A method for testing and reinforcing geotechnical medium parameters

Abstract

The application discloses a kind of geotechnical medium parameter test and reinforcement method, contain following steps, step 1: multiple in-situ test holes and geophysical prospecting test lines are laid in slope, foundation or filling site, in-situ test and geophysical prospecting test are carried out, including in-situ borehole shear test, permeability test and pore pressure observation, to obtain the mechanical properties of soil layer, permeability and pore water pressure parameter;Step 2: according to the obtained soil layer parameter, the initial pressure, flow and slurry concentration required for grouting are calculated;Step 3: using grouting equipment, slurry is injected into soil layer, and the deformation of soil layer, pressure change and permeability change are monitored during grouting process;Step 4: according to the real-time monitoring data, adjust the grouting pressure and flow, to ensure the uniformity of grouting and reinforcement effect, compared with traditional empirical grouting method, the application can significantly improve the uniformity of reinforcement and anti-seepage effect, reduce energy consumption and risk, with good universality and engineering application value.

Description

Technical Field

[0001] This invention belongs to the field of rock and soil foundation reinforcement technology, specifically a method for testing and reinforcing rock and soil media parameters. Background Technology

[0002] In geotechnical engineering, ground reinforcement and seepage prevention are crucial for ensuring the safety and stability of underground structures. Grouting reinforcement technology, due to its flexibility and adaptability, is widely used in projects such as subway tunnels, deep foundation pits, slopes, hydraulic structures, and mine roadways. This technology improves the strength, deformation characteristics, and permeability of the ground by injecting a solidifiable grout into it. However, traditional grouting methods rely heavily on experience to set parameters such as grouting pressure, flow rate, and time, lacking real-time perception and feedback adjustment of the actual ground conditions. Due to the complex and variable nature of the ground and the uneven distribution of parameters, the grouting process often encounters problems such as uneven diffusion, localized leakage, overpressure damage, or insufficient reinforcement, leading to uncertainty in the reinforcement effect and safety risks.

[0003] In existing technologies, the acquisition of geotechnical parameters mainly relies on in-situ testing and laboratory experiments. In-situ testing methods, such as in-situ borehole shear tests, in-situ direct shear tests, pressuremeter tests, permeability tests, and pore pressure observations, can reflect the local mechanical and seepage characteristics of strata, but the number of test points is limited, making it difficult to reflect the overall continuous distribution of strata. Geophysical techniques, such as seismic wave testing and resistivity tomography, can obtain information over a wide area, but their results largely depend on empirical models and are greatly affected by geological structures and water-bearing conditions, resulting in limited accuracy when used alone. Both types of methods have their advantages in spatial resolution and parameter accuracy, but there has long been a lack of effective means to integrate and apply them.

[0004] In terms of grouting construction control, existing systems mostly employ constant pressure or staged pressurization methods, adjusting grouting parameters based on preset curves or manual judgment. These methods are ill-suited to rapidly changing formation conditions, resulting in significant lag in the control process. If formation properties undergo abrupt changes, the system cannot respond promptly, easily leading to grout leakage, localized damage, or energy waste. Some studies have attempted to introduce numerical simulations or empirical formulas to predict the grouting diffusion range, but these models are mostly based on static assumptions, failing to achieve dynamic coupling with field monitoring data and unable to be corrected in real time during construction.

[0005] With the development of intelligent sensing, machine learning, and data fusion technologies, data-driven methods for stratum identification and control have gradually emerged in the field of geotechnical engineering. However, most existing research focuses on single aspects, such as the optimization of inversion algorithms or the improvement of control algorithms, and has not yet formed a complete closed-loop system from stratum identification to grouting control. Existing methods lack the ability to collaboratively analyze multi-source test data, cannot identify changes in stratum parameters in real time and accurately during grouting, and also lack intelligent control mechanisms that can automatically adjust construction strategies based on stratum response.

[0006] Therefore, current technologies still have significant shortcomings in terms of formation parameter testing, data fusion and inversion, and grouting control: First, the information acquisition dimensions are limited, resulting in low accuracy in formation inversion; second, the control system relies on human experience and lacks real-time response capabilities; and third, data and control do not form an effective feedback loop, making it difficult to reduce construction energy consumption and risks. These problems directly affect the uniformity, construction safety, and economy of formation reinforcement, necessitating a new method and system that can organically combine in-situ testing with geophysical information to achieve real-time identification of formation parameters and intelligent control of grouting, thereby improving reinforcement quality and construction efficiency and promoting the intelligent development of geotechnical engineering. Summary of the Invention

[0007] The purpose of this invention is to provide a method for testing and reinforcing soil and rock media parameters. By combining multi-source information fusion, parameter inversion, and intelligent control, it realizes the transformation of soil reinforcement from experience-based construction to intelligent and precise construction. This method not only improves the predictability and consistency of reinforcement effects but also significantly reduces construction risks and energy consumption. It demonstrates superior comprehensive effects in terms of technical rationality, economy, and safety, and has significant innovative and engineering promotion value.

[0008] The technical solution adopted in this invention is as follows:

[0009] A method for testing and reinforcing soil and rock media parameters includes the following steps:

[0010] Step 1: Set up multiple in-situ test holes and geophysical exploration lines on the slope, foundation or fill site to conduct in-situ tests and geophysical exploration tests, including in-situ borehole shear tests, permeability tests and pore pressure observations, in order to obtain the mechanical properties, permeability characteristics and pore water pressure parameters of the soil layer;

[0011] Step 2: Based on the obtained soil parameters, calculate the required initial grouting pressure, flow rate, and grout concentration;

[0012] Step 3: Use grouting equipment to inject grout into the soil layer, and monitor the deformation, pressure changes and permeability changes of the soil layer during the grouting process;

[0013] Step 4: Adjust the grouting pressure and flow rate based on real-time monitoring data to ensure grouting uniformity and reinforcement effect, and optimize the soil reinforcement process.

[0014] Preferably, in step 1, in-situ test boreholes and geophysical exploration lines are laid out in the target reinforcement area, with a borehole spacing of 1.5~3.0m; the testing equipment includes an in-situ borehole shearing apparatus, an in-situ direct shear apparatus, a pressure gauge, a pore pressure gauge, a permeameter, a vane shearing apparatus, and strain gauges; a resistivity tomography and seismic wave detection system are simultaneously deployed; the collected data include pressure Pt, deformation u, pore water pressure pw, permeability km, wave velocity Vp, and resistivity ρ; the data sampling frequency is not less than 1Hz; the in-situ test and geophysical exploration data are normalized and preprocessed, input into a hybrid inversion neural network, and output formation parameters: elastic modulus E (MPa), Poisson's ratio ν, permeability k (cm / s), cohesion c (kPa), and internal friction angle. (°);

[0015] In step 2, based on the inversion results, the initial grouting pressure P and flow rate Q are calculated using the following formula:

[0016] ;

[0017] ;

[0018] : These are the stratigraphic parameters obtained through inversion; : This refers to the design target value or initial baseline value; : This is an empirical coefficient, ranging from 0.2 to 1.5; Initial grouting pressure (MPa); Grouting flow rate (L / min);

[0019] In step 4, during the grouting process, grouting pressure, flow rate, pore pressure changes, deformation, acoustic wave diffusion signal, and resistivity changes are collected in real time to form a state vector:

[0020] ;

[0021] Input the reinforcement learning controller and update the grouting parameters as follows:

[0022] ;

[0023] in: For the reward function; Energy consumption for grouting (kJ); To set an upper limit for energy consumption (kJ); These represent the relative changes in the inversion parameters; These are weighting coefficients; the controller employs a deep deterministic policy gradient algorithm, based on... Value optimization and To achieve adaptive control of grouting;

[0024] Grouting will automatically stop when the following condition is met: formation stiffness increase rate. Permeability coefficient decrease rate Furthermore, the rate of change of formation volumetric strain .

[0025] Preferably, the hybrid inversion neural network adopts a 3-layer fully connected residual structure: the input layer has 8 nodes, and the input parameters are... The hidden layer has 128 nodes each; the output layer has 4 nodes (E, v, k, c); the activation function is ReLU, and the optimizer is Adam; the number of training samples is no less than 500. Among these: Shear strength (kPa); For depth (m).

[0026] Preferably, the resistivity tomography electrode spacing for geophysical acquisition is 0.5m, the detection depth is 10-30m, the seismic wave detection sampling interval is 0.2m, and the ground-penetrating radar center frequency is 200MHz, used to identify changes in the medium interface before and after grouting.

[0027] Preferably, the grouting slurry is a cement-water glass dual-liquid system: cement slurry concentration 30-45wt%; water glass concentration 10-20wt%; water-cement ratio 0.6-1.0; reaction time 5-15min; the system automatically adjusts the mixing ratio according to the permeability, and the control error does not exceed ±3%.

[0028] Preferably, this method is applicable to weak clay, gravel layers and fault fracture zones, with grouting hole depth of 3 to 5 m and hole spacing of 1.0 to 2.5 m; it can form a closed loop of "inversion prediction - real-time reinforcement - retesting and verification" within 5 m in front of the subway tunnel face.

[0029] A geotechnical medium parameter testing and intelligent grouting reinforcement system for implementing this method includes an in-situ testing unit, a geophysical inversion unit, a data fusion and inversion calculation unit, a grouting execution and monitoring unit, and an intelligent control and optimization decision-making unit. Each unit is connected to a central server via an industrial Ethernet. The central server runs an inversion algorithm module, a reinforcement learning control module, and a three-dimensional visualization module, and enables remote monitoring and parameter synchronization updates via a mobile terminal.

[0030] Preferably, the grouting execution and monitoring unit includes:

[0031] Automatic pressure controlled grouting pump (range 0~8MPa, accuracy ±0.05MPa);

[0032] Flow meter (0~100L / min);

[0033] Online slurry concentration monitor (accuracy 0.1%).

[0034] Acoustic wave monitoring probe (20-100kHz);

[0035] All sensors are connected to the central control system via Industrial Ethernet (TCP / IP protocol);

[0036] The in-situ test unit connects to the central server via a LoRa wireless module, with a communication frequency of 433MHz, a transmission distance of ≥200m, and automatic calibration and temperature compensation functions.

[0037] Preferably, the state update cycle of the reinforcement learning control module is 5s, and the action adjustment range is: pressure change ΔP: ±0.5MPa; flow rate change ΔQ: ±10L / min; and the control step size is automatically adjusted based on the Reward value trend.

[0038] Preferably, the 3D visualization module displays the inversion parameters through voxel interpolation and color isosurface rendering. , , , The spatial distribution of the data, updated at a frequency of ≥1 time / minute, is used to determine the uniformity of reinforcement.

[0039] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0040] The beneficial effects of this invention are reflected in three aspects: testing, identification, and control, and are present throughout the entire reinforcement process.

[0041] From a mechanistic perspective, it establishes a dynamic perception capability of the true state of the formation through a combination of in-situ tests and geophysical inversion. In-situ tests provide information on local stress and permeability of the formation, while geophysical exploration reflects large-scale structural changes. The fusion of these two data sets forms a comprehensive understanding of the overall characteristics of the formation. This complementary information allows the elasticity, permeability, and cohesion properties of the formation to be identified and tracked in real time during the reinforcement process, providing an accurate basis for subsequent control.

[0042] During the grouting control phase, this invention couples the inverted formation parameters with a reinforcement learning algorithm, enabling the grouting system to possess self-learning and adaptive capabilities. The system can automatically adjust the grouting pressure, flow rate, and grout concentration based on changes in formation stiffness, permeability, and deformation compatibility, thereby maintaining the formation within a safe and effective reinforcement range. Under different formation conditions, the controller optimizes the grouting strategy through real-time feedback, avoiding formation uplift due to overpressure or insufficient reinforcement due to underpressure, achieving steady-state control throughout the entire process.

[0043] Compared to traditional manual experience-based control, the significant advantages of this invention lie in higher reinforcement uniformity, lower energy consumption, and reduced construction risk. Because the system can identify the formation condition in real time and adjust control parameters accordingly, the grout diffusion within the formation is more uniform, and the reinforcement range is more controllable. The formation's impermeability and strength are improved more stably, and the construction process no longer relies on the experience of a single operator but achieves self-optimization through data-driven processes. Simultaneously, energy and material usage are more rational, reducing ineffective grouting and energy waste, resulting in significant economic benefits.

[0044] Furthermore, this invention is highly adaptable and can operate stably in various geological environments, including clay layers, gravel layers, permeable interlayers, fault fracture zones, and slopes. It can automatically match control strategies based on the response of various strata without requiring changes to the control model, demonstrating broad applicability and a high level of intelligence. The digitization and visualization of the entire reinforcement process make the construction status monitorable and controllable, and the project quality quantifiable and assessable.

[0045] In summary, this invention, through the combination of multi-source information fusion, parameter inversion, and intelligent control, realizes the transformation of ground reinforcement construction from experience-based to intelligent and precise construction. It not only improves the predictability and consistency of reinforcement effects but also significantly reduces construction risks and energy consumption. It demonstrates superior comprehensive effects in terms of technical rationality, economy, and safety, possessing significant innovation and engineering promotion value. Attached Figure Description

[0046] Figure 1 This is a schematic flowchart of the method of the present invention;

[0047] Figure 2 This is a schematic diagram illustrating the specific process of the method of the present invention;

[0048] Figure 3 This is a schematic diagram of the system architecture of the present invention. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0050] See Figures 1 to 3 This embodiment is a method for testing soil and rock medium parameters and intelligent grouting reinforcement based on the coupling of in-situ testing and geophysical inversion. It is applicable to grouting reinforcement projects in soft strata, gravel layers and fault fracture zones, and has significant effects, especially in the face of subway tunnels, deep foundation pit support and underground structure seepage prevention and reinforcement.

[0051] This method forms a closed-loop control system through "in-situ testing - geophysical inversion - parameter identification - intelligent grouting - feedback correction", which enables dynamic inversion of formation parameters, automatic adjustment of grouting parameters, and adaptive optimization of the reinforcement process.

[0052] The system mainly consists of the following five units:

[0053] In-situ testing unit: includes in-situ borehole shearing instrument, in-situ direct shearing instrument, pressure gauge, pore pressure gauge, permeameter, vane shearing instrument and strain gauge; adopts LoRa wireless communication module (433MHz, transmission distance ≥200m), automatic calibration accuracy is ±0.1%FS, sampling frequency ≥1Hz.

[0054] Geophysical inversion unit: including resistivity tomography (ERT), seismic wave detection system and ground-penetrating radar (GPR): electrode spacing 0.5m; detection depth 10-30m; P-wave frequency 150Hz, S-wave frequency 75Hz; GPR center frequency 200MHz, used for reflection interface detection.

[0055] Data fusion and inversion computing unit: Composed of an industrial computer and a GPU acceleration module, running a hybrid inversion neural network with the following parameters: Input layer 8 nodes: ; 3 hidden layers, each with 128 nodes; ReLU activation function; Output layer: ; Optimizer Adam (learning rate) Loss function MSE; 500 training samples. Output parameter definition: : Formation equivalent elastic modulus (MPa); Poisson's ratio (dimensionless); : Permeability coefficient (cm / s); Cohesion (kPa), internal friction angle (°).

[0056] Grouting execution and monitoring unit: includes an automatic pressure-controlled grouting pump, flow meter, online grout concentration monitor, and acoustic monitoring probe; pressure control range 0-8MPa, accuracy ±0.05MPa; flow range... L / min; slurry concentration measurement accuracy 0.1%; acoustic detection frequency kHz. Each device is connected to the central server via Industrial Ethernet (TCP / IP).

[0057] Intelligent control and optimization decision-making unit: Runs a reinforcement learning controller using the Deep Deterministic Policy Gradient (DDPG) algorithm, with a state update cycle of 5 seconds, a learning rate of 0.01, and a discount factor of 0.95. The action output is the grouting pressure adjustment amount. With flow adjustment .

[0058] Each unit is connected to a central server via an industrial Ethernet network. The central server runs an inversion algorithm module, a reinforcement learning control module, and a 3D visualization module, and enables remote monitoring and parameter synchronization via mobile terminals. The reinforcement learning control module has a status update cycle of 5 seconds, and its action adjustment range is: pressure change ΔP: ±0.5MPa; flow rate change ΔQ: ±10L / min; and it automatically adjusts the control step size based on the reward value trend. The 3D visualization module displays the inversion parameters through voxel interpolation and color isosurface rendering. , , , The spatial distribution of the data, updated at a frequency of ≥1 time / minute, is used to determine the uniformity of reinforcement.

[0059] The implementation steps of the method are as follows:

[0060] S1. Combined in-situ testing and geophysical exploration data acquisition

[0061] Test boreholes and geophysical probes were deployed within a 5m area in front of the subway tunnel face: the in-situ test boreholes were 4m deep and spaced 2m apart; pressure was collected. Deformation amount pore pressure Permeability coefficient Wave speed resistivity The sampling frequency is 1Hz, and the data is transmitted in real time to the central server via a LoRa wireless module.

[0062] S2. Data Fusion and Parameter Inversion

[0063] After normalization and Kalman filtering, the input is HybridInversionNet.

[0064] The following is an example of outputting formation parameters:

[0065]

[0066] S3. Calculation of initial parameters for grouting

[0067] Based on the inversion results and design target parameters:

[0068]

[0069]

[0070] Substitute into the following formula to calculate:

[0071] ;

[0072] Pick ;

[0073] ;

[0074] ;

[0075] set up ;

[0076] ;

[0077] Therefore, the initial grouting pressure is set to... Traffic settings are as follows .

[0078] S4. Intelligent Grouting and Feedback Correction

[0079] The system collects pore pressure and slurry diffusion radius in real time. resistivity change rate With formation deformation This forms a state vector:

[0080] ;

[0081] The reinforcement learning controller bases its decisions on the reward function:

[0082]

[0083] in:

[0084] , , , , ;

[0085] For pump energy consumption, kJ.

[0086] The system updates every 5 seconds, based on Value optimization , ;

[0087] For example, within a certain period:

[0088] , , , kPa, Substituting into the calculation, we get This indicates that the reinforcement effect is good, and the system automatically reduces [the size / weight]. MPa, keep the flow rate constant to avoid excessive energy consumption.

[0089] The grouting slurry is a cement-water glass dual-liquid system: cement slurry concentration 30-45wt%; water glass concentration 10-20wt%; water-cement ratio 0.6-1.0; reaction time 5-15min; the system automatically adjusts the mixing ratio according to the permeability, and the control error does not exceed ±3%.

[0090] S5. Determination of the Reinforcement End Point

[0091] Grouting will automatically terminate when the following conditions are met: .

[0092] Upon termination, the system automatically outputs a reinforcement report, including: grouting pressure-time curve; flow rate change curve; inversion parameter evolution diagram; and three-dimensional visualization of the reinforcement zone range.

[0093] After reinforcement is completed, the system automatically reconstructs the three-dimensional model of the formation: using voxel interpolation... Mapped to a 3D mesh (0.25m resolution), the distribution of ground reinforcement is displayed through colored isosurfaces.

[0094] The results show that the radius of the reinforced area is approximately 2.3m; An average increase of 76%; The energy consumption of grouting was reduced by 92% compared to the design value; the energy consumption of grouting was reduced by 8% compared to the design value. Subsequent pressure meter tests verified that the formation deformation modulus increased from 85MPa to 150MPa, with an error of less than 5% compared to the inversion results.

[0095] This method can be applied to subway tunnels, slope protection, underground stations, and seepage prevention projects for hydraulic structures.

[0096] In practical engineering, this system can be implemented using standard equipment, and the control algorithm can run on an industrial PC or embedded controller without requiring additional high-end hardware. After three consecutive engineering verifications, the system has demonstrated high stability, with an inversion error of less than ±8%, enabling fully automated grouting control.

[0097] This invention is based on the dynamic sensing and feedback control principle of multi-field coupled behavior of soil and rock media. Through the combined application of in-situ tests and geophysical inversion, it achieves real-time identification of formation parameters and adaptive adjustment of the grouting process. The entire system constitutes a closed-loop control system consisting of "sensing—cognition—decision—execution—feedback". Its core mechanism lies in coupling the formation physical response and grouting control variables into the same information chain, transforming the reinforcement process from experience-based construction to data-driven intelligent control.

[0098] In the information acquisition phase, this invention simultaneously collects multi-dimensional stratigraphic information by deploying in-situ test boreholes and geophysical exploration lines. In-situ testing provides point-like mechanical response data, such as pressure, pore water pressure, deformation, and permeability, reflecting the local bearing capacity and deformation characteristics of the strata. Geophysical data, through resistivity and seismic wave velocity, reflects the structural continuity and pore water-bearing characteristics of the strata. These two types of data are complementary in scale, and their fusion can yield a unified description of the macroscopic structure and microscopic response of the strata.

[0099] After data fusion, a hybrid inversion neural network is employed to comprehensively learn and identify parameters from signals originating from different sources. This network establishes a nonlinear mapping relationship between physical quantities such as seismic wave velocity, resistivity, permeability rate, and pore pressure response, and constitutive parameters of the soil and rock through a large number of samples, thereby retrieving the equivalent elastic modulus, Poisson's ratio, permeability coefficient, cohesion, and internal friction angle of the formation. These parameters reflect the formation stiffness, volumetric deformation compatibility, permeability, and shear strength, and are core indicators characterizing the reinforcement status. The network continuously updates its parameter output during real-time operation, achieving dynamic identification of formation characteristics over time and space.

[0100] Grouting control relies on these inversion parameters. The elastic modulus represents formation stiffness, Poisson's ratio reflects deformation compatibility, the permeability coefficient determines grout diffusion capacity, and cohesion and internal friction angle affect the strength and stability after reinforcement. The system analyzes the relative deviations of these parameters to comprehensively assess whether the formation has reached the optimal reinforcement state. If the formation stiffness is too low or the permeability is too high, the system automatically increases the grouting pressure and flow rate to enhance permeability; if the formation deformation rate is too high or the energy consumption is too high, the pressure is reduced to prevent structural damage. This control process forms an adaptive matching mechanism based on physical response, ensuring that the grouting zone remains in a balanced range of reinforcement without excessive disturbance.

[0101] The control algorithm employs reinforcement learning principles, using a reward and punishment mechanism to evaluate the effectiveness of each parameter adjustment. The system sets comprehensive performance indicators, taking formation stiffness increase, permeability reduction, energy consumption level, and deformation stability as optimization objectives. The algorithm automatically corrects the control strategy through multiple iterations, gradually approaching the optimal grouting parameter combination, achieving real-time self-learning and dynamic optimization. Unlike traditional static control, the algorithm of this invention can adaptively update when formation conditions change, thereby ensuring a continuous, stable, and efficient reinforcement process.

[0102] During grouting, the system monitors signals such as pore pressure, resistivity changes, acoustic propagation time, and grouting energy consumption in real time through a sensor network. These signals reflect the formation density and grout diffusion state. The monitoring data is fused and analyzed to determine the reinforcement effect. When stiffness growth stabilizes, permeability decreases to a set ratio, deformation rate approaches zero, and energy consumption no longer increases significantly, the system determines that reinforcement is complete and automatically terminates grouting. This process forms a real-time "detection-judgment-control-verification" cycle, ensuring optimal utilization of grouting energy and materials.

[0103] Physically, grouting causes the grout to seep and diffuse through porous media, altering the formation's pore structure and stress distribution. As the grout solidifies, the effective stress of the formation increases, porosity decreases, and the elastic modulus and cohesion increase, while the permeability coefficient decreases. This invention accurately tracks these changes through real-time parameter inversion, enabling the control system to respond instantly to even minor changes in formation properties, achieving dynamic optimization of the reinforcement effect.

[0104] To further clarify the technical solution and beneficial effects of the present invention, the implementation process of the present invention is described in detail below with reference to specific engineering examples. The embodiments described are based on the same technical principles and system architecture, and their application under different geological conditions verifies the universality, stability, and engineering effectiveness of the present invention. The implementation methods of the present invention are not limited to the following examples; all formation testing and grouting reinforcement performed using the technical ideas and control principles of the present invention should be considered to fall within the protection scope of the present invention.

[0105] In each embodiment, a combined acquisition method of in-situ testing and geophysical exploration is employed. The in-situ testing portion utilizes equipment such as in-situ borehole shear gauges, in-situ direct shear gauges, pressure gauges, pore pressure gauges, permeameters, and strain gauges to obtain the local stress-strain response and permeability characteristics of the formation. The geophysical exploration portion obtains the electrical structure and wave velocity distribution of the formation through resistivity tomography and seismic wave testing. All data are fused and input into a hybrid inversion neural network, which outputs inversion parameters such as the formation's equivalent elastic modulus, Poisson's ratio, permeability coefficient, cohesion, and internal friction angle. The system automatically sets the initial grouting pressure and flow rate based on the deviation between these parameters and the design target, and adjusts the control quantities in real time during the grouting process using a reinforcement learning algorithm to keep the formation reinforcement process within the optimal stable range.

[0106] Each embodiment includes four stages: geological conditions, testing and inversion process, grouting control strategy, and reinforcement effect verification. The parameters, equipment models, and operating conditions used are all actual values ​​within the feasible range of the project. By comparing the results with traditional constant pressure or manually controlled grouting methods, the adaptability and superiority of this invention in different types of strata are systematically demonstrated. The embodiments cover five typical working conditions: silty clay layer, gravel layer, permeable interlayer in deep foundation pit, fault fracture zone, and slope seepage prevention, respectively verifying the comprehensive performance of this invention in seepage prevention, reinforcement, pressure stabilization, energy saving, and deformation control.

[0107] Example 1: Grouting reinforcement in front of the tunnel face in a subway section

[0108] During the construction of a subway tunnel in a certain city, the stratum ahead consisted of silty clay interbedded with silty sand, exhibiting high permeability and sensitivity to deformation. To prevent face collapse and leakage, the in-situ testing combined with geophysical inversion of the soil and rock medium parameters and the grouting reinforcement system of this invention were employed. Before construction, four in-situ test boreholes and two geophysical probes were laid out within a 5-meter radius ahead of the tunnel face. Ground response data were simultaneously collected using an in-situ borehole shear tester, an in-situ direct shear tester, a pressure gauge, a pore pressure gauge, a permeameter, and seismic wave and resistivity testing instruments. The formation elastic modulus was determined to be 85 MPa and the permeability coefficient was [missing information]. The system has a Poisson's ratio of 0.31 and a cohesion of 28 kPa. The initial grouting pressure is automatically set to 2.3 MPa and the flow rate to 42 L / min based on the deviation relationship.

[0109] During the grouting process, a reinforcement learning control algorithm continuously adjusted grouting parameters based on changes in pore pressure, resistivity, and grout energy consumption. When the system detected a slowdown in the formation stiffness growth rate and a decrease in permeability to near the target value, it gradually reduced the pressure to prevent excessive seepage. The entire process involved twenty rounds of parameter adjustments, resulting in a 14% reduction in grouting energy consumption compared to manual control. After grouting was completed, sampling tests showed that the elastic modulus increased to 150 MPa, and the permeability coefficient decreased to [missing value]. The reinforced zone had a radius of 2.3 meters. Subsequent monitoring showed that the stratum deformation was stable and leakage was completely eliminated. This embodiment verifies that the present invention can effectively achieve adaptive pressure control and improve reinforcement uniformity in weak silty clay layers, increasing construction efficiency by approximately 15%.

[0110] Example 2: Grouting for seepage prevention in subway tunnel sections with sand and gravel layers

[0111] When traversing a highly permeable gravel layer in a subway section, traditional constant-pressure grouting often results in grout leakage and pipe blockage. By employing the combined testing and intelligent control method of this invention, five in-situ test boreholes were deployed around the tunnel, and a three-dimensional geophysical survey line was arranged. The geological distribution characteristics were obtained through permeability tests and resistivity tomography. The inversion results show an elastic modulus of 110 MPa, a Poisson's ratio of 0.27, and a permeability coefficient of... The system calculates the initial grouting pressure as 1.9 MPa and the flow rate as 38 L / min.

[0112] During construction, the reinforcement learning control system automatically identified the grouting stage based on the formation permeability reduction rate and pore pressure change trend. In the initial stage, the system determined that the grout diffusion was sufficient and the formation permeability was high, maintaining a low-viscosity grout to facilitate diffusion. In the later stage, when the permeability decreased significantly and the pore pressure stabilized, the system automatically switched to a high-concentration grout to form a dense curtain. No grout leakage or pipe blockage occurred during the grouting process. After reinforcement, the elastic modulus increased to 190 MPa, and the permeability decreased to [missing value]. Energy consumption is reduced by 18% compared to traditional manual control. This example demonstrates that the present invention can automatically and in stages regulate slurry characteristics in highly permeable formations through a self-learning algorithm, achieving efficient seepage prevention and uniform reinforcement.

[0113] Example 3: Reinforcement of permeable interlayer between retaining piles in deep foundation pit

[0114] In the foundation pit construction of an underground commercial complex, a permeable sandy interlayer exists between the retaining piles, with a depth of approximately 18 meters. Traditional manual grouting methods are prone to overpressure, causing pile displacement. By applying the system of this invention, test holes are laid out between the retaining piles, and pore pressure gauges and electrode arrays are installed. Real-time ground data is obtained through pressure-side and resistivity tests. The inversion results show a permeability coefficient of... The cohesion is 38 kPa. The system automatically determines the initial grouting pressure to be 1.5 MPa and the flow rate to be 35 L / min. During the grouting process, the intelligent control system updates every 5 seconds and automatically fine-tunes the pressure according to the changes in pore pressure, keeping the fluctuation range within ±0.05 MPa.

[0115] No sudden pressure or grout leakage occurred during the entire grouting process. Post-completion testing showed that the permeability coefficient decreased to [value missing]. The cohesion was increased to 60 kPa, the average pore pressure distribution was uniform, and the maximum surface settlement was only 3 mm, lower than the design allowable value. Compared with adjacent traditional construction areas, the stratum disturbance was reduced by about 50%, and the seepage prevention effect was significant. This embodiment proves that the present invention can maintain pressure stability and prevent overpressure damage under high water pressure and weak interlayer conditions, achieving safe and efficient retaining reinforcement.

[0116] Example 4: Pre-grouting of tunnels in fault fracture zones

[0117] During the construction of a railway tunnel traversing a fault fracture zone, the fault width was approximately 6 meters, with well-developed joints and fissures and abundant groundwater. Traditional grouting methods struggled to control the grout diffusion range, easily leading to localized grout inrushes. Using the system of this invention, three geophysical probes were deployed for real-time monitoring, and multiple in-situ test boreholes were arranged to determine pore pressure and permeability. Inversion yielded an elastic modulus of 45 MPa and a permeability coefficient... Poisson's ratio is 0.25. The system automatically calculates the initial pressure of 1.6 MPa based on the parameters and initiates intelligent control.

[0118] As grouting progressed, the system detected an accelerated rate of increase in formation stiffness and a decrease in permeability, automatically reducing the flow rate to maintain stable energy consumption. The reward value of the control algorithm stabilized at 0.72 after multiple iterations, indicating that the control system had achieved energy efficiency balance. After reinforcement, the elastic modulus increased to 120 MPa, and the permeability coefficient decreased to... The deformation convergence rate of the surrounding rock in the fault zone is increased by 3 times, and the energy consumption is reduced by 20% compared with the traditional method. This embodiment verifies that under complex geological conditions, the present invention can adaptively balance pressure and energy consumption, ensuring grouting stability and controllable reinforcement of the fault zone.

[0119] Example 5: Slope seepage prevention and stabilization reinforcement

[0120] In a mudstone and sandstone slope engineering project, the slope was prone to seepage and sliding during the rainy season due to well-developed fissures. To improve seepage prevention and anti-sliding capabilities, the method of this invention was used for seepage prevention grouting. The system obtained stratigraphic data through geophysical exploration and in-situ testing, and the inversion parameters were an elastic modulus of 95 MPa and a permeability coefficient. The cohesion is 35 kPa. The control system is set to an initial grouting pressure of 1.8 MPa, monitors the slope deformation rate in real time, and automatically adjusts the pressure and flow rate based on feedback. During construction, the system detects a decrease in the deformation rate and determines that the slope is stabilizing, gradually reducing the pressure to complete the reinforcement.

[0121] Post-grouting test results showed that the permeability coefficient decreased to [value missing]. The cohesion was increased to 52 kPa, the surface seepage flow of the slope was reduced by 90%, and the overall anti-sliding safety factor increased from 1.15 to 1.35. The energy consumption during the entire construction was controlled within the design range, and no surface heave or crack propagation occurred. This embodiment illustrates that the present invention can achieve synergistic control of "deformation monitoring—pressure regulation—seepage inhibition" in slope-like strata, significantly improving seepage prevention and stability.

[0122] As demonstrated by the five embodiments above, this invention can operate stably in various complex geological environments, including silty clay layers, gravel layers, permeable interlayers, fault fracture zones, and slopes. The system can automatically identify changes in formation parameters and dynamically match grouting pressure and flow rate, achieving uniform reinforcement and energy consumption optimization. Compared with traditional methods, the elastic modulus is increased by an average of 70%, the permeability coefficient is reduced by 90%, energy consumption is reduced by approximately 18%, reinforcement uniformity is improved by 50%, and the construction cycle is shortened by approximately 15%. The results show that the intelligent grouting control mechanism based on the coupling of in-situ testing and geophysical inversion proposed in this invention has significant engineering adaptability and technical advantages. It can achieve self-learning, self-adjustment, and high-efficiency energy saving in the formation reinforcement process, demonstrating significant creativity and practicality.

[0123] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for testing and reinforcing parameters of soil and rock media, characterized in that, Includes the following steps: Step 1: Set up multiple in-situ test holes and geophysical lines on the slope, foundation or fill site to conduct in-situ tests, including in-situ borehole shear tests, permeability tests and pore pressure observations, in order to obtain the mechanical properties, permeability characteristics and pore water pressure parameters of the soil layer; Step 2: Based on the obtained soil parameters, calculate the required initial grouting pressure, flow rate, and grout concentration; Step 3: Use grouting equipment to inject grout into the soil layer, and monitor the deformation, pressure changes and permeability changes of the soil layer during the grouting process; Step 4: Adjust the grouting pressure and flow rate based on real-time monitoring data to ensure grouting uniformity and reinforcement effect, and optimize the soil reinforcement process; In step 1, in-situ test holes and geophysical probes are laid out in the target reinforcement area, with a spacing of 1.5~3.0m between the in-situ test holes. The testing equipment includes an in-situ borehole shearing apparatus, an in-situ direct shear apparatus, a pressure gauge, a pore pressure gauge, a permeameter, a vane shearing apparatus, and strain gauges. A resistivity tomography and seismic wave detection system are deployed simultaneously. The collected data include pressure Pt, deformation u, pore water pressure pw, permeability coefficient km, wave velocity Vp, and resistivity ρ. The data sampling frequency is no less than 1Hz; the in-situ test and geophysical data are normalized and preprocessed, then input into a hybrid inversion neural network, and the output formation parameters are: elastic modulus E (MPa), Poisson's ratio ν, permeability coefficient k (cm / s), cohesion c (kPa), and internal friction angle. (°); In step 2, based on the inversion results, the initial grouting pressure P and flow rate Q are calculated using the following formula: ； ； : These are the stratigraphic parameters obtained through inversion; : This refers to the design target value or initial baseline value; : This is an empirical coefficient, ranging from 0.2 to 1.5; Initial grouting pressure (MPa); Grouting flow rate (L / min); In step 4, during the grouting process, grouting pressure, flow rate, pore pressure changes, deformation, acoustic wave diffusion signal, and resistivity changes are collected in real time to form a state vector: ； Input the reinforcement learning controller and update the grouting parameters as follows: ； in: For the reward function; Energy consumption for grouting (kJ); To set an upper limit for energy consumption (kJ); These represent the relative changes in the inversion parameters; These are weighting coefficients; the controller employs a deep deterministic policy gradient algorithm, based on... Value optimization and To achieve adaptive control of grouting; Grouting will automatically stop when the following condition is met: formation stiffness increase rate. Permeability coefficient decrease rate Furthermore, the rate of change of formation volumetric strain .

2. The method for testing and reinforcing soil and rock media parameters according to claim 1, characterized in that: The hybrid inversion neural network adopts a 3-layer fully connected residual structure: the input layer has 8 nodes, and the input parameters are... The hidden layer has 128 nodes each; the output layer has 4 nodes (E, v, k, c); the activation function is ReLU, and the optimizer is Adam; the number of training samples is no less than 500; among which: Shear strength (kPa); For depth (m).

3. The method for testing and reinforcing soil and rock media parameters according to claim 1, characterized in that: The resistivity tomography electrode spacing for geophysical acquisition is 0.5m, the detection depth is 10-30m, the seismic wave detection sampling interval is 0.2m, and the ground-penetrating radar center frequency is 200MHz, which is used to identify changes in the medium interface before and after grouting.

4. The method for testing and reinforcing soil and rock media parameters according to claim 1, characterized in that: The grout used for grouting is a cement-water glass dual-liquid system: cement grout concentration 30-45wt%; water glass concentration 10-20wt%; water-cement ratio 0.6-1.0; reaction time 5-15min; the system automatically adjusts the mixing ratio according to the permeability, and the control error does not exceed ±3%.

5. The method for testing and reinforcing soil and rock media parameters according to claim 1, characterized in that: This method is applicable to weak clay, gravel layers and fault fracture zones, with grouting hole depths of 3–5 m and hole spacing of 1.0–2.5 m; it can form a closed loop of "inversion prediction—real-time reinforcement—retesting and verification" within 5 m in front of the subway tunnel face.

6. A soil and rock medium parameter testing and intelligent grouting reinforcement system for implementing the soil and rock medium parameter testing and reinforcement method according to any one of claims 1 to 5, characterized in that: It includes an in-situ testing unit, a geophysical inversion unit, a data fusion and inversion calculation unit, a grouting execution and monitoring unit, and an intelligent control and optimization decision-making unit. Each unit is connected to a central server via an industrial Ethernet. The central server runs an inversion algorithm module, a reinforcement learning control module, and a 3D visualization module, and enables remote monitoring and parameter synchronization updates via mobile terminals.

7. The geotechnical medium parameter testing and intelligent grouting reinforcement system according to claim 6, characterized in that, The grouting execution and monitoring unit includes: Automatic pressure controlled grouting pump; Flow meter; Online slurry concentration monitoring instrument; Acoustic wave monitoring probe; All sensors are connected to the central control system via industrial Ethernet; The in-situ test unit connects to the central server via a LoRa wireless module, with a communication frequency of 433MHz, a transmission distance of ≥200m, and automatic calibration and temperature compensation functions.

8. The soil and rock medium parameter testing and intelligent grouting reinforcement system according to claim 6, characterized in that: The state update cycle of the reinforcement learning control module is 5 seconds, and the action adjustment range is: pressure change ΔP: ±0.5MPa; flow rate change ΔQ: ±10L / min; and the control step size is automatically adjusted based on the Reward value trend.

9. The soil and rock medium parameter testing and intelligent grouting reinforcement system according to claim 6, characterized in that: The 3D visualization module displays the inversion parameters through voxel interpolation and colored isosurface rendering. , , , The spatial distribution of the data, updated at a frequency of ≥1 time / minute, is used to determine the uniformity of reinforcement.

Images