A supergravity model test device and method for simulating progressive instability of shield excavation face

Abstract

A supergravity model test device for simulating the progressive instability of a shield excavation face, comprising a model box, a shield tunnel model, a servo loading control system and a data acquisition system; the servo control device comprises a servo motor, a planetary roller screw electric cylinder, and a loading rod, and the data acquisition system comprises a displacement sensor, an axial force meter, a pore pressure sensor, an earth pressure sensor, and an industrial camera; the servo loading control system is connected to the excavation panel through the loading rod, and controls the excavation panel to translate forward and backward along the axial direction of the tunnel model at a low speed to achieve the simulation of the progressive instability process of the excavation face, and the data acquisition system can obtain the instability range of the excavation face, the three-dimensional deformation of the surface, and the ultimate support pressure. The present invention also includes a supergravity model test method for simulating the progressive instability of a shield excavation face. The present invention can achieve the simulation of the instability of the shield excavation face under various working conditions such as different strata and different burial depths, explore the law of the progressive instability of the shield excavation face, and provide theoretical support for the design of the ultimate support force of the shield excavation face.

Description

Technical Field

[0001] The invention belongs to the field of tunnel engineering, and in particular relates to a supergravity model test device and method for simulating progressive instability of a shield excavation face. Background Art

[0002] With the increasing intensification of urban traffic congestion and land space resource conflicts, vigorously developing urban rail transit has become an important way to solve the problems of big cities. The shield method has the advantages of fast construction speed, high degree of mechanization, and little impact on the surrounding environment. It has been widely used in urban rail transit construction. During the tunneling process of the shield machine, the support force provided by the soil compartment or pressurized mud water is used to balance the soil and water pressure in front of the excavation face and reduce the disturbance of the shield excavation to the surrounding strata. However, when the support pressure provided by the shield machine is less than or greater than the water and soil pressure in front of the excavation face, the soil in front of the excavation face will be actively or passively unstable, causing the surface to sink or uplift, and may even cause an instability accident of the excavation face, threatening the safety of the use of existing buildings (structures). Therefore, determining a reasonable limit support pressure is crucial to maintaining the stability of the shield excavation face, which can effectively guide the control of shield construction parameters and reduce the risk of instability of the shield excavation face.

[0003] Since the instability of the shield excavation face is a destructive condition of underground rock and soil, it cannot be analyzed through actual engineering observations. At present, most of the research is done by indoor model tests. Normal gravity scaled model tests cannot restore the site stress and cannot accurately reproduce the mechanical behavior of the soil in front during the shield excavation instability process. The high gravity model test places the model on a high-speed rotating geotechnical centrifuge to obtain a stress state similar to the prototype, thereby simulating the mechanical behavior of the prototype soil. It has become the mainstream test method for studying the stability of the shield excavation face.

[0004] At present, the existing centrifugal test devices for shield excavation face instability mainly include stress-controlled and displacement-controlled types. The stress-controlled device supports the shield excavation face by arranging latex bags filled with compressed air or liquid in the shield shell, gradually reducing the pressure in the latex bags to cause the shield excavation face to become unstable; the displacement-controlled device connects the rigid plate through the loading device to provide support pressure for the excavation face, and controls the rigid plate to move forward and backward to simulate the instability process of the excavation face. However, the support force provided by compressed air is evenly distributed, which does not conform to the characteristics of the change of stratum stress along the depth, resulting in the error of the test results being greatly affected by the depth of the tunnel, especially for deep buried tunnels. Although liquid with the same weight as the foundation soil can also be used to provide the excavation face with support pressure that changes along the depth, since the liquid is in a hydrostatic pressure state, it is often impossible to simultaneously balance the vertical and lateral soil pressures of the stratum. Therefore, the stress-controlled shield excavation face instability device cannot accurately obtain the law of support pressure change during the shield excavation face instability process. Compared with the stress-controlled type, the displacement-controlled device can more accurately reflect the change of support pressure, so it is more suitable for simulating the instability process of the shield excavation face. In the displacement-controlled device, foam strips are generally used to fill the gap between the shield shell and the rigid excavation panel. However, practice has shown that this sealing form is not effective and will cause leakage on the shield excavation face, affecting the accuracy of the test results. In addition, the failure of the seal of the model box will also affect the accuracy of the instability mode of the shield excavation face.

[0005] On the other hand, the current experimental research mainly focuses on the instability of the shield excavation face and the changing law of the support pressure, and lacks in-depth research on the impact of excavation face instability on surface settlement and soil stress transfer mechanism. This is because the surface settlement of the model in the existing experiment is monitored by LVDT displacement sensors or laser displacement sensors, which can only collect settlement data from limited monitoring points, making it difficult to obtain the key characteristics of surface settlement, and contact deformation measurement has a certain interference with the development of settlement. Summary of the invention

[0006] In order to solve the problems existing in the existing shield excavation face stability test research, the present invention proposes a supergravity model test device and test method for simulating the progressive instability of the shield excavation face, which uses a servo motor to drive the precise control of the excavation panel movement, optimizes the sealing method to solve the shield excavation panel sealing problem, and uses multi-eye stereo vision technology to monitor the surface settlement of the model, so as to more accurately obtain the progressive instability characteristics and impact range of the excavation face. This is specifically achieved through the following technical solutions:

[0007] A supergravity model test device for simulating progressive instability of a shield excavation face, characterized in that it comprises a model box, a shield tunnel model, a servo loading control system and a data acquisition system; the shield tunnel model is arranged inside the model box, the servo loading control system is arranged outside the model box, and the data acquisition system is distributed inside and outside the model box; the specific connection method of each system is as follows: the shield tunnel model is fixed to the inner wall of the model box by bolts, the loading control system is connected to the excavation panel by a loading rod, and the sensor of the data acquisition system is connected to other systems by embedding, connecting and fixing.

[0008] The model box includes a rigid box body, a sealing plate, and a camera installation frame; the rigid box body is connected by front and rear side panels and an intermediate load-bearing frame through high-strength bolts, the box body is lined with a stainless steel layer, and the sealing of the box body is ensured by welding. The front side panels are made of high-strength transparent organic glass (for easy visual monitoring), and the other side panels are made of aluminum alloy; a circular hole is opened on the left side of the rigid box body, and a screw hole is provided on the left outer wall. The sealing plate is fixed to the left side panel of the rigid box body by bolts, and serves as a sealing component of the loading rod to achieve the purpose of waterproofing the model box; the camera installation frame is fixed to the top of the model box by bolts, and 4 longitudinal beams are arranged in the frame for installing an industrial camera array for monitoring the surface settlement of the model.

[0009] The shield tunnel model includes a shield shell, a rigid excavation panel, and a support block; the shield shell is a semi-cylindrical shell with a connecting plate with an opening at the tail, so that it can be connected to the side plate of the model box by bolts, and the connecting plate is provided with a sealing ring; the rigid excavation panel is a semi-cylindrical shell with a connecting screw hole and a sensor groove in the middle of the soil-facing surface, the screw hole is used to connect the loading rod, and the sensor groove is used to install the soil pressure sensor and the pore pressure sensor; the rigid excavation panel has two transverse grooves at the tail and is covered with two O-rings; the top shape of the support block is consistent with the shield shell, and is arranged at the bottom of the shield shell to play a role in supporting the shield tunnel model.

[0010] The servo loading control system includes a servo motor, a screw, a planetary roller, a nut, a driven shaft and a loading rod connected in sequence; the circular motion of the screw is converted into the reciprocating linear motion of the nut through the planetary roller, the nut is connected to the driven shaft to realize the linear motion of the driven shaft, the driven shaft is connected to the loading rod through an axial force meter, and the linear motion is transmitted to the loading rod, driving the rigid excavation panel to move forward and backward; further, the servo loading control system accurately controls the low-speed movement of the excavation panel through the excavation panel displacement signal fed back by the LVDT displacement sensor, and realizes the simulation of active instability and passive instability conditions of the shield excavation face according to the forward or backward movement of the loading rod.

[0011] The data acquisition system includes an LVDT displacement sensor, an axial force meter, a pore pressure sensor, a soil pressure sensor, and an industrial camera. The LVDT displacement sensor is fixed to the lower part of the servo motor to monitor the displacement of the rigid excavation panel; the axial force meter has threads at both ends and is installed between the driven shaft and the loading rod through threaded connection to monitor the support pressure during the instability of the excavation face; the pore pressure sensor and the soil pressure sensor are installed in the sensor groove on the rigid excavation panel, and the pore pressure sensor and the soil pressure sensor are buried in the soil in front of the excavation face to monitor the pore pressure and soil pressure of the soil in front; the industrial camera array is fixed to the camera mounting frame by a special bracket to obtain a three-dimensional deformation image of the model surface; the front industrial camera is also arranged in front of the model box to shoot the instability image of the excavation face through high-strength organic glass, and the particle image velocimetry technology is used to analyze the two-dimensional deformation of the soil in front of the excavation face.

[0012] The industrial camera monitors the three-dimensional displacement of the model surface based on the principle of multi-eye stereo vision. It is characterized in that it is necessary to arrange marking points on the surface of the model soil, and install a calibration plate (with equidistant marking points on the surface of the calibration plate) and an LED light strip on the side wall of the model box, which are used for image calibration and lighting compensation respectively.

[0013] The industrial camera monitors the two-dimensional deformation of the soil in front of the excavation face, and is characterized in that equidistant reference points and control points need to be set inside the high-strength organic glass, and for textureless soils such as clay, fine sand marking points need to be arranged on the side surface of the soil.

[0014] The various sensors and industrial cameras are connected to the geotechnical centrifuge data acquisition module via data transmission cables, and are controlled and analyzed by the host in the centrifuge control room.

[0015] A hypergravity model test method for simulating progressive instability of an excavation face specifically comprises the following steps:

[0016] Step 1: Determine the tunnel depth ratio and the g value of excess gravity, lay a 30 mm thick sand cushion and geotextile at the bottom of the model box, and lead out a drainage pipeline as a drainage channel to accelerate the consolidation of the model soil;

[0017] Step 2: Install the model tunnel, sealing plate and servo loading control system, connect the axial force meter, driven shaft and loading rod, install two pore pressure sensors and two soil pressure sensors in the groove of the excavation panel and seal them with structural adhesive, and then tighten the loading rod and the excavation panel with bolts;

[0018] Step 3: In the set hypergravity environment, the servo loading control system drives the excavation panel to move forward and backward at the speed set in the test, records the axial force meter readings during the movement and takes the average value, calibrates the friction between the excavation panel and the shield shell, and the actual support pressure of the shield excavation face is equal to the difference between the axial force meter reading and the friction force;

[0019] Step 4: According to the designed tunnel burial depth ratio, fill soil or fill soil is poured into the model box to a specified height, and 9 pore pressure sensors and 6 soil pressure sensors are buried. After the soil sample is prepared, the front side plate is removed and fine sand marking points are arranged on the surface of the soil sample and on the side in contact with the plexiglass, and then the front side plate is installed;

[0020] Step 5: Paste the LED light strip on the side wall of the model box, install the calibration plate near the model soil, tighten the four industrial cameras on the camera mounting frame with bolts, adjust the lens focal length, angle and other parameters, and complete the camera calibration;

[0021] Step 6: Hang the model box into the centrifuge basket, install the industrial camera in front of the model box, and complete the camera calibration. Finally, protect and fix the various sensor signal lines on the basket and centrifuge arm, and connect them to the data acquisition module.

[0022] Step 7: Conduct a centrifugal test, increase the centrifuge acceleration to the specified g value and stabilize it for a period of time, then control the loading rod to drive the excavation panel to move at a constant speed at the set speed, so that the excavation surface gradually becomes unstable, and synchronously control the industrial camera to continuously capture unstable images and record the data of each sensor. When the excavation panel moves to the designed stroke, stop loading and reduce the centrifugal acceleration to stop;

[0023] Step 8: Use particle image velocimetry technology and multi-eye stereo vision technology to analyze the side and surface deformation images of the model foundation respectively. Combined with the changes in soil pressure and pore pressure in front of the excavation face, the progressive instability mode of the excavation face and the development characteristics of surface settlement are clarified, and the ultimate support pressure is determined based on the excavation panel displacement-support pressure relationship.

[0024] Furthermore, the specific steps of using multi-eye stereo vision technology to analyze the three-dimensional deformation of the model surface in step 8 are:

[0025] (1) Camera calibration: An array camera is installed on the top of the model box, and a calibration plate and LED light strip are installed along the side wall of the model box. The internal and external parameters of the camera are calibrated using the Levenber-Marquard method.

[0026] (2) Image processing: The captured images are preprocessed using point operations, neighborhood operations, etc. to improve the image signal-to-noise ratio; the scale-invariant feature transformation method is used to extract feature points from the landmarks in the calibration plate to determine the correspondence between the same object points in the model box in different images; then, the obtained image points are stereo matched using the belief propagation and dynamic programming algorithms to establish the correspondence between the same soil particles in multiple images.

[0027] (3) Three-dimensional reconstruction: Based on the corresponding coordinates of a point in space in multiple images and combined with the camera parameter matrix, the disparity depth coordinates of each point in the image are obtained. Then, the deep learning algorithm is used to optimize the disparity depth map of each point, thereby reconstructing the spatial coordinates of each pixel in the image and obtaining the three-dimensional displacement field of the model surface.

[0028] The advantages of the present invention are as follows:

[0029] (1) The device of the present invention optimizes the design of the shield excavation panel and the waterproofing of the model box, thereby avoiding the influence of leakage on the pore pressure change of the soil in front of the excavation face, and especially improves the accuracy of simulating the instability of the shield excavation face in soft clay strata;

[0030] (2) Compared with the hydraulic servo drive device, the present invention realizes the precise control of the low-speed uniform motion of the excavation panel by adopting the planetary roller screw servo electric cylinder, which can effectively improve the stability of the shield excavation face progressive instability supergravity model test;

[0031] (3) The present invention uses multi-eye stereo vision technology to monitor the three-dimensional displacement of the model surface, which has the advantages of high accuracy, low interference, and wide measurement range. It overcomes the shortcomings of contact measurement such as traditional LVDT displacement sensors, and realizes non-contact monitoring of the three-dimensional deformation of the surface during the instability of the shield excavation face, thereby revealing the influence mechanism of the instability of the excavation face on the surface settlement;

[0032] (4) The various parts of the device of the present invention are easy to disassemble and install, and can simulate the instability of the shield excavation face under various working conditions such as different strata and different burial depths, thereby exploring the law of progressive instability of the shield excavation face, and providing an experimental basis for the design of shield excavation face support pressure in engineering practice. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a front view of the test device in the present invention;

[0034] Figure 2 It is a side view of the test device in the present invention;

[0035] Figure 3 A top view of the test device of the present invention;

[0036] Figure 4 It is a schematic diagram of the shield tunnel model in the present invention.

[0037] Among them, 1. Rigid box; 2. Sealing plate; 3. Camera mounting frame; 4. Shield; 5. Rigid excavation panel; 5-1. O-ring; 5-2. Sensor groove; 5-3. Connecting screw hole; 6. Support block; 7. Servo motor; 8. Screw; 9. Planetary roller; 10. Nut; 11. Driven shaft; 12. Cylinder; 13. Loading rod; 14. Base; 15. LVDT displacement sensor; 16. Axis force meter; 17. Borehole pressure sensor; 18. Soil pressure sensor; 19. Industrial camera array; 20. Special bracket; 21. LED light strip; 22. Calibration plate; 23. Front industrial camera. DETAILED DESCRIPTION

[0038] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0039] Example 1

[0040] like Figure 1 , Figure 2 and Figure 3 As shown, a supergravity model test device for simulating progressive instability of shield excavation face is characterized by: comprising a model box, a shield tunnel model, a servo loading control system and a data acquisition system; the shield tunnel model is arranged inside the model box, the servo loading control system is arranged outside the model box, and the data monitoring system mainly monitors key physical quantities such as excavation panel displacement, support pressure, pore pressure and soil pressure of the soil in front of the excavation face, and side and surface deformation of the soil.

[0041] The model box consists of a rigid box 1, a sealing plate 2 and a camera mounting frame 3. The main material of the rigid box is aluminum alloy. In order to better observe the instability process of the shield excavation face, the front side plate of the model box is designed to be a high-strength transparent glass material, so that the front industrial camera 23 in front of the model box can obtain the instability image. A circular hole is opened on the left side of the rigid box 1, and the loading rod 13 passes through the circular hole and is connected to the rigid excavation panel 5 by a thread. Considering the waterproof problem of the model box, an O-ring is arranged in the sealing plate 2, which is fixed in the middle of the rectangular plate by bolts and fastened to the left side of the rigid box 1. The camera mounting frame 3 is provided with 4 longitudinal beams, and screw holes are opened on the longitudinal beams for fixing 4 industrial cameras. The base of the camera mounting frame 3 is fixed to the top of the model box by bolts.

[0042] The shield tunnel model includes a shield shell, a rigid excavation panel, and a support block; the shield shell 4 is designed in a semi-cylindrical shell shape, and its tail has a connecting plate with screw holes as a connection method with the left side plate of the model box; Figure 4As shown, the rigid excavation panel 5 is located on the inner right side of the shield shell 4, and two O-rings 5-1 are arranged in the middle to waterproof the excavation panel; the panel facing the soil is provided with connecting screw holes 5-3 and sensor grooves 5-2, and the bolts connect the excavation panel and the loading rod through the screw holes. The sensor groove 5-2 can be installed with a pore pressure sensor 17 and a soil pressure sensor 18 for monitoring the pore pressure and soil pressure of the soil in front of the excavation face; the support block 6 is arranged at the bottom of the shield shell 4 to support the shield tunnel model.

[0043] The loading control system includes a servo motor, a planetary roller screw electric cylinder, and a loading rod; the planetary roller screw electric cylinder includes a cylinder body 12 and a screw 8, a planetary roller 9, a nut 10 and a driven shaft 11 connected in sequence, the planetary roller 9 converts the circular motion of the screw into the reciprocating linear motion of the nut 10, and transmits the linear motion to the driven shaft 11 and the loading rod 13, so as to realize the forward and backward translation of the rigid excavation panel 5 along the axial direction of the tunnel; the simulation of the instability of the shield excavation face is displacement controlled, that is, the LVDT displacement sensor 15 is used to obtain the displacement signal of the rigid excavation panel 5 to accurately control the movement of the panel, and a stroke threshold is set, and the passive instability of the shield excavation face is simulated by moving the rigid excavation panel 5 forward, and the active instability of the shield excavation face is simulated by moving the rigid excavation panel 5 backward.

[0044] In order to accurately evaluate the progressive instability characteristics of the excavation face and the scope of surface settlement, the data acquisition system includes an LVDT displacement sensor 15, an axial force meter 16, a pore pressure sensor 17, an earth pressure sensor 18, an industrial camera array 19, and a front industrial camera 23.

[0045] The LVDT displacement sensor 15 is located at the lower part of the servo motor 7, and its probe is tightened on the iron sheet on the upper part of the loading rod 13 with a nut, which is used to monitor the displacement of the rigid excavation panel 5; the axial force meter 16 is set between the driven shaft 11 and the loading rod 13 through a threaded connection, which is used to monitor the support pressure of the excavation face; two pore pressure sensors 17 and two earth pressure sensors 18 are installed on the rigid excavation panel 5, and nine pore pressure sensors 17 and six earth pressure sensors 18 are buried equidistantly in the soil in front of the excavation face in the horizontal and longitudinal directions to monitor the changes in pore pressure and earth pressure of the soil in front.

[0046] Four industrial cameras form an industrial camera array 19, which is fixed on the camera mounting frame 3 by a special bracket 20 and arranged on the top of the rigid box 1. Another front industrial camera 23 is arranged in front of the rigid box 1. As a preferred method, a calibration plate 22 (with equidistant marking points) is installed on the wall of the model box and an LED light strip 21 is attached for camera calibration and lighting compensation; for non-textured soil samples such as clay, fine sand marking points need to be arranged on the surface of the model soil.

[0047] During the instability of the shield excavation face, the soil deformation process is continuously photographed by synchronously controlling the industrial camera, and the multi-eye stereo vision technology and particle image velocimetry technology are used to extract the three-dimensional displacement field of the surface soil and the two-dimensional displacement field of the soil in front of the excavation face respectively.

[0048] The various sensors and industrial cameras are connected to the geotechnical centrifuge data acquisition module via data transmission cables, and are controlled and analyzed by the host computer in the control room.

[0049] A hypergravity model test method for simulating progressive instability of an excavation face specifically comprises the following steps:

[0050] Step 1: Determine the tunnel depth ratio and the g value of excess gravity, lay a 30 mm thick sand cushion and geotextile at the bottom of the model box, and lead out a drainage pipeline as a drainage channel to accelerate the consolidation of the model soil;

[0051] Step 2: Install the model tunnel, sealing plate and servo loading control device, connect the axial force meter, driven shaft and loading rod, install two pore pressure sensors and two soil pressure sensors in the sensor groove of the excavation panel and seal them with structural adhesive, and then tighten the loading rod and the excavation panel with bolts;

[0052] Step 3: In the set hypergravity environment, the servo loading control system drives the excavation panel to move forward and backward at the speed set in the test, records the axial force meter readings during the movement and takes the average value, calibrates the friction between the excavation panel and the shield shell, and the actual support pressure of the shield excavation face is equal to the difference between the axial force meter reading and the friction force;

[0053] Step 4: According to the designed tunnel burial depth ratio, fill soil or fill soil is poured into the model box to a specified height, and 9 pore pressure sensors and 6 soil pressure sensors are buried. After the soil sample is prepared, the front side plate is removed and fine sand marking points are arranged on the surface of the soil sample and on the side in contact with the plexiglass, and then the front side plate is installed;

[0054] Step 5: Paste the LED light strip on the side wall of the model box, install the calibration plate near the model soil, tighten the four industrial cameras on the camera mounting frame with bolts, adjust the lens focal length, angle and other parameters, and complete the camera calibration;

[0055] Step 6: Hang the model box into the centrifuge basket, install the industrial camera in front of the model box, and complete the camera calibration. Finally, protect and fix the various sensor signal lines on the basket and centrifuge arm, and connect them to the data acquisition module.

[0056] Step 7: Conduct a centrifugal test, increase the centrifuge acceleration to the specified g value and stabilize it for a period of time, then control the loading rod to drive the excavation panel to move at a constant speed at the set speed, so that the excavation surface gradually becomes unstable, and synchronously control the industrial camera to continuously capture unstable images and record the data of each sensor. When the excavation panel moves to the designed stroke, stop loading and reduce the centrifugal acceleration to stop;

[0057] Step 8: Use particle image velocimetry and multi-eye stereo vision technology to analyze the side and surface deformation images of the model foundation respectively. Combined with the changes in soil pressure and pore pressure in front of the excavation face, the progressive instability mode of the excavation face and the development characteristics of surface settlement are clarified, and the ultimate support pressure is determined based on the excavation plate displacement-support pressure relationship.

[0058] In the actual experiment, the specific steps of using multi-eye stereo vision technology to analyze the three-dimensional deformation of the model surface in step 8 are as follows:

[0059] (1) Camera calibration: An array camera is installed on the top of the model box, and a calibration plate and LED light strip are installed along the side wall of the model box. The internal and external parameters of the camera are calibrated using the Levenber-Marquard method.

[0060] (2) Image processing: The captured images are preprocessed using point operations, neighborhood operations, etc. to improve the image signal-to-noise ratio; the scale-invariant feature transformation method is used to extract feature points from the landmarks in the calibration plate to determine the correspondence between the same object points in the model box in different images; then, the obtained image points are stereo matched using the belief propagation and dynamic programming algorithms to establish the correspondence between the same soil particles in multiple images.

[0061] (3) Three-dimensional reconstruction: Based on the corresponding coordinates of a point in space in multiple images and combined with the camera parameter matrix, the disparity depth coordinates of each point in the image are obtained. Then, the deep learning algorithm is used to optimize the disparity depth map of each point, thereby reconstructing the spatial coordinates of each pixel in the image and obtaining the three-dimensional displacement field of the model surface.

[0062] The contents described in the embodiments of this specification are merely an enumeration of implementation forms of the inventive concept. The protection scope of the present invention should not be regarded as limited to the specific forms described in the embodiments. The protection scope of the present invention also includes equivalent technical means that can be thought of by those skilled in the art based on the inventive concept.

Claims

1. A high gravity model test device for simulating progressive instability of shield excavation face, characterized in that: It includes a model box, a shield tunnel model, a servo loading control system and a data acquisition system; the shield tunnel model is arranged inside the model box, the servo loading control system is arranged outside the model box, and the data acquisition system is distributed inside and outside the model box; the model box includes a rigid box body, a sealing plate, and a camera mounting frame, the shield tunnel model includes a shield shell, a rigid excavation panel, and a support block, the loading control system includes a servo motor, a planetary roller screw electric cylinder, a loading rod and a base, and the data acquisition system includes an LVDT displacement sensor, an axial force meter, a pore pressure sensor, an earth pressure sensor, and an industrial camera; The shield shell adopts a semi-cylindrical shell design, and a connection plate with an opening is set at the tail as a connection method with the left side plate of the model box; the rigid excavation panel is located on the inner right side of the shield shell, and two O-rings are set at its tail to waterproof the excavation surface; there are connecting screw holes and sensor grooves in the middle of the panel facing the soil, which are used to connect the excavation panel and the loading rod, and install the soil pressure sensor and the pore pressure sensor respectively; The planetary roller screw electric cylinder includes a cylinder body and a screw, planetary rollers, nuts and driven shaft connected in sequence. The planetary rollers convert the circular motion of the screw into linear motion of the nut and driven shaft, and transmit the linear motion to the loading rod, thereby realizing the forward and backward translation of the excavation panel along the axial direction of the tunnel.

2. The high gravity model test device for simulating progressive instability of shield excavation face according to claim 1, characterized in that: The front side panel of the model box is made of high-strength transparent plexiglass, which is convenient for observing the instability process of the shield excavation face; a circular hole is opened on the left side of the model box for installing a sealing plate and passing a loading rod as a waterproof measure for the side panel of the model box.

3. The high gravity model test device for simulating progressive instability of shield excavation face according to claim 2, characterized in that: The camera mounting frame is fixed on the top of the model box through screw holes, and four longitudinal beams are arranged in the frame to fix the industrial camera.

4. The high gravity model test device for simulating progressive instability of shield excavation face according to claim 1, characterized in that: The LVDT displacement sensor is used to obtain the movement of the rigid excavation panel displacement and accurately control the movement of the control panel. The travel threshold is set and the passive and active instability of the shield excavation face are simulated by moving the excavation panel forward or backward.

5. The high gravity model test device for simulating progressive instability of shield excavation face according to claim 1, characterized in that: Four industrial cameras are installed on the camera mounting frame by a special bracket and arranged on the top of the model box. One front industrial camera is arranged in front of the model box to obtain deformation images of the soil surface and side during the instability process. A calibration plate is installed on the wall of the model box and an LED light strip is pasted for camera calibration and lighting compensation.

6. A high gravity model test method for simulating progressive instability of shield excavation face, characterized in that: The test device is the test device according to any one of claims 1 to 5, and specifically comprises the following steps: Step 1: Determine the tunnel depth ratio and the excess gravity g value, lay a 30 mm thick sand cushion and geotextile at the bottom of the model box, and lead out a drainage pipeline as a drainage channel to accelerate the consolidation of the model soil; Step 2: Install the model tunnel, sealing plate and servo loading control system, connect the axial force meter, driven shaft and loading rod, install two pore pressure sensors and two soil pressure sensors in the groove of the excavation panel and seal them with structural adhesive, and then tighten the loading rod and the excavation panel with bolts; Step 3: In the set hypergravity environment, the servo loading control system drives the excavation panel to move forward and backward at the speed set in the test, records the axial force meter readings during the movement and takes the average value, calibrates the friction between the excavation panel and the shield shell, and the actual support pressure of the shield excavation face is equal to the difference between the axial force meter reading and the friction force; Step 4: According to the designed tunnel burial depth ratio, fill soil or fill soil is poured into the model box to a specified height, and 9 pore pressure sensors and 6 soil pressure sensors are buried. After the soil sample is prepared, the front side plate is removed and fine sand marking points are arranged on the surface of the soil sample and on the side in contact with the plexiglass, and then the front side plate is installed; Step 5: Paste the LED light strip on the side wall of the model box, install the calibration plate near the model soil, tighten the four industrial cameras on the camera mounting frame with bolts, adjust the lens focal length, angle and other parameters, and complete the camera calibration; Step 6: Hang the model box into the centrifuge basket, install the industrial camera in front of the model box, and complete the camera calibration. Finally, protect and fix the various sensor signal lines on the basket and centrifuge arm, and connect them to the data acquisition module. Step 7: Conduct a centrifugal test, increase the centrifuge acceleration to the specified g value and stabilize it for a period of time, then control the loading rod to drive the excavation panel to move at a constant speed at the set speed, so that the excavation surface gradually becomes unstable, and synchronously control the industrial camera to continuously capture unstable images and record the data of each sensor. When the excavation panel moves to the designed stroke, stop loading and reduce the centrifugal acceleration to stop; Step 8: Use particle image velocimetry technology and multi-eye stereo vision technology to analyze the side and surface deformation images of the model foundation respectively. Combined with the changes in soil pressure and pore pressure in front of the excavation face, the progressive instability mode of the excavation face and the development characteristics of surface settlement are clarified, and the ultimate support pressure is determined based on the excavation panel displacement-support pressure relationship.

7. A high gravity model test method for simulating progressive instability of shield excavation face according to claim 6, characterized in that: The process of using multi-eye stereo vision technology to analyze the three-dimensional deformation of the model surface in step 8 is as follows: (1) Camera calibration: Install an array camera on the top of the model box, install a calibration plate and LED light strip along the side wall of the model box, and use the Levenber-Marquard method to calibrate the internal and external parameters of the camera; (2) Image processing: Use point operations, neighborhood operations, etc. to pre-process the captured images to improve the image signal-to-noise ratio; The scale-invariant feature transformation method is used to extract feature points from the landmarks in the calibration plate, and the corresponding relationship between the same object point in the model box in different images is determined. Then, the obtained image points are stereo matched using the belief propagation and dynamic programming algorithms to establish the corresponding relationship between the same soil particles in multiple images. (3) Three-dimensional reconstruction: Based on the corresponding coordinates of a point in space in multiple images and combined with the camera parameter matrix, the disparity depth coordinates of each point in the image are obtained. Then, the deep learning algorithm is used to optimize the disparity depth map of each point, thereby reconstructing the spatial coordinates of each pixel in the image and obtaining the three-dimensional displacement field of the model surface.

Images