer. Steel pads are installed on the crossbars using fasteners to press the geogrid tightly against the surface of the fill layer. After each reinforcement layer is laid and anchored, a permeable crushed stone layer with a thickness of 0.2m to 0.3m is laid on top of the reinforcement layer. The specifications and construction methods of other fill layers above the first fill layer are the same as those of the first fill layer.

[0007] S5. Monitoring and Feedback Control: Deep horizontal displacement monitoring points are set up at the centerline of the roadbed and the toes of both slopes. The monitoring frequency is once after every 2.0m thick layer of fill material is filled. When the monitoring data shows that the horizontal displacement increment of any monitoring point during the single-layer filling period exceeds 20mm, or the cumulative horizontal displacement exceeds 60mm, feedback control measures are initiated. The feedback control measures include suspending filling, adding a layer of reinforced geogrid in the displacement area, reducing the thickness of the subsequent three layers of fill material to 0.2m to 0.3m, and increasing the number of compaction passes to 10 to 12.

[0008] Preferably, a layer of needle-punched nonwoven geotextile is also laid on the surface of the permeable crushed stone layer as a filter layer, wherein the unit area mass of the needle-punched nonwoven geotextile is 300 g / m². 2 Up to 400g / m 2 The vertical permeability coefficient is not less than 1.0×10⁻⁶. - 2 cm / s, and during laying, a 5% to 8% elongation allowance should be reserved in the cross-sectional direction of the roadbed.

[0009] Preferably, after laying the needle-punched nonwoven geotextile, a layer of coarse sand with a thickness of 50mm to 80mm is first laid and statically compacted, and then the upper filler layer is laid.

[0010] Preferably, the crushed stone soil filler is continuously graded, with a maximum particle size not exceeding 50 mm and a uniformity coefficient C. u ≥5, curvature coefficient C e Between 1 and 3; before laying, its moisture content is adjusted to be within the range of -1% to +2% of the optimal moisture content.

[0011] Preferably, the two legs of the portal anchor have threads or textured surfaces, and are coated with cement slurry before installation to enhance the grip and anchoring effect between the anchor and the filler.

[0012] Preferably, a gravel filter bag is provided at the top of the vertical drainage system where it contacts the permeable crushed stone layer. The gravel filter bag is composed of stones with a particle size of 5 mm to 15 mm filled in geotextile and has a diameter of 300 mm to 400 mm.

[0013] Preferably, during the compaction process, when the ambient temperature is above 30℃ or the wind speed is greater than 3m / s, the surface of the compacted filler should be cured with water mist, with the spraying rate controlled at 0.5L / m. 2 Up to 1.0L / m 2 This is to maintain the moisture content of the filler surface and ensure uniform and stable compaction.

[0014] Preferably, for areas with localized strong weathered granite grooves or fissures, the bottom is first cleaned, then C15~C20 lean concrete is used for backfilling until it is flush with the surrounding rock surface, and then the entire surface is roughened; the maximum particle size of the aggregate in the lean concrete does not exceed 25mm. The crushed stone filler incorporates 0.3% to 0.5% polypropylene mesh fiber with a fiber length of 12mm to 19mm. The fiber is added evenly during the mixing of the filler to suppress drying shrinkage cracks and improve the integrity of the filler.

[0015] Preferably, during the monitoring process in step S4, settlement monitoring points are also set up simultaneously; based on the data collected from the deep horizontal displacement monitoring points and settlement monitoring points, combined with the roadbed filling height, slope geometric parameters and fill physical and mechanical parameters, a real-time safety factor assessment model for the overall stability of the roadbed is established with the simplified Bishop method as the core. During the filling process, after each layer of fill material is placed, the evaluation model is automatically run once to calculate the real-time safety factor F of the current roadbed. s Set two-level warning thresholds: when 1.30 ≤ F sWhen the tensile strength is less than 1.50, a primary warning is issued, and a first-level performance compensation measure is initiated: in the next planned reinforcement layer, the tensile strength of the geogrid is temporarily increased by 20% to 30% based on the original design; when 1.20 ≤ F s When the value is less than 1.30, a high-level warning is issued and a secondary performance compensation measure is initiated: on the basis of implementing the primary performance compensation measure, sandbags are simultaneously piled at the toe of the slope in the corresponding area of ​​the roadbed, and a temporary counter-pressure load of 20 kPa to 35 kPa is applied. Within 24 hours after each completion of the aforementioned feedback control measures, the monitoring frequency is increased to once every 6 hours. Using the displacement and settlement data collected during this period, the parameters of packing cohesion c and internal friction angle φ in the evaluation model are inverted and analyzed. The model parameters are then fine-tuned based on the inversion results for more accurate predictions in the future.

[0016] Preferably, the process of fine-tuning the model parameters based on the inversion analysis and inversion results includes: During the 24-hour intensive monitoring period following the completion of primary or secondary performance compensation measures, the horizontal displacement increment sequence {Δd} of each monitoring point within the specified time period is obtained. i} and the settlement increment sequence {Δs i}, and at the same time record the newly added filling thickness ΔH during this period; Using the horizontal displacement increment sequence and settlement increment sequence as the system response, an inverse analysis objective function is constructed with the packing cohesion c and internal friction angle φ as optimization variables. The objective function is the sum of squared residuals between the calculated and monitored values, i.e.: Minimize: Σ[ (Δd i,cal - Δd i,mea ) 2 + (Δs i,cal - Δs i,mea ) 2 ],(d i,cal , Δs i,cal ) is the calculated value, (Δd) i,mea , Δs i,mea ) represents the monitored value; among which, the calculated value (Δd) i,cal , Δs i,cal The displacement prediction model is obtained by iteratively calculating the result by substituting the candidate (c, φ) parameter set into the simplified Bishop method extended displacement prediction model; The objective function above is solved using a Monte Carlo-steepest descent hybrid algorithm to obtain a set of optimal (c, φ) parameters that make the calculated response closest to the actual monitoring response; the optimal (c, φ) parameters obtained by this inversion are used to update the original packing mechanical parameters in the evaluation model.

[0017] The present invention has at least the following beneficial effects: 1. Improve the bonding and interface strength of the foundation: By roughening the surface of the granite strata and laying a cement-bentonite slurry transition layer and manufactured sand, the bonding force between the roadbed fill and the rock strata is significantly enhanced, and the interface shear resistance is effectively improved.

[0018] 2. Ensure smooth internal drainage and stable consolidation: The vertical drainage system, combined with a permeable crushed stone layer, creates an effective drainage path, accelerates the drainage and consolidation of the foundation soil, helps reduce post-construction settlement, and improves the long-term stability of the roadbed.

[0019] 3. Achieving layered reinforcement and overall strengthening: A reinforcement layer consisting of geogrids and portal anchors is laid and anchored in layers, and polypropylene fibers are incorporated into the fill material to form a multi-layered, three-dimensional reinforcement system, which effectively improves the integrity, deformation resistance and crack resistance of the fill material.

[0020] 4. Enhance dynamic monitoring and proactive control during construction: By deploying displacement and settlement monitoring points, establishing a real-time safety factor assessment model, and setting early warning thresholds and performance compensation measures, dynamic assessment and feedback control of the roadbed stability are achieved. This method can adjust construction parameters and reinforcement strategies in a timely manner based on monitoring data, transforming passive handling into proactive intervention, thus enhancing the controllability and safety of the construction process.

[0021] 5. Optimize model accuracy and predictive capability: Utilize encrypted monitoring data for parameter inversion analysis, continuously revise and evaluate the mechanical parameters of the fill material in the model, so that the model can more accurately reflect the actual working state of the subgrade, thereby providing a more reliable basis for stability prediction and decision-making in the subsequent construction stage.

[0022] In summary, this invention, through the comprehensive application of a series of technical means such as foundation treatment, drainage system, layered reinforcement, intelligent monitoring and feedback control, forms a complete and effective construction system, which is of positive significance for ensuring the engineering quality and long-term safe and stable operation of high-fill roadbeds on granite strata.

[0023] Other advantages, objectives and features of the present invention will become apparent in part from the following description, and in part from those skilled in the art through study and practice of the invention. Detailed Implementation

[0024] The present invention will be further described in detail below with reference to embodiments, so that those skilled in the art can implement it based on the description.

[0025] It should be noted that, unless otherwise specified, the experimental methods described in the following implementation plan are all conventional methods, and the reagents and materials described are all commercially available unless otherwise specified.

[0026] This invention provides a method for multi-stage reinforcement of high-fill roadbeds in granite strata, comprising the following steps: S1. Foundation treatment: Clean and roughen the surface of the granite strata to a depth of 8mm to 15mm to form a rough bonding surface. S2. Transition layer construction: A layer of cement-bentonite slurry transition layer is laid on the roughened rock surface. Before the slurry initially sets, a layer of manufactured sand with a particle size of 5mm to 10mm is evenly spread to form an enhanced bonding interface. S3. First Layer Filling and Drainage System Setup: After laying and compacting the first layer of fill material, a vertical drainage system is set up; the fill material layer is 0.3m to 0.5m thick crushed stone soil, which is compacted 6 to 8 times using a vibratory roller at a speed of 2km / h to 4km / h, achieving a compaction degree of 95% to 97%; the vertical drainage system consists of multiple plastic drainage strips with filter membranes arranged in a square grid, with the insertion depth penetrating the fill material layer and contacting the granite strata; S4. Layered Filling and Reinforcement: Starting from the first fill layer, a reinforcement layer, a permeable crushed stone layer, and the fill layer are laid and compacted, repeated multiple times until the design elevation is reached. The reinforcement layer includes geogrid and portal anchors. The geogrid has a tensile strength of 80kN / m to 120kN / m, is tensioned and fixed during laying, and has an overlap width of 0.4m to 0.5m. Immediately after laying, portal anchors are used for anchoring. The portal anchors are spaced 1.5m to 2.0m apart, with their two legs driven into the fill layer. Steel plates are installed on the crossbars using fasteners to press the geogrid tightly against the fill layer surface. After each reinforcement layer is laid and anchored, a permeable crushed stone layer with a thickness of 0.2m to 0.3m is laid on top of the reinforcement layer. The specifications and construction methods of other fill layers above the first fill layer are the same as those of the first fill layer. S5. Monitoring and Feedback Control: Deep horizontal displacement monitoring points are set up at the centerline of the roadbed and the toes of both slopes. The monitoring frequency is once after every 2.0m thick layer of fill material is filled. When the monitoring data shows that the horizontal displacement increment of any monitoring point during the single-layer filling period exceeds 20mm, or the cumulative horizontal displacement exceeds 60mm, feedback control measures are initiated. The feedback control measures include suspending filling, adding a layer of reinforced geogrid in the displacement area, reducing the thickness of the subsequent three layers of fill material to 0.2m to 0.3m, and increasing the number of compaction passes to 10 to 12.

[0027] Specifically, during the foundation treatment stage, the surface of the granite strata is cleaned to remove loose materials and impurities. The roughening depth can be selected as 8mm, 10mm, 12mm, or 15mm to create a rough bonding surface. During the transition layer construction, a cement-bentonite slurry transition layer is laid on the roughened rock surface. Before the slurry initially sets, manufactured sand is evenly spread. The particle size of the manufactured sand can be selected as 5mm, 6mm, 8mm, or 10mm. The cement-bentonite slurry can be prepared by mixing ordinary Portland cement and sodium-based bentonite, with the water-cement ratio controlled between 0.5 and 0.6. The manufactured sand can be sourced from a local quarry and produced using mechanical crushing. During construction, the manufactured sand is evenly spread using manual or mechanical spreading equipment to ensure complete coverage of the rock surface. The assembly location is on the surface of the granite strata, and the thickness of the transition layer is controlled between 10mm and 20mm. The work process includes first cleaning the rock surface, then spraying the slurry, and then immediately spreading the manufactured sand to form a reinforced bonding interface.

[0028] In the first-floor filling and drainage system, the fill layer thickness can be selected as 0.3m, 0.4m, or 0.5m, using crushed stone and soil fill. The crushed stone and soil can be natural gravel or crushed rock, with a maximum particle size not exceeding 50mm and a uniformity coefficient C. u ≥5, curvature coefficient C e The moisture content should be between 1 and 3. Before laying the fill material, adjust the moisture content by watering or sun-drying to within the range of -1% to +2% of the optimum moisture content. A self-propelled vibratory roller can be used for compaction, with a rolling speed of 2 km / h, 3 km / h, or 4 km / h, and 6, 7, or 8 passes, achieving a compaction degree of 95% to 97%. The vertical drainage system consists of plastic drainage strips with filter membranes. Commercially available composite drainage materials can be used, arranged in a square grid with a grid spacing of 1.0m to 1.5m. The insertion depth should penetrate the fill material layer and contact the granite stratum, using a specialized insertion machine. The assembly location is inside the first fill material layer, with a gravel filter bag placed at the contact point between the top of the drainage strip and the permeable crushed stone layer. The work process includes laying the fill material layer, compacting it, inserting the drainage strip, and finally checking the drainage effect.

[0029] During the layered filling and reinforcement stage, the reinforcement layer includes geogrids and portal anchors. The geogrid tensile strength can be selected from 80kN / m, 100kN / m, or 120kN / m. It is tensioned and fixed during installation, with overlap widths of 0.4m, 0.45m, or 0.5m. The portal anchor spacing can be selected from 1.5m, 1.8m, or 2.0m, with threads or textured surfaces on both legs. Cement grout is applied before installation to enhance bond strength. Portal anchors can be made of carbon steel, and ordinary steel plates can be used for the steel pads. The thickness of the permeable crushed stone layer can be selected from 0.2m, 0.25m, or 0.3m, with a crushed stone particle size of 5mm to 20mm. The assembly position is located on the surface of the fill layer; the geogrid is laid on the fill layer, and the portal anchors are driven into the fill layer. The work process includes first laying geogrid, then installing portal anchors and securing the pads, followed by laying a permeable crushed stone layer, repeating this process until the design elevation is reached. The construction of the fill layer is the same as the first layer, ensuring consistent compaction of each layer.

[0030] In monitoring and feedback control, deep horizontal displacement monitoring points are set up at the centerline of the roadbed and the toes of both slopes. Inclinometers or inclinometer tubes can be used as monitoring equipment, installed inside boreholes. Monitoring frequency is once after every 2.0m of fill thickness. The threshold is set when the horizontal displacement increment during a single layer exceeds 20mm, or the cumulative horizontal displacement exceeds 60mm. Feedback control measures include suspending filling, adding a layer of reinforcing geogrid in the displacement area, and reducing the thickness of the subsequent three fill layers to 0.2m, 0.25m, or 0.3m, while increasing the number of compaction passes to 10, 11, or 12. The reinforcing geogrid can be made of materials with higher tensile strength, such as 150kN / m. The assembly location is in the roadbed area with significant displacement. The work process includes monitoring data first, suspending construction when the threshold is exceeded, implementing reinforcement measures, and then resuming filling. Parameter settings are based on field monitoring data, and construction parameters are adjusted through real-time analysis.

[0031] This embodiment can improve the stability and durability of high-fill roadbeds in granite strata. Through multi-level reinforcement and real-time monitoring and control, it can reduce uneven settlement and horizontal displacement, and enhance the overall performance of the roadbed.

[0032] In another embodiment, a layer of needle-punched nonwoven geotextile is further laid on the surface of the permeable gravel layer as a filter layer, wherein the mass per unit area of ​​the needle-punched nonwoven geotextile is 300 g / m². 2 Up to 400g / m 2 The vertical permeability coefficient is not less than 1.0×10⁻⁶. -2 cm / s, and during laying, a 5% to 8% elongation allowance should be reserved in the cross-sectional direction of the roadbed.

[0033] Specifically, needle-punched nonwoven geotextiles can be commercially available products made from polyester or polypropylene. Their unit area weight can be selected as 300 g / m². 2 350g / m 2 Or 400g / m 2 The vertical permeability coefficient is not less than 1.0 × 10⁻⁶. -2 cm / s, the specific value can be obtained from 1.2×10 -2 cm / s or 1.5×10 -2 The product achieves a flow rate of cm / s. Materials can be sourced from common industrial geosynthetic suppliers. The geotextile is fitted onto the top surface of the permeable gravel layer, directly covering it. Its working process involves the geotextile acting as a filter as water flows downwards from the upper filler layer, preventing fine soil particles from entering and clogging the underlying gravel layer, while its high permeability ensures that water can drain smoothly.

[0034] During the laying process, the geotextile retains a certain elongation allowance in the cross-sectional direction of the roadbed, specifically 5%, 6%, or 8%. This allowance can be controlled by manual estimation or by marking with measuring ropes. During construction, manual spreading or a dedicated geotextile laying machine can be used. During laying, the roll of geotextile is first unrolled transversely along the roadbed, smoothly covering the surface of the crushed stone layer. A calculated slack is reserved at the side slopes to avoid excessive tension due to filler compaction or subsequent deformation. The assembly position is located between the permeable crushed stone layer and the upper filler layer, serving as an independent filter layer. Its working process is similar to that of a conventional geotextile filter layer, relying on its fiber structure to form a filter mesh to achieve the reverse filtration function.

[0035] To verify its effectiveness, indoor functional tests can be conducted on the geotextile samples used. The test subject is the planned 300g / m² geotextile. 2 Up to 400g / m 2 Needle-punched nonwoven geotextiles. Testing methods can refer to relevant standards, including gradient ratio tests or permeability tests, to confirm that their filtration performance and vertical permeability coefficient meet a requirement of not less than 1.0 × 10⁻⁶. -2 The requirement is cm / s. Through this type of test, materials that meet the technical requirements can be selected for use in engineering projects.

[0036] This embodiment can form a reliable filter isolation layer to prevent the fine particles of the upper filler from being lost into the lower permeable layer, thereby effectively maintaining the long-term function of the drainage system and enhancing the stability of the roadbed internal structure.

[0037] In another embodiment, after the needle-punched nonwoven geotextile is laid, a layer of coarse sand with a thickness of 50 mm to 80 mm is first laid and statically compacted, and then the upper filler layer is laid.

[0038] Specifically, the thickness of the coarse sand protective layer can be selected as 50mm, 60mm, 70mm, or 80mm. The coarse sand used can be natural river sand or manufactured sand, with a particle size range controlled between 1mm and 5mm and a mud content of less than 3%. The material source can be a qualified local sand and gravel supplier. This protective layer is directly assembled onto the upper surface of the needle-punched nonwoven geotextile, completely covering the geotextile and forming a physical buffer layer. The process involves immediately transporting the coarse sand to the site by truck after the geotextile installation has passed inspection, and then spreading it using a combination of manual labor and mechanical equipment to form a uniform sand layer of the designed thickness.

[0039] During compaction, static rollers, such as small to medium-sized double-drum static rollers, can be used. The number of compaction passes is typically 1 to 2, and the roller's speed can be controlled between 1.5 km / h and 3.0 km / h. During construction, the roller rolls from the edge of the subgrade towards the centerline in an alternating pattern to ensure uniform compaction of the entire protective layer. The assembly location is clearly defined between the geotextile filter layer and the upper fill layer. Its working process is similar to that of conventional thin sand cushion layer compaction, using static pressure to rearrange coarse sand particles to form a flat and stable working surface, while avoiding potential tearing or damage to the underlying geotextile caused by vibration compaction.

[0040] To verify the quality of the protective layer installation, a simple on-site functional test can be conducted. The test object is the compacted coarse sand protective layer. Test methods may include using the ring cutter method or sand cone method to test its dry density after compaction, and using a level to measure whether its installation thickness meets the design requirement of 50mm to 80mm. Through this type of test, it can be confirmed whether the protective layer provides uniform support and effectively isolates the heavy construction equipment above from the geotextile below.

[0041] This embodiment provides effective mechanical protection for the vulnerable geotextile filter layer by setting and compacting a coarse sand protective layer, preventing the upper gravel and soil filler from directly puncturing or damaging the geotextile, thereby ensuring the integrity and durability of the filter layer and drainage system.

[0042] In another embodiment, the crushed stone soil filler is continuously graded, with a maximum particle size not exceeding 50 mm and a uniformity coefficient C. u ≥5, curvature coefficient C e Between 1 and 3; before laying, its moisture content is adjusted to be within the range of -1% to +2% of the optimal moisture content.

[0043] Specifically, the maximum particle size of the crushed stone soil filler should not exceed 50mm, with 30mm, 40mm, or 50mm chosen as the upper limit. Its uniformity coefficient C... u Not less than 5, C can be selected in actual engineering. uGradation of 5, 7, or 10. Curvature coefficient C e The value should be between 1 and 3, for example, 1.2, 2.0, or 2.8. The fill material can be machine-made crushed stone soil obtained by mechanically crushing and screening lithologies such as granite and limestone, or natural gravel soil that meets the gradation requirements. The material source can be a local quarry or a qualified sand and gravel supply base. This fill material is assembled in each layer of the roadbed, starting from the first layer and continuing up to the design elevation. The process involves crushing and screening the stone at the quarry to form a mixture that meets the design gradation requirements, then transporting it to the site for spreading and compaction.

[0044] Before laying the fill material, its moisture content needs to be controlled. The optimal moisture content is determined through a standard compaction test. The control target is to keep the moisture content of the fill material within the range of -1% to +2% of this optimal value, for example, -0.5%, +1%, or +2%. If the moisture content is too low, a water truck can be used for spraying to increase humidification; if the moisture content is too high, plowing and drying or using a rotary dryer can be used for dehydration. The control operation is usually carried out in the material yard or the paving area on site. The process is as follows: first, a rapid moisture content test is performed on the incoming material or the fill material on site; then, based on the difference between the test result and the target value, the required amount of water to be added or the dehydration time is calculated; and then the appropriate machinery is used to ensure that the moisture content of the fill material is uniform and within the suitable range.

[0045] To verify whether the gradation and moisture content of the fill material meet the requirements, on-site testing is necessary. The test object is the crushed stone soil that will be used for filling. The testing methods include sieving tests on the fill material samples using standard sieve sets to confirm whether its gradation curve is continuous and the maximum particle size, C... u Value, C e The values ​​meet the requirements; simultaneously, the moisture content is tested using a drying method or a rapid moisture content meter to ensure it is within the range of -1% to +2% of the optimum moisture content. This type of testing ensures the quality of the filler material used, laying the foundation for subsequent compaction operations.

[0046] This embodiment creates favorable conditions for compaction operations by controlling the gradation and moisture content of the fill material, which helps to improve the compaction degree and uniformity of the fill, thereby enhancing the overall strength and stability of the roadbed and reducing the risk of post-construction settlement.

[0047] In another embodiment, the two legs of the portal anchor have threads or textured surfaces, and are coated with cement slurry before installation to enhance the grip and anchoring effect between the anchor and the filler.

[0048] Specifically, the two legs of the portal anchor can be threaded, with a pitch of 2mm or 3mm and a depth of 0.5mm to 1.0mm; or they can be machined into regular undulating patterns, such as a height difference of 1mm to 1.5mm between troughs and crests. The anchor body can be made of Q235 or Q345 carbon steel. The cement grout can be made by mixing ordinary Portland cement with water, with a water-cement ratio of 0.45 or 0.50; to improve the grout performance, bentonite can be added at 2% to 3% of the cement weight. Commercially available PO 42.5 grade cement can be used. The anchor is installed in the fill layer of the roadbed, with its two legs driven into the fill material, and the crossbar located on the surface of the fill layer and pressing against the geogrid. Its working process is that the textured structure of the anchor legs forms a mechanical interlock with the surrounding fill material.

[0049] Before installation, the two legs of the anchor bolt need to be coated with cement grout. The grout can be prepared on-site using a small vertical mixer. Coating can be done manually with a brush or by dipping the bolt in the grout, ensuring that the grooves and depressions are fully filled with grout to form a uniform coating layer. The coating thickness should be controlled between 1mm and 2mm. Coating is usually carried out in the anchor bolt storage area or near the installation point. The process involves evenly applying the pre-mixed cement grout to the clean, dry surface of the anchor bolt legs, and then quickly installing the anchor bolt before the grout initially sets, ensuring a tight bond between the grout-coated anchor bolt and the filler.

[0050] To verify the effectiveness of this reinforcement measure, on-site anchor pull-out tests can be conducted. The test subjects are portal anchors installed using this process. The testing method can refer to relevant specifications, using hydraulic jacks and reaction devices to perform pull-out tests on randomly selected anchors, recording their ultimate pull-out force, and comparing it with anchors without grout coating. Through this type of test, the effect of cement grout coating on improving anchor bond strength and anchoring effectiveness can be visually evaluated.

[0051] This embodiment significantly increases the interfacial resistance between the anchor bolt and the surrounding filler by combining the surface texture structure with the bonding effect of cement grout, thereby improving the overall anchoring reliability of the reinforcement system and helping to maintain the long-term stability of the geogrid reinforcement effect.

[0052] In another embodiment, a gravel filter bag is provided at the top of the vertical drainage system where it contacts the permeable crushed stone layer. The gravel filter bag is composed of stones with a particle size of 5 mm to 15 mm filled in a geotextile wrapping, and its diameter is 300 mm to 400 mm.

[0053] Specifically, the particle size of the gravel filter bags can be selected from 5mm, 8mm, 10mm, or 15mm. These stones can be hard river gravel, or mechanically crushed granite or limestone. The geotextile wrapping the stones can be selected with a unit area mass of 200g / m². 2 Up to 300g / m 2 The material is needle-punched nonwoven geotextile. The aggregate can be sourced from local gravel pits, and the geotextile is a commercially available, common geosynthetic material. The filter pack is fitted on top of each plastic drainage strip, positioned at the interface between the drainage strip outlet and the overlying permeable gravel layer. Its working process is as follows: when water carrying a small amount of soil particles exits the drainage strip, the filter pack, through its gradation and the filtering effect of the geotextile, allows water to pass freely while trapping solid particles, preventing clogging.

[0054] The diameter of the filter bag can be selected from 300mm, 350mm, or 400mm. During fabrication, the geotextile is first cut into squares, then filled with stones of the selected particle size. Finally, the opening is secured with wire or binding tape, forming an approximately spherical enclosure. Filling can be done on-site at the prefabrication yard or near the work site. The process is similar to that of conventional filter pipe wrapping: stones are placed manually or using simple molds onto the spread geotextile, then gathered, wrapped, and secured to form the finished filter bag. It is then placed on top of the drainage strip, ensuring good contact with the upper permeable gravel layer.

[0055] To verify the filtration efficiency of the filter bag, a simple on-site functional observation can be conducted. The observation focuses on the installed filter bag and the drainage conditions around it. Observation methods include checking the drainage outlet for blockages after continuous rainfall or submersion of the roadbed, observing for significant fine soil accumulation around the filter bag, and verifying the permeability of the overlying gravel layer. These observations can determine whether the filter bag has effectively fulfilled its protective function.

[0056] This embodiment sets up a physical barrier at a key node of the drainage system, which effectively prevents the surrounding fine-grained filler from intruding into and clogging the drainage strip. This helps maintain the long-term drainage efficiency of the vertical drainage system, thereby accelerating the consolidation of the subgrade soil and improving its stability.

[0057] In another embodiment, during the compaction process, when the ambient temperature is higher than 30°C or the wind speed is greater than 3 m / s, the surface of the compacted filler is treated with water mist curing, with the spraying rate controlled at 0.5 L / m. 2 Up to 1.0L / m 2 This is to maintain the moisture content of the filler surface and ensure uniform and stable compaction.

[0058] The trigger thresholds for maintenance measures are defined as an ambient temperature of 30℃ or a wind speed of 3m / s. Maintenance must be implemented when on-site monitoring data meets or exceeds either condition. The spraying rate should be controlled at 0.5L / m³. 2 Up to 1.0L / m 2 Between these values, you can specifically choose 0.5L / m. 2 0.7L / m 2 Or 1.0L / m 2 Spraying equipment can include backpack sprayers, push-type sprayers, or atomizing nozzle systems mounted on water trucks. Spraying water can be sourced from standard tap water or clean river or lake water. The installation location for this maintenance operation is on the surface of the compacted and inspected roadbed filler layer. The process involves on-site personnel first monitoring environmental parameters using thermometers, hygrometers, and anemometers. When the parameters exceed threshold values, the equipment is then used to uniformly spray water mist onto the designated area.

[0059] During spraying, operators can hold the sprayer boom or control the spraying system on the water truck, moving at a constant speed to ensure the spray volume per unit area reaches the preset value. Spraying is usually carried out immediately after the compaction process, or supplemented as needed when the compacted surface shows signs of drying, depending on environmental conditions. The process is similar to spraying in agriculture or dust suppression, where highly atomized water is evenly distributed over the target surface, effectively replenishing moisture while preventing water erosion or localized water accumulation that could damage the compacted surface.

[0060] To evaluate the effectiveness of the curing process, sampling tests can be conducted on the treated areas. The test object is the surface layer of the filler material after water mist spraying. Testing methods include measuring the moisture content of the filler material at a depth of 20mm to 30mm below the surface using a ring cutter or a rapid moisture meter at certain time intervals after curing, and comparing this moisture content with a control area that has not undergone curing. This type of testing verifies the actual effectiveness of water mist curing in maintaining surface moisture content and inhibiting excessively rapid evaporation.

[0061] This embodiment replenishes the compacted surface with necessary moisture under dry and windy conditions, slowing down the rate of water loss from the filler surface. This helps maintain the filler near its optimum moisture content, thereby ensuring the uniformity of compaction and the overall stability of the subgrade structure. In another embodiment, for locally existing strongly weathered granite grooves or fissures, the bottom is first cleaned, then C15~C20 lean concrete is used for backfilling until it is flush with the surrounding rock surface, and then the whole surface is roughened; the maximum particle size of the aggregate in the lean concrete does not exceed 25mm. The crushed stone filler incorporates 0.3% to 0.5% polypropylene mesh fiber with a fiber length of 12mm to 19mm. The fiber is added evenly during the mixing of the filler to suppress drying shrinkage cracks and improve the integrity of the filler.

[0062] For the treatment of defects in strongly weathered rock masses, the cleaning depth must reach the solid bedrock surface. The lean concrete used for backfilling can be of strength grade C15, C18, or C20. The maximum aggregate size should not exceed 25mm, specifically 20mm or 25mm. Ready-mixed concrete can be used, or it can be mixed on-site using a mobile mixer. PO 42.5 grade cement can be used. This backfill material is installed within the cleaned bedrock grooves or fissures, with its top surface flush with the surrounding intact granite surface. The process involves first removing the loose weathered rock mass using a pneumatic pick or manual excavation. After acceptance, lean concrete is poured, compacted using an immersion vibrator, and finally, the top surface is smoothed.

[0063] Regarding filler modification, the dosage of polypropylene mesh fibers can be selected as 0.3%, 0.4%, or 0.5% (based on the dry weight of the filler). Fiber lengths can be selected as 12mm, 15mm, or 19mm. Commercially available mesh polypropylene fibers specifically designed for concrete or soil reinforcement can be used. The mixing equipment can be a forced mixer from a field mixing plant or a rotary tiller for road mixing. The fiber assembly in the filler is achieved through the mixing process, ensuring its uniform dispersion throughout the crushed stone and soil mixture. The process involves calculating the required fiber weight per batch according to the design dosage, and simultaneously adding the fiber manually or via a dedicated feeder to the mixer along with the filler and other components (such as crushed stone and soil), mixing until homogeneous.

[0064] To verify the dispersibility and function of the fibers, on-site sampling inspection can be conducted on the mixed fiber-modified filler. The inspection object is the crushed stone-soil mixture at the discharge port. The inspection method includes randomly sampling from different batches, sieving with water, observing whether the fibers clump together, counting the number of fibers per unit mass of filler, and assessing its dispersion uniformity. Through this type of inspection, it can be confirmed whether the fibers are uniformly distributed in the filler as required.

[0065] This embodiment addresses local defects in the foundation by backfilling with lean concrete, ensuring the integrity and flatness of the base. By uniformly adding fibers to the fill material and utilizing their three-dimensional network structure, the tendency of the fill material to crack due to drying shrinkage is effectively suppressed, thereby enhancing the integrity and long-term durability of the high-fill roadbed.

[0066] In another embodiment, during the monitoring process in step S4, settlement monitoring points are also deployed simultaneously; based on the data collected from the deep horizontal displacement monitoring points and settlement monitoring points, combined with the roadbed filling height, slope geometric parameters and fill physical and mechanical parameters, a real-time safety factor assessment model for the overall stability of the roadbed is established with the simplified Bishop method as the core. During the filling process, after each layer of fill material is placed, the evaluation model is automatically run once to calculate the real-time safety factor F of the current roadbed. s Set two-level warning thresholds: when 1.30 ≤ F s When the tensile strength is less than 1.50, a primary warning is issued, and a first-level performance compensation measure is initiated: in the next planned reinforcement layer, the tensile strength of the geogrid is temporarily increased by 20% to 30% based on the original design; when 1.20 ≤ F s When the value is less than 1.30, a high-level warning is issued and a secondary performance compensation measure is initiated: on the basis of implementing the primary performance compensation measure, sandbags are simultaneously piled at the toe of the slope in the corresponding area of ​​the roadbed, and a temporary counter-pressure load of 20 kPa to 35 kPa is applied. Within 24 hours after each completion of the aforementioned feedback control measures, the monitoring frequency is increased to once every 6 hours. Using the displacement and settlement data collected during this period, the parameters of packing cohesion c and internal friction angle φ in the evaluation model are inverted and analyzed. The model parameters are then fine-tuned based on the inversion results for more accurate predictions in the future.

[0067] Settlement monitoring points can be deployed along the roadbed centerline and shoulders, forming a supporting observation network with deep horizontal displacement monitoring points. Monitoring equipment can include automatic total stations or hydrostatic leveling systems. Based on the collected displacement and settlement data, combined with parameters such as fill height, slope ratio, fill material density, and shear strength, a simplified Bishop method is used to construct an overall roadbed stability assessment model. This model can run on conventional computers or embedded industrial control systems. The process involves the system automatically reading the latest monitoring data and fill parameters after each layer of fill is placed, running the assessment model, and calculating the real-time safety factor F of the current roadbed. s .

[0068] Two warning thresholds are set. The primary warning threshold is when 1.30 ≤ F s <1.50, the advanced warning threshold is 1.20≤F s<1.30. When a primary warning is triggered, a Level 1 performance compensation measure is initiated: In the next planned reinforcement layer, the tensile strength of the geogrid is temporarily increased by 20% to 30% based on the original design strength. For example, the original design strength of 100 kN / m can be temporarily replaced with a geogrid of 120 kN / m or 130 kN / m. When a high-level warning is triggered, a Level 2 performance compensation measure is initiated: Based on the implementation of the Level 1 measures, sandbags are simultaneously piled at the toe of the slope in the corresponding area of ​​the roadbed. The sandbags can be made of polypropylene woven bags filled with medium or coarse sand, and the applied temporary counter-pressure load is 20 kPa to 35 kPa, for example, 25 kPa or 30 kPa can be selected. The sandbags are assembled close to the outside of the toe line of the roadbed slope. The working process is that after the warning is triggered, the construction personnel follow the instructions and either use mechanical hoisting or manual stacking of sandbags until the calculated load value is reached.

[0069] Within 24 hours of each completion of the aforementioned feedback control measures, the monitoring frequency is increased to once every 6 hours. Using the displacement and settlement increment data collected during this period, combined with the increased fill thickness, an inversion analysis is performed on the fill cohesion *c* and internal friction angle *φ* parameters in the evaluation model. The inversion process uses the monitored displacement and settlement sequence as the system response, and an optimization algorithm is used to find the optimal (c, φ) parameter set. This result is then used to update the original parameters in the evaluation model. The process involves the system collecting data during the intensive monitoring period, automatically running the inversion analysis program, and correcting the model parameters to make subsequent safety factor predictions more accurate.

[0070] This embodiment establishes a dynamic, closed-loop monitoring-evaluation-early warning-compensation-feedback control system, which can adjust the construction plan and reinforcement measures in a timely manner according to the actual response status of the roadbed, thereby improving the risk management capability and long-term stability guarantee level of the high-fill roadbed construction process.

[0071] In another embodiment, the process of fine-tuning the model parameters using the inversion analysis and inversion results includes: During the 24-hour intensive monitoring period following the completion of primary or secondary performance compensation measures, the horizontal displacement increment sequence {Δd} of each monitoring point within the specified time period is obtained. i} and the settlement increment sequence {Δs i}, and at the same time record the newly added filling thickness ΔH during this period; Using the horizontal displacement increment sequence and settlement increment sequence as the system response, an inverse analysis objective function is constructed with the packing cohesion c and internal friction angle φ as optimization variables. The objective function is the sum of squared residuals between the calculated and monitored values, i.e.: Minimize: Σ[ (Δd i,cal - Δd i,mea ) 2 + (Δs i,cal - Δsi,mea ) 2 ],(d i,cal , Δs i,cal ) is the calculated value, (Δd) i,mea , Δs i,mea ) represents the monitored value; among which, the calculated value (Δd) i,cal , Δs i,cal The displacement prediction model is obtained by iteratively calculating the result by substituting the candidate (c, φ) parameter set into the simplified Bishop method extended displacement prediction model; The objective function above is solved using a Monte Carlo-steepest descent hybrid algorithm to obtain a set of optimal (c, φ) parameters that make the calculated response closest to the actual monitoring response; the optimal (c, φ) parameters obtained by this inversion are used to update the original packing mechanical parameters in the evaluation model.

[0072] During the intensive monitoring period, it is necessary to obtain the horizontal displacement increment sequence {Δd} of each monitoring point. i} and the settlement increment sequence {Δs i These data can be collected using an automatic total station or hydrostatic leveling system, with a monitoring frequency of once every 6 hours. Simultaneously, the newly added fill thickness ΔH during this period needs to be recorded. This value can be calculated by measuring the number of fill layers and the design thickness of each layer, for example, adding a 0.3m layer of fill. This data is transmitted and stored in a central data processing unit, typically a computer or server deployed on-site. The process involves various sensors collecting data at set frequencies and transmitting it to the data center via wired or wireless networks. This data is then integrated synchronously with construction records to form a standard dataset for back-analysis.

[0073] Based on the above data, an inverse analysis objective function is constructed with the filler cohesion c and internal friction angle φ as optimization variables. The objective function is in the form of the sum of squared residuals between the calculated and monitored values, i.e., Minimize: Σ[ (Δd i,cal - Δd i,mea ) 2 + (Δs i,cal - Δs i,mea ) 2 Among them, the calculated value (Δd) i,cal , Δs i,cal The displacement prediction model needs to be iteratively calculated by substituting the candidate (c, φ) parameter set into a displacement prediction model based on a simplified Bishop method extension. This displacement prediction model and inversion calculation program can run on a computer with numerical calculation software installed. The process involves initializing a set of (c, φ) parameters, calling the displacement prediction model to calculate the theoretical response of the subgrade under the current fill thickness ΔH, comparing the theoretically calculated displacement settlement increment sequence with the measured sequence, and calculating the objective function value.

[0074] The objective function described above is solved using a hybrid Monte Carlo-steepest descent algorithm. The Monte Carlo part employs global random sampling, with the initial sampling count set to 1000 to 5000 times. The steepest descent method is used for a local fine-grained search within the region of relatively optimal solutions found in the Monte Carlo method. This optimization algorithm can be implemented using existing functions from numerical computation libraries or a custom program. The algorithm first performs extensive random sampling within a pre-defined reasonable range of (c, φ) to evaluate the objective function and select the parameter regions with better performance. Then, it switches to the steepest descent method, iteratively updating the (c, φ) values ​​along the negative gradient direction until it finds a set of optimal (c, φ) parameters that minimizes the objective function, i.e., the calculated response most closely approximates the actual monitored response. Finally, the optimal (c, φ) parameters obtained through this inversion are used to update the original packing mechanical parameters in the evaluation model.

[0075] This embodiment uses systematic inversion analysis to feed back the actual deformation response of the roadbed into the mechanical model, thereby achieving self-correcting estimation of key soil parameters, improving the accuracy and predictive ability of the stability assessment model, and providing a more reliable basis for subsequent construction decisions.

[0076] Although embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention. For those skilled in the art, other modifications can be easily made. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details and embodiments shown and described herein.

Claims

1. A method for multi-stage reinforcement of high-fill roadbeds in granite strata, characterized in that, Includes the following steps: S1. Foundation treatment: The surface of the granite strata is cleaned and roughened to a depth of 8mm to 15mm to form a rough bonding surface; S2. Transition layer construction: A layer of cement-bentonite slurry transition layer is laid on the roughened rock surface. Before the slurry initially sets, a layer of manufactured sand with a particle size of 5mm to 10mm is evenly spread to form a reinforced bonding interface; S3. First Layer Filling and Drainage System Setup: After laying and compacting the first layer of fill material, a vertical drainage system is set up; the fill material layer is 0.3m to 0.5m thick crushed stone soil, which is compacted 6 to 8 times using a vibratory roller at a speed of 2km / h to 4km / h, achieving a compaction degree of 95% to 97%; the vertical drainage system consists of multiple plastic drainage strips with filter membranes arranged in a square grid, with the insertion depth penetrating the fill material layer and contacting the granite strata; S4. Layered Filling and Reinforcement: Starting from the first fill layer, a reinforcement layer, a permeable crushed stone layer, and the fill layer are laid and compacted, repeated multiple times until the design elevation is reached. The reinforcement layer includes geogrid and portal anchors. The geogrid has a tensile strength of 80kN / m to 120kN / m, is tensioned and fixed during laying, and has an overlap width of 0.4m to 0.5m. Immediately after laying, portal anchors are used for anchoring. The portal anchors are spaced 1.5m to 2.0m apart, with their two legs driven into the fill layer. Steel plates are installed on the crossbars using fasteners to press the geogrid tightly against the fill layer surface. After each reinforcement layer is laid and anchored, a permeable crushed stone layer with a thickness of 0.2m to 0.3m is laid on top of the reinforcement layer. The specifications and construction methods of other fill layers above the first fill layer are the same as those of the first fill layer. S5. Monitoring and Feedback Control: Deep horizontal displacement monitoring points are set up at the centerline of the roadbed and the toes of both slopes. The monitoring frequency is once after every 2.0m thick layer of fill material is filled. When the monitoring data shows that the horizontal displacement increment of any monitoring point during the single-layer filling period exceeds 20mm, or the cumulative horizontal displacement exceeds 60mm, feedback control measures are initiated. The feedback control measures include suspending filling, adding a layer of reinforced geogrid in the displacement area, reducing the thickness of the subsequent three layers of fill material to 0.2m to 0.3m, and increasing the number of compaction passes to 10 to 12.

2. The multi-stage reinforcement construction method for high-fill roadbeds in granite strata as described in claim 1, characterized in that, A layer of needle-punched nonwoven geotextile is also laid on the surface of the permeable crushed stone layer as a filter layer. The unit area mass of the needle-punched nonwoven geotextile is 300 g / m². 2 Up to 400g / m 2 The vertical permeability coefficient is not less than 1.0×10⁻⁶. -2 cm / s, and during laying, a 5% to 8% elongation allowance should be reserved in the cross-sectional direction of the roadbed.

3. The multi-stage reinforcement construction method for high-fill roadbeds in granite strata as described in claim 2, characterized in that, After laying the needle-punched nonwoven geotextile, a layer of coarse sand with a thickness of 50mm to 80mm is first laid and statically compacted, and then the upper filler layer is laid.

4. The multi-stage reinforcement construction method for high-fill roadbeds in granite strata as described in claim 1, characterized in that, The crushed stone soil filler is continuously graded, with a maximum particle size not exceeding 50 mm and a uniformity coefficient C. u ≥5, curvature coefficient C e Between 1 and 3; before laying, its moisture content is adjusted to be within the range of -1% to +2% of the optimal moisture content.

5. The multi-stage reinforcement construction method for high-fill roadbeds in granite strata as described in claim 3, characterized in that, The two legs of the portal anchor have threads or raised patterns on their surfaces, and are coated with cement slurry before installation to enhance the grip and anchoring effect between the anchor and the filler.

6. The multi-stage reinforcement construction method for high-fill roadbeds in granite strata as described in claim 1, characterized in that, At the top of the vertical drainage system where it contacts the permeable gravel layer, a gravel filter bag is installed. The gravel filter bag is made of stones with a particle size of 5mm to 15mm filled in geotextile and has a diameter of 300mm to 400mm.

7. The multi-stage reinforcement construction method for high-fill roadbeds in granite strata as described in claim 6, characterized in that, During the compaction process, when the ambient temperature exceeds 30℃ or the wind speed exceeds 3m / s, water mist curing should be applied to the surface of the compacted filler, with the spraying rate controlled at 0.5L / m. 2 Up to 1.0L / m 2 This is to maintain the moisture content of the filler surface and ensure uniform and stable compaction.

8. The multi-stage reinforcement construction method for high-fill roadbeds in granite strata as described in claim 1, characterized in that, For areas with localized strong weathered granite grooves or fissures, first clean the bottom, then backfill with C15~C20 lean concrete until it is flush with the surrounding rock surface, and then roughen the overall surface; the maximum particle size of the aggregate in the lean concrete shall not exceed 25mm. The crushed stone filler incorporates 0.3% to 0.5% polypropylene mesh fiber with a fiber length of 12mm to 19mm. The fiber is added evenly during the mixing of the filler to suppress drying shrinkage cracks and improve the integrity of the filler.

9. The multi-stage reinforcement construction method for high-fill roadbeds in granite strata as described in claim 1, characterized in that, During the monitoring process in step S4, settlement monitoring points were also set up simultaneously. Based on the data collected from the deep horizontal displacement monitoring points and settlement monitoring points, combined with the roadbed filling height, slope geometric parameters and fill physical and mechanical parameters, a real-time safety factor assessment model for the overall stability of the roadbed was established with the simplified Bishop method as the core. During the filling process, after each layer of fill material is placed, the evaluation model is automatically run once to calculate the real-time safety factor F of the current roadbed. s Set two-level warning thresholds: when 1.30 ≤ F s When the tensile strength is less than 1.50, a primary warning is issued, and a first-level performance compensation measure is initiated: in the next planned reinforcement layer, the tensile strength of the geogrid is temporarily increased by 20% to 30% based on the original design; when 1.20 ≤ F s When the value is less than 1.30, a high-level warning is issued and secondary performance compensation measures are initiated: on the basis of implementing the aforementioned primary performance compensation measures, sandbags are simultaneously piled at the toe of the slope in the corresponding area of ​​the roadbed, and a temporary counter-pressure load of 20 kPa to 35 kPa is applied. Within 24 hours after each completion of the aforementioned feedback control measures, the monitoring frequency is increased to once every 6 hours. Using the displacement and settlement data collected during this period, the parameters of packing cohesion c and internal friction angle φ in the evaluation model are inverted and analyzed. The model parameters are then fine-tuned based on the inversion results for more accurate predictions in the future.

10. The multi-stage reinforcement construction method for high-fill roadbeds in granite strata as described in claim 9, characterized in that, The process of fine-tuning the model parameters based on the inversion analysis and inversion results includes: During the 24-hour intensive monitoring period following the completion of primary or secondary performance compensation measures, the horizontal displacement increment sequence {Δd} of each monitoring point within the specified time period is obtained. i } and the settlement increment sequence {Δs i }, and at the same time record the newly added filling thickness ΔH during this period; Using the horizontal displacement increment sequence and settlement increment sequence as the system response, an inverse analysis objective function is constructed with the packing cohesion c and internal friction angle φ as optimization variables. The objective function is the sum of squared residuals between the calculated and monitored values, i.e.: Minimize: Σ[ (Δd i,cal - Δd i,mea ) 2 + (Δs i,cal - Δs i,mea ) 2 ],(d i,cal , Δs i,cal ) is the calculated value, (Δd) i,mea , Δs i,mea ) represents the monitored value; among which, the calculated value (Δd) i,cal , Δs i,cal The displacement prediction model is obtained by iteratively calculating the result by substituting the candidate (c,φ) parameter set into the simplified Bishop method extended displacement prediction model. The objective function above is solved using a Monte Carlo-steepest descent hybrid algorithm to obtain a set of optimal (c, φ) parameters that make the calculated response closest to the actual monitoring response; the optimal (c, φ) parameters obtained by this inversion are used to update the original packing mechanical parameters in the evaluation model.