Tunnel surrounding rock crustal stress simulation system based on three-dimensional loading feedback control

Abstract

The invention relates to the field of geological engineering, and discloses a tunnel surrounding rock crustal stress simulation system based on three-dimensional loading feedback control, and the system comprises a mechanical loading module which simulates a tunnel surrounding rock crustal stress field by assembling a steel plate structure, arranging a steel strand in a penetrating manner and applying a three-dimensional tensioning force through a hollow jack; the data acquisition module is used for acquiring stress and surrounding rock response parameters of the test piece in all directions in real time; the data modeling module is used for receiving the collected data and establishing a mathematical model between stress and response; the feedback control module is used for calculating an error and outputting a correction loading control instruction based on the mathematical model and the actual response value; and the visualization and interaction module is used for displaying a loading state, stress distribution and a response result, and providing an interface for a user to set loading parameters and regulate and control a loading process. According to the invention, a loading adjustment technology based on feedback control is adopted, the loading force can be adjusted in real time according to the response of the surrounding rock, the high fitting of the loading precision and the stress field in the experiment process is ensured, and the reliability and scientificity of experiment data are improved.

Description

Technical Field

[0001] This invention relates to the field of geological engineering, specifically to a tunnel surrounding rock stress simulation system based on three-dimensional loading feedback control. Background Technology

[0002] In geological engineering projects such as tunnels, mines, and underground facilities, accurate simulation and control of surrounding rock stress are crucial for ensuring construction safety and optimizing design schemes. Especially when facing complex geological environments, simulation experiments become a key means of predicting surrounding rock behavior and analyzing stress distribution. Traditional tunnel surrounding rock stress simulation systems have played a vital role in assisting engineers in stress analysis and developing reasonable construction plans.

[0003] Existing tunnel surrounding rock stress simulation systems typically rely on static loading methods for experiments. These systems simulate the stress environment by setting a fixed loading force, simplifying the loading process and providing relatively stable experimental conditions. The advantages of this method are its simplicity, low cost, and ability to provide effective stress simulation results under certain geological conditions. Furthermore, existing systems incorporate optimization techniques, such as using simple objective function optimization to control the loading force, which can improve loading accuracy to some extent and meet basic experimental requirements. These techniques can be helpful for some standardized tunnel designs, especially in conventional geological environments, where their application is relatively stable.

[0004] However, existing technologies still have many shortcomings. First, traditional static loading methods cannot be dynamically adjusted according to the actual response of the surrounding rock, resulting in poor loading accuracy and an inability to cope with stress changes in complex geological environments. Second, single-objective optimization schemes lack multi-dimensional control and cannot simultaneously consider the coordination between multiple loading parameters, thus affecting the overall accuracy and stability of the loading process. Third, existing visualization methods mostly rely on two-dimensional data charts, lacking three-dimensional technology for intuitively displaying complex data, making it difficult for operators to fully understand the experimental progress. Finally, existing loading devices have relatively simple structures when simulating complex geological stress states, failing to provide accurate three-dimensional loading, and the disassembly process is prone to generating significant mechanical impacts, leading to equipment damage and distorted experimental data. Therefore, those skilled in the art propose a tunnel surrounding rock stress simulation system based on three-dimensional loading feedback control to address these problems. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a tunnel surrounding rock stress simulation system based on three-dimensional loading feedback control, which solves problems such as insufficient loading accuracy, limited response control, unintuitive visualization, and simple loading device structure in existing technologies.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a tunnel surrounding rock stress simulation system based on three-dimensional loading feedback control, comprising:

[0007] The mechanical loading module is used to apply a three-dimensional loading force to the test piece by assembling a steel plate structure, threading steel strands, and applying tension with the help of hollow jacks, thereby simulating the stress field of the surrounding rock of the tunnel.

[0008] The data acquisition module is used to collect the stress on the test specimen in all directions and the resulting surrounding rock response parameters in real time during the loading process.

[0009] The data modeling module is used to receive stress and response data transmitted from the data acquisition module and to establish a mathematical relationship model between stress and response using a fitting method.

[0010] The feedback control module receives the mathematical model and actual response value output by the data modeling module, calculates the error between the target response and the current response, and outputs the corrected loading control command to the mechanical loading module.

[0011] The visualization and interaction module is used to receive status information from each module, realize the visualization display of loading status, stress distribution and response results, and provide an interactive interface for users to set loading parameters and control the loading process.

[0012] Preferably, the steel plate structure consists of three sets of mutually perpendicular rigid steel plates, which are used to form loading surfaces in the X, Y, and Z directions, respectively.

[0013] Each group of steel plate structures is provided with several through holes for threading steel strands;

[0014] A limiting plate is sleeved on the outside of the steel strand, and a hollow jack is sleeved on the outside of the steel strand.

[0015] The hollow jack is equipped with anchors on both the upper and lower sides, and the anchors are equipped with clamps inside. The hollow jack is driven by a hydraulic pump to apply tension, which drives the steel strand to be loaded in the corresponding direction.

[0016] The test piece is fixed inside by a steel plate structure, and a triaxial tensile force is applied to its outer surface.

[0017] Preferably, the loading process of the mechanical loading module includes:

[0018] The test piece was placed in the central space of the steel plate structure and fixed in place by three sets of mutually perpendicular rigid steel plates;

[0019] The steel strands are threaded through the through holes in the three sets of rigid steel plates of the steel plate structure, and the limiting plate, hollow jack and anchor are sequentially sleeved on the outside of them.

[0020] The steel strands are locked in place by the internal clamps of the anchorage, forming a two-way anchoring structure;

[0021] The hydraulic pump is started to drive the hollow jack to apply tension, thereby loading the test piece with the steel strand in the X, Y, and Z directions.

[0022] Preferably, the data acquisition process of the data acquisition module includes:

[0023] Strain gauges are symmetrically arranged on the surface of the test piece to collect strain signals in each loading direction;

[0024] Pressure sensors are installed on the inner side of each steel plate structure near the contact surface of the test piece to accurately measure the applied stress.

[0025] Displacement sensors are placed in the anchorage areas at both ends of the steel strand to monitor the displacement during the tensioning process.

[0026] Synchronously collect the propulsion speed V of the propulsion equipment d With rotational pressure P r Surrounding rock response parameters.

[0027] Preferably, the data modeling module establishes a linear mathematical model between stress and response, and the modeling process includes:

[0028] Triaxial stress σ x , σ y , σ z As an input variable, the propulsion speed V d With rotational pressure P r As an output variable;

[0029] The least squares method was used to perform multiple linear regression to fit the functional relationship between stress and response.

[0030] Output the linear model parameters a1 to a4 and b1 to b4;

[0031] The linear model includes:

[0032] V d =a1·σ x +a2·σ y +a3·σ z +a4;

[0033] P r =b1·σ x +b2·σ y +b3·σ z +b4;

[0034] Where: σ x , σ y , σ zV represents the stress in the loading direction, expressed in MPa. d The propulsion speed is expressed in mm / min; P r The rotational pressure is expressed in MPa; a i b i (i = 1 to 4) are the fitting coefficients.

[0035] Preferably, when the linear correlation of the linear model is insufficient, the model can adopt a multinomial regression form, establishing the following mathematical expression:

[0036]

[0037] Where: σ1=σ x , σ2=σ y , σ3=σ z c ij d ij These are the fitting coefficients, with units of output variable units / MPa. 2 c0 and d0 are constant terms.

[0038] Preferably, the correction control process of the feedback control module includes:

[0039] Receive the propulsion speed in the current loading state With rotational pressure and the set target speed Target turning pressure Compare;

[0040] Calculation error

[0041] Construct the optimization objective function J(F) x ,F y ,F z )=w1·(ΔV d ) 2 +w2·(ΔP r ) 2 ;

[0042] Among them, F x ,F y ,F z Tension force in three directions, unit: kN; w1, w2 are weighting coefficients;

[0043] Error optimization calculation is used to iteratively solve the objective function and calculate the optimal combination of loading forces.

[0044] Preferably, the error optimization calculation is performed using Newton's iteration method, and the iteration process is as follows:

[0045] F k+1 =Fk -J -1 (F k )·Δr;

[0046] Wherein: F k =[F x ,F y ,F z ] T J is the current tension; -1 The Jacobian inverse matrix of the response error with respect to the applied force; Δr=[ΔV d ,ΔP r ] T This is the error vector.

[0047] Preferably, the visualization and interaction module includes:

[0048] The display unit is used to graphically display the triaxial loading force, stress distribution status and surrounding rock response changes in real time.

[0049] The control input unit is used to set the target loading value and control the loading process;

[0050] The data export unit is used to store loading and response data during the experiment and output experiment reports;

[0051] The 3D view generation unit is used to construct a 3D deformation image of the test piece based on the strain distribution and dynamically display the contour map and isoline map during the loading process.

[0052] Preferably, after the system simulation loading experiment is completed, its disassembly process includes:

[0053] Release the hydraulic pressure in the hollow jacks in sequence to release the tension of the steel strand;

[0054] Remove the connection between the steel strands and the anchorages, and remove the steel plate structure;

[0055] Remove the test piece and perform subsequent processing and analysis based on the recorded data.

[0056] This invention provides a tunnel surrounding rock stress simulation system based on three-dimensional loading feedback control. It has the following beneficial effects:

[0057] 1. This invention adopts a loading adjustment technology based on feedback control, which can adjust the loading force in real time according to the response of the surrounding rock, ensuring the loading accuracy and high fit of the stress field during the experiment. Compared with the traditional static loading method in the prior art, this technical solution can dynamically adjust the loading, solve the problems of experimental error and stress field instability caused by inaccurate loading, and greatly improve the reliability and scientificity of experimental data.

[0058] 2. This invention introduces a multi-objective optimization algorithm, which uses the error minimization objective function to dynamically adjust the loading direction and force value. This technical solution effectively avoids the response deviation problem in the traditional single-objective control scheme, and can simultaneously optimize multiple parameters such as propulsion speed and rotation pressure, thereby improving the overall coordination and accuracy of the loading process and significantly reducing the accumulation of errors in the experiment.

[0059] 3. The three-dimensional view generation unit of the present invention enhances the intuitive display effect of experimental data by visualizing the deformation and stress distribution of the test piece. Compared with the existing technology that only relies on numerical curves, the visualization technology of the present invention not only makes the experimental data easier to understand, but also provides an effective auxiliary tool for subsequent analysis and decision-making, which helps to quickly identify weak areas in the structure.

[0060] 4. The mechanical loading module of this invention adopts a hollow jack, steel strand, and steel plate structure, which effectively simulates a complex geostress environment and can accurately apply triaxial tension, improving the accuracy and reliability of stress simulation. Compared with the simple loading device in the prior art, this structure can better control the distribution of loading force and realistically reproduce the stress state in the geological environment. During disassembly, the hollow jack can smoothly release the tension, and the steel strand and steel plate structure can be quickly disassembled, avoiding the mechanical impact and equipment damage in traditional disassembly methods, and ensuring the safety and efficiency of the disassembly process. Attached Figure Description

[0061] Figure 1 This is a three-dimensional structural diagram of the mechanical loading module of the present invention;

[0062] Figure 2 This is a schematic diagram of the anchor structure of the present invention;

[0063] Figure 3 This is a schematic diagram of the steel plate structure of the present invention;

[0064] Figure 4 This is a schematic diagram of the system architecture of the present invention;

[0065] Figure 5 This is a schematic diagram of the visualization and interaction module architecture of the present invention.

[0066] Among them, 1. steel plate structure; 2. steel strand; 201. limiting plate; 3. hollow jack; 301. anchor; 4. test piece. Detailed Implementation

[0067] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0068] Please see the appendix Figure 1 -Appendix Figure 5 This invention provides a tunnel surrounding rock stress simulation system based on three-dimensional loading feedback control, comprising:

[0069] The mechanical loading module is used to apply a three-dimensional loading force to the test piece 4 by assembling the steel plate structure 1, threading the steel strands 2 and applying tension with the help of the hollow jacks 3, so as to simulate the stress field of the surrounding rock of the tunnel.

[0070] Specifically, the mechanical loading module, as the core of the tunnel surrounding rock stress simulation system of this invention, undertakes the fundamental task of applying multi-directional spatial loading forces to the test specimen and constructing an equivalent three-dimensional stress field. This module is closely related to the subsequent data acquisition module, and its loading accuracy and stability directly affect the accuracy of the surrounding rock response data and the reliability of the fitting model. Therefore, the structural design, loading method, and configuration arrangement of the mechanical loading module in three-dimensional space must possess high stability, adjustability, and structural coordination.

[0071] Generally, geostress simulation systems need to achieve independent loading in the X, Y, and Z directions, and the tensioning process needs to be controllable, measurable, and repeatable. To address this requirement, this invention designs a three-dimensional loading device consisting of a rigid structure, prestressed loading components, and tension control elements. This device uses physical simulation to represent the multiaxial stress state experienced by the surrounding rock of a tunnel at deep underground.

[0072] In this embodiment, the mechanical loading module includes components such as a steel plate structure 1, steel strands 2, a limiting plate 201, a hollow jack 3, and an anchor 301.

[0073] Specifically, the steel plate structure 1 consists of three sets of rigid steel plates arranged perpendicularly to each other, which are used to form loading reaction surfaces in the X, Y, and Z directions, respectively, forming a spatially enclosed loading frame. Each set of steel plate structures 1 has several through holes at equal intervals for threading steel strands 2. These steel plates are made of high-strength low-alloy steel with a thickness of not less than 20mm, which has good rigidity and load-bearing capacity to prevent bending or instability during loading.

[0074] In one possible implementation, the steel strand 2 is a multi-strand steel wire stranded structure with a diameter of 15.2 mm, wrapped with an anti-corrosion protective layer, which has high prestress performance and shear resistance. The steel strand 2 enters from one side of the steel plate structure 1, passes through the through hole in the steel plate and penetrates the external space of the test piece 4, and at its other end is sequentially fitted with a limiting plate 201, a hollow jack 3 and an anchor 301.

[0075] A limiting plate 201 is positioned close to the inner side of the steel plate structure 1, adhering tightly to the steel plate to prevent the position of the steel strand 2 from shifting during tensioning. A hollow jack 3 is positioned outside the middle section of the steel strand 2; it is a unidirectional tensioning hydraulic jack, with both ends fixedly connected to the steel strand via anchors 301, forming an effective tensioning channel. The anchor 301 adopts a clamp-type structure with wedge-shaped clamps inside, which can self-lock the steel strand during tensioning.

[0076] Alternatively, the hollow jack 3 is connected to a hydraulic pump system to apply tension. During loading, the tension displacement is locked by anchors 301 at both ends, forming a stable and adjustable prestressed system. The direction of force application by the jack is consistent with the normal direction of the steel plate structure 1, ensuring no significant deviation in the loading path.

[0077] In some embodiments, to ensure the uniformity and synchronization of the three-dimensional loading forces, the loading operation follows a phased step-by-step control strategy. The initial loading in each direction is set to increase linearly with a small amplitude, and the loading amplitude is then adjusted step by step after observing the system stiffness and stress state.

[0078] During the tensioning process, to ensure the mappability between the loading amount and the target stress, the mechanical loading module introduces a stress mapping formula, which converts the stress using the equivalent relationship between the tension force of the steel strand and the loading stress. The equivalent stress expression is as follows:

[0079]

[0080] Where: σ i F represents the equivalent loading stress in the i-th direction, in MPa. i This indicates the tension force applied by the hollow jack 3 in this direction, in kN; A i This represents the area of ​​the loaded surface on the corresponding steel plate structure 1, in mm². 2 i = x, y, z, representing the X, Y, and Z directions respectively.

[0081] By regulating F i The size of the stress can be adjusted to regulate the triaxial equivalent stress applied to the outer surface of the test piece 4, thereby achieving precise control of the stress level.

[0082] Under normal circumstances, the tensioning process will generate a small additional eccentric force. In order to reduce the cumulative effect of such errors, the steel strands 2 are arranged in an axisymmetric manner, with at least four steel strands in each direction, evenly distributed along the boundary to ensure uniform stress distribution.

[0083] As an extended technical implementation, when the material of test piece 4 is a non-homogeneous material, such as rock mass with structural weak surfaces, the tensioning process can be coordinated with the fine adjustment of the position of anchor 301 to form a non-uniform stress field, further simulating the stress distribution mode of complex surrounding rock.

[0084] This embodiment achieves three-dimensional loading control of test piece 4 through the coordinated operation of the aforementioned components. This mechanical loading module provides a controllable and stable stress source for subsequent data response acquisition and model construction, which is the fundamental guarantee for the normal operation of the system.

[0085] The data acquisition module is used to collect the stress on the test specimen 4 in each direction and the surrounding rock response parameters caused by it in real time during the loading process;

[0086] Specifically, after the tensioning deployment of the mechanical loading module is completed, the system immediately enters the data response monitoring phase. The data acquisition module in this phase acts as a bridge connecting the physical loading behavior with the subsequent data modeling process. The stress state and deformation response exhibited by the test piece during loading must be fully captured by a high-precision sensing system. This module not only requires real-time performance but also high spatial recognition accuracy and multi-source synchronous processing capabilities. Therefore, the design of the data acquisition module directly determines the physical reliability and structural accuracy of the subsequent model input data, making it an indispensable component of the entire system.

[0087] In this embodiment, the data acquisition module mainly includes: strain gauges, pressure sensors, displacement sensors, and a response parameter acquisition subsystem. The sensor nodes are rationally distributed between the test piece 4 and the steel plate structure 1, forming a complete stress-response sensing network.

[0088] Specifically, strain gauges are symmetrically arranged along the X, Y, and Z directions on the outer surface of test piece 4. At least three sets of measuring points are arranged in each direction, located in the central and edge regions respectively. This point arrangement strategy can effectively capture the local deformation changes and overall deformation trends caused by loading, and is suitable for inversion calculations of spatial strain fields.

[0089] Alternatively, when the test specimen material has a layered structure, strain gauges can also be placed along the interlayer interface to monitor strain gradients and response changes in the microcrack propagation zone. The strain gauges are resistance strain gauges with a signal acquisition frequency of 500Hz or higher, connected to a high-precision static strain acquisition device.

[0090] In one possible implementation, a pressure sensor is installed on the inner side of the steel plate structure 1, i.e., in the area directly in contact with the test piece 4. The sensor is positioned corresponding to the center area of ​​the through hole through which the steel strand 2 passes, to obtain the stress transmission state at the loading point. This type of sensor is a miniature piezoelectric diaphragm type, which has high sensitivity, small size, and does not interfere with the loading path.

[0091] In addition, to achieve full-process tracking and monitoring of the steel strand loading path, displacement sensors are installed on the outer edges of both sides of the anchor 301. Magnetostrictive displacement gauges or laser displacement gauges are selected as the sensors, with a measuring range of not less than 50 mm. They are used to measure the minute elongation changes of the steel strand 2 during tensioning, indirectly reflecting the loading speed and synchronicity.

[0092] The data acquisition module also includes a response parameter acquisition channel. This channel is directly connected to the propulsion device and acquires the propulsion velocity V via a digital adapter interface. d With rotational pressure P r These parameters are among the key indicators for judging the internal stress disturbance of the test piece, and must be obtained in real time, synchronously, and non-destructively.

[0093] In some embodiments, to unify the acquisition rhythm of various signals, the system includes an edge data synchronization module. Its function is to convert analog signals from various sensors into a unified digital format, add a unified timestamp, and then package and upload them, providing a standardized data stream for subsequent modeling and feedback.

[0094] To ensure the standardization and physical consistency of the collected data, this embodiment introduces a physical verification formula as a basis for real-time verification before data transmission. For example:

[0095] Δl i =ε i ·L i ;

[0096] Where: Δl i ε represents the absolute displacement increment caused by strain in the i-th direction, in mm. i The micro-strain measured by the strain gauge is dimensionless; L i The initial distance between the sensor's measuring point and the reference point is in mm; i = x, y, z correspond to the three directions respectively.

[0097] The above verification formula enables real-time error assessment and measurement point anomaly identification, avoiding system control deviations caused by data drift.

[0098] Under normal circumstances, the refresh rate of the data acquisition module is controlled above 100Hz to ensure that continuous responses during the fast loading phase are fully recorded. The data caching system is configured with a cyclic write strategy to prevent overflow due to data accumulation during high-frequency acquisition.

[0099] As an extension, if embedded structures are permitted inside test piece 4, fiber optic strain gauges can be pre-installed in its core area to achieve internal strain profile monitoring, thereby enhancing the spatial coverage of the data dimension.

[0100] The data acquisition module in this embodiment, through the coordinated deployment of various high-sensitivity sensing elements, achieves quantitative perception of the entire process of loading path, stress transmission and surrounding rock response, providing real and complete data support for the subsequent modeling module.

[0101] The data modeling module is used to receive stress and response data transmitted from the data acquisition module and to establish a mathematical relationship model between stress and response using a fitting method.

[0102] Specifically, after completing the real-time acquisition of triaxial loading and surrounding rock response parameters for test specimen 4, the system enters the modeling stage. At this point, a large amount of raw data has been transmitted through the data acquisition module, covering physical quantities such as loading stress, strain information, tension displacement, propulsion velocity, and rotational pressure. To transform this discrete data into a mathematical structure that can be used for predictive control and error correction, this system has set up an independent data modeling module, aiming to establish a quantitative mapping relationship between stress and surrounding rock response. This module not only requires the modeling results to have sufficient fitting accuracy but also needs to meet the requirements of model differentiability, parameter stability, and adaptability. Therefore, the constructed model will serve as the core basis for the feedback control module to execute tension correction control.

[0103] In this embodiment, the data modeling module receives all raw data from the data acquisition module and first completes the data cleaning and preprocessing process. Generally, this step includes outlier removal, missing value imputation, physical quantity unit conversion, and sampling frequency standardization. After data preprocessing, the main modeling process begins.

[0104] Specifically, in this embodiment, a linear regression method is preferentially used to construct the stress-response mapping model. In this model, the input variable is the triaxial loading stress value σ. x σ y σ z The output variable is the surrounding rock response index, namely the propulsion speed V. d With rotational pressure P r .

[0105] Based on the principle of least squares, the fitted model expression is as follows:

[0106] V d =a1·σ x +a2·σ y +a3·σ z +a4;

[0107] P r =b1·σx +b2·σ y +b3·σ z +b4;

[0108] Where: σ x , σ y , σ z V represents the applied stress in the X, Y, and Z directions, respectively, with units of MPa; d P represents propulsion speed, measured in mm / min; r This indicates the slewing pressure, expressed in MPa; a i b i (i = 1 to 4) are the fitting coefficients obtained by fitting the model, and have no units.

[0109] The coefficients are solved by minimizing the sum of squared errors, and the objective function is constructed and solved by normal equations or gradient descent.

[0110] As an alternative, if the system detects that the linear model cannot accurately reflect the surrounding rock response within a certain loading area, i.e., the residual is greater than a set threshold, the system will automatically switch to a quadratic polynomial regression model to improve the fitting ability of the nonlinear response.

[0111] Specifically, the following extended model form is adopted:

[0112]

[0113] Where: σ1=σ x , σ2=σ y , σ3=σ z c ij d ij These are the second-order fitting coefficients, with units of mm / min / MPa. 2 with MPa / MPa 2 c0 and d0 are constant terms, with units consistent with the corresponding output variables.

[0114] This form has a higher degree of freedom in fitting and can simulate complex nonlinear coupling effects, making it particularly suitable for soft rock or jointed surrounding rock conditions.

[0115] In one possible implementation, the system performs cross-validation on the model after modeling to ensure its generalization ability. Validation methods include K-fold cross-validation or leave-one-out cross-validation to quantify the model's performance on unknown data.

[0116] To enhance the efficiency of the subsequent feedback control module in utilizing the model, the system will also output the fitted model in a parameterized structure and encapsulate it into a function interface that the control module can directly call.

[0117] In some embodiments, the modeling module also has a model update function, which allows for incremental model iteration based on historical data after multiple loading experiments, thereby continuously improving fitting accuracy and system adaptability.

[0118] In addition, to improve the computational efficiency of the model, the data modeling module normalizes all physical quantities before modeling, using a zero-mean, unit-variance standardization method to avoid the impact of variable scale differences on the model convergence speed.

[0119] The data modeling module in this embodiment lays the foundation for subsequent intelligent decision-making in load control by rationally constructing the mathematical mapping relationship between stress and response. The completeness and interpretability of the model parameters also ensure the traceability and transparency of the system's control behavior.

[0120] The feedback control module receives the mathematical model and actual response value output by the data modeling module, calculates the error between the target response and the current response, and outputs the corrected loading control command to the mechanical loading module.

[0121] Specifically, after establishing the stress-response model, the system enters the loading control phase based on predictive feedback. This phase is executed by the feedback control module, whose main function is to dynamically adjust the loading force during actual loading based on the real-time acquired surrounding rock response parameters and the mathematical relationship model output by the modeling module, in order to approximate the set target response value. This module realizes closed-loop control logic through error calculation, objective function construction, and loading command generation, and is a key link in ensuring loading accuracy and consistency of stress field fitting.

[0122] In this embodiment, the feedback control module mainly includes an error analysis unit, an optimization solution unit, and a loading command output unit. These units work together to achieve continuous adjustment of the tensioning process.

[0123] Specifically, the system first receives the actual response parameters under the current loading state, including the propulsion speed. With rotational pressure Simultaneously, the target propulsion speed is obtained from the user-defined or modeling module. Target turning pressure

[0124] Generally, the response error is calculated using the following formula:

[0125]

[0126]

[0127] in: These are the set target propulsion speed and target rotation pressure, respectively, in mm / min and MPa; The actual response value provided in real time by the data acquisition module; ΔV d ΔP r The values ​​represent the response deviation, in mm / min and MPa, respectively.

[0128] Subsequently, the system constructs and loads an optimization objective function, with error minimization as the control objective. In this embodiment, the optimization function is selected in the following form:

[0129] J(F x ,F y ,F z )=w1·(ΔV d ) 2 +w2·(ΔP r ) 2 ;

[0130] Wherein: F x F y F z , respectively, represent the tension forces applied in the X, Y, and Z directions, in kN; w1 and w2 are error weighting coefficients used to adjust the relative importance of propulsion speed and rotational pressure in the optimization objective, and are dimensionless; J(F x ,F y ,F z ) represents the value of the multi-objective minimization function.

[0131] The optimization unit outputs the update direction and step size of the currently loaded instruction based on the objective function.

[0132] In one possible implementation, the optimization process employs Newton's iteration method to correct the applied force vector. The specific iterative expression is as follows:

[0133] F k+1 =F k -J -1 (F k )·Δr;

[0134] Wherein: F k =[F x ,F y ,F z ] T This represents the three-dimensional tension vector in the k-th iteration, in kN; Δr = [ΔV d ,ΔP r ] T J represents the response bias vector; -1 (F k ) is the inverse of the Jacobian matrix of the error function with respect to the loading force, with a dimension of 3×2, used to describe the sensitivity of the response change to the loading change.

[0135] The calculation process uses numerical methods to quickly correct the applied force vector in the direction of the gradient of the error function. The convergence condition is set as the target error being lower than a set threshold or the maximum number of iterations reaching the upper limit.

[0136] As an alternative, the feedback control module introduces a dynamic limiting strategy before performing load correction, setting a maximum change ΔF for the load correction increment in each direction. max This prevents loading oscillations or overshoot caused by system response delays or modeling deviations.

[0137] In some embodiments, to enhance the stability of the control system, the feedback control module also embeds PID parameter adjustment logic, which outputs fine-tuning signals in a proportional adjustment manner when the error is small, and responds in an integral or derivative manner when the error changes abruptly, forming loading strategies with different response levels.

[0138] Specifically, the following adjustment formula can be set:

[0139] F i,new =F i,old +K p ·Δ i +K i ·∑Δ i +K d ·(Δ i -Δ i-1 );

[0140] Wherein: F i,new For a new round of loading force; F i,old The current applied force; Δ i K represents the current error. p K i K d These are the proportional, integral, and differential coefficients, respectively; i = x, y, z.

[0141] The above control methods can be automatically switched according to the system status, thus balancing control sensitivity and load stability.

[0142] The loading command output unit sends the correction results to the hydraulic control system of the hollow jack 3 in real time, realizing three-dimensional tension adjustment and forming a real-time closed loop.

[0143] The feedback control module in this embodiment uses error calculation and multi-objective optimization mechanisms to continuously correct the loading accuracy, providing dynamic and stable control assurance for geostress simulation experiments.

[0144] The visualization and interaction module is used to receive status information from each module, realize the visualization display of loading status, stress distribution and response results, and provide an interactive interface for users to set loading parameters and control the loading process.

[0145] Specifically, after the feedback control module completes the dynamic adjustment of the loading force and outputs the final loading command, the system enters the continuous response phase. At this point, the user needs to fully grasp the loading progress, stress distribution changes, and surrounding rock response trends, and can manually intervene or adjust parameters according to actual test requirements. To meet these functional requirements, the system is equipped with a visualization and interaction module. This module connects information display, control input, and data output, serving as the direct interface between the system's operating status and the user. Its design must balance real-time performance, completeness, and operability, and requires the ability to synchronously process multi-source information throughout the entire loading cycle.

[0146] In this embodiment, the visualization and interaction module includes a display unit, a control input unit, a data export unit, and a 3D view generation unit, and the functional units communicate with each other via a bus structure.

[0147] Specifically, the display unit is used to present the real-time operating status of each module in the system and the curves showing the changes in major physical quantities. During actual operation, users can intuitively view the time-varying trends of the X, Y, and Z axial loading forces, the surface strain contour map of test specimen 4, and surrounding rock response indicators such as the propulsion velocity V on the display interface. d Rotational pressure P r The synchronization curve.

[0148] Generally, the loading force information is presented in the form of a line graph, with the force value in kN, the vertical axis representing the loading amplitude, and the horizontal axis representing the test time; the stress isopleths are rendered using a pseudo-color image to achieve visual differentiation of stress distribution in different areas.

[0149] As an option, the system features a multi-window interface that presents the raw data and the fitted model results separately, making it easier for users to judge the stability and effectiveness of the model fit.

[0150] The control input unit is located in the operation panel area, through which the user can set the target load value, such as setting σ. x =2.0MPa, adjust the tensioning rate unit kN / min, set whether the loading direction is enabled simultaneously, or start and stop the loading process.

[0151] In one possible implementation, the control input unit works in conjunction with a touch display panel to achieve human-computer interaction through graphical controls such as sliders, selection boxes, and switches, thereby improving the efficiency of test configuration.

[0152] The data export unit has dual-channel output capabilities for both local and remote data. During the loading process, the system synchronously writes the raw collected data, model regression results, loading command records, and changes in response indicators to the data cache. After the experiment, users can export reports in .csv or .excel format, or call built-in functions to generate experimental data graphs or statistical analysis results.

[0153] In some embodiments, the system has a built-in report template engine that automatically fills in core information such as the test number, loading parameters, model coefficients, and error distribution into a preset format to form a complete loading test report document.

[0154] This embodiment further includes a 3D view generation unit, used to map the spatial strain information in the experimental data into a 3D solid model. This module supports a deformation amplification algorithm based on surface strain to construct a 3D deformation image of the deformed test piece 4.

[0155] Specifically, the system calculates the spatial deformation vector of each measuring point according to the following relationship:

[0156] Δu i =ε i ·l i ;

[0157] Where: Δu i ε is the displacement vector at the i-th measuring point, in mm; i The strain value at the measuring point is dimensionless; i This is the initial position vector between the measuring point and the origin, in mm.

[0158] Based on this, the system uses the three-dimensional geometric model before deformation as a reference, superimposes the deformation vector field, generates dynamic simulation images of deformation during the loading process, and supports outputting the complete loading and deformation process in the form of animation.

[0159] The 3D view module also supports the overlay display of stress cloud maps and contour maps, and uses color to distinguish the stress intensity levels of different areas, enabling rapid identification of weak parts of the structure.

[0160] As an extension, to improve data traceability, the system has a database interface. All loading events and response data are stored in real time, and metadata such as timestamps, loading status tags, and test numbers are automatically added to ensure that the data is manageable in the long term.

[0161] The visualization and interaction module in this embodiment provides users with a comprehensive and intuitive operation and analysis environment throughout the loading process through multi-source information synchronous display, input control, and graphic reconstruction functions, supporting the efficient and stable operation of the system.

[0162] After the system completes the simulated loading experiment and reaches the preset target, the entire experimental process enters the disassembly phase. The disassembly process includes releasing the load, removing the prestressed components, removing the test specimens, and recovering and repositioning the components, ensuring the safety and structural integrity of the entire experimental process. The detailed steps of the disassembly process are as follows.

[0163] First, the operator gradually releases the three-way loading state through the control system. Specifically, this involves shutting off the hydraulic supply path of the hollow jack 3 and slowly releasing the hydraulic pressure until the tension in the tensioning system drops to a safe threshold. A pressure-limiting and throttling device can be installed in the hydraulic circuit to ensure a smooth unloading process.

[0164] After confirming that there is no residual tension in the steel strand 2, the operator loosens the locking nut on the anchor 301, removes the clamp assembly, and disconnects the mechanical connection between the steel strand 2 and the steel plate structure 1. Then the steel strand 2 is removed and recycled.

[0165] Next, the operators removed the limiting plate 201 and its fasteners one by one and placed it into the storage unit. Subsequently, the connecting bolts on the steel plate structure 1 were removed, and the steel plates were separated. The steel plates are spliced ​​using a sliding groove nesting structure. During disassembly, the side plates were removed first, followed by the top plate, to avoid unnecessary disturbance to the test piece 4.

[0166] Finally, test specimen 4 was slowly removed and sent to the subsequent processing stage for excavation, crack analysis, or porosity analysis, in order to further study the response changes of the surrounding rock.

[0167] The entire disassembly process is monitored in real time by the control system, which records the disassembly time, stress release status, and test piece removal status at each stage to ensure the accuracy and completeness of the experimental data.

[0168] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A tunnel surrounding rock stress simulation system based on three-dimensional loading feedback control, characterized in that, include: The mechanical loading module is used to apply a three-dimensional loading force to the test piece (4) by assembling a steel plate structure (1), threading steel strands (2) and applying tension with the help of a hollow jack (3), so as to simulate the stress field of the surrounding rock of the tunnel. The data acquisition module is used to collect the stress on the test piece (4) in each direction and the surrounding rock response parameters caused by it in real time during the loading process; The data modeling module is used to receive stress and response data transmitted from the data acquisition module and to establish a mathematical relationship model between stress and response using a fitting method. The feedback control module receives the mathematical model and actual response value output by the data modeling module, calculates the error between the target response and the current response, and outputs the corrected loading control command to the mechanical loading module. The visualization and interaction module is used to receive status information from each module, realize the visualization display of loading status, stress distribution and response results, and provide an interactive interface for users to set loading parameters and control the loading process.

2. The tunnel surrounding rock stress simulation system based on three-dimensional loading feedback control according to claim 1, characterized in that, The steel plate structure (1) consists of three sets of mutually perpendicular rigid steel plates, which are used to form loading surfaces in the X, Y and Z directions respectively; Each group of steel plate structures (1) is provided with several through holes for threading steel strands (2); The steel strand (2) is fitted with a limiting plate (201), and a hollow jack (3) is fitted on the outside of the steel strand (2). The hollow jack (3) is provided with anchors (301) on both the upper and lower sides. The anchors (301) are provided with clamps inside. The hollow jack (3) is driven by a hydraulic pump to apply tension force, which drives the steel strand (2) to be loaded in the corresponding direction. The test piece (4) is fixed inside by the steel plate structure (1), and the triaxial tension is applied to its outer surface.

3. The tunnel surrounding rock stress simulation system based on three-dimensional loading feedback control according to claim 1, characterized in that, The loading process of the mechanical loading module includes: The test piece (4) is placed in the middle space of the steel plate structure (1) and fixed by three sets of mutually perpendicular rigid steel plates; The steel strands (2) are threaded through the through holes in the three sets of rigid steel plates of the steel plate structure (1), and the limiting plate (201), hollow jack (3) and anchor (301) are sequentially sleeved on the outside of them; The steel strand (2) is locked by the internal clamps of the anchor (301) to form a two-way anchoring structure; Start the hydraulic pump to drive the hollow jack (3) to apply tension, so that the steel strand (2) loads the test piece (4) in the X, Y and Z directions.

4. The tunnel surrounding rock stress simulation system based on three-dimensional loading feedback control according to claim 1, characterized in that, The data acquisition process of the data acquisition module includes: Strain gauges are symmetrically arranged on the surface of the test piece (4) to collect strain signals in each loading direction; Pressure sensors are installed on the inner side of each steel plate structure (1) near the contact surface of the test piece (4) to accurately measure the loaded stress; Displacement sensors are arranged in the anchorage (301) area at both ends of the steel strand (2) to monitor the displacement during the tensioning process of the steel strand; Synchronously collect the propulsion speed V of the propulsion equipment d With rotational pressure P r Surrounding rock response parameters.

5. The tunnel surrounding rock stress simulation system based on three-dimensional loading feedback control according to claim 1, characterized in that, The data modeling module establishes a linear mathematical model between stress and response. The modeling process includes: Triaxial stress σ x , σ y , σ z As an input variable, the propulsion speed V d With rotational pressure P r As an output variable; The least squares method was used to perform multiple linear regression to fit the functional relationship between stress and response. Output the linear model parameters a1 to a4 and b1 to b4; The linear model includes: V d =a1·σ x +a2·s y +a3·s z +a4; P r =b1·s x +b2·s y +b3·s z +b4; Where: σ x , σ y , σ z V represents the stress in the loading direction, expressed in MPa. d The propulsion speed is expressed in mm / min; P r The rotational pressure is expressed in MPa; a i b i (i = 1 to 4) are the fitting coefficients.

6. The tunnel surrounding rock stress simulation system based on three-dimensional loading feedback control according to claim 5, characterized in that, When the linear correlation of the linear model is insufficient, the model can adopt a multinomial regression form, and the following mathematical expression can be established: Where: σ1=σ x , σ2=σ y , σ3=σ z c ij d ij These are the fitting coefficients, with units of output variable units / MPa. 2 c0 and d0 are constant terms.

7. The tunnel surrounding rock stress simulation system based on three-dimensional loading feedback control according to claim 1, characterized in that, The correction control process of the feedback control module includes: Receive the propulsion speed in the current loading state With rotational pressure and the set target speed Target turning pressure Compare; Calculate the error amount Construct the optimization objective function J(F) x ,F y ,F z )=w1·(ΔV d ) 2 +w2·(ΔP r ) 2 ; Among them, F x ,F y ,F z Tension force in three directions, unit: kN; w1, w2 are weighting coefficients; ΔV d ΔP r Indicates response deviation; Error optimization calculation is used to iteratively solve the objective function and calculate the optimal combination of loading forces.

8. The tunnel surrounding rock stress simulation system based on three-dimensional loading feedback control according to claim 7, characterized in that, The error optimization calculation is performed using Newton's iteration method, and the iteration process is as follows: F k+1 =F k -J -1 (F k )·Δr; Wherein: F k =[F x ,F y ,F z ] T J is the current tension; -1 The Jacobian inverse matrix of the response error with respect to the applied force; Δr=[ΔV d ,ΔP r T is the error vector.

9. The tunnel surrounding rock stress simulation system based on three-dimensional loading feedback control according to claim 1, characterized in that, The visualization and interaction module includes: The display unit is used to graphically display the triaxial loading force, stress distribution status and surrounding rock response changes in real time. The control input unit is used to set the target loading value and control the loading process; The data export unit is used to store loading and response data during the experiment and output experiment reports; The 3D view generation unit is used to construct a 3D deformation image of the test piece based on the strain distribution and dynamically display the contour map and isoline map during the loading process.

10. The tunnel surrounding rock stress simulation system based on three-dimensional loading feedback control according to claim 1, characterized in that, After the system simulation loading experiment is completed, its disassembly process includes: Release the hydraulic pressure in the hollow jack (3) in sequence to release the tension of the steel strand (2); Remove the connection between the steel strand (2) and the anchor (301), and remove the steel plate structure (1); Remove the test piece (4) and perform subsequent processing and analysis based on the recorded data.

Images