A Dynamic Calculation Method for Pipe Jacking Resistance Based on Real-Time Mud Sensing Data

Abstract

This invention relates to the field of tunnel and underground engineering technology, and particularly to a dynamic calculation method for pipe jacking resistance based on real-time mud sensing data. The method includes the following steps: S1. Constructing a pipe-slurry-soil interface friction force database; S2. Calculating the jacking resistance based on real-time mud sensing data; S3. Optimizing the data-driven friction force database; S4. Calculating the hydraulic cylinder thrust based on the real-time jacking resistance. This method directly determines the "pipe-slurry-soil" contact state by real-time detection of the drag-reducing mud thickness and distribution, and calls upon the dynamically optimized interface friction force database to achieve real-time, accurate, and unitized calculation of jacking resistance. Utilizing machine learning technology, the database parameters can self-learn and adaptively correct based on field measurement data, ensuring the long-term accuracy of the calculation model. Based on the real-time calculated and predicted resistance distribution, an intelligent control system is constructed to achieve adaptive matching and control of the thrust of the jacking cylinder and the articulated cylinder.

Description

Technical Field

[0001] This invention relates to the field of tunnel and underground engineering technology, and in particular to a dynamic calculation method for pipe jacking resistance driven by real-time mud sensing data. Background Technology

[0002] Pipe jacking, with its minimal impact on ground traffic and the environment and strong adaptability to different sections, has become an important trenchless technology for short-distance tunnel construction. However, the frictional resistance between the jacking machine head and pipe sections and the ground during pipe jacking is crucial for long-distance jacking, reducing energy consumption, and minimizing ground disturbance. Accurate calculation of the jacking resistance is fundamental to the precise control of pressure parameters in the launching shaft's thrust cylinders and the articulated cylinders between the front and tail shields during construction. This significantly improves the jacking machine head's attitude control and reduces energy consumption. However, in actual jacking operations, the thrust of the thrust cylinders and articulated cylinders is often adjusted based on experience. The jacking resistance around the pipe is constantly changing due to variations in ground conditions and the distribution of drag-reducing mud, making it difficult to accurately adjust the thrust distribution of the thrust cylinders and articulated cylinders to achieve balance with the frictional resistance around the pipe and thus achieve precise control of the jacking attitude and position. Therefore, accurate calculation of the frictional resistance distribution around the pipe during construction is urgently needed.

[0003] Publication No. CN105512411A discloses a method for calculating the thrust of pipe jacking based on the spatiotemporal effect of surrounding rock deformation. This method considers the spatiotemporal effect of surrounding rock deformation during construction, analyzes the loads at different locations of the pipe jacking, and calculates the jacking resistance. Publication No. CN117436263A discloses a method for calculating the thrust of underground pipe jacking tunnels. This method calculates the jacking resistance by analyzing the changes in the contact state between the pipe section and the soil / mud caused by the offset of the pipe jacking machine.

[0004] Existing methods for calculating pipe jacking resistance, such as CN105512411A and CN117436263A, are constantly improving in terms of theoretical calculation. However, they are essentially methods based on idealized assumptions and cannot accurately reflect the real dynamic changes in the pipe-soil contact state and friction parameters during pipe jacking construction. They also cannot provide effective guidance for the refined control of the thrust of the jacking and articulated cylinders.

[0005] For example, the key to calculating the frictional resistance of pipe jacking in publication CN105512411A lies in the judgment of the "pipe-soil" contact state, which relies on the mathematical derivation and calculation of indirect physical quantities such as soil deformation and pipe section displacement. It also assumes that the friction coefficient remains constant under both pipe-soil and pipe-mud contact conditions, failing to reflect the changing characteristics of the frictional resistance under the influence of geological conditions, grouting pressure, and thickness. Therefore, the calculated distribution of the frictional resistance is inaccurate. Furthermore, publication CN105512411A uses sine or cosine functions to assume the offset of the pipe jacking machine during the jacking process, which does not reflect the uncertainty caused by the influence of geological conditions and construction parameters on the attitude deviation of the pipe jacking machine during actual construction. In summary, existing pipe jacking resistance calculation methods based on ideal assumptions cannot consider the actual situation of geological changes, grouting pressure and thickness changes, and pipe alignment deviations during dynamic pipe jacking construction, nor can they provide scientific guidance for the real-time and precise control of the jacking / articulation cylinders during pipe jacking construction. Summary of the Invention

[0006] This invention provides a dynamic calculation method for pipe jacking resistance based on real-time mud sensing data, aiming to overcome the problems of inaccurate jacking force calculation results caused by existing jacking force calculation methods based on ideal assumptions, which rely on indirect parameters and static theoretical calculation models, and cannot provide real-time guidance for jacking and articulated cylinder thrust control.

[0007] This invention provides a dynamic calculation method for pipe jacking resistance based on real-time mud sensing data, characterized by the following steps:

[0008] S1. Construct a database of frictional force at the pipe-grout-soil interface: Collect undisturbed soil samples around the pipe based on geological survey information; determine the unit area frictional force between the steel plate and concrete pipe section of the pipe jacking machine head and different strata under different grout pressures and thicknesses through pipe-grout-soil interface tests, and form a database of unit area frictional force at the pipe-grout-soil interface during pipe jacking construction.

[0009] S2. Calculate the jacking resistance of real-time mud detection: Based on the geometric arrangement parameters of the mud detection device around the pipe, divide the head of the pipe jacking machine and the surface around the pipe section into several units, and call the friction force per unit area of ​​the pipe-slurry-soil interface under the corresponding conditions in the database. By summing the product of all units and the friction force, the distribution state of friction resistance around the pipe jacking is obtained.

[0010] S3. Optimize the data-driven friction resistance database: Calculate the jacking resistance based on the friction resistance around the pipe, compare and analyze the jacking resistance with the jacking thrust data of the jacking cylinder through machine learning, and correct the unit friction force database of the pipe-grout-soil interface in real time in the test section.

[0011] S4. Cylinder thrust calculated in real time based on jacking resistance: Based on the obtained frictional resistance around the jacking pipe and the monitored soil pressure at the working face, the equilibrium equation of the force system of the machine head-pipe section in the jacking direction is established, and the cylinder thrust distribution is obtained based on the jacking parameters.

[0012] As a further improvement of the present invention, S1 specifically includes:

[0013] S11. Collecting soil samples: Based on the geological survey information of the pipe jacking project, collect undisturbed soil samples around the pipe that can truly reflect the mechanical properties of the strata;

[0014] S12. Obtain pipe jacking parameters: Obtain the interface roughness parameters of the pipe jacking machine head and the four sides of the pipe section that are in contact with the formation;

[0015] S13. Constructing an interface test and database: Prepare interface friction specimens with the same roughness parameters as the pipe jacking machine head and pipe section interface, conduct pipe-grout-soil interface friction tests, and measure the unit area friction force between the machine head steel plate, concrete pipe section and various strata under different grout pressure and thickness conditions. Based on the test results, construct a database of the unit area friction force of the pipe-grout-soil interface at the initial stage of pipe jacking construction.

[0016] As a further improvement of the present invention, the frictional force per unit area is f(G,S,P,T), where G represents the frictional interface type between the pipe jacking and the soil, S represents the soil type, P represents the drag-reducing mud pressure, and T represents the drag-reducing mud thickness.

[0017] As a further improvement of the present invention, S2 specifically includes:

[0018] S21. Division of typical stratigraphic sections: According to the distribution characteristics of the strata around the pipe jacking tunnel, the pipe jacking tunnel is divided into different typical stratigraphic sections along the jacking mileage direction, and the coordinate area values ​​of different typical stratigraphic sections are recorded;

[0019] S22. Identification of contact areas between the jacking machine and soil, and between the pipe and soil: The outer surface of the jacking machine head and each pipe section is divided into several calculation units. Each unit on the outer surface of the pipe section corresponds to the detection range of a mud detection device and has a defined surface area A. i ; Pipe jacking at a certain mileage Z i Based on the stratigraphic distribution obtained from the survey, the surface area A of different surface units of the machine head and different pipe sections was obtained. i The geological formation type S that the jacking pipe is in contact with. i Contact type G i ;

[0020] S23. Calculation of frictional resistance around the drill head and pipe sections based on real-time sensing of drag-reducing mud around the pipe: Using a mud detection device, the drag-reducing mud pressure P at the center of each unit is obtained in real time. i and thickness data Ti For the surface area A of each unit i According to its G i S i P i T i The parameter calls the corresponding unit area frictional resistance f from the initial database. i Calculate the frictional resistance of this unit:

[0021] ,

[0022] Among them, f (i) This represents the frictional force f per unit area corresponding to element i;

[0023] By obtaining the frictional resistance of each unit, the distribution of frictional resistance around the pipe jacking machine head and pipe section is obtained.

[0024] As a further improvement of the present invention, in S21, different typical stratigraphic sections include an initiation reinforcement zone, a receiving reinforcement zone, stratigraphic type 1, stratigraphic type 2, and stratigraphic type 3, and the coordinate region value of different typical stratigraphic sections is recorded as s. h ∈(x h ,y h ,z h ), where the origin is the center of the starting portal, x is the width direction, y is the height direction, z is the jacking mileage direction, and h represents one of the typical stratigraphic sections.

[0025] As a further improvement of the present invention, the contact type G i This includes contact type G1 between the machine head interface and the soil, and contact type G2 between the pipe section interface and the soil.

[0026] As a further improvement of the present invention, S3 specifically includes:

[0027] S31. Data preparation: Construct a friction resistance database optimization module to continuously collect and store multi-dimensional data during the jacking process to form a training data pool;

[0028] S32. Model Training and Parameter Correction: A machine learning regression model is used to establish a quantitative relationship between the theoretical jacking resistance calculated based on the initial friction resistance database and the actual measured jacking thrust. The quantitative relationship is then used to correct the parameters of the initial friction resistance database until the calculated jacking resistance value is basically consistent with the actual value. At this point, the database optimization machine learning model training is mature.

[0029] As a further improvement of the present invention, in S31, the multidimensional data during the jacking process includes mud thickness and pressure, spatial position of the jacking head and pipe sections, formation condition, initial frictional resistance database, and actual jacking force.

[0030] As a further improvement of the present invention, S32 specifically includes:

[0031] Input and output definitions: The theoretical resistance components calculated from the initial frictional resistance database are used as the model input features, and the actual total thrust collected synchronously is used as the model training target;

[0032] Model fitting discrepancies: Regression models are trained on a large amount of historical or real-time data to capture the systematic deviation patterns between theoretical calculations and actual measurements;

[0033] Parameter back-calibration: After the model training is mature, the actual jacking force corresponding to the given theoretical resistance calculation value is predicted, and the underlying parameters in the initial friction resistance database are adjusted in reverse so that the theoretical resistance value calculated based on the new parameters is consistent with the actual jacking thrust after model mapping.

[0034] Closed-loop iterative optimization: New data is continuously input, the model is constantly retrained to drive the database parameters and continuously update them iteratively.

[0035] As a further improvement of the present invention, S4 specifically includes:

[0036] S41. Based on the well-trained friction resistance database model and jacking resistance calculation method, the friction resistance of multiple units of the jacking pipe is accumulated along the jacking direction to obtain the friction resistance distribution value F between the jacking head and the pipe circumference. f ;

[0037] S42. The pressure distribution F at the tunnel face is obtained based on the pressure monitoring of the tunnel head soil chamber. c ;

[0038] S43. Establish the force equilibrium equations for pipe jacking:

[0039] ,

[0040] ,

[0041] Where F D M is the thrust of the hydraulic cylinder. J M c These are the hydraulic cylinder thrust and the earth pressure at the working face, respectively; M f This represents the torque generated by the frictional resistance of the die head on the die head, or the torque generated by the frictional resistance between the die head and the pipe circumference on the die head; i represents the i-th unit, j represents the number of push cylinders; F f This indicates the frictional resistance between the turbine head and the formation, or the frictional resistance between the turbine head and the formation plus the pipe section.

[0042] The beneficial effects of this invention are as follows: This method directly determines the contact state of the "pipe-slurry-soil" by real-time detection of the thickness and distribution of the drag-reducing mud, and calls upon a dynamically optimized interface friction database to achieve real-time, accurate, and unitized calculation of jacking resistance; furthermore, by utilizing machine learning technology, the database parameters can self-learn and adaptively correct based on on-site measured data, ensuring the long-term accuracy of the calculation model; finally, based on the real-time calculated and predicted resistance distribution, an intelligent control system is constructed to achieve adaptive matching and control of the thrust of the jacking cylinder and the articulated cylinder, thereby providing scientific guidance from real-time perception to intelligent decision-making for the refined control of the thrust of the pipe jacking construction cylinder. Attached Figure Description

[0043] Figure 1 This is a flowchart of the dynamic calculation method for pipe jacking resistance based on real-time mud sensing data of the present invention.

[0044] Figure 2 This is a diagram of the intelligent control system architecture for the jacking cylinder based on real-time detection of jacking resistance, as described in this invention.

[0045] Figure 3 This is a plan view of the jacking pipe of the present invention;

[0046] Figure 4 This is a geological longitudinal section view of the pipe jacking process of this invention;

[0047] Figure 5 This is a schematic diagram of the cross-section of the pipe jacking and the arrangement of the mud detection equipment of the present invention;

[0048] Figure 6 This is a diagram showing the area distribution of the mud detection equipment unit around the pipe jacking machine head and pipe section of this invention;

[0049] Figure 7 This is a force analysis diagram of the pipe jacking process of the present invention;

[0050] Figure 8 This is a diagram showing the distribution of the articulated hydraulic cylinders in the pipe jacking machine of this invention;

[0051] Figure 9 This is a distribution diagram of the jacking cylinders acting on the pipe section in the starting well of the pipe jacking project of this invention. Detailed Implementation

[0052] 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.

[0053] The present invention provides a dynamic calculation method for pipe jacking resistance based on real-time mud sensing data, comprising the following steps:

[0054] S1. Constructing a database of frictional forces at the pipe-grout-soil interface: This database is based on on-site sampling of undisturbed strata, parameters of drag-reducing slurry pressure and thickness, and the roughness characteristics of the pipe jacking perimeter. Specifically, this includes collecting undisturbed soil samples from the surrounding strata based on geological survey information; and determining the unit area frictional force between the pipe jacking machine's head steel plate and concrete pipe sections and different strata under different slurry pressures and thicknesses through pipe-grout-soil interface tests, thus forming a database of unit area frictional forces at the pipe-grout-soil interface during pipe jacking construction.

[0055] S2. Calculation of jacking resistance based on real-time mud detection: The dynamic calculation method for jacking resistance based on real-time mud detection specifically includes dividing the surface of the jacking machine head and pipe sections into several units according to the geometric arrangement parameters of the mud detection device around the pipe, and calling the friction force per unit area of ​​the pipe-slurry-soil interface under the corresponding conditions in the database. By summing the products of all units and the friction force, the distribution state of friction resistance around the jacking pipe is obtained. This calculation method dynamically adjusts the jacking force distribution of the hinged hydraulic cylinder between the front and rear shields and the jacking hydraulic cylinder in the starting well through the distribution state of friction resistance around the jacking pipe, so as to realize the control of the jacking attitude and the optimized distribution of jacking force.

[0056] S3. Optimize the data-driven friction resistance database: A dynamic optimization method based on the data-driven friction force database per unit area around the pipe jacking pipe is used. Specifically, this includes calculating the jacking resistance based on the friction resistance around the pipe jacking pipe, comparing and analyzing the jacking resistance with the jacking thrust data of the jacking cylinder through machine learning, and correcting the unit friction force database of the pipe-grout-soil interface in real time in the test section. Based on the high-fidelity data measured on site, the database parameters are calibrated and continuously optimized in real time.

[0057] S4. Cylinder thrust based on real-time calculation of jacking resistance: Adaptive control of jacking cylinder and articulated cylinder, specifically including establishing the force balance equation of the machine head-pipe joint force system in the jacking direction based on the obtained frictional resistance around the jacking pipe and the monitored soil pressure at the working face, and obtaining the cylinder thrust distribution based on the jacking parameters, including the thrust distribution of the jacking cylinder or the articulated cylinder of the machine head, to achieve adaptive dynamic and precise control of the jacking construction parameters.

[0058] The method for constructing the pipe-grout-soil interface friction database in step S1: This method aims to provide accurate initial friction parameters for the system, and is characterized by including the following steps:

[0059] S11. Collect soil samples from the strata: such as Figure 4 Based on the geological survey information of the pipe jacking project, undisturbed soil samples around the pipe are collected to truly reflect the mechanical properties of the strata.

[0060] S12. Obtain pipe jacking parameters: Use a roughness measuring instrument to obtain the interface roughness parameters of the pipe jacking machine head and the four sides of the pipe section that are in contact with the stratum.

[0061] S13. Establishing an Interface Test and Database: Interfacial friction specimens with the same roughness parameters as the pipe jacking machine head and pipe sections were fabricated. Pipe-grout-soil interface friction tests were conducted to determine the unit area friction force between the machine head steel plate, concrete pipe sections, and various soil strata under different slurry pressures and thicknesses, denoted as f(G,S,P,T), where G represents the type of the pipe-soil friction interface, such as the machine head steel plate or concrete pipe section; S represents the soil type, such as sand, clay, silt, etc.; P represents the drag-reducing slurry pressure; and T represents the drag-reducing slurry thickness. Based on the test results, a database of unit area friction force at the pipe-grout-soil interface at the initial stage of pipe jacking construction was established.

[0062] Step S2: Dynamic Calculation Method for Jacking Resistance Based on Real-time Mud Detection: The core of this method lies in dynamically determining the pipe-slurry-soil contact state using real-time mud detection data, and calculating the jacking resistance based on the initial pipe-slurry-soil interface friction database. This includes the following steps:

[0063] S21. Typical Stratigraphic Section Division: Based on the engineering geological survey report, the tunnel is divided into different typical stratigraphic sections along the jacking mileage direction according to the strata distribution characteristics around the tunnel, such as the starting reinforcement zone, receiving reinforcement zone, stratigraphic type 1, stratigraphic type 2, etc. The coordinate values ​​of different typical stratigraphic sections are recorded. h ∈(x h ,y h ,z h ), where the origin is the center of the starting portal, x is the width direction, y is the height direction, z is the jacking mileage direction, and h represents one of the typical stratigraphic sections.

[0064] Specifically, such as s h ∈(x1,x2;y1,y2;z1,z2), where x h ∈[x1,x2], where x1 and x2 represent two coordinate points located to the left and right of the originating portal center, respectively, and the distance between the two coordinate points is the width of the jacking pipe or the stratum; y h ∈[y1,y2], where y1 and y2 represent two coordinate points located above and below the origin of the tunnel portal, and the distance between the two coordinate points is the height of the jacking pipe or the stratum; z h ∈[z1,z2], where y1 and y2 represent the forward and backward coordinates of the jacking direction. For the geological formation, they are used to determine the geological formation of the contact section along the jacking path; for the pipe jacking, they are used to determine the jacking distance. Through x... h ,y h ,z h These three intervals can determine S h The location and strata of the region.

[0065] S22. Identification of contact areas between the jacking machine and soil, and between pipe and soil: (e.g., ...) Figure 6 The outer surface of the machine head and each pipe section is divided into several calculation units. Each unit on the outer surface of the pipe section corresponds to the detection range of a mud detection device and has a defined surface area A. i ;like Figure 4 Pipe jacking at a certain mileage Z i Based on the stratigraphic distribution obtained from the survey, the surface area A of different surface units of the machine head and different pipe sections was obtained. i The geological formation type S that the jacking pipe is in contact with. i Contact type G i Contact type G i They are divided into two categories: G1, the contact type between the machine head interface and the soil, and G2, the contact type between the pipe section interface and the soil.

[0066] S23. Calculation of frictional resistance around the drill head and pipe sections based on real-time sensing of drag-reducing mud around the pipe: Using a mud detection device, the drag-reducing mud pressure P at the center of each unit is obtained in real time. i and thickness data T i For the surface area A of each unit i According to its G i S i P i T i The parameter calls the corresponding unit area frictional resistance f from the initial database. i Calculate the frictional resistance of this unit:

[0067] ,

[0068] Among them, f (i) f(G,S,P,T) represents the frictional force per unit area corresponding to element i.

[0069] By obtaining the frictional resistance of each unit, the distribution of frictional resistance around the pipe jacking machine head and pipe section is obtained.

[0070] Step S3, a data-driven dynamic optimization method based on a friction resistance database, aims to overcome the discrepancy between laboratory data and actual field conditions by establishing a self-learning parameter calibration system, including:

[0071] S31. Data Preparation: Construct a friction resistance database optimization module to continuously collect and store multi-dimensional data during the jacking process, forming a training data pool; the multi-dimensional data during the jacking process includes mud thickness and pressure, spatial position of the jacking head and pipe sections, formation condition, initial friction resistance database, actual jacking force, etc.

[0072] S32. Model Training and Parameter Correction: A machine learning regression model is used to analyze the mapping relationship between the jacking resistance obtained from the initial friction resistance database and the actual jacking thrust. With the accumulation of long-term jacking data, the initial friction resistance database is continuously corrected until the database optimization machine learning model training matures, that is, the calculated jacking resistance value is basically consistent with the actual value, realizing the real-time calibration and parameter correction function of the friction resistance database.

[0073] The machine learning regression model is either an ensemble learning model or a neural network model; preferably, a random forest regression model is used. Since the amount of historical data available for training is limited in the early stages of pipe jacking construction, random forests are robust under small sample conditions and have high training efficiency, meeting the needs of real-time optimization on-site. More importantly, the random forest model can output feature importance indices, which can intuitively reflect the influence of factors such as mud pressure P, thickness T, and formation type S on the unit frictional resistance f. This provides a clear direction for correcting the physical meaning of database parameters, making the "reverse correction" process not only data-driven but also engineering interpretable.

[0074] The specific process of model training and parameter correction involves establishing a quantitative relationship between the theoretical jacking resistance calculated from the initial friction resistance database and the actual measured jacking thrust through a machine learning regression model, and then using this relationship to correct the parameters of the initial friction resistance database. The core process is as follows: Input and Output Definition: The theoretical resistance components calculated from the initial friction resistance database (such as pipe-soil friction, face resistance, etc.) are used as the model input features, and the synchronously collected actual total jacking thrust is used as the model training target (output true value). Model Fit Difference: A regression model (such as a neural network) is used to train on a large amount of historical or real-time data. Through learning, the model captures the systematic deviation between theoretical calculations and actual measurements; this deviation reflects the difference between the initial database parameters and the actual working conditions. Parameter Back-Calibration: After the model training matures, it can accurately predict the corresponding actual jacking thrust for a given theoretical resistance calculation value. The system uses an optimization algorithm to back-adjust the underlying parameters (such as friction coefficient, earth pressure coefficient, etc.) in the initial friction resistance database so that the theoretical resistance value calculated based on the new parameters, after model mapping, is highly consistent with the actual jacking thrust. Closed-loop iterative optimization: With the continuous input of new data, the model is constantly retrained, driving the database parameters to be continuously updated iteratively, so that the theoretical calculations can directly and accurately reflect the actual working conditions, forming a self-calibrating closed loop.

[0075] This mapping relationship enables machine learning models to automatically discover the correction relationship between theoretical and measured values ​​through data-driven modeling, and to use this relationship to automatically invert and correct key parameters in the theoretical model, thereby achieving self-evolution and precise calibration of the database.

[0076] Step S4: Intelligent Decision-Making Method for Cylinder Thrust Based on Real-Time Calculation of Jacking Resistance: This system constructs an intelligent decision-making mechanism for cylinder thrust based on the aforementioned method, including the following steps:

[0077] S41. Based on the well-trained friction resistance database model and jacking resistance calculation method, the friction resistance of multiple units of the jacking pipe is accumulated along the jacking direction to obtain the friction resistance distribution value F between the jacking head and the pipe circumference. f Specifically, such as Figure 5 The left side of the nose shield has four units. By summing the frictional resistance of these four units, the distribution value of the nose frictional resistance on the left side pipe wall of the nose shield / rear shield can be obtained. Similarly, ... Figure 5 Alternatively, the frictional resistance distribution values ​​of the nose cone can be obtained by accumulating the values ​​of multiple units on the same side, corresponding to the upper, right, and lower sides of the nose cone's front / rear shield. Accumulating the frictional resistance values ​​of multiple units in these four locations yields the overall nose cone frictional resistance distribution value for the entire nose cone / rear shield. The same accumulation method can be used to obtain the corresponding circumferential frictional resistance distribution values ​​for multiple units on the pipe section.

[0078] S42. The pressure distribution F at the tunnel face is obtained based on the pressure monitoring of the tunnel head soil chamber. c The pressure distribution at the tunnel face is obtained by multiplying the values ​​monitored by pressure gauges placed at different positions at the tunnel jacking machine head by the corresponding area and then summing them up.

[0079] S43. Establish the force equilibrium equations for pipe jacking:

[0080]

[0081]

[0082] Where F D M is the thrust of the hydraulic cylinder. J M c M f These represent the forces exerted on the jacking cylinder by the hydraulic cylinder thrust, the face earth pressure, and the torque generated by the jacking head and / or pipe friction, respectively. 'i' represents the i-th element, and 'j' represents the number of jacking cylinders. The force equilibrium equation can be used to calculate the thrust of the jacking cylinders in the launching well or the thrust of the jacking head articulated cylinders. When calculating the thrust of the jacking cylinders in the launching well, F... D M is the thrust of the jacking cylinder. J M c M f These are the thrust of the jacking cylinder, the earth pressure at the tunnel face, and the torque generated by the frictional resistance between the jacking head and the pipe circumference on the jacking head, respectively. f This represents the frictional resistance between the machine head and the formation; when the computer head articulates the hydraulic cylinder thrust, F D For the thrust of the articulated hydraulic cylinder, M J M c Mf These are the torques generated by the articulated hydraulic cylinder thrust, the soil pressure at the tunnel face, and the frictional resistance at the tunnel head on the tunnel head, respectively. f This indicates the frictional resistance between the nose cone and the ground.

[0083] By calculating the relevant bending moment using relevant forces, and obtaining the relevant parameters, the hydraulic cylinders at the construction site are adjusted to prevent deviation during pipe jacking. When the pipe's peripheral friction and soil pressure change, the hydraulic cylinder thrust distribution value can be calculated in real time using the above formula, providing scientific guidance for decisions on adjusting pipe jacking construction parameters.

[0084] like Figure 1 As shown, the execution flow of the dynamic calculation method for pipe jacking resistance based on real-time mud sensing data driven by this invention during construction is as follows: k1. Take soil samples on site and test the frictional resistance of pipe-soil and machine-soil units under different drag-reducing mud thicknesses and pressures in the laboratory; k2. Input the frictional resistance measured in the laboratory into the storage medium to form an initial frictional resistance database; k3. Install a drag-reducing mud detection device inside the pipe, and record its spatial position, mud pressure, and thickness in the storage medium; k4. Pipe jacking construction, inject drag-reducing mud, perform mud detection to obtain mud thickness and pressure, and transmit it to the storage medium; k5. Call the frictional resistance database to calculate the pipe circumferential frictional resistance; k6. Perform force balance analysis on the pipe jacking machine head and pipe sections, and calculate the thrust distribution of the jacking / articulated hydraulic cylinders; k7. Adjust the construction parameters according to the recommended values ​​of hydraulic cylinder thrust, and continuously cycle the construction of jacking-mud detection-jacking resistance calculation-jacking parameter optimization until the jacking is completed.

[0085] like Figure 2As shown, this invention implements a dynamic calculation method for pipe jacking resistance driven by real-time mud sensing data through an intelligent control system for jacking cylinders based on real-time detection of jacking resistance. The intelligent control system for jacking cylinders based on real-time detection of jacking resistance includes a friction resistance database module, a drag-reducing mud detection module, a wireless data transmission module, a data storage and analysis module, a jacking resistance calculation module, a jacking cylinder thrust decision module, and an articulated cylinder thrust decision module. The friction resistance database module stores the unit area friction force between the steel plate and concrete pipe section of the pipe jacking machine head and different strata under different mud pressure and thickness conditions, as determined by pipe-slurry-soil interface tests. The drag-reducing mud detection module is uniformly embedded around the pipe jacking machine to record parameters such as the spatial location of the mud detection module, mud pressure, and mud thickness. The data storage and analysis module analyzes, calculates, and stores the detected parameters. The wireless data transmission module wirelessly transmits the analyzed data (three-dimensional spatial coordinate information, drag-reducing mud pressure, and mud thickness) to the cloud and displays a real-time distribution map of the drag-reducing mud around the pipe during pipe jacking construction. The jacking resistance calculation module calculates the jacking resistance based on real-time mud detection. The jacking cylinder thrust decision module calculates the jacking cylinder thrust based on the real-time jacking resistance and controls the jacking cylinder thrust accordingly. The articulated cylinder thrust decision module calculates the articulated cylinder thrust based on the real-time jacking resistance and controls the articulated cylinder thrust accordingly.

[0086] The following example further illustrates the application of the present invention in pipe jacking engineering.

[0087] Example 1: Application of a dynamic calculation and control system for pipe jacking resistance based on real-time mud sensing in a rectangular pipe jacking project, combined with... Figures 3 to 9 .

[0088] 1. Project Background:

[0089] A subway parking line is constructed using the rectangular pipe jacking method. The external dimensions of each pipe section are 12m (width) × 9m (height), with a single section length of 1.5m. The designed total jacking length is 200m, and the overburden thickness is 10m. The strata traversed by the pipe jacking tunnel from top to bottom are: ① clay (4m thick), ② sand (5m thick), a cement-soil mixing pile reinforcement zone at the starting shaft end (10m long), and a reinforcement zone at the receiving shaft end (10m long).

[0090] 2. System Deployment and Initial Database Construction:

[0091] (1) Installation and integration of sensing hardware:

[0092] According to the design drawings, drag-reducing mud detection devices are evenly pre-embedded and installed on the outer walls of the jacking head and the outer walls of the jacking pipe sections. Each device corresponds to a calculation unit on the pipe wall surface, with an area of ​​1.5 × 3 = 4.5 m². 2It connects to the ground-based data receiving and processing center via a built-in wireless transmission module. Simultaneously, signals from the jacking machine's thrust cylinder pressure sensors, articulated cylinder pressure sensors, jacking guide system, and headstock soil chamber pressure sensors are all connected to this processing center, creating an integrated data acquisition network.

[0093] (2) Constructing an initial database of interfacial friction forces at the pipe-grout-soil interface:

[0094] Soil sampling and specimen preparation: Based on the geological survey report, undisturbed soil samples of silty clay and sand were collected at the engineering site. In the laboratory, corresponding friction interface specimens were prepared according to the surface treatment process of the pipe jacking machine head steel plate and the actual roughness of the outer wall of the concrete pipe section.

[0095] Interface Friction Test: A pipe-grout-soil interface friction test was conducted to measure the unit area friction force between the jacking head steel plate, concrete pipe section, and the soil under different drag-reducing grout pressures (e.g., 0.15, 0.2, 0.25 MPa) and different grout thicknesses (e.g., 0, 5, 10 mm) in the two aforementioned strata and the reinforced strata. This friction force is denoted as f(G,S,P,T), where G represents the surface properties of the jacking pipe (G1 is the jacking head steel plate, G2 is the concrete pipe section surface); S represents the stratum type (S1 is clay, S2 is sand, S3 is reinforced soil); P represents the grout pressure around the pipe (P1=0.15MPa, P2=0.2MPa, P3=0.25MPa); and T represents the grout thickness (T1=0mm, T2=5mm, T3=10mm). A typical database table obtained from the test is shown below:

[0096]

[0097] Database formation: All the above test results are entered into the system software using f(G, S, P, T) as the index to form an initial friction force database, which serves as the benchmark for subsequent calculations.

[0098] 3. Dynamic calculation and data-driven optimization during construction:

[0099] (1) Real-time data sensing and dynamic resistance calculation:

[0100] After the jacking begins, the system starts the real-time calculation module.

[0101] Step a1: Spatial and State Mapping: Based on the jacking mileage, the system automatically determines the geological section (e.g., reinforced zone, clay layer, and sand layer) where the current jacking head and the jacked sections are located. Simultaneously, each mud detection device returns a set of pressure values ​​(P) to its corresponding calculation unit every minute. i Thickness T i )data.

[0102] Step a2: Calculation of unit friction: For the calculation unit numbered i (surface area Ai = 4.5 m²) 2 The system uses its preset interface type G. i (Head or pipe section), the geological formation Si of the section, and combined with real-time sensing (P) i , T i The corresponding unit friction force fi is retrieved from the initial database. For example, in a clay layer, P is measured on the surface element of a certain pipe section. i =0.25MPa, T i =7.5 mm, the system retrieves f from the database via interpolation. i ≈ 1.025 kPa. Therefore, the instantaneous frictional force of this unit is: F i = f i × A i = 1.025 × 4.5 = 4.6125 kN.

[0103] Step a3: Total resistance integral: Fa of the system with respect to all elements (e.g., currently 1000 effective elements). i Summing these values ​​yields the current total frictional resistance F. f This is then superimposed with the frontal resistance F obtained from the earth pressure sensor. c That is, the real-time total jacking resistance is: F total = F f + F c .

[0104] (2) Data-driven dynamic optimization of database parameters:

[0105] In the initial stage of jacking (such as the first 50 meters), the system synchronously starts the optimization module.

[0106] Data preparation: The optimization module continuously collects all input features within a time window (e.g., the past 30 minutes), including: the (G) of each unit. i , S i , P i , T i , initial f i It also obtains operating data such as jacking speed and pipe section number from the hydraulic system. Simultaneously, it acquires the measured average total jacking force F within the same time window from the hydraulic system. m .

[0107] Model training and calibration:

[0108] The system is based on the formula μ 实测 = (F m - F c ) / (All Units A) iThe sum of these factors is used to calculate the macroscopic "measured average unit friction force" for that period.

[0109] The collected multidimensional feature data and μ 实测 The samples are used as training data and input into a random forest regression model for training. The goal of the model is to learn the mapping relationship from complex field conditions to real friction forces.

[0110] Once the model training is stable, it will output a calibration coefficient k for new real-time data. For example, under specific working conditions of clay, if the model finds that the initial database values ​​are systematically too small, it will output k=1.15.

[0111] The system then batch updates the f-values ​​in the database for the corresponding conditions (silty clay, specific pressure thickness range): f 优化后 = k × f 初始 For example, the original 1.025 kPa is revised to 1.179 kPa.

[0112] Results: As the jacking progresses and data accumulates, the optimization module continuously outputs calibration coefficients under different geological conditions and working conditions, gradually bringing the parameter values ​​in the database closer to the actual conditions of the current project. The calculated F... total Compared with the measured F m The error decreased from the initial ±20% and stabilized within ±5%.

[0113] 4. Intelligent decision-making and control of hydraulic cylinder thrust based on real-time calculation:

[0114] (1) Real-time thrust matching decision:

[0115] The system control module uses the latest calculated, high-fidelity friction distribution F f (Including each unit F) i Spatial location information) and frontal resistance F c Force system analysis is performed.

[0116] Top thrust equilibrium equation: F total = F f + F c ,

[0117] Torque balance equation: ,

[0118] Solve the above equations to obtain the thrust distribution of the jacking machine's pushing cylinder and articulated cylinder at different construction times in real time.

[0119] For example, calculations revealed that the total frictional force on the left side was 150 kN·m greater than that on the right side, which could have an adverse effect on attitude control. The control module then calculated the corrective torque value that needed to be generated in the head articulated cylinder system.

[0120] Command generation and issuance: Converting the corrective torque value into specific commands, such as... Figure 8 Add left-side articulated hydraulic cylinder F J1 Pressure 50 kN, while reducing the pressure of the right-side articulated cylinder F. J2 "Pressure 50 kN". This command is sent in real time to the hydraulic controller of the pipe jacking machine via the industrial bus.

[0121] 5. Implementation Results

[0122] Through the application of this system, the following was achieved in this pipe jacking project:

[0123] The jacking resistance shows that the distribution and total value of the tube circumferential friction resistance are dynamically monitored at a frequency of minutes throughout the process, and the error between the calculated result and the measured thrust is less than 5%.

[0124] Construction control automation: The adjustment of the jacking force and the correction torque basically requires no manual intervention. The system responds automatically and keeps the horizontal and vertical deviation of the machine head within ±20mm.

[0125] Parameter self-evolution: After being trained with data from approximately 100 meters of jacking, the parameters of the friction database have become highly adapted to the specific geological and mud conditions of this project, providing valuable initial data for subsequent similar projects.

[0126] Safety and efficiency improvement: Through predictive control, two potential risks of sudden increases in jacking force were successfully avoided, and the overall jacking efficiency was improved by about 15% compared with the traditional manual control method.

[0127] This invention proposes a dynamic calculation method for pipe jacking resistance based on real-time mud sensing data. Compared with existing technologies, the technical solution of this invention brings the following significant advantages:

[0128] (1) The calculation results of jacking resistance based on direct detection of drag-reducing mud are more realistic and reliable: This invention determines the contact state by directly detecting the mud thickness and calls the database for unitized calculation, replacing the traditional indirect method that relies on theoretical inversion, so that the resistance calculation results can truly reflect the actual changes in construction.

[0129] (2) Automatic and precise control of the construction process is realized: The system can automatically convert the resistance distribution data calculated in real time into thrust adjustment commands for the jacking and articulated cylinders, thereby automatically matching the jacking resistance and correcting the attitude deviation, which significantly improves the automation and precision of the control.

[0130] (3) The model parameters can be self-corrected and are highly adaptable: Through machine learning algorithms, the system can continuously calibrate and optimize the core friction database parameters using on-site measured data, so that the calculation model can automatically adapt to complex strata and construction conditions, ensuring the accuracy of long-term calculations.

[0131] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A dynamic calculation method for pipe jacking resistance based on real-time mud sensing data, characterized in that, Includes the following steps: S1. Construct a database of frictional force at the pipe-grout-soil interface: Collect undisturbed soil samples around the pipe based on geological survey information; determine the unit area frictional force between the steel plate and concrete pipe section of the pipe jacking machine head and different strata under different grout pressures and thicknesses through pipe-grout-soil interface tests, and form a database of unit area frictional force at the pipe-grout-soil interface during pipe jacking construction. S2. Calculate the jacking resistance of real-time mud detection: Based on the geometric arrangement parameters of the mud detection device around the pipe, divide the head of the pipe jacking machine and the surface around the pipe section into several units, and call the friction force per unit area of ​​the pipe-slurry-soil interface under the corresponding conditions in the database. By summing the product of all units and the friction force, the distribution state of friction resistance around the pipe jacking is obtained. S3. Optimize the data-driven friction resistance database: Calculate the jacking resistance based on the friction resistance around the pipe, compare and analyze the jacking resistance with the jacking thrust data of the jacking cylinder through machine learning, and correct the unit friction force database of the pipe-grout-soil interface in real time in the test section. S4. Cylinder thrust calculated in real time based on jacking resistance: Based on the obtained frictional resistance around the jacking pipe and the monitored soil pressure at the working face, establish the equilibrium equation of the force system of the machine head-pipe section in the jacking direction, and obtain the cylinder thrust distribution based on the jacking parameters. S2 specifically includes: S21. Division of typical stratigraphic sections: According to the distribution characteristics of the strata around the pipe jacking tunnel, the pipe jacking tunnel is divided into different typical stratigraphic sections along the jacking mileage direction, and the coordinate area values ​​of different typical stratigraphic sections are recorded; S22. Identification of contact areas between the jacking machine and soil, and between the pipe and soil: The outer surface of the jacking machine head and each pipe section is divided into several calculation units. Each unit on the outer surface of the pipe section corresponds to the detection range of a mud detection device and has a defined surface area A. i ; Pipe jacking at a certain mileage Z i Based on the stratigraphic distribution obtained from the survey, the surface area A of different surface units of the machine head and different pipe sections was obtained. i The geological formation type S that the jacking pipe is in contact with. i Contact type G i ; S23. Calculation of frictional resistance around the drill head and pipe sections based on real-time sensing of drag-reducing mud around the pipe: Using a mud detection device, the drag-reducing mud pressure P at the center of each unit is obtained in real time. i and thickness data T i For the surface area A of each unit i According to its G i S i P i T i The parameter calls the corresponding unit area frictional resistance f from the initial database. i Calculate the frictional resistance of this unit: ， Among them, f (i) This represents the frictional force f per unit area corresponding to element i; By obtaining the frictional resistance of each unit, the distribution of frictional resistance around the pipe jacking machine head and pipe section is obtained.

2. The method for dynamic calculation of pipe jacking resistance based on real-time mud sensing data as described in claim 1, characterized in that, S1 specifically includes: S11. Collecting soil samples: Based on the geological survey information of the pipe jacking project, collect undisturbed soil samples around the pipe that can truly reflect the mechanical properties of the strata; S12. Obtain pipe jacking parameters: Obtain the interface roughness parameters of the pipe jacking machine head and the four sides of the pipe section that are in contact with the formation; S13. Constructing an interface test and database: Prepare interface friction specimens with the same roughness parameters as the pipe jacking machine head and pipe section interface, conduct pipe-grout-soil interface friction tests, and measure the unit area friction force between the machine head steel plate, concrete pipe section and various strata under different grout pressure and thickness conditions. Based on the test results, construct a database of the unit area friction force of the pipe-grout-soil interface at the initial stage of pipe jacking construction.

3. The method for dynamically calculating pipe jacking resistance based on real-time mud sensing data as described in claim 2, characterized in that, The frictional force per unit area is f(G,S,P,T), where G represents the friction interface type between the pipe jacking and the soil, S represents the soil type, P represents the drag-reducing mud pressure, and T represents the drag-reducing mud thickness.

4. The method for dynamic calculation of pipe jacking resistance based on real-time mud sensing data as described in claim 1, characterized in that, In S21, different typical stratigraphic sections include the initiation reinforcement zone, the receiving reinforcement zone, stratigraphic type 1, stratigraphic type 2, and stratigraphic type 3. The coordinate region value of different typical stratigraphic sections is recorded as s. h ∈(x h ,y h ,z h ), where the origin is the center of the starting portal, x is the width direction, y is the height direction, z is the jacking mileage direction, and h represents one of the typical stratigraphic sections.

5. The method for dynamically calculating pipe jacking resistance based on real-time mud sensing data as described in claim 1, characterized in that, The contact type G i This includes contact type G1 between the machine head interface and the soil, and contact type G2 between the pipe section interface and the soil.

6. The method for dynamic calculation of pipe jacking resistance based on real-time mud sensing data as described in claim 1, characterized in that, S3 specifically includes: S31. Data preparation: Construct a friction resistance database optimization module to continuously collect and store multi-dimensional data during the jacking process to form a training data pool; S32. Model Training and Parameter Correction: A machine learning regression model is used to establish a quantitative relationship between the theoretical jacking resistance calculated based on the initial friction resistance database and the actual measured jacking thrust. The quantitative relationship is then used to correct the parameters of the initial friction resistance database until the calculated jacking resistance value is basically consistent with the actual value. At this point, the database optimization machine learning model training is mature.

7. The method for dynamic calculation of pipe jacking resistance based on real-time mud sensing data as described in claim 6, characterized in that, In S31, the multidimensional data during the jacking process includes mud thickness and pressure, spatial position of the jacking head and pipe sections, formation condition, initial frictional resistance database, and actual jacking force.

8. The method for dynamic calculation of pipe jacking resistance based on real-time mud sensing data as described in claim 6, characterized in that, Specifically, S32 includes: Input and output definitions: The theoretical resistance components calculated from the initial frictional resistance database are used as the model input features, and the actual total thrust collected synchronously is used as the model training target; Model fitting discrepancies: Regression models are trained on a large amount of historical or real-time data to capture the systematic deviation patterns between theoretical calculations and actual measurements; Parameter back-calibration: After the model training is mature, the actual jacking force corresponding to the given theoretical resistance calculation value is predicted, and the underlying parameters in the initial friction resistance database are adjusted in reverse so that the theoretical resistance value calculated based on the new parameters is consistent with the actual jacking thrust after model mapping. Closed-loop iterative optimization: New data is continuously input, the model is constantly retrained to drive the database parameters and continuously update them iteratively.

9. The method for dynamically calculating pipe jacking resistance based on real-time mud sensing data as described in claim 1, characterized in that, S4 specifically includes: S41. Based on the well-trained friction resistance database model and jacking resistance calculation method, the friction resistance of multiple units of the jacking pipe is accumulated along the jacking direction to obtain the friction resistance distribution value F between the jacking head and the pipe circumference. f ; S42. The pressure distribution F at the tunnel face is obtained based on the pressure monitoring of the tunnel head soil chamber. c ; S43. Establish the force equilibrium equations for pipe jacking: ， ， Where F D M is the thrust of the hydraulic cylinder. J M c These are the hydraulic cylinder thrust and the earth pressure at the working face, respectively; M f This represents the torque generated by the frictional resistance of the die head on the die head, or the torque generated by the frictional resistance between the die head and the pipe circumference on the die head; i represents the i-th unit, j represents the number of push cylinders; F f This indicates the frictional resistance between the turbine head and the formation, or the frictional resistance between the turbine head and the formation plus the pipe section.

Images