Test method and system of indoor model for ultra-large-span cavern excavation support and medium

Abstract

The invention discloses a test method and system of an indoor model for ultra-large-span cavern excavation support and a medium, and relates to the technical field of geotechnical engineering physical model tests. Through the counter-force frame and the three-dimensional gradient active loading system, a real non-uniform three-dimensional crustal stress field is precisely restored. By means of a numerical control mechanical arm carrying a head-replaceable end effector, a finish-milling cutting head and a negative pressure suction system, extra disturbance of excavation to surrounding rock is reduced. Excavation and support triggering conditions are preset to control actuator tool switching, and excavation-support time sequence linkage is achieved. Through cooperative acquisition and analysis of a pre-embedded sensor, monitoring equipment around a model box and a multi-source data acquisition cabinet, full-dimensional data is obtained, and the finally constructed four-dimensional surrounding rock failure evolution comprehensive map completely presents the whole process from microscopic crack initiation to macroscopic instability. And a high-fidelity test basis is provided for risk pre-judgment and design optimization of a super-large-span cavern project.

Description

Technical Field

[0001] This application relates to the field of physical model testing technology in geotechnical engineering, specifically to a test method, system, and medium for an indoor model of excavation and support of an ultra-large span tunnel. Background Technology

[0002] Currently, the excavation of large underground caverns generally adopts a "layered, segmented, and phased" construction method. This method is relatively mature in the application of small and medium span caverns. By breaking down large cross-sections into smaller cross-sections for phased construction, the deformation of the surrounding rock during construction is effectively controlled, and the construction risks are manageable. However, with the continuous increase in the scale of projects in fields such as hydropower, transportation, and national defense, the number of super-large span caverns with spans of 30 meters and above is gradually increasing, and the span of underground projects may exceed 70 meters. The above-mentioned traditional construction methods have revealed obvious limitations in such projects.

[0003] Specifically, when the span of the tunnel reaches 70m, its mechanical response exhibits a qualitative difference. First, the step-by-step excavation process becomes more complex, with a significant increase in excavation steps, leading to dozens or even hundreds of severe redistributions of the surrounding rock stress state, resulting in an exceptionally prominent cumulative deformation effect. Second, before a stable load-bearing arch is formed, the stability of the roof arch and high sidewalls of the ultra-large span tunnel is highly dependent on the timeliness and effectiveness of the support system after each excavation step. The traditional "central guide tunnel + enlargement excavation" method cannot ensure that the surrounding rock can form a complete and effective "arch effect" in a timely manner under such a large span, thus making it difficult to control the settlement of the roof arch and the convergence of the sidewalls, significantly increasing the risk of structural deformation, and even potentially causing safety hazards such as local collapse or overall instability.

[0004] To anticipate the aforementioned risks and optimize excavation and support schemes before construction, physical model testing has become an indispensable research method. Traditional techniques typically involve pouring similar materials into a model test trench and applying boundary loads using jacks to simulate ground stress, followed by tunnel excavation tests. However, these traditional model testing methods have fundamental shortcomings when simulating 70m-class super-large span tunnels. First, the laboratory size limits the model scale, resulting in an excessively large geometric similarity ratio. This makes it extremely difficult to accurately simulate the complex step-by-step excavation process on a small model. Conventional "manual excavation" methods cause significant disturbance and cannot reproduce the low-disturbance mechanical construction process. Second, traditional loading techniques struggle to reproduce the complex gradient ground stress field of the engineering site within the model, distorting the initial stress state of the model. Third, the lack of micro-support system installation technology that can be linked with precise excavation steps prevents the realization of a true "one step excavation, one step support" construction sequence, resulting in the inability to accurately reflect the interaction mechanism between the support system and the surrounding rock. Therefore, traditional indoor model tests for excavation and support of ultra-large span tunnels are prone to distortion, resulting in low accuracy of test results. Summary of the Invention

[0005] The purpose of this application is to provide a test method, system and medium for an indoor model of excavation and support of ultra-large span tunnels, in order to solve the problem that the test results of traditional indoor model tests of ultra-large span tunnels are difficult to guarantee a high degree of matching with the stress state and failure law of actual engineering projects.

[0006] To achieve the above objectives, the first aspect of this application provides a test method for an indoor model of excavation and support of an ultra-large span tunnel, comprising: In the indoor model of the excavation and support of the super-large span tunnel, which is mounted in the model box within the reaction frame, a non-uniform three-dimensional initial geostress field is established using a three-dimensional gradient active loading system. The three-dimensional initial geostress field includes vertical stress and gradient-distributed horizontal stress. Import the preset three-dimensional outline of the cavern and the layered excavation sequence file. The CNC robotic arm equipped with a replaceable end effector drives the precision milling cutting head on the replaceable end effector to excavate layer by layer along the preset path. At the same time, the negative pressure suction system matched with the precision milling cutting head is started to remove the excavation debris in real time. Preset excavation and support trigger conditions. When the trigger conditions are detected during the excavation process, control the CNC robotic arm to switch the tool type of the interchangeable head end effector, and sequentially complete the micro anchor bolt implantation, prestressed anchor cable tensioning and shotcrete layer laying. After the support is completed, the excavation operation is resumed. Multi-source data on internal strain, surface displacement, and internal micro-damage of the indoor model are simultaneously collected by sensors embedded in the indoor model and monitoring equipment arranged around the model box. The multi-source data is transmitted to a multi-source data acquisition cabinet, and spatiotemporal alignment and overlay analysis are performed on the multi-source data to obtain a comprehensive map of the evolution of surrounding rock failure.

[0007] The second aspect of this application provides a test system for an indoor model of excavation and support of an ultra-large span tunnel, comprising: A module is established to create a non-uniform three-dimensional initial geostress field in an indoor model of excavation and support of an ultra-large span tunnel within a model box mounted in a reaction frame, using a three-dimensional gradient active loading system. The three-dimensional initial geostress field includes vertical stress and gradient-distributed horizontal stress. The startup module is used to import the preset three-dimensional outline of the cavern and the layered and block excavation sequence file. The CNC robotic arm equipped with a replaceable end effector drives the precision milling cutting head on the replaceable end effector to excavate layer by layer along the preset path. At the same time, the negative pressure suction system matched with the precision milling cutting head is started to remove the excavation debris in real time. The support module is used to preset the excavation support trigger conditions. When the trigger conditions are detected during the excavation process, the CNC robotic arm is controlled to switch the tool type of the interchangeable head end effector to complete the micro anchor bolt implantation, prestressed anchor cable tensioning and shotcrete layer laying in sequence. After the support is completed, the excavation operation is resumed. The analysis module is used to synchronously collect multi-source data of internal strain, surface displacement and internal micro-damage of the indoor model through sensors embedded in the indoor model and monitoring equipment arranged around the model box. The multi-source data is transmitted to the multi-source data acquisition cabinet, and the multi-source data is subjected to spatiotemporal alignment and overlay analysis to obtain a comprehensive map of the evolution of surrounding rock failure.

[0008] A third aspect of this application provides a computer-readable storage medium storing a program that can be loaded by a processor and executed to perform the above-described test method for an indoor model of excavation and support of an ultra-large span tunnel.

[0009] The beneficial effects of this application are: This application accurately recreates the real non-uniform three-dimensional geostress field of ultra-large span caverns through a reaction frame and a three-dimensional gradient active loading system, solving the problem of distortion in traditional experimental stress environments. Utilizing a CNC robotic arm equipped with a replaceable end effector, a precision milling cutting head, and a negative pressure suction system, it achieves low-disturbance layered excavation with millimeter-level precision, reducing additional disturbance to the surrounding rock. By controlling the switching of actuator tools through preset excavation and support trigger conditions, it achieves sequential linkage between excavation and support, recreating the construction logic of excavation and support as it occurs on-site. Combined with the collaborative acquisition and analysis of pre-embedded sensors, monitoring equipment around the model box, and multi-source data acquisition cabinets, it obtains comprehensive data covering internal strain, surface displacement, and internal micro-damage. The resulting four-dimensional comprehensive map of surrounding rock failure evolution fully presents the entire process from microcrack initiation to macroscopic instability, providing high-fidelity experimental evidence for risk prediction and design optimization in ultra-large span cavern projects.

[0010] Other features and advantages of this application will be described in detail in the following detailed description section. Attached Figure Description

[0011] Figure 1 This is a schematic diagram illustrating the application scenario of an indoor model test method for excavation and support of an ultra-large span tunnel provided in this application embodiment; Figure 2 This is a flowchart illustrating an indoor test method for an ultra-large span tunnel excavation and support system provided in this application embodiment. Figure 3 This is a schematic diagram of the structure of an indoor model test system for excavation and support of an ultra-large span tunnel provided in the embodiments of this application.

[0012] Explanation of reference numerals in the attached figures 1. Reaction frame; 2. Model box; 3. Hydraulic oil source; 4. Flexible pressure transmission-rigid pressure bearing composite loading plate; 5. Multi-source data acquisition cabinet; 6. Pre-embedded distributed optical fiber; 7. Acoustic probe; 8. CNC robotic arm; 9. Interchangeable end effector. Detailed Implementation

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0014] In the description of this application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "a plurality of" means two or more, unless otherwise explicitly specified. Details are set forth in the following description for illustrative purposes. It should be understood that those skilled in the art will recognize that this application can be implemented without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid unnecessarily obscuring the description of this application. Therefore, this application is not intended to be limited to the embodiments shown, but rather to be consistent with the broadest scope of the principles and features disclosed herein.

[0015] Figure 1 This is a schematic diagram illustrating the application scenario of an indoor model test method for excavation and support of an ultra-large span tunnel, as provided in the embodiments of this application. For example... Figure 1 As shown, the test method for the indoor model of excavation and support of ultra-large span tunnels is applied to the indoor model test device for ultra-large span tunnels. The indoor model test device may include a reaction frame 1, a model box 2, a hydraulic oil source 3, a flexible pressure transmission-rigid pressure bearing composite loading plate 4, a multi-source data acquisition cabinet 5, a pre-embedded distributed optical fiber 6, an acoustic probe 7, a CNC robotic arm 8, and a replaceable head end effector 9.

[0016] The reaction frame 1 is a stable support structure assembled from rigid profiles. It provides rigid reaction force support for the model box 2 and the loading system, reducing deformation of the device during loading and serving as the mechanical support skeleton of the entire experimental system. The model box 2 is a sealed box structure fixed inside the reaction frame 1, used to simulate an ultra-large span cavern model. In one example, the model box 2 can support an ultra-large span cavern model configured at a 1:100 similarity ratio (i.e., simulating a 70-meter-class prototype cavern), constraining the spatial boundaries of the model and ensuring the stability of the model forming and experimental process.

[0017] Hydraulic oil source 3 is a hydraulic power supply device that provides high-pressure oil to the hydraulic servo actuator of the three-dimensional gradient active loading system, serving as the power source for applying ground stress. The flexible pressure transmission-rigid bearing composite loading plate 4 is a loading component that combines flexible uniform pressure transmission with rigid deformation resistance. It is used to uniformly transfer the load from the hydraulic servo actuator to the model surface (flexible pressure transmission, avoiding local stress concentration), while simultaneously bearing large loads without deformation (rigid bearing), thus achieving precise application of three-dimensional gradient ground stress. Hydraulic oil source 3 is connected to the hydraulic servo actuator behind the flexible pressure transmission-rigid bearing composite loading plate 4 via hydraulic pipelines, providing power to the plate.

[0018] The multi-source data acquisition cabinet 5 is an integrated device for receiving and temporarily storing multi-channel data. It is used to centrally receive data (such as strain and micro-fracture signals) collected by pre-embedded sensors and external monitoring devices, providing data storage and transmission support for subsequent spatiotemporal alignment and fusion analysis. The pre-embedded distributed optical fiber 6 is a distributed sensing element pre-embedded inside the model, used to continuously monitor the strain distribution inside the model at a set sampling frequency (e.g., 10Hz) and capture changes in the mechanical response inside the surrounding rock. The acoustic emission probe 7 is an acoustic emission signal acquisition device, located at the top of the model box 2. It can acquire acoustic emission signals generated by micro-fractures inside the model with a threshold value of 40dB, locating the initiation and propagation locations of micro-cracks. The pre-embedded distributed optical fiber 6 and the acoustic emission probe 7 are connected to the multi-source data acquisition cabinet 5 via data cables, transmitting the collected strain and micro-fracture data to the multi-source data acquisition cabinet 5.

[0019] The CNC robotic arm 8 is a programmable multi-degree-of-freedom robotic arm used to carry a replaceable end effector 9. It performs layered excavation and synchronous support actions according to a preset path, serving as the operational unit for the test process. The replaceable end effector 9 is a switchable tool actuator installed at the end of the CNC robotic arm. It is used to switch between tools such as a milling head (excavation), an anchor installer (anchoring), and a micro-spray nozzle (spraying layer), achieving integrated excavation and support operations. The replaceable end effector 9 is installed at the end of the CNC robotic arm 8 and is driven by the CNC robotic arm 8.

[0020] The aforementioned hydraulic oil source 3, CNC robotic arm 8, and multi-source data acquisition cabinet 5 are all communicatively connected to the central controller (not shown in the figure). The central controller is used to achieve coordinated control such as ground stress loading parameter control, excavation path and tool switching control, and synchronous triggering of monitoring data.

[0021] In this embodiment, a model box 2 is used as a carrier to prepare similar materials and embed sensing elements to construct a cavern model that matches the prototype. A reaction frame 1 provides support, and a composite loading plate is driven by a hydraulic oil source 3 to apply vertical stress and gradient horizontal stress to the model, accurately reproducing the non-uniform underground stress field. A CNC robotic arm 8, equipped with a replaceable end effector 9, excavates layer by layer along a preset path using a precision milling cutting head (with simultaneous negative pressure slag removal to reduce disturbance). Tool switching enables the sequential linkage of excavation triggering → support operation → resumption of excavation, replicating the on-site construction logic of excavation and support as needed. Throughout the process, pre-embedded distributed optical fibers 6 and acoustic probes 7 collect multi-dimensional data on internal strain, surface displacement, and micro-fractures. This data is transmitted to a multi-source data acquisition cabinet for fusion analysis, ultimately obtaining the complete failure evolution law from micro-cracks to macro-instability, providing a high-fidelity experimental reference for actual ultra-large span cavern projects. This application embodiment constructs a closed-loop test system in the laboratory to faithfully reproduce the entire process of a real engineering project. By simulating real ground stress, low-disturbance excavation, excavation and support as it is excavated, and full-dimensional monitoring, the stress and failure process of ultra-large span tunnels can be reproduced.

[0022] Figure 2 This is a schematic flowchart illustrating an indoor test method for excavation and support of an ultra-large span tunnel, as provided in an embodiment of this application. Figure 2 As shown, this test method can execute steps 201-204 based on the central controller in the above application scenario, which will be described in detail below.

[0023] Step 201: In the indoor model of the excavation and support of the ultra-large span tunnel, which is mounted in the model box within the reaction frame, a non-uniform three-dimensional initial geostress field is established using a three-dimensional gradient active loading system.

[0024] The three-dimensional gradient active loading system is a loading device that integrates a reaction frame, hydraulic servo array, etc., and can accurately apply non-uniform three-dimensional geostress. The non-uniform three-dimensional initial geostress field simulates the stress state of a real underground environment, including vertical self-weight stress and horizontal tectonic stress that varies from top to bottom. It can include vertical stress and gradient-distributed horizontal stress.

[0025] In one example, similar materials can be prepared first and then poured and cured in a model box. The similar materials can be prepared by mass ratio as follows: 48% barite powder, 30% quartz sand, 12% gypsum, and 10% water.

[0026] Then, a three-dimensional gradient active loading system, consisting of a reaction frame, a hydraulic servo array, a flexible pressure-transmitting / rigid pressure-bearing composite loading plate, and a hydraulic oil source, is used. The flexible pressure-transmitting / rigid pressure-bearing composite loading plate applies vertical stress and gradient-distributed horizontal stress to the indoor model, establishing a non-uniform initial ground stress field that matches the engineering scenario. For example, the geometric similarity ratio of the indoor model can be 1:100, cast in a model box with dimensions of 2.9m × 1m × 1.8m (length × width × height). The geometric similarity ratio is the proportional relationship between the dimensions of the indoor model and the engineering prototype, ensuring that the spatial form of the model is consistent with the prototype.

[0027] During the casting process inside the model box, distributed fiber optic sensors and acoustic probes can be pre-embedded simultaneously. During casting, similar materials are filled into the model box in three layers and compacted and leveled to ensure uniform density of the indoor model. Specifically, in a 2.9m × 1m × 1.8m model box, casting can be done in layers, compacted and leveled, with distributed fiber optic sensors pre-embedded simultaneously.

[0028] By using similar materials in specific proportions, the mechanical properties of the model are ensured to be proportionally similar to those of the prototype surrounding rock, reducing discrepancies between experimental results and actual engineering conditions. Layered pouring and compaction leveling ensure uniform model density, eliminating experimental errors caused by localized loosening. Traditional tests can only monitor the surface and cannot obtain internal data; however, by synchronously embedding sensors, continuous monitoring of the model's internal state can be achieved.

[0029] After the casting process is completed, the model box can be cured for no less than 7 days under conditions of ambient temperature 25±2℃ and relative humidity greater than 90%, until the uniaxial compressive strength of the similar material reaches the set value of 0.3-0.5 MPa. Uniaxial compressive strength is the ultimate strength of a material under uniaxial compression, and it is a core indicator for evaluating whether the mechanical properties of the similar material match those of the prototype surrounding rock. Stable temperature and humidity conditions and sufficient curing time allow the mechanical properties of the similar material to gradually stabilize, reducing fluctuations in material strength over time during the test. Controlling the uniaxial compressive strength at 0.3-0.5 MPa precisely matches the strength characteristics of the prototype surrounding rock at a 1:100 scale, ensuring that the test stress response is consistent with the actual engineering.

[0030] Next, vertical stress can be applied slowly in multiple stages using a hydraulic servo actuator array positioned on top of the model box. This can be done in 5-10 stages. The hydraulic servo actuator array is a loading device composed of multiple high-precision hydraulic power units, and the load size and rate can be precisely adjusted by the control system. Vertical stress simulates the vertical pressure generated by the self-weight of rock mass in an underground environment and is one of the core initial stresses of ultra-large span caverns. The application rate is controlled within a first set rate, and each load level is held for a first set time until the target value is reached. The first set rate and the first set time are both set thresholds that match the actual scenario of vertical stress application. For example, the first set rate can be 0.05 MPa / min, and the first set time can be 10 minutes. The target value is the set value that the vertical stress needs to reach, such as 0.5-1.0 MPa. By using multi-stage slow loading and holding each stage, the possibility of localized damage to the model due to instantaneous load application can be reduced, ensuring that the vertical stress is uniformly transmitted to the interior of the model. Precise control of the application rate and target value restores the magnitude and loading process of the actual underground self-weight stress, providing a realistic initial vertical stress environment for subsequent excavation and support tests.

[0031] Furthermore, horizontal stress can be applied in stages using hydraulic servo actuator arrays arranged on both sides of the model box. By independently controlling the pressure of actuators at different positions, the ratio of vertical stress to horizontal stress (σh / σv) gradually changes from a first ratio at the top of the model to a second ratio at the bottom, with the first ratio being less than the second ratio (e.g., the first ratio could be 0.5, and the second ratio could be 0.8). This non-uniform state of the ratio of horizontal stress to vertical stress varying with the spatial position of the model (from top to bottom) is a typical characteristic of stress in real underground structures. Each load level is held for a second set time, which is a threshold value matching the real scenario of applying horizontal stress, for example, 10 minutes. After the deformation of the indoor model stabilizes, the next level of horizontal stress is applied. The horizontal stress can be transferred to the indoor model through flexible pressure-transmitting and rigid pressure-bearing composite loading plates set on both sides of the model box.

[0032] By applying vertical and horizontal stresses, the limitations of traditional uniform horizontal loading in tests can be overcome. By simulating the stepwise stress distribution in real underground structures through stress ratio gradients, this method solves the problem that traditional tests can only apply uniform or uniaxial stress. It accurately recreates the real stress environment of the arch foot and sidewalls of ultra-large span tunnels, providing an initial stress foundation that closely matches engineering realities for subsequent excavation and support tests. Furthermore, independent control of actuator pressure and stable load-bearing operation ensures the precise and stable application of horizontal stress gradients, making the model's stress state highly consistent with the engineering site.

[0033] Step 202: Import the preset three-dimensional contour of the cavern and the file of the excavation sequence for layered and block-by-block excavation. The numerically controlled robotic arm equipped with a replaceable end effector drives the precision milling cutter head on the replaceable end effector to excavate layer by layer along the preset path, and simultaneously starts the negative pressure suction system supporting the precision milling cutter head to remove the excavation debris in real time.

[0034] The precision milling cutter head is a numerically controlled high-precision cutting component, which can achieve low-disturbance excavation with millimeter-level precision. The negative pressure suction system is a debris removal device supporting the precision milling head, which avoids the extrusion of the surrounding rock by the excavation debris through negative pressure.

[0035] In one example, the pre-designed three-dimensional contour of the cavern (1:100 scale) and the file of the excavation sequence for layered and block-by-block excavation with the pilot heading in the middle and then the side expansion excavation can be imported into the central controller. The layered and block-by-block excavation is a step-by-step excavation process simulating the on-site construction method to restore the stress redistribution during construction.

[0036] Then start the numerically controlled robotic arm equipped with a replaceable end effector to drive the precision milling cutter head at the end to operate along the preset path. Among them, the milling depth of the precision milling cutter head can be 1 - 3 mm to achieve refined layered excavation. The milling linear velocity can be 100 - 200 mm / min, and the rotational speed of the milling head can be maintained at 3000 - 5000 rpm to ensure the balance between cutting precision and efficiency. The negative pressure value of the negative pressure suction system can be stabilized between -60 kPa and -80 kPa to suck out the debris generated by the excavation in real time.

[0037] After completing the single-layer excavation of one partition, the central controller can automatically record the excavation progress and control the numerically controlled robotic arm to drive the precision milling cutter head to transfer to the next partition or the next layer for excavation until all the excavation procedures are completed. In this way, it can replace the traditional manual or simple excavation method, achieve low-disturbance and millimeter-level precision step-by-step excavation, reduce the additional damage to the surrounding rock during excavation, and ensure the stability of the initial stress state of the surrounding rock. The combination of the numerically controlled robotic arm and the precision milling cutter head accurately reproduces the preset three-dimensional contour of the cavern, reduces the shape deviation of the traditional excavation, and ensures the morphological consistency between the model cavern and the prototype. Traditional tests are usually overall excavations. The layered and block-by-block process of the pilot heading in the middle and then the side expansion excavation simulates the step-by-step unloading process in the actual engineering construction, and can truly restore the cumulative effect of stress redistribution during the on-site excavation. The negative pressure suction system removes the debris in real time, avoids the accumulation and extrusion of the debris on the model surrounding rock, ensures that the stress state of the surrounding rock after excavation is consistent with the actual project, and improves the accuracy of the test data. Step 203: Preset excavation and support trigger conditions. When the trigger conditions are met during excavation, control the CNC robotic arm to switch the tool type of the interchangeable-head end effector, and sequentially complete the micro-anchor insertion, prestressed anchor tensioning, and shotcrete layer laying. After the support is completed, the excavation operation resumes. The micro-anchors, anchors, and shotcrete are all support components scaled down to a 1:100 similarity ratio, simulating the anchors, prestressed anchors, and shotcrete on-site.

[0038] The preset excavation and support triggering conditions in this embodiment are: the cumulative excavation depth reaches a set depth (e.g., 20mm), or a target part in the indoor model is exposed. The target part may include the tunnel arch foot and the roof arch. The set depth and the target part are preset thresholds that match the on-site excavation and support logic, and are signal conditions for triggering support operations. For example, two types of trigger signals can be preset in the central controller: one is a quantification condition signal, where the cumulative excavation depth reaches 20mm (matching the engineering excavation advance at a 1:100 similarity ratio); the other is a location condition signal, where key stress-bearing parts such as the tunnel arch foot and the roof arch in the indoor model are exposed. The controller receives the excavation progress data and model contour recognition signals from the CNC robotic arm in real time to determine whether the triggering conditions are met.

[0039] The CNC robotic arm switches the tool head of the interchangeable-head end effector from the milling head to the anchor installer. The anchor installer picks up a miniature anchor bolt of a set diameter (e.g., 1-2 mm) and inserts it into the predetermined position on the indoor model. When the trigger condition is met, the central controller instructs the CNC robotic arm to stop excavation and switch the tool head of the interchangeable-head end effector from the milling head to the anchor installer. The installer picks up a miniature anchor bolt (made of copper-plated steel wire or fiberglass rod) with a diameter of 1-2 mm (1:100 scale) and inserts it into the predetermined position on the model according to the preset coordinates through pre-drilling or self-tapping, thus completing the anchor bolt fixation.

[0040] After the anchor bolt is inserted, the CNC robotic arm switches the tool head of the interchangeable end effector to a miniature tensioning jack. The miniature tensioning jack applies a predetermined prestress to the prestressed anchor cable (matching the prototype anchor cable tension at a similar ratio, such as 5-20N). The jack is immediately locked after tensioning, and tensioning data is recorded simultaneously. The miniature tensioning jack is a high-precision tensioning device adapted to scaled-down anchor cables, capable of accurately applying and locking the predetermined prestress, simulating the active support effect of the anchor cable in the field.

[0041] After the anchor cables are tensioned, the CNC robotic arm switches the tool head of the interchangeable end effector to a micro-spray nozzle. Through the micro-spray nozzle, a fast-setting material is sprayed at a set pressure (e.g., 0.2-0.4 MPa) to form a uniform spray layer of a set thickness (e.g., 2-5 mm) at the corresponding location on the indoor model. The fast-setting material is a low-strength, rapid-forming spray material specifically designed for scaled-down models, ensuring rapid setting of the spray layer and providing temporary support strength to match the time-sensitive nature of the testing process. The sprayed material can be prepared using a mass ratio of gypsum:water:retarder = 1:0.45:0.001.

[0042] By pre-setting trigger conditions and automatically switching tools, the system rigorously replicates the construction logic of excavating and supporting sections on-site, solving the problem of distorted support timeliness caused by traditional tests that involve excavation followed by concentrated support. This accurately reflects the constraint effect of support on the stress release of surrounding rock. Automated support operations driven by CNC robotic arms avoid positioning deviations and uneven force caused by manual operation, ensuring precise control over the placement of micro-anchors, anchor cable prestress values, and shotcrete thickness, thus improving the repeatability and reliability of test data. Micro-anchors, anchor cables, and shotcrete layers are all designed at a 1:100 similarity ratio, with their dimensions and mechanical parameters proportionally matched to the model's surrounding rock. This avoids distortions in stress response caused by incompatibility between support components and the model, ensuring that test results effectively reflect the prototype project. The system sequentially completes the composite support process of anchor bolt anchoring, anchor cable prestressing, and shotcrete layer sealing, simulating the working mode of multi-method collaborative support on-site. This allows for precise research on the contribution of different support stages to the stability of ultra-large span tunnels, providing direct evidence for optimizing engineering support schemes. By realizing automated time-series linkage between excavation and support, the impact of support timeliness on the stability of ultra-large span tunnels is accurately simulated, and the synergistic force effect between support and surrounding rock is restored.

[0043] Step 204: Simultaneously collect multi-source data on internal strain, surface displacement, and internal micro-damage of the indoor model using sensors embedded in the indoor model and monitoring equipment arranged around the model box. Transmit the multi-source data to a multi-source data acquisition cabinet, and perform spatiotemporal alignment and overlay analysis on the multi-source data to obtain a comprehensive map of surrounding rock failure evolution. Spatiotemporal alignment and overlay analysis is a processing method that unifies monitoring data from different dimensions and times into a unified spatiotemporal reference and integrates them into a complete dataset. The comprehensive map of surrounding rock failure evolution presents a data map showing the entire process of cavern failure from microcrack initiation to macroscopic instability. This overcomes the limitations of traditional point-based / single-dimensional monitoring, enabling full-dimensional data acquisition from internal to surface and from micro to macro levels, constructing a complete failure evolution pattern, and providing high-fidelity experimental evidence for risk prediction and design optimization of ultra-large span cavern projects.

[0044] In one example, sensors embedded in the indoor model may include distributed fiber optic sensors for monitoring internal strain and acoustic probes for monitoring internal micro-fractures, directly acquiring internal state data of the model. Monitoring equipment arranged around the model enclosure may include two high-speed cameras from a 3D-DIC system. The 3D-DIC system is a three-dimensional digital image correlation technology system that uses two high-speed cameras to capture images of the model surface and combines this with algorithms to calculate displacement changes at various points on the surface, achieving non-contact surface displacement monitoring. The distributed fiber optic sensors, acoustic probes, and the high-speed cameras of the 3D-DIC system are all networked with a central controller, which sends synchronization trigger signals to each device to ensure consistent data timestamps.

[0045] Specifically, the internal strain distribution of the indoor model can be continuously monitored using distributed fiber optic sensors at a set sampling frequency (e.g., 10 Hz). Surface displacement data of the indoor model is acquired using two high-speed cameras of a 3D-DIC system at a second set rate (e.g., 1 frame / second). Internal micro-fracture events of the indoor model are continuously acquired by using an acoustic probe with a set decibel (e.g., 40 dB) threshold value, acquiring only micro-fracture signals above this threshold.

[0046] After the experiment, internal strain data from the distributed fiber optic sensor, surface displacement data from the 3D-DIC system, and micro-damage localization data from the acoustic probe were retrieved from the multi-source data acquisition cabinet. The internal strain data, surface displacement data, and micro-damage localization data were spatiotemporally aligned and overlaid to construct a comprehensive map of the surrounding rock failure evolution, encompassing three-dimensional space and time dimensions. This map fully presents the entire process of the indoor model from microcrack initiation to macroscopic instability. For example, the three types of data can be unified to the same spatial coordinate system (matching the model's 1:100 scale) and time axis (based on a unified timestamp of synchronous triggering), eliminating data deviations in the spatiotemporal dimensions. Overlay analysis of the aligned data establishes the correlation between internal strain, surface displacement, and micro-damage. Based on the fused data, a comprehensive map containing three-dimensional space (each part of the model) and time (the entire experimental process) is generated, fully presenting the entire process from microcrack initiation and propagation to macroscopic deformation and eventual instability.

[0047] By combining pre-embedded sensors (internal) and peripheral equipment (surface), three types of data—internal strain, surface displacement, and internal micro-damage—are simultaneously acquired, achieving full-dimensional coverage of internal-surface and mechanical-damage aspects (traditional tests can only monitor discrete points or single surface data, failing to reflect the overall failure pattern). The synchronous triggering function of the central control system ensures that all data timestamps are consistent, and then data deviations are eliminated through spatiotemporal alignment processing, forming a strong correlation between internal strain changes, surface displacement trends, and micro-damage locations, accurately pinpointing where damage occurs first, how deformation occurs, and why instability occurs. The comprehensive map of surrounding rock failure evolution clearly presents the entire process from microcrack initiation (acoustic data capture), internal strain concentration (fiber optic data reflection), to significant surface displacement (3D-DIC data recording), and finally macroscopic instability, solving the problem that traditional tests can only see the final failure result and cannot trace the source and development path of failure. The comprehensive and full-process test data can be used to quantitatively analyze the weak points of ultra-large span tunnels (such as the tendency for strain concentration to occur at the arch foot and the tendency for cracks to develop in the top arch) and the effectiveness of support measures (such as whether anchor bolts suppress strain concentration), providing high-fidelity test data for risk prediction and support scheme optimization in prototype projects.

[0048] The following is a typical engineering case study, simulating the construction process of the underground powerhouse (spanning approximately 70 meters) of the Baihetan Hydropower Station, to illustrate the specific implementation of this invention in detail.

[0049] First, based on similarity theory, the geometric similarity ratio was determined to be 1:100. Accordingly, the model cavern span was designed to be 0.7 meters, and a consolidation material consisting of barite powder, quartz sand, gypsum, and a retarder mixed in a specific mass ratio was selected as the rock mass similarity material. The entire experiment was conducted within the modular device described above, with each subsystem integrated and linked through central control software.

[0050] The implementation process began with model preparation and the establishment of the geostress field. In a model box equipped with internal monitoring optical fibers, similar materials were poured and vibrated in layers, and cured for 7 days in a constant temperature and humidity environment to stabilize their strength. Subsequently, a three-dimensional gradient loading system was activated: vertical loads up to 0.8 MPa were applied in stages via a top plate actuator array to simulate the self-weight of the overlying rock mass; simultaneously, horizontal loads were applied independently via side plate actuator arrays, with the ratio increasing from 0.5 at the top of the model to 0.8 at the bottom, to accurately reproduce the non-uniform tectonic stress field reported in the engineering geological report. The entire load application process was controlled in a closed-loop manner by sensor network feedback to ensure that the initial stress state established within the model was consistent with the design target, which is the foundation for all subsequent high-fidelity simulations.

[0051] After the ground stress field stabilized, the excavation and support linkage simulation phase began. The central control system invoked a pre-programmed digital program of "first excavating the central pilot tunnel, then expanding the excavation layer by layer on both sides," driving the integrated work platform to begin operation. The robotic arm, equipped with a precision milling-suction composite head, milled the central pilot tunnel area strictly according to the path at a speed of 150 mm per minute and a depth of 2 mm per layer. Simultaneously, the negative pressure system used a suction force of -70 kPa to instantly remove debris, achieving near-zero disturbance excavation. When the central pilot tunnel excavation reached the predetermined advance (20 mm depth), the program automatically paused, and the robotic arm switched to a micro support tool head. At predetermined locations on the exposed tunnel wall, the robotic arm installed copper-plated steel wire with a diameter of 1.5 mm to simulate anchor bolts, and applied 10 N of prestress to the fiber rods at key locations on the arch foot using micro tensioners. Subsequently, a 3 mm thick layer of gypsum-based quick-setting material was sprayed to simulate the initial concrete support. This process follows the sequential cycle of "excavation-surveying-support" until the entire cavern outline is formed, truly replicating the mechanical process of dynamic construction on site.

[0052] To verify the effectiveness of this embodiment, a distributed fiber optic sensing system, a binocular high-speed camera (i.e., a 3D-DIC system), and an acoustic emission probe were used for synchronous monitoring throughout the process. Experimental results show that the system successfully captured the formation and evolution of the unique "arch effect" in ultra-large span caverns under gradient stress: the DIC displacement cloud map clearly shows the spatiotemporal distribution of the arch settlement and sidewall convergence; fiber optic strain data reveals the depth and development trend of the plastic zone in the surrounding rock; and the location of acoustic emission events highly matches the potential slip surface predicted by numerical simulation. Finally, through multi-source data fusion analysis, this embodiment not only quantitatively verifies the feasibility of the excavation and support scheme but also accurately reveals the crucial role of support timing in controlling surrounding rock deformation, providing direct and reliable experimental evidence for optimizing the design and construction of the prototype project.

[0053] Compared to traditional technical challenges, the embodiments of this application have the following beneficial effects.

[0054] 1. It has achieved ultra-high fidelity reproduction of complex mechanical environments.

[0055] The three-dimensional gradient active loading system overcomes the limitation of traditional model tests that can only apply unidirectional loads. It can accurately reproduce the complex non-uniform stress field (including vertical self-weight stress and horizontal structural stress) of the engineering site in the laboratory, providing the model with highly realistic initial mechanical boundary conditions. Combined with a flexible pressure-transmitting and rigid pressure-bearing composite loading plate, it ensures the uniform distribution of stress on the surface of large-size models, significantly improving the physical similarity of the test and the reliability of the results.

[0056] 2. Millimeter-level precision control and ultra-low disturbance simulation of the excavation process were achieved.

[0057] By utilizing an integrated milling-vacuum excavation device, CNC milling replaces traditional manual excavation, achieving millimeter-level precision control over the tunnel outline, layer thickness, and block size. Synchronous vacuum suction technology ensures that excavation debris is removed immediately, fundamentally reducing the compression and disturbance to the surrounding rock caused by debris accumulation. This allows for a true and clear reflection of the dynamic redistribution of stress in the surrounding rock caused by each step of excavation.

[0058] 3. A dynamic linkage mechanism for excavation and support based on the actual construction sequence was established.

[0059] The time-programmable support linkage mechanism deeply integrates excavation and support procedures through a central controller, forcibly realizing an engineering process where excavation and support are carried out in quick succession. This mechanism ensures that support structures such as micro-anchors, prestressed anchors, and shotcrete can be intervened at the precise timing required by the design, realistically simulating the interaction between the support and the surrounding rock at different construction stages. It provides an irreplaceable technical means for studying the timeliness of support and optimizing support parameters.

[0060] 4. We obtained full-dimensional, full-field evolution data from microscopic damage to macroscopic destruction.

[0061] Relying on a multi-source information fusion monitoring system, this system integrates three advanced monitoring technologies: distributed fiber optic sensing (internal strain), 3D-DIC (surface full-field displacement), and acoustic emission (internal micro-fractures). This integrated solution breaks through the limitations of traditional point-based monitoring, enabling the synchronous and continuous capture of multi-dimensional information on the entire process of surrounding rock, from the initiation of micro-cracks to their expansion into the macro-plastic zone and ultimately to instability. It constructs a complete digital profile of the failure evolution, greatly deepening the scientific understanding of the failure mechanism of ultra-large span caverns.

[0062] 5. A standardized and repeatable advanced experimental research paradigm has been formed.

[0063] By combining highly integrated hardware devices with programmed control methods, a complete, standardized, and highly automated testing system has been formed. This paradigm greatly reduces the uncertainty of manual operation, ensures a high degree of repeatability and comparability of the testing process and results, and enables it to serve as a reliable research tool for widely used to verify and optimize different excavation methods (such as the CD method and CRD method) and support design parameters. It has direct engineering application value for guiding actual engineering design and reducing construction risks.

[0064] Figure 3 This is a schematic diagram of the structure of an indoor test system 300 for excavation and support of an ultra-large span tunnel, as provided in this embodiment of the application. Figure 3 As shown, the test system may include a setup module 301, a startup module 302, a support module 303, and an analysis module 304.

[0065] The module 301 is used to establish a non-uniform three-dimensional initial geostress field in the indoor model of the excavation and support of the ultra-large span tunnel in the model box mounted in the reaction frame, using a three-dimensional gradient active loading system. The three-dimensional initial geostress field includes vertical stress and gradient-distributed horizontal stress.

[0066] The startup module 302 is used to import the preset three-dimensional outline of the cavern and the layered excavation sequence file. The CNC robotic arm equipped with a replaceable end effector drives the precision milling cutting head on the replaceable end effector to excavate layer by layer along the preset path. At the same time, the negative pressure suction system matched with the precision milling cutting head is started to remove the excavation debris in real time.

[0067] The support module 303 is used to preset the excavation support trigger conditions. When the trigger conditions are detected during the excavation process, the CNC robotic arm is controlled to switch the tool type of the interchangeable head end effector and sequentially complete the micro anchor bolt implantation, prestressed anchor cable tensioning and shotcrete layer laying. After the support is completed, the excavation operation is resumed.

[0068] The analysis module 304 is used to synchronously collect multi-source data of internal strain, surface displacement and internal micro-damage of the indoor model through sensors embedded in the indoor model and monitoring equipment arranged around the model box. The multi-source data is transmitted to the multi-source data acquisition cabinet, and the multi-source data is spatiotemporally aligned and superimposed to obtain a comprehensive map of the surrounding rock failure evolution.

[0069] The establishment module 301, the startup module 302, the support module 303, and the analysis module 304 can be used to execute steps 201-204 in the corresponding embodiments of the test method for the indoor model of the above-mentioned ultra-large span tunnel excavation and support. For the specific implementation methods of these modules and more details, please refer to the corresponding method section, which will not be elaborated here.

[0070] This application also provides a computer-readable storage medium storing a program that can be loaded by a processor and executed as in any of the embodiments of this application, a test method for an indoor model of excavation and support of an ultra-large span tunnel.

[0071] Those skilled in the art will understand that all or part of the functions of the various methods in the above embodiments can be implemented by hardware or by computer programs. When all or part of the functions in the above embodiments are implemented by computer programs, the program can be stored in a computer-readable storage medium, which may include: read-only memory, random access memory, disk, optical disk, hard disk, etc., and the program is executed by a computer to achieve the above functions. For example, the program can be stored in the memory of a device, and when the program in the memory is executed by the processor, all or part of the above functions can be achieved. In addition, when all or part of the functions in the above embodiments are implemented by computer programs, the program can also be stored in a server, another computer, disk, optical disk, flash drive, or external hard drive, etc., and can be downloaded or copied to the memory of a local device, or the system of the local device can be updated. When the program in the memory is executed by the processor, all or part of the functions in the above embodiments can be achieved.

[0072] The above examples illustrate this application only to aid understanding and are not intended to limit its scope. Those skilled in the art to which this application pertains can make various simple deductions, modifications, or substitutions based on the ideas presented.

Claims

1. A test method for an indoor model of excavation and support of an ultra-large span tunnel, characterized in that, include: In the indoor model of the excavation and support of the super-large span tunnel, which is mounted in the model box within the reaction frame, a non-uniform three-dimensional initial geostress field is established using a three-dimensional gradient active loading system. The three-dimensional initial geostress field includes vertical stress and gradient-distributed horizontal stress. Import the preset three-dimensional outline of the cavern and the layered excavation sequence file. The CNC robotic arm equipped with a replaceable end effector drives the precision milling cutting head on the replaceable end effector to excavate layer by layer along the preset path. At the same time, the negative pressure suction system matched with the precision milling cutting head is started to remove the excavation debris in real time. Preset excavation and support trigger conditions. When the trigger conditions are detected during the excavation process, control the CNC robotic arm to switch the tool type of the interchangeable head end effector, and sequentially complete the micro anchor bolt implantation, prestressed anchor cable tensioning and shotcrete layer laying. After the support is completed, the excavation operation is resumed. Multi-source data on internal strain, surface displacement, and internal micro-damage of the indoor model are simultaneously collected by sensors embedded in the indoor model and monitoring equipment arranged around the model box. The multi-source data is transmitted to a multi-source data acquisition cabinet, and spatiotemporal alignment and overlay analysis are performed on the multi-source data to obtain a comprehensive map of the evolution of surrounding rock failure.

2. The test method for the indoor model of excavation and support of ultra-large span tunnels according to claim 1, characterized in that, Within the indoor model of the excavation and support of the ultra-large span tunnel, housed in a model box within a reaction frame, a non-uniform three-dimensional initial geostress field is established using a three-dimensional gradient active loading system, including: The process involves preparing similar materials and then casting and curing them in the mold box. A three-dimensional gradient active loading system, consisting of a reaction frame, a hydraulic servo array, a flexible pressure-transmitting-rigid pressure-bearing composite loading plate, and a hydraulic oil source, is used to apply vertical stress and gradient-distributed horizontal stress to the indoor model through the flexible pressure-transmitting-rigid pressure-bearing composite loading plate, thereby establishing a non-uniform initial ground stress field that matches the engineering scene.

3. The test method for the indoor model of excavation and support of ultra-large span tunnels according to claim 2, characterized in that, The similar materials are configured in the following mass ratios: 48% barite powder, 30% quartz sand, 12% gypsum, and 10% water. During the casting process in the model box, distributed fiber optic sensors and acoustic probes are simultaneously embedded in the indoor model. After the casting process is completed, the model box is placed in an environment with a temperature of 25±2℃ and a relative humidity of more than 90% for curing for no less than 7 days until the uniaxial compressive strength of the similar material reaches the set value of 0.3-0.5MPa.

4. The test method for the indoor model of excavation and support of ultra-large span tunnels according to claim 2, characterized in that, The application of vertical stress and gradient-distributed horizontal stress to the indoor model through the flexible pressure-transmitting-rigid pressure-bearing composite loading plate includes: Vertical stress is applied slowly in multiple stages by an array of hydraulic servo actuators arranged on the top of the model box. The application rate is controlled within a first set rate, and each load is held for a first set time until the target value is reached. The horizontal stress is applied in stages by an array of hydraulic servo actuators arranged on both sides of the model box. By independently controlling the pressure of the actuators at different positions, the ratio of the vertical stress to the horizontal stress gradually changes from a first ratio at the top of the model to a second ratio at the bottom. Each load is held for a second set time. After the deformation of the indoor model stabilizes, the next level of horizontal stress is applied. The first ratio is less than the second ratio. The horizontal stress is transmitted to the indoor model through flexible pressure-transmitting and rigid pressure-bearing composite loading plates arranged on both sides of the model box.

5. The test method for the indoor model of excavation and support of ultra-large span tunnels according to claim 1, characterized in that, The geometric similarity ratio of the indoor model is 1:100, and it is poured into the model box with a length × width × height of 2.9m × 1m × 1.8m. During pouring, the similar material is filled into the model box in three layers and tamped and leveled to ensure that the density of the indoor model is uniform.

6. The test method for the indoor model of excavation and support of ultra-large span tunnels according to claim 1, characterized in that, The milling depth of the precision milling cutting head is 1-3mm, the milling linear speed is 100-200mm / min, the milling head rotation speed is 3000-5000rpm, and the negative pressure value of the negative pressure suction system is stable between -60kPa and -80kPa; the order of the layered and block excavation sequence is that the central guide tunnel is excavated first, followed by the expansion excavation on both sides. The process of using a CNC robotic arm equipped with a replaceable end effector to drive a precision milling cutting head on the replaceable end effector to excavate in layers along a preset path includes: After completing the excavation of a single layer in a zone, the excavation progress is automatically recorded, and the CNC robotic arm is controlled to drive the precision milling head to the next zone or the next layer for excavation, until all excavation processes are completed.

7. The test method for the indoor model of excavation and support of ultra-large span tunnels according to claim 1, characterized in that, The preset excavation support triggering conditions are: the cumulative excavation depth reaches the set depth, or the target part in the indoor model is exposed, the target part includes the cave arch foot and the top arch; The sequential completion of micro-anchor bolt implantation, prestressed anchor cable tensioning, and shotcrete layer laying includes: The CNC robotic arm is controlled to switch the tool head of the interchangeable end effector from the milling head to the anchor bolt installer. The anchor bolt installer then picks up a micro anchor bolt of a set diameter and implants the micro anchor bolt at a predetermined position on the indoor model. The CNC robotic arm is controlled to switch the tool head of the interchangeable end effector to a miniature tensioning jack, and the miniature tensioning jack is used to apply a predetermined prestress to the prestressed anchor cable and lock it in place. The CNC robotic arm is controlled to switch the tool head of the interchangeable end effector to a micro-spray nozzle, and the micro-spray nozzle sprays fast-setting material at a set pressure to form a uniform spray layer of a set thickness at the corresponding position of the indoor model.

8. The test method for the indoor model of excavation and support of ultra-large span tunnels according to claim 1, characterized in that, The sensors embedded in the indoor model include distributed fiber optic sensors and acoustic probes, and the monitoring equipment arranged around the model box includes two high-speed cameras of the 3D-DIC system. Multi-source data on the internal strain, surface displacement, and internal micro-damage of the indoor model are simultaneously collected using sensors embedded in the indoor model and monitoring devices arranged around the model box. These data include: The internal strain distribution of the indoor model is continuously monitored using the distributed fiber optic sensors at a set sampling frequency. The surface displacement data of the indoor model is acquired by two high-speed cameras of the 3D-DIC system at a second set rate. The acoustic probe is used to continuously collect internal micro-fracture events of the indoor model with a set decibel threshold value. The distributed fiber optic sensor, the acoustic probe, and the high-speed camera of the 3D-DIC system are all connected to the central controller, which sends synchronization trigger signals to each device to ensure that the timestamps of the data are consistent. The process of performing spatiotemporal alignment and overlay analysis on the multi-source data to obtain a comprehensive map of surrounding rock failure evolution includes: After the experiment, the internal strain data of the distributed optical fiber sensor, the surface displacement data of the 3D-DIC system, and the micro-damage location data of the acoustic probe were retrieved from the multi-source data acquisition cabinet, and the internal strain data, the surface displacement data, and the micro-damage location data were spatiotemporally aligned and superimposed. A comprehensive map of surrounding rock failure evolution, including three-dimensional spatial and temporal dimensions, is constructed to fully present the entire process of the indoor model from microcrack initiation to macroscopic instability.

9. A test system for an indoor model of excavation and support of an ultra-large span tunnel, characterized in that, include: A module is established to create a non-uniform three-dimensional initial geostress field in an indoor model of excavation and support of an ultra-large span tunnel within a model box mounted in a reaction frame, using a three-dimensional gradient active loading system. The three-dimensional initial geostress field includes vertical stress and gradient-distributed horizontal stress. The startup module is used to import the preset three-dimensional outline of the cavern and the layered and block excavation sequence file. The CNC robotic arm equipped with a replaceable end effector drives the precision milling cutting head on the replaceable end effector to excavate layer by layer along the preset path. At the same time, the negative pressure suction system matched with the precision milling cutting head is started to remove the excavation debris in real time. The support module is used to preset the excavation support trigger conditions. When the trigger conditions are detected during the excavation process, the CNC robotic arm is controlled to switch the tool type of the interchangeable head end effector to complete the micro anchor bolt implantation, prestressed anchor cable tensioning and shotcrete layer laying in sequence. After the support is completed, the excavation operation is resumed. The analysis module is used to synchronously collect multi-source data of internal strain, surface displacement and internal micro-damage of the indoor model through sensors embedded in the indoor model and monitoring equipment arranged around the model box. The multi-source data is transmitted to the multi-source data acquisition cabinet, and the multi-source data is subjected to spatiotemporal alignment and overlay analysis to obtain a comprehensive map of the evolution of surrounding rock failure.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that can be loaded by a processor and executed as a test method for an indoor model of excavation and support of an ultra-large span tunnel as described in any one of claims 1 to 8.

Images