A shield underpass shallow foundation house active compensation system and method

Abstract

The application discloses a shield under shallow foundation house active compensation system and method, and relates to the field of shield tunnel construction equipment.The application arranges a multi-source sensing matrix around a shallow foundation house, and obtains ground foundation displacement field and foundation inclination angle data in real time; a stratum foundation coupling settlement prediction model is established to output a predicted settlement field distribution; the size and application timing of a compensation force are determined according to a differential settlement gradient, a mixed compensation execution array composed of a hydraulic jack and a shape memory alloy driver is driven to implement active compensation; during the compensation process, a compensation correction amount is calculated based on real-time monitoring data and closed-loop iterative adjustment is performed, so that the deviation converges to a preset allowable threshold; after the shield passes through the influence area, the compensation force is gradually released according to a preset unloading gradient, and the compensation effect is evaluated according to the foundation rebound amount.The application can realize active control of differential settlement of a shallow foundation house, and improve compensation accuracy and stability.

Description

Technical Field

[0001] This invention relates to the field of tunnel boring machine (TBM) construction equipment, and in particular to an active compensation system and method for TBMs passing under shallow foundation buildings. Background Technology

[0002] In urban rail transit construction, the shield tunneling method is widely used in subway tunnel construction due to its advantages such as high construction efficiency and minimal impact on surface traffic. When a shield tunnel passes under existing shallow-foundation buildings, the redistribution of ground stress and soil disturbance caused by the tunnel boring can lead to settlement of the overlying strata, further causing deformation of the building foundation. When there are significant spatial differences in foundation settlement, it can easily cause structural damage such as wall cracking, door and window deformation, and foundation tilting, and in severe cases, even affect the normal use of the building. Therefore, effective settlement control of shallow-foundation buildings during shield tunneling is an important technical measure to ensure the safety of existing buildings.

[0003] Currently, protective measures for shield tunneling under shallow foundation buildings mainly include grouting reinforcement, foundation replacement, and construction parameter control. Grouting reinforcement reduces settlement by improving soil mechanical properties, but the grout diffusion range is greatly affected by geological conditions, resulting in some uncertainty in the reinforcement effect. Furthermore, it can easily generate additional disturbances during construction. Foundation replacement can improve the bearing capacity of the building foundation, but it has a long construction period, high project costs, and can affect the normal use of existing buildings. Construction parameter control mainly reduces construction disturbance by adjusting shield thrust, surface pressure, or tunneling rate, but due to the lack of direct feedback on the deformation state of the building foundation, it is difficult to accurately control localized differential settlement.

[0004] In summary, most existing shield tunneling technologies for protecting houses under shallow foundations adopt overall control or post-construction compensation methods, lacking an active compensation mechanism that can dynamically adjust according to the settlement development of the foundation. This makes it difficult to effectively control differential settlement of the foundation in a timely manner, resulting in the protection effect of the house being greatly affected by changes in construction conditions, and making it difficult to balance compensation accuracy, real-time performance, and stability. Summary of the Invention

[0005] To address the above problems, this invention provides an active compensation system and method for shield tunneling under shallow foundation buildings.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] On one hand, this invention discloses an active compensation method for shield tunneling under shallow foundation buildings, comprising:

[0008] Step 1: Deploy a multi-source sensing matrix consisting of fiber optic grating sensors and tilt sensors around the shallow foundation houses in the preset influence area in front of the tunnel boring machine to collect the foundation displacement field data and foundation tilt angle data of the shallow foundation houses in real time.

[0009] Step 2: Input the thrust parameters, surface pressure parameters, tunneling rate parameters, and stratum physical and mechanical parameters during the shield tunneling process into the pre-established stratum-foundation coupled settlement prediction model, and output the predicted settlement field distribution within the future preset time window;

[0010] Step 3: Calculate the differential settlement gradient at each compensation point based on the predicted settlement field distribution, and determine the required compensation force and application sequence at each compensation point based on the differential settlement gradient.

[0011] Step 4: Activate the hybrid compensation execution array deployed below the foundation of the shallow foundation house. The hybrid compensation execution array includes hydraulic jacks and shape memory alloy actuators, and applies compensation force to each compensation point according to the magnitude of the compensation force and the application sequence.

[0012] Step 5: During the application of the compensation force, the multi-source sensing matrix is ​​used to collect the foundation displacement field data and foundation tilt angle data in real time, and the deviation between the foundation displacement field data and the predicted settlement field distribution is calculated to obtain the compensation correction amount.

[0013] Step 6: Based on the compensation correction amount, perform closed-loop iterative adjustment of the compensation force at each compensation point in the hybrid compensation execution array until the deviation between the foundation displacement field data and the predicted settlement field distribution converges to a preset allowable threshold.

[0014] Step 7: After the tunnel boring machine passes through the preset influence area, the compensation force of the hybrid compensation execution array is gradually released according to the preset unloading gradient, and the basic rebound amount is monitored to evaluate the compensation effect.

[0015] Furthermore: Step one includes:

[0016] Multiple fiber optic grating sensors are installed at preset intervals around the perimeter of the outer wall foundation of the shallow foundation building. Each fiber optic grating sensor is buried in the soil layer below the bottom of the foundation along the longitudinal direction of the foundation to collect vertical displacement data at various locations of the foundation as ground displacement field data.

[0017] Inclination sensors are installed at each corner of the shallow foundation building. Each inclination sensor is fixed to the upper surface of the foundation to collect the tilt angle data at each corner of the foundation as the foundation tilt angle data.

[0018] The data collected by each fiber Bragg grating sensor and each tilt sensor are aggregated to the data receiving terminal through data transmission cables to form a multi-source sensing matrix.

[0019] Furthermore, step two includes:

[0020] The real-time thrust parameters, surface pressure parameters, and tunneling rate parameters during the shield tunneling process are obtained, as well as the physical and mechanical parameters of the strata obtained from the exploration, including soil compression modulus, Poisson's ratio, internal friction angle, and cohesion.

[0021] Using thrust parameters, surface pressure parameters, and tunneling rate parameters during shield tunneling as dynamic input variables and geological physical and mechanical parameters as static geological parameters, a stratum-foundation coupled settlement prediction model is constructed. The stratum-foundation coupled settlement prediction model is a hybrid prediction model based on the fusion of finite element numerical simulation and long short-term memory neural network.

[0022] The dynamic input variables and static geological parameters are input into the strata-foundation coupled settlement prediction model, and the settlement prediction values ​​of each preset spatial grid node within the future preset time window are output to form the predicted settlement field distribution.

[0023] Furthermore, step three includes:

[0024] For each preset spatial grid node in the predicted settlement field distribution, the settlement difference between adjacent preset spatial grid nodes is calculated, and the ratio of the settlement difference to the horizontal distance between adjacent preset spatial grid nodes is used as the differential settlement gradient.

[0025] The required compensation force at each compensation point is calculated based on the product of the differential settlement gradient and the foundation stiffness. The direction of the compensation force is opposite to the direction of the differential settlement gradient, so as to generate a reverse torque to counteract the differential settlement.

[0026] Based on the tunnel boring machine's (TBM) excavation rate and the distance from the TBM's current position to each compensation point, the time it takes for the TBM to reach the corresponding stratum position at each compensation point is calculated, and this time is used as the application sequence.

[0027] Furthermore, step four includes:

[0028] At each compensation point below the foundation of a shallow foundation house, a hydraulic jack and a shape memory alloy actuator are installed. The hydraulic jack and the shape memory alloy actuator are arranged side by side and connected to the bottom surface of the foundation.

[0029] The magnitude of the compensation force is decomposed into a basic compensation force component and a fine adjustment compensation force component, wherein the basic compensation force component is provided by a hydraulic jack and the fine adjustment compensation force component is provided by a shape memory alloy actuator.

[0030] According to the application sequence, the hydraulic jack is first activated to apply the basic compensation force component, and then the shape memory alloy actuator applies the fine adjustment compensation force component. The resultant force of the basic compensation force component and the fine adjustment compensation force component is equal to the magnitude of the compensation force.

[0031] Furthermore, step five includes:

[0032] During the application of the compensation force, the foundation displacement field data and foundation tilt angle data are collected in real time using a multi-source sensing matrix at a preset sampling frequency. The foundation displacement field data are mapped to each preset spatial grid node in the predicted settlement field distribution to obtain the actual settlement value of each preset spatial grid node.

[0033] For each preset spatial grid node, the difference between the actual settlement value and the predicted settlement value of the corresponding node in the predicted settlement field distribution is calculated, and the difference is used as the settlement deviation value.

[0034] The compensation correction amount at each compensation point is calculated based on the settlement deviation value and the preset proportional coefficient. When the settlement deviation value is greater than zero, the compensation correction amount is positive to increase the compensation force, and when the settlement deviation value is less than zero, the compensation correction amount is negative to reduce the compensation force.

[0035] Furthermore, step six includes:

[0036] The compensation correction amount is added to the current compensation force magnitude of each compensation point to update the compensation force magnitude of each compensation point, and the hybrid compensation execution array is controlled to reapply the compensation force according to the updated compensation force magnitude.

[0037] During the application of the updated compensation force, the foundation displacement field data is re-acquired using a multi-source sensing matrix, and the deviation between the foundation displacement field data and the predicted settlement field distribution is calculated to obtain the updated settlement deviation value.

[0038] Determine whether the absolute value of the updated settlement deviation is less than or equal to the preset allowable threshold. If so, stop the iterative adjustment.

[0039] If not, return to the step of adding the compensation correction amount to the current compensation force magnitude of each compensation point until the absolute value of the updated settlement deviation value is less than or equal to the preset allowable threshold.

[0040] Furthermore: Step seven includes:

[0041] Based on the relative distance between the tunnel boring machine's location and the boundary of the preset influence zone, determine whether the tunnel boring machine has passed through the preset influence zone;

[0042] After the tunnel boring machine passes through the preset influence area, the compensation force of the hybrid compensation execution array is gradually released in stages according to the preset unloading gradient. In each unloading stage, the compensation force is reduced by a preset proportion at a preset time interval.

[0043] During each unloading phase, the base rebound amount is continuously monitored using a multi-source sensing matrix, and the deviation between the base rebound amount and the expected elastic recovery amount is calculated.

[0044] After all unloading stages are completed, the compensation effect is evaluated based on the deviation between the final base rebound amount and the expected elastic recovery amount. If the deviation is within the preset evaluation threshold, the compensation effect is deemed qualified.

[0045] If the deviation exceeds the preset evaluation threshold, an abnormal compensation effect report will be generated.

[0046] On the other hand, this invention discloses an active compensation system for shield tunneling under shallow foundation buildings, comprising:

[0047] The sensing module is used to deploy a multi-source sensing matrix composed of fiber optic grating sensors and tilt sensors around the shallow foundation houses in the preset influence area in front of the shield tunneling, so as to collect the foundation displacement field data and foundation tilt angle data of the shallow foundation houses in real time.

[0048] The prediction module is used to input the thrust parameters, surface pressure parameters, tunneling rate parameters, and stratum physical and mechanical parameters during the shield tunneling process into a pre-established stratum-foundation coupled settlement prediction model, and output the predicted settlement field distribution within a preset future time window.

[0049] The calculation module is used to calculate the differential settlement gradient at each compensation point based on the predicted settlement field distribution, and to determine the required compensation force and application sequence at each compensation point based on the differential settlement gradient.

[0050] An execution module is used to activate a hybrid compensation execution array deployed below the foundation of the shallow foundation house. The hybrid compensation execution array includes hydraulic jacks and shape memory alloy actuators, and applies compensation force to each compensation point according to the magnitude of the compensation force and the application sequence.

[0051] The feedback module is used to collect foundation displacement field data and foundation tilt angle data in real time using the multi-source sensing matrix during the application of compensation force, and to calculate the deviation between the foundation displacement field data and the predicted settlement field distribution to obtain the compensation correction amount.

[0052] The adjustment module is used to perform closed-loop iterative adjustment of the compensation force of each compensation point in the hybrid compensation execution array based on the compensation correction amount, until the deviation between the foundation displacement field data and the predicted settlement field distribution converges to a preset allowable threshold.

[0053] The unloading evaluation module is used to gradually release the compensation force of the hybrid compensation execution array according to the preset unloading gradient after the tunnel boring machine passes through the preset influence area, and monitor the basic rebound amount to evaluate the compensation effect.

[0054] The technological advancements achieved by this invention compared to existing technologies are as follows:

[0055] This invention establishes an active compensation mechanism for the tunnel boring machine (TBM) tunneling process. It dynamically determines the compensation strategy based on the settlement development of the building foundation and continuously adjusts the compensation process in conjunction with real-time monitoring results. This enables the compensation effect to adaptively adjust with changes in foundation deformation, achieving active control over differential settlement of shallow foundation buildings. It improves the real-time performance, accuracy, and stability of compensation, effectively reduces the adverse effects of TBM tunneling on existing shallow foundation buildings, and enhances the protection of buildings. Attached Figure Description

[0056] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0057] In the attached diagram:

[0058] Figure 1 This is a flowchart of the present invention;

[0059] Figure 2 This is a system structure diagram of the present invention. Detailed Implementation

[0060] The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of the present invention will now be described with reference to the accompanying drawings.

[0061] Example 1

[0062] Combination Figure 1 and Figure 2 This invention discloses an active compensation method for shield tunneling under shallow foundation buildings, comprising:

[0063] Step 1: Deploy a multi-source sensing matrix consisting of fiber optic grating sensors and tilt sensors around the shallow foundation houses in the preset influence area in front of the tunnel boring machine to collect the foundation displacement field data and foundation tilt angle data of the shallow foundation houses in real time.

[0064] Step 2: Input the thrust parameters, surface pressure parameters, tunneling rate parameters, and stratum physical and mechanical parameters during the shield tunneling process into the pre-established stratum-foundation coupled settlement prediction model, and output the predicted settlement field distribution within the future preset time window;

[0065] Step 3: Calculate the differential settlement gradient at each compensation point based on the predicted settlement field distribution, and determine the required compensation force and application sequence at each compensation point based on the differential settlement gradient.

[0066] Step 4: Activate the hybrid compensation execution array deployed below the foundation of the shallow foundation house. The hybrid compensation execution array includes hydraulic jacks and shape memory alloy actuators, and applies compensation force to each compensation point according to the magnitude of the compensation force and the application sequence.

[0067] Step 5: During the application of the compensation force, the multi-source sensing matrix is ​​used to collect the foundation displacement field data and foundation tilt angle data in real time, and the deviation between the foundation displacement field data and the predicted settlement field distribution is calculated to obtain the compensation correction amount.

[0068] Step 6: Based on the compensation correction amount, perform closed-loop iterative adjustment of the compensation force at each compensation point in the hybrid compensation execution array until the deviation between the foundation displacement field data and the predicted settlement field distribution converges to a preset allowable threshold.

[0069] Step 7: After the tunnel boring machine passes through the preset influence area, the compensation force of the hybrid compensation execution array is gradually released according to the preset unloading gradient, and the basic rebound amount is monitored to evaluate the compensation effect.

[0070] Specifically, the purpose of step one is to establish a multi-source sensing matrix that can continuously reflect the overall stress state and local deformation state of shallow foundation buildings. Before the tunnel boring machine enters the influence range of the building, it can obtain the initial foundation response information of the building foundation area in advance, and output the foundation displacement field data and foundation tilt angle data in real time during the continuous tunnel boring process. This provides continuous and stable data input for the subsequent stratum foundation coupled settlement prediction model, and provides real-time feedback basis for the hybrid compensation execution array.

[0071] In practice, based on the tunneling direction, tunneling depth, geological conditions, and the influence range of the tunnel, a pre-determined influence area corresponding to the shallow foundation building is established. This pre-determined influence area refers to the spatial range within which the tunneling process may cause stress redistribution and foundation settlement changes. A multi-source sensing matrix is ​​deployed within this area. The entire multi-source sensing matrix consists of multiple fiber optic grating sensors and multiple tilt sensors. These two types of sensors are responsible for acquiring vertical displacement information and foundation attitude change information, respectively. The two types of data complement each other, forming a dataset that describes the spatial deformation state of the foundation.

[0072] Multiple fiber Bragg grating sensors are deployed at predetermined intervals around the perimeter of the shallow foundation building's exterior wall foundation. Each fiber Bragg grating sensor is embedded in the soil layer below the foundation bottom along the longitudinal direction of the foundation, ensuring that the sensor's sensitive axis aligns with the main settlement direction of the foundation. When foundation settlement occurs, the surrounding soil undergoes compressive or tensile deformation simultaneously, and the fiber Bragg grating experiences axial strain, causing a change in its reflection center wavelength. The data receiving terminal calculates the vertical displacement at the corresponding location based on the change in reflected wavelength, thereby obtaining settlement data for each monitoring location.

[0073] For the The vertical displacement of a fiber Bragg grating sensor at a corresponding position can be expressed as:

[0074]

[0075] In the formula: Indicates the first Vertical displacement at each monitoring location; This indicates the change in the center wavelength of the corresponding fiber grating reflection; This represents the fiber grating displacement conversion coefficient, which indicates the vertical displacement corresponding to a unit wavelength change.

[0076] Using the above relationship, discrete displacement values ​​obtained from multiple fiber Bragg grating sensors can be combined into a set of vertical displacement data for each position of the foundation.

[0077] To comprehensively reflect changes in the overall attitude of the foundation, tilt sensors are installed at each corner of the shallow foundation building. Since the corners are typically the most sensitive locations to overall tilt and uneven settlement, they can reflect changes in the overall stress on the foundation at an earlier stage. Each tilt sensor is fixed to the upper surface of the foundation, with its sensing direction corresponding to two orthogonal directions of the building. This allows for continuous acquisition of the foundation's tilt angle changes in both directions, generating foundation tilt angle data. When differential settlement occurs in the foundation, the tilt angle at each corner will change, and the data from multiple tilt sensors can reflect the overall spatial attitude changes of the foundation.

[0078] Fiber Bragg grating sensors acquire information on the local vertical displacement of the foundation, reflecting the degree of settlement at different locations. Tilt sensors acquire information on the overall attitude change of the foundation, reflecting whether the foundation is tilting and the direction of tilt. Since these two types of data correspond to local deformation and overall deformation respectively, they can jointly describe the spatial deformation state of shallow foundation buildings, avoiding the situation where a single monitoring data is insufficient to fully reflect the stress characteristics of the foundation.

[0079] All fiber Bragg grating sensors and tilt sensors are connected to the data receiving terminal via data transmission cables. The data receiving terminal synchronously receives data output from each sensor according to a unified sampling period and establishes a correspondence based on the unique number of each sensor, organizing data from different sources according to spatial location. Each monitoring location corresponds to a unique spatial coordinate, and the data collected at each moment corresponds to a unique time stamp, thereby constructing a multi-source sensing matrix that includes spatial and temporal dimensions.

[0080] The system is deployed in total. A fiber grating sensor and If there are several tilt sensors, the data generated at any sampling time can be expressed as:

[0081]

[0082] In the formula: Indicates the sampling time The corresponding multi-source sensing matrix; to This represents the ground displacement field data corresponding to each fiber Bragg grating sensor; to This represents the basic tilt angle data corresponding to each tilt sensor.

[0083] This multi-source sensing matrix not only saves the data values ​​of each monitoring location, but also preserves the spatial correspondence between the data. This allows subsequent steps to directly complete settlement prediction, compensation force calculation, and deviation analysis based on the mapping relationship between the monitoring location and the spatial grid nodes, without having to re-establish the spatial correspondence, thus improving the efficiency of subsequent data processing.

[0084] As the tunnel boring machine (TBM) continues to advance, the multi-source sensing matrix is ​​continuously updated according to a preset sampling frequency. Each sampling generates new ground displacement field data and foundation tilt angle data. Since all monitoring data are acquired synchronously according to a unified time reference, it can accurately reflect the development process of shallow foundation building foundation deformation during TBM advancement. This provides continuous dynamic input data for the ground foundation coupled settlement prediction model in step two, and provides real-time feedback data for deviation calculation in step five and closed-loop iterative adjustment in step six. This ensures that the entire intelligent active compensation process is always based on real-time sensing, achieving continuous tracking and dynamic monitoring of foundation deformation status.

[0085] Step two is based on the multi-source sensing matrix formed in step one. Its purpose is to use shield tunneling parameters, stratum physical and mechanical parameters, and the foundation displacement field data and foundation tilt angle data obtained in real time in step one to construct a stratum-foundation coupled settlement prediction model that can reflect the coupling relationship between shield tunneling, stratum disturbance and shallow foundation building settlement. This model can predict the possible settlement of shallow foundation buildings at various locations within a preset time window in advance, form a predicted settlement field distribution, and provide a basis for the compensation force calculation in the subsequent step three. This realizes the transformation from passive compensation after settlement occurs to active pre-compensation before settlement occurs.

[0086] In practice, the data receiving terminal continuously receives the multi-source sensing matrix generated in step one while simultaneously acquiring real-time construction parameters output by the tunnel boring machine (TBM) construction control system. These real-time construction parameters include thrust parameters, surface pressure parameters, and tunneling rate parameters. The thrust parameter reflects the axial force exerted by the cutterhead and shield on the ground ahead during TBM advancement; the surface pressure parameter reflects the earth pressure balance within the excavation chamber, directly affecting the stability of the tunnel face and the stress release level of the surrounding strata; and the tunneling rate parameter represents the distance the TBM advances per unit time, determining the disturbance propagation speed and the time process of settlement development. All three parameters are dynamic input variables that change continuously with TBM tunneling, reflecting the real-time construction status of the TBM.

[0087] Simultaneously, the geological physical and mechanical parameters obtained during the engineering survey phase are imported into the prediction model. These parameters include soil compression modulus, Poisson's ratio, internal friction angle, and cohesion. Specifically, the soil compression modulus describes the soil's ability to resist compressive deformation; Poisson's ratio describes the relationship between lateral and longitudinal deformation; the internal friction angle reflects the frictional characteristics between soil particles; and cohesion reflects the cohesive capacity between soil particles. Since these parameters remain relatively stable during shield tunneling, they are used as static geological parameters in the prediction model calculations to determine the deformation response characteristics corresponding to different geological strata.

[0088] To ensure that the prediction results not only reflect the tunnel boring machine's construction conditions but also the actual response state of the building foundation, the multi-source sensing matrix formed in step one is synchronously used as the state input to the prediction model. Specifically, the foundation displacement field data reflects the settlement already occurring at each monitored location of the foundation, while the foundation tilt angle data reflects the overall attitude change trend of the foundation. The prediction model continuously corrects its current calculation state using the aforementioned real-time monitoring data, ensuring that the prediction process is always based on the latest monitoring information, rather than solely relying on theoretical models for extrapolation. This improves the consistency between the predicted settlement field distribution and the actual development trend.

[0089] In the model construction process, the finite element numerical simulation model is integrated with the long short-term memory neural network to form a coupled settlement prediction model for the foundation. The finite element numerical simulation model is used to establish the physical relationship between the redistribution of ground stress, soil deformation, and changes in foundation stress caused by tunnel boring machine (TBM) excavation. Based on the current position of the TBM, TBM construction parameters, and the physical and mechanical parameters of the ground, it calculates the changes in the stress and displacement fields within the current ground, describing the impact of TBM construction on shallow foundation building foundations from a spatial perspective.

[0090] The foundation settlement response output by the finite element model can be expressed as:

[0091]

[0092] In the formula: This represents the foundation settlement response calculated using the finite element model. This represents the dynamic input variables consisting of thrust parameters, surface pressure parameters, and tunneling rate parameters; This represents static geological parameters consisting of soil layer compression modulus, Poisson's ratio, internal friction angle, and cohesion. This indicates the finite element numerical calculation process.

[0093] The finite element method (FEM) calculation results described above can reflect the settlement development trend at various locations under different construction parameters and geological conditions. However, since shield tunneling is a continuous dynamic process, relying solely on FEM calculations is insufficient to fully reflect the temporal continuity of settlement development. Therefore, the FEM calculation results and the multi-source sensing matrix obtained in real time in step one are input into a long short-term memory (LSTM) neural network, enabling the network to learn the law of continuous settlement change over time during shield tunneling.

[0094] Long Short-Term Memory (LSTM) neural networks correlate and analyze the current construction status, monitoring status, and finite element analysis results from multiple historical moments to extract the temporal evolution characteristics of settlement development. Because the network has the ability to remember historical states, it can identify the settlement growth rate, settlement change trends, and the continuous impact of changes in tunnel boring machine (TBM) construction parameters on subsequent settlement development, thereby enabling the prediction of settlement status within a preset time window.

[0095] The fusion prediction result is formed by the combined effect of the spatial mechanical relationships provided by the finite element model and the temporal evolution relationships provided by the long short-term memory neural network. The fusion prediction result retains the ability of the finite element model to describe the geological mechanical mechanism and has the learning ability of the long short-term memory neural network to learn about complex nonlinear temporal changes. Therefore, it can more accurately reflect the development trend of shallow foundation settlement of buildings during shield tunneling.

[0096] Subsequently, a pre-defined spatial grid is established based on the foundation area of ​​the shallow foundation building, dividing the entire foundation area into multiple spatial grid nodes. Each spatial grid node corresponds to one or more monitoring locations from step one. After the prediction model completes its calculations, the predicted settlement values ​​corresponding to each spatial grid node within a future preset time window are output.

[0097] The basic region is divided into: With a preset spatial grid of nodes, the predicted settlement field distribution can be represented as:

[0098]

[0099] In the formula: This indicates the predicted subsidence field distribution; to These represent the predicted settlement values ​​corresponding to each preset spatial grid node within a future preset time window.

[0100] The predicted settlement field distribution described above not only includes the potential future settlement of each spatial grid node but also preserves the spatial relationships between the nodes. Therefore, it can comprehensively describe the future settlement development of the entire shallow foundation building foundation area, rather than just the settlement value at a single point. The differences in predicted settlement values ​​between different spatial grid nodes directly reflect the future trend of differential settlement and provide a direct data basis for step three to calculate the differential settlement gradient at each compensation point, determine the magnitude of the compensation force, and determine the application sequence.

[0101] As the tunnel boring machine (TBM) advances, the multi-source sensing matrix generated in step one is continuously updated, the TBM construction parameters change continuously, and the ground foundation coupled settlement prediction model synchronously re-executes prediction calculations according to the preset calculation cycle, continuously outputting new predicted settlement field distributions. This ensures that the preset time window is always updated synchronously with the TBM's progress. This forms a continuous dynamic prediction mechanism, allowing subsequent compensation to be adjusted based on the latest prediction results, achieving proactive prediction and control of settlement in shallow foundation buildings.

[0102] Step three is based on the multi-source sensing matrix formed in step one and the predicted settlement field distribution output in step two. Its purpose is to further convert the predicted settlement field into an executable active compensation scheme. Since the damage to the building structure mainly comes from differential settlement rather than overall synchronous settlement, this step does not directly determine the compensation strategy based on the predicted settlement value. Instead, it analyzes the settlement change trend between different spatial locations in the predicted settlement field, calculates the differential settlement gradient within the foundation area, and determines the magnitude of the compensation force to be applied at each compensation point and the timing of compensation initiation based on the differential settlement gradient. This allows the hybrid compensation execution array to apply compensation before the stratum deformation is transmitted to the foundation, thus actively offsetting any future uneven settlement.

[0103] In practice, the predicted settlement field distribution output in step two has been divided into multiple preset spatial grid nodes according to the basic area. Each preset spatial grid node corresponds to a fixed spatial location within the basic area and maintains a unified spatial coordinate system with the multi-source sensing matrix in step one. Therefore, each spatial grid node can form a one-to-one correspondence with its corresponding monitoring location. The system first reads the predicted settlement values ​​corresponding to each preset spatial grid node within a future preset time window and establishes a settlement change sequence according to the spatial connection relationship between adjacent spatial grid nodes.

[0104] Subsequently, settlement difference calculations are performed on every two adjacent pre-set spatial grid nodes. The settlement difference refers to the change between the predicted settlement values ​​of two adjacent spatial grid nodes, reflecting whether the settlement variation in a local area is uniform. When the predicted settlement values ​​of two adjacent spatial grid nodes are close, it indicates that the area is settling synchronously with minimal foundation bending. When the difference between the predicted settlement values ​​of the two spatial grid nodes gradually increases, it indicates that the foundation will experience significant bending deformation locally, and this area is the key location requiring subsequent compensation.

[0105] The settlement difference between adjacent preset spatial grid nodes can be expressed as:

[0106]

[0107] In the formula: Indicates adjacent preset spatial grid nodes and The difference in settlement between them; Indicates the first Predicted settlement values ​​corresponding to each preset spatial grid node; Indicates the first The predicted settlement value corresponding to each preset spatial grid node.

[0108] The aforementioned settlement difference only represents the amount of settlement change between adjacent locations, but not the rate of change. Therefore, the differential settlement gradient is further calculated by combining the horizontal distance between adjacent preset spatial grid nodes. The differential settlement gradient represents the degree of settlement change per unit horizontal distance. The larger the value, the more severe the local deformation of the foundation, and the more obvious the adverse impact on the overall structure of the building.

[0109] Differential settlement gradient can be expressed as:

[0110]

[0111] In the formula: This represents the differential settlement gradient between adjacent preset spatial grid nodes; This represents the settlement difference between corresponding nodes; This indicates the horizontal distance between adjacent preset spatial grid nodes.

[0112] By calculating the differential settlement gradients between all adjacent spatial grid nodes in the entire predicted settlement field, a complete differential settlement gradient distribution within the foundation area is formed. The system further analyzes multiple differential settlement gradients within the neighborhood of each compensation point based on its spatial location, thereby determining the degree of local deformation corresponding to each compensation point. If a large differential settlement gradient continuously appears around a compensation point, it indicates a significant future foundation bending trend in that area, and active compensation should be prioritized. If a differential settlement gradient is small in a certain area, it indicates that the foundation as a whole tends to settle synchronously, and only a small compensation force is needed to maintain foundation stability.

[0113] After calculating the differential settlement gradient, the system combines the structural stiffness of the foundation itself to calculate the magnitude of the compensation force required at each compensation point. Foundation stiffness reflects the foundation's ability to resist bending deformation. When the foundation stiffness is large, a larger compensation force is required to offset the same differential settlement gradient; when the foundation stiffness is small, the corresponding compensation force requirement decreases. The foundation stiffness is the equivalent stiffness parameter for the bending deformation resistance of shallow foundation building foundations, which can be determined based on the elastic modulus of the foundation material. and moment of inertia of cross section and the equivalent span or calculated length between adjacent compensation points. Calculated, i.e. For existing buildings, the data can be obtained directly from design or survey data. Therefore, the compensation force depends not only on the settlement gradient but also on the mechanical properties of the foundation structure itself.

[0114] The magnitude of the compensating force corresponding to each point of application can be expressed as:

[0115]

[0116] In the formula: This indicates the magnitude of the compensating force corresponding to the point of application of the compensation. Indicates the foundation stiffness; This represents the differential settlement gradient of the area corresponding to the compensation point.

[0117] The calculated compensation force only represents the magnitude of the compensation force; its direction is determined based on the direction of the differential settlement gradient. When the predicted settlement in a certain area of ​​the foundation is greater than that in adjacent areas, the compensation force is directed towards the uplift direction of the foundation, generating a reverse moment opposite to the bending direction of the foundation to counteract potential future differential settlement. When the predicted settlement on the other side of the foundation is larger, the direction of the compensation force is adjusted synchronously, so that each compensation point forms a stress state opposite to the predicted deformation trend, thereby reducing the overall bending degree of the foundation and maintaining a more uniform stress state.

[0118] Since tunnel boring is a continuous process, the time at which different locations are disturbed by the tunnel boring machine (TBM) is not consistent. Therefore, compensation points cannot be activated simultaneously; instead, the optimal compensation timing must be determined based on the TBM's advancing position. The system reads the TBM tunneling rate parameters and the current TBM position synchronously acquired in step two in real time, and calculates the estimated time for the TBM disturbance to propagate to the corresponding area by combining this information with the geological location of each compensation point.

[0119] Let the distance from the current position of the tunnel boring machine to the corresponding stratum position of a certain compensation point be . The current tunneling rate of the shield is The estimated time for the tunnel boring machine to reach the corresponding location can be expressed as:

[0120]

[0121] In the formula: Indicates the application timing corresponding to the compensation application point; This indicates the distance from the current position of the tunnel boring machine to the corresponding stratum position of the compensation point; This indicates the current tunneling rate of the shield machine.

[0122] The system uses the calculated time as the application sequence of the corresponding compensation points. When the tunnel boring machine's (TBM) advancing speed changes, the application sequence is updated synchronously. When the TBM's speed decreases or increases during advancement, the activation time of each compensation point is advanced or delayed accordingly, ensuring that the compensation action is always synchronized with the propagation process of the TBM disturbance. This prevents situations where compensation is too early, leading to premature attenuation of the compensation force, or where compensation is too late, resulting in settlement already occurring.

[0123] Finally, the system generates compensation control information for each compensation point, including compensation location, compensation force magnitude, compensation direction, and application timing, and establishes a compensation point control sequence based on spatial location. This control sequence maintains consistency with the spatial coordinates of the multi-source sensing matrix in step one, enabling subsequent feedback analysis of the compensation effect using real-time monitoring data. Furthermore, it keeps synchronized with the predicted settlement field distribution in step two. When the predicted settlement field changes, the system recalculates the differential settlement gradient, compensation force magnitude, and application timing, allowing the entire compensation scheme to be dynamically adjusted according to the tunnel boring machine's construction status.

[0124] Thus, step three realizes the conversion of the predicted settlement field into compensation control parameters, further quantifying the future settlement development trend obtained in step two into the magnitude and timing of the compensation force that the hybrid compensation execution array can directly execute, providing a complete control basis for the active compensation in step four, while ensuring that the compensation action is always based on the combined effect of the real-time monitoring data in step one and the dynamic prediction results in step two.

[0125] Step four is based on the multi-source sensing matrix formed in step one, the predicted settlement field distribution output in step two, and the compensation control information generated in step three. Its purpose is to convert the magnitude and timing of the compensation force calculated in step three into actual compensation actions. By deploying a hybrid compensation execution array under the foundation of shallow foundation buildings, compensation forces are actively applied to each compensation point, so that the foundation forms a stress state opposite to the predicted differential settlement before the shield tunneling disturbance arrives. This weakens the foundation deformation trend caused by shield tunneling in advance and provides an adjustable execution basis for the real-time feedback correction in the subsequent step five.

[0126] In practical implementation, based on the compensation action point control sequence established in step three, a hydraulic jack and a shape memory alloy actuator are installed at each compensation action point below the shallow foundation of the building. These two actuators are arranged side-by-side along the bottom surface of the foundation and connected to the same stress position on the foundation bottom surface, ensuring that both actuators act on the same compensation action point. The spatial positional relationship between each compensation action point remains consistent with the preset spatial grid established in step two, thereby ensuring that the compensation location maintains a unified coordinate system with the predicted settlement field distribution and the multi-source sensing matrix formed in step one, enabling a spatial correspondence between the monitoring location, the predicted location, and the compensation location.

[0127] Among them, the hydraulic jack, as the main load-bearing actuator, has a large output force and can quickly establish the main support force required for the foundation; the shape memory alloy actuator, which uses the phase change characteristics of shape memory alloy material to achieve mechanical drive, serves as an auxiliary adjustment actuator with high output displacement resolution, enabling continuous micro-adjustment of the compensation force. When both act together at the same compensation point, they can meet the overall compensation requirements of the foundation while also taking into account local compensation accuracy, giving the entire hybrid compensation actuator array both large compensation capacity and high adjustment accuracy.

[0128] The compensation force calculated in step three is not directly borne by a single actuator, but is decomposed into a basic compensation force component and a fine adjustment compensation force component according to the characteristics of different actuators. Among them, the basic compensation force component undertakes the majority of the compensation task and is used to quickly establish the overall reverse support of the foundation; the fine adjustment compensation force component is used to eliminate local stress errors generated during the foundation compensation process, making the compensation effect more gradual and continuous, and avoiding new stress concentrations in local areas of the foundation.

[0129] Let the magnitude of the compensating force corresponding to a certain point of application of the compensation be calculated in step three. Then it satisfies:

[0130]

[0131] In the formula: This indicates the magnitude of the compensating force corresponding to the point of application of the compensation. Indicates the basic compensation force components; This indicates the fine adjustment of the compensation force component.

[0132] The basic compensation force component is provided by a hydraulic jack, while the fine adjustment compensation force component is provided by a shape memory alloy actuator. Together, they form the target compensation force determined in step three.

[0133] After the compensation force decomposition is completed, each compensation point is activated sequentially according to the application timing calculated in step three. Before the tunnel boring machine advances to the influence range of the corresponding compensation point, the system first controls the hydraulic jacks to start working. The hydraulic jacks slowly output the foundation compensation force component vertically along the foundation, so that the bottom surface of the foundation gradually forms a reverse support force. Because the compensation process adopts a gradual loading method, the stress state of the foundation can change continuously, avoiding sudden changes in local stress of the foundation caused by instantaneous loading of the compensation force.

[0134] As the basic compensation force components are gradually established, the overall stress state of the foundation has basically reached the target level calculated in step three. At this point, the system further activates the shape memory alloy actuator. The shape memory alloy actuator continues to apply finely adjustable compensation force components to the bottom surface of the foundation through its own length changes, refining the compensation force established by the hydraulic jack, making the compensation effect closer to the target compensation force calculated in step three.

[0135] Since both actuators act on the same compensation point, they form a shared stress state throughout the compensation process. The hydraulic jack continuously bears the main compensation load, while the shape memory alloy actuator continuously performs small-scale compensation corrections. Both maintain synchronized output, ensuring that the actual compensation force applied to the foundation bottom is equal to the compensation force determined in step three. Simultaneously, the compensation direction is always opposite to the differential settlement gradient direction determined in step three, thus forming a reverse support effect within the foundation to counteract future differential settlement development.

[0136] As the tunnel boring machine (TBM) continues to advance, the application timing of different compensation points arrives sequentially. Each compensation point enters the working state one by one according to the control sequence, rather than all starting simultaneously. Compensation points closer to the current position of the TBM establish compensation force first, while compensation points farther away remain in standby mode. As the TBM advances further, subsequent compensation points are activated sequentially, causing the entire hybrid compensation execution array to form an active compensation area that moves continuously along the TBM's advance direction. This ensures that the compensation effect is always concentrated in the area where the TBM disturbance is most obvious, while also reducing ineffective compensation and improving compensation efficiency.

[0137] Simultaneously, the multi-source sensing matrix established in step one continuously collects foundation displacement field data and foundation tilt angle data at each monitoring location; step two continuously updates the predicted settlement field distribution; and step three synchronously updates the magnitude and application sequence of the compensation force corresponding to each compensation point. Once new compensation control information is generated, the hybrid compensation execution array immediately adjusts the current output state of each compensation point based on the updated control information. This ensures that the hydraulic jacks already in operation continue to maintain their main compensation function, the shape memory alloy actuators synchronously adjust and finely regulate the compensation force components, and the compensation points that have not yet been activated re-determine their activation time according to the new application sequence. This ensures that the entire compensation execution process remains synchronized with the tunnel boring machine's construction status and the foundation settlement development trend.

[0138] Through the above execution process, each compensation point can complete active compensation according to the compensation force magnitude and application sequence determined in step three. A reverse force state corresponding to the predicted differential settlement is established in advance within the foundation, ensuring that the settlement trend caused by the ground disturbance generated during shield tunneling is actively offset. Thus, step four completes the conversion from the predicted compensation scheme to the actual compensation action, and unifies the real-time monitoring of step one, the dynamic prediction of step two, and the compensation control of step three within the hybrid compensation execution array. This provides a stable and continuous execution target for step five, which involves compensation correction based on real-time monitoring results.

[0139] Step 5 is based on the multi-source sensing matrix formed in Step 1, the predicted settlement field distribution output in Step 2, the compensation control information generated in Step 3, and the hybrid compensation execution array implemented in Step 4. Its purpose is to obtain the actual deformation state of the foundation after active compensation in real time, continuously compare the actual settlement state with the predicted settlement state, calculate the settlement deviation generated during the compensation process, and generate the compensation correction amount accordingly. This provides real-time feedback for the closed-loop iterative adjustment in Step 6, and transforms the entire active compensation process from predictive control to closed-loop control that combines prediction and feedback.

[0140] In practice, during step four, each compensation point begins outputting compensation force according to its corresponding application sequence, and the hybrid compensation execution array continuously applies reverse support to the foundation bottom surface. Simultaneously, the multi-source sensing matrix established in step one maintains continuous operation, synchronously acquiring ground displacement field data output by each fiber optic grating sensor and foundation tilt angle data output by each tilt sensor according to a preset sampling frequency. Since the multi-source sensing matrix and the preset spatial grid established in step two use a unified spatial coordinate system, each set of real-time monitoring data can directly correspond to the corresponding spatial location in the predicted settlement field, ensuring a unified data foundation for both the prediction and actual monitoring results.

[0141] The system continuously outputs vertical displacement data at different locations on the foundation, while the tilt sensors continuously output information on the overall attitude change of the foundation. The ground displacement field data reflects the actual settlement at each location on the foundation, and the foundation tilt angle data determines whether the foundation still exhibits a sustained tilting trend. When local settlement is controlled but the overall tilt continues to increase, it indicates that the distribution of compensation force in some areas remains uneven. The foundation tilt angle data helps determine the development direction of the actual settlement field, providing a reference for subsequent compensation corrections. This ensures that the feedback analysis not only focuses on local settlement changes but also considers overall foundation attitude changes.

[0142] As the monitoring data is continuously updated, the system maps the real-time collected foundation displacement field data to the corresponding preset spatial grid nodes based on the preset spatial grid established in step two. Since each fiber optic grating sensor corresponds to a fixed monitoring position, the vertical displacement data collected at each monitoring position can be converted into the actual settlement value of the corresponding preset spatial grid node, thus forming the actual foundation settlement field at the current moment.

[0143] The basic region is divided into: Given a preset spatial grid of nodes, the actual settlement field formed at the current moment can be represented as:

[0144]

[0145] In the formula: This indicates the actual settlement field of the foundation at the current moment; to These represent the actual settlement values ​​corresponding to each preset spatial grid node.

[0146] After establishing the actual settlement field, the system further retrieves the predicted settlement field distribution at the same time point output in step two and compares it one by one according to the spatial grid nodes. Since the predicted settlement field and the actual settlement field have the same spatial nodes, there is no need to re-establish the spatial mapping relationship, and the deviation can be directly calculated for each node.

[0147] For each preset spatial grid node, the system calculates the difference between the actual settlement value and the predicted settlement value of the corresponding node, and uses this difference as the settlement deviation value.

[0148] The settlement deviation value can be expressed as:

[0149]

[0150] In the formula: Indicates the first Settlement deviation values ​​corresponding to each preset spatial grid node; Indicates the first The actual settlement value corresponding to each preset spatial grid node; Indicates the first The predicted settlement value corresponding to each preset spatial grid node.

[0151] The settlement deviation value mentioned above can directly reflect the difference between the actual deformation state and the predicted deformation state of the foundation after active compensation. When the settlement deviation value is close to zero, it means that the compensation force implemented in step four is basically consistent with the target compensation force calculated in step three, and the actual settlement state of the foundation is consistent with the predicted state. When the settlement deviation value gradually increases, it means that a deviation has occurred in the compensation process, and further adjustment of the compensation output is required.

[0152] When the settlement deviation value is greater than zero, it means that the actual settlement value is greater than the predicted settlement value, that is, the actual settlement of the foundation exceeds the prediction result, indicating that the current compensation effect is insufficient and the compensation force provided by the hybrid compensation execution array is too small. The compensation force output at the corresponding compensation point should be further increased. When the settlement deviation value is less than zero, it means that the actual settlement value is less than the predicted settlement value, that is, the current compensation effect has exceeded the target requirement, indicating that the compensation force output is too large. The compensation force at the corresponding compensation point should be appropriately reduced to avoid excessive local lifting of the foundation.

[0153] To convert settlement deviation into control parameters that can directly affect the hybrid compensation execution array, the system calculates the compensation correction amount corresponding to each compensation point based on a preset proportional coefficient. This preset proportional coefficient establishes the correspondence between settlement deviation values ​​and compensation correction amounts, ensuring that a larger settlement deviation corresponds to a larger compensation correction amount, and a smaller settlement deviation corresponds to a smaller compensation correction amount, thereby guaranteeing that the compensation adjustment range remains consistent with the actual deviation.

[0154] The compensation correction amount can be expressed as:

[0155]

[0156] In the formula: Indicates the amount of compensation correction; Indicates the preset scaling factor; This represents the settlement deviation value of the area where the corresponding compensation point is located.

[0157] Specifically, a positive compensation correction indicates that the current output compensation force at the compensation point needs to be increased; a negative compensation correction indicates that the current output compensation force at the compensation point needs to be decreased. Since the compensation correction directly originates from actual monitoring data, it accurately reflects the current foundation stress state, rather than relying solely on theoretical predictions for adjustment, thus enabling the entire compensation control process to have real-time feedback capabilities.

[0158] To ensure that the compensation correction amount accurately corresponds to each compensation point, the system, based on the compensation point control sequence established in step three, performs correlation analysis on the settlement deviation values ​​of the corresponding spatial grid nodes around each compensation point, and allocates the compensation correction amount formed in the corresponding area to the corresponding compensation point. When multiple preset spatial grid nodes within the influence range of a compensation point exhibit significant positive settlement deviations, the compensation correction amount corresponding to that compensation point increases synchronously; when multiple spatial grid nodes exhibit negative settlement deviations, the compensation correction amount decreases accordingly, ensuring that the compensation adjustment range remains consistent with the actual deformation area of ​​the foundation.

[0159] Meanwhile, the foundation tilt angle data from step one continues to participate in the feedback analysis. When the settlement deviation values ​​of the areas corresponding to multiple adjacent compensation points are small, but the foundation tilt angle continues to change, it indicates that there is still an unbalanced trend in the overall foundation stress. At this time, the system maintains the corresponding compensation correction amount to participate in the subsequent closed-loop adjustment, so that the local settlement control and the overall attitude control are consistent, and avoids ignoring the overall deformation trend of the foundation by only compensating based on the local settlement value.

[0160] As the tunnel boring machine continues to advance, the multi-source sensing matrix formed in step one continuously outputs new foundation displacement field data and foundation tilt angle data. Step two synchronously updates the predicted settlement field distribution. Step three continuously updates the compensation control information. Step four continuously implements active compensation. Step five continuously recalculates the settlement deviation values ​​of each preset spatial grid node and the compensation correction amount corresponding to each compensation action point, so that the entire feedback analysis process forms a continuous dynamic update mechanism. This provides a real-time, continuous, and spatially corresponding control basis for step six to implement closed-loop iterative adjustment based on the compensation correction amount, realizing the continuous integration of predicted information and real-time monitoring information in the active compensation process.

[0161] Step six is ​​based on the multi-source sensing matrix formed in step one, the predicted settlement field distribution output in step two, the compensation control information generated in step three, the hybrid compensation execution array implemented in step four, and the compensation correction amount calculated in step five. Its purpose is to use the compensation correction amount obtained in step five to continuously correct the compensation force currently output at each compensation point, so that the hybrid compensation execution array can continuously adjust the compensation output according to the real-time deformation state of the foundation until the deviation between the actual settlement field and the predicted settlement field gradually decreases and stabilizes within the preset allowable threshold, thereby forming a complete closed-loop active compensation process.

[0162] In practice, step five has already calculated the corresponding compensation correction amount for each compensation point. The system first reads the current output compensation force magnitude of each compensation point in step four, and simultaneously reads the compensation correction amount generated in step five. This compensation correction amount is then added to the current compensation force magnitude to generate a new target compensation force. Since the compensation correction amount itself includes both positive and negative values, it can increase or decrease the compensation force according to the actual deformation of the foundation, ensuring that the compensation adjustment continuously optimizes around the target compensation state determined in step three, rather than establishing a new compensation scheme.

[0163] The updated compensation force at each compensation point can be expressed as:

[0164]

[0165] In the formula: Indicates the current compensation point of action. The magnitude of the compensation force corresponding to the next iteration; This represents the compensation correction amount calculated in step five; This indicates the magnitude of the compensation force after the compensation correction.

[0166] When the compensation correction is positive, it indicates that the current compensation effect is insufficient, and the updated compensation force will increase accordingly; when the compensation correction is negative, it indicates that the current compensation effect is too large, and the updated compensation force will decrease accordingly. Through this continuous superposition method, the compensation force can gradually approach the optimal compensation state, without causing drastic changes in the foundation stress due to an excessively large adjustment in one step.

[0167] After the compensation force is updated, the system immediately controls the hybrid compensation actuator array to re-output the compensation force. The hydraulic jacks at each compensation point continue to perform the main compensation function, and their output basic compensation force component adjusts synchronously with the updated compensation force. The shape memory alloy actuator simultaneously and finely adjusts the compensation force component, allowing both actuators to jointly output a new compensation force magnitude, which continues to act on the foundation surface. Because the compensation adjustment is based on the compensation established in step four, the entire adjustment process is a continuous correction, rather than a reloading of the compensation force, resulting in a smoother change in foundation stress.

[0168] As the updated compensation force is reapplied, the actual stress state of the foundation changes accordingly. At this time, the multi-source sensing matrix established in step one continues to collect the latest foundation displacement field data and foundation tilt angle data at a preset sampling frequency, and maps the real-time foundation displacement field data to each preset spatial grid node in the same data processing method as in step five, and re-establishes the actual settlement field at the current moment.

[0169] Subsequently, the system calls upon the predicted settlement field distribution obtained from step two, which is continuously updated, and recalculates the corresponding settlement deviation value for each preset spatial grid node. Because the predicted settlement field distribution is updated synchronously during the tunnel boring machine's advancement, each deviation calculation corresponds to the current construction status, rather than using fixed prediction results for comparison, ensuring that the entire closed-loop adjustment is always based on the latest predicted status.

[0170] Let the first The settlement deviation value corresponding to the next iteration is:

[0171]

[0172] In the formula: This represents the settlement deviation value after updating the compensation force; This represents the actual settlement value after the compensation force is updated; This represents the predicted settlement value corresponding to the predicted settlement field at the same time.

[0173] By continuously recalculating the settlement deviation value, the system can determine in real time whether the compensation adjustment has achieved the expected effect. When the settlement deviation value decreases significantly after the compensation adjustment, it indicates that the compensation correction direction is correct and the actual settlement state of the foundation is gradually approaching the predicted state; if the settlement deviation value is still large, it indicates that the current compensation output is still insufficient or over-compensated, and the next round of compensation adjustment needs to be implemented.

[0174] After calculating the new settlement deviation value, the system performs threshold judgment on each preset spatial grid node. The system pre-sets a preset allowable threshold, which represents the maximum allowable deviation range between the actual settlement state and the predicted settlement state of the foundation. When the absolute value of the settlement deviation value of the corresponding preset spatial grid node is less than or equal to the preset allowable threshold, it indicates that the actual settlement at that location has met the active compensation requirements, and the corresponding compensation point does not need to be adjusted further. When the absolute value of the settlement deviation value is still greater than the preset allowable threshold, it indicates that the compensation at that location has not yet reached the target state, and it is necessary to continue calculating a new compensation correction amount and enter the next round of compensation adjustment.

[0175] The judgment condition can be expressed as:

[0176]

[0177] In the formula: This indicates the settlement deviation value of the area corresponding to the current compensation point; This indicates the preset allowable threshold.

[0178] When the above conditions are met, it indicates that the actual settlement field has a high degree of consistency with the predicted settlement field. The system maintains the current output state of the compensation point and no longer adjusts the compensation force at that compensation point. When the above conditions are not met, the system calls back step five, recalculates the compensation correction amount based on the latest settlement deviation value, and executes the compensation force update, compensation output, real-time monitoring and deviation calculation again to form a complete closed-loop iterative adjustment process.

[0179] Throughout the closed-loop adjustment process, the compensation points do not operate independently. Since the foundation is a continuously stressed structure, a change in the compensation force at one compensation point will alter the stress state of the corresponding area at adjacent compensation points. Therefore, after each compensation adjustment, the system does not merely update a single compensation point; instead, it reacquires the ground displacement field data and foundation tilt angle data for the entire foundation area, re-establishes the actual settlement field of the entire foundation area, and recalculates the settlement deviation values ​​corresponding to all preset spatial grid nodes. This ensures that all compensation points are always coordinated and adjusted according to the latest stress state of the entire foundation, thereby preventing new differential settlements caused by local compensation.

[0180] Meanwhile, Step 1 continuously outputs real-time monitoring data, Step 2 continuously updates the predicted settlement field distribution, Step 3 continuously keeps the compensation control information consistent with the current prediction state, Step 4 continuously executes compensation output, Step 5 continuously calculates the compensation correction amount, and Step 6 continuously completes the iterative cycle between compensation correction, compensation execution, monitoring feedback and deviation judgment, so that the entire intelligent active compensation system forms a closed-loop control mechanism with continuous linkage of four links: prediction, execution, monitoring and correction.

[0181] As the tunnel boring machine (TBM) advances, the system continues the aforementioned closed-loop iteration. When the settlement deviation values ​​corresponding to each preset spatial grid node in the foundation area all converge stably to within the preset allowable threshold, it indicates that the active compensation has made the actual settlement state of the foundation consistent with the predicted settlement state. The hybrid compensation execution array enters a stable holding state, no longer adjusting the compensation force, but only maintaining the current compensation output, and continues to receive real-time monitoring data from step one, providing a stable foundation stress state for the gradual release of compensation force after the TBM passes through the preset influence area in step seven.

[0182] Step 7 is based on the multi-source sensing matrix formed in Step 1, the predicted settlement field distribution output in Step 2, the compensation control information generated in Step 3, the active compensation implemented in Step 4, and the closed-loop iterative adjustment completed in Step 6. Its purpose is to gradually release the compensation force applied by the hybrid compensation execution array according to the preset unloading gradient after the shield tunneling leaves the main influence range of the shallow foundation building, so that the foundation stress state can smoothly transition from the active compensation state to the natural stress state. By continuously monitoring the foundation rebound amount, the entire active compensation process is evaluated to determine whether it has achieved the expected protection effect, thus providing the final evaluation basis for the structural safety of shallow foundation buildings.

[0183] In practice, after the closed-loop iterative adjustment is completed in step six, the compensation force output by each compensation point has been stably maintained at the corresponding target value, and the settlement deviation between the actual settlement field and the predicted settlement field of the foundation has been stably converged to within the preset allowable threshold. At this time, the hybrid compensation execution array continues to maintain the current compensation state. The system continuously reads the current position of the shield machine output by the shield machine construction control system, and, in conjunction with the preset influence area boundary established in step two, calculates the relative distance between the current position of the shield machine and the preset influence area boundary in real time to determine whether the shield machine has passed through the main influence area corresponding to the shallow foundation building.

[0184] When the shield tunneling is still within the preset influence area, it indicates that the shield tunneling is still causing disturbance to the foundation. The system maintains the compensation output state determined in step six and does not execute the compensation force release. When the shield tunneling has crossed the boundary of the preset influence area, it indicates that the main ground disturbance caused by the shield tunneling has gradually moved away from the foundation area, and the foundation enters the stress recovery stage. At this time, the system starts the compensation force unloading process.

[0185] To avoid instantaneous changes in foundation stress caused by a one-time release of the compensating force, the entire unloading process is carried out in stages. The system pre-sets a preset unloading gradient and divides the entire unloading process into multiple continuous unloading stages according to preset time intervals. At each unloading stage, the system reduces the current output compensating force at each compensation point according to a preset ratio, so that the foundation stress gradually returns to its natural stress state, without causing new settlement fluctuations or reverse deformation of the foundation due to the sudden disappearance of the compensating force.

[0186] Let the current compensating force at a certain point of application be... The uninstallation ratio corresponding to each uninstallation stage is: Then the compensating force after the unloading phase ends can be expressed as:

[0187]

[0188] In the formula: This indicates the compensation force at the point of application of the compensation at the start of the current unloading phase; This represents the compensating force after the current unloading phase is completed; This indicates the preset uninstallation ratio corresponding to the current uninstallation stage, and its value is greater than zero and less than one.

[0189] The above method ensures that the compensation force decreases gradually according to a fixed gradient, rather than dropping directly to zero, thus guaranteeing a continuous and smooth change in the foundation stress. Each compensation point releases its compensation force synchronously at the same unloading rhythm. The hydraulic jack gradually reduces the foundation compensation force component, while the shape memory alloy actuator synchronously reduces the fine adjustment compensation force component. Both actuators work together to complete the entire compensation force release process, gradually exiting the hybrid compensation state established in step four.

[0190] During each unloading phase, the multi-source sensing matrix established in step one continues continuous monitoring. Each fiber optic grating sensor continuously collects ground displacement field data at various locations on the foundation, and each tilt sensor continuously collects foundation tilt angle data. Based on the continuously collected ground displacement field data, the system calculates the current actual settlement value of each preset spatial grid node of the foundation and compares it with the actual settlement value of the corresponding node before unloading began, thereby obtaining the foundation rebound amount.

[0191] The basic springback represents the degree to which the foundation recovers its deformation during the gradual release of the compensating force, and can be expressed as:

[0192]

[0193] In the formula: Indicates the first The basic rebound amount corresponding to each preset spatial grid node; This indicates the actual settlement value of the node corresponding to the current unloading phase; This indicates the actual settlement value of the corresponding node before unloading begins.

[0194] The rebound amount reflects the recovery of the foundation after the compensating force is gradually released. When the rebound amount gradually approaches the theoretical recovery value, it indicates that the foundation has not produced significant residual deformation during the active compensation process, and the foundation stress recovery is relatively stable. If the rebound amount deviates significantly from the normal recovery trend, it indicates that there may still be local plastic deformation inside the foundation or that there may be overcompensation or undercompensation during the compensation process.

[0195] To further evaluate the compensation effect, the system combines the predicted settlement field distribution from step two with the foundation stress state after stable convergence from step six, and calculates the expected elastic recovery for each preset spatial grid node. The expected elastic recovery represents the theoretical deformation value that the foundation should recover after the compensation force is completely removed. It is compared with the foundation rebound value currently monitored to obtain the foundation recovery deviation.

[0196] The baseline recovery deviation can be expressed as:

[0197]

[0198] In the formula: Indicates the basic recovery deviation; This represents the baseline rebound amount obtained from actual monitoring; This indicates the expected elastic recovery at the corresponding location.

[0199] The system continuously calculates the foundation recovery deviation at each unloading stage and combines this with the foundation tilt angle data to determine the overall foundation attitude recovery. When the foundation rebound gradually approaches the expected elastic recovery amount, and the foundation tilt angle tends to stabilize, it indicates that the overall foundation recovery process is normal. If the foundation rebound meets the requirements but the foundation tilt angle continues to change, it indicates that the overall foundation still has an uneven recovery trend. The system continues to maintain the current unloading stage and does not enter the next stage until the overall foundation attitude tends to stabilize before releasing the compensation force to ensure that the local foundation recovery is consistent with the overall foundation recovery.

[0200] As each unloading stage is completed sequentially, the compensation force output at each compensation point gradually decreases until it is completely released, and the hybrid compensation execution array exits the working state. At this point, the system summarizes the foundation recovery deviations corresponding to each preset spatial grid node throughout the unloading process and performs a final evaluation based on a pre-set preset evaluation threshold. When the foundation recovery deviations corresponding to all preset spatial grid nodes are less than or equal to the preset evaluation threshold, it indicates that the active compensation process has effectively offset the differential settlement of the foundation caused by the tunnel boring machine, the foundation recovery status meets the expected requirements, the system determines that the compensation effect is qualified, and ends the active compensation process.

[0201] If the foundation recovery deviation corresponding to any preset spatial grid node exceeds the preset evaluation threshold, it indicates that there is a significant difference between the foundation recovery status in that area and the theoretical recovery state, and the system generates a compensation effect anomaly report. The compensation effect anomaly report includes the location of the abnormal compensation application point, the corresponding foundation recovery deviation, the foundation rebound amount change process, the foundation tilt angle change process, and the corresponding unloading stage information, which is used to completely record the operating status of each stage in the compensation process, providing a basis for subsequent analysis of the foundation recovery status and optimization of subsequent shield tunneling construction parameters.

[0202] Thus, step seven completes the smooth transition of the active compensation process from a stable compensation state to a natural stress state, achieving the orderly release of compensation force, continuous monitoring of the foundation recovery state, and final evaluation of the compensation effect. Combining the real-time sensing, dynamic prediction, active compensation, feedback correction, and closed-loop iteration mechanism formed in steps one through six, the entire intelligent active compensation method forms a complete protection process covering the entire shield tunneling process. It not only enables active pre-compensation before shield disturbance occurs but also smoothly releases the compensation effect after the shield passes and quantitatively evaluates the final protection effect, thereby improving the overall safety and compensation control accuracy of shallow foundation buildings during shield tunneling construction.

[0203] Example 2

[0204] The sensing module is used to deploy a multi-source sensing matrix composed of fiber optic grating sensors and tilt sensors around the shallow foundation houses in the preset influence area in front of the shield tunneling, so as to collect the foundation displacement field data and foundation tilt angle data of the shallow foundation houses in real time.

[0205] The prediction module is used to input the thrust parameters, surface pressure parameters, tunneling rate parameters, and stratum physical and mechanical parameters during the shield tunneling process into a pre-established stratum-foundation coupled settlement prediction model, and output the predicted settlement field distribution within a preset future time window.

[0206] The calculation module is used to calculate the differential settlement gradient at each compensation point based on the predicted settlement field distribution, and to determine the required compensation force and application sequence at each compensation point based on the differential settlement gradient.

[0207] An execution module is used to activate a hybrid compensation execution array deployed below the foundation of the shallow foundation house. The hybrid compensation execution array includes hydraulic jacks and shape memory alloy actuators, and applies compensation force to each compensation point according to the magnitude of the compensation force and the application sequence.

[0208] The feedback module is used to collect foundation displacement field data and foundation tilt angle data in real time using the multi-source sensing matrix during the application of compensation force, and to calculate the deviation between the foundation displacement field data and the predicted settlement field distribution to obtain the compensation correction amount.

[0209] The adjustment module is used to perform closed-loop iterative adjustment of the compensation force of each compensation point in the hybrid compensation execution array based on the compensation correction amount, until the deviation between the foundation displacement field data and the predicted settlement field distribution converges to a preset allowable threshold.

[0210] The unloading evaluation module is used to gradually release the compensation force of the hybrid compensation execution array according to the preset unloading gradient after the tunnel boring machine passes through the preset influence area, and monitor the basic rebound amount to evaluate the compensation effect.

[0211] The modules in Embodiment 2 are used to implement the functions in Embodiment 1. This embodiment can be implemented by a system including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement the active compensation method for shield tunneling under shallow foundation buildings according to Embodiment 1 of this application. The system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface. Their settings and functions are known in the art and will not be described in detail here.

[0212] In this application, the aforementioned memory can be any tangible medium that includes or stores a program that can be used or combined with an instruction execution system, system, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store required information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device. Any application or module described in this application can be implemented using computer-readable / executable instructions that can be stored or otherwise retained by such a computer-readable medium.

[0213] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for active compensation when a shield tunnel passes under a shallow foundation building, characterized in that, Includes the following steps: Step 1: Deploy a multi-source sensing matrix consisting of fiber optic grating sensors and tilt sensors around the shallow foundation houses in the preset influence area in front of the tunnel boring machine to collect the foundation displacement field data and foundation tilt angle data of the shallow foundation houses in real time. Step 2: Input the thrust parameters, surface pressure parameters, tunneling rate parameters, and stratum physical and mechanical parameters during the shield tunneling process into the pre-established stratum-foundation coupled settlement prediction model, and output the predicted settlement field distribution within the future preset time window; Step 3: Calculate the differential settlement gradient at each compensation point based on the predicted settlement field distribution, and determine the required compensation force and application sequence at each compensation point based on the differential settlement gradient. Step 4: Activate the hybrid compensation execution array deployed below the foundation of the shallow foundation house. The hybrid compensation execution array includes hydraulic jacks and shape memory alloy actuators, and applies compensation force to each compensation point according to the magnitude of the compensation force and the application sequence. Step 5: During the application of the compensation force, the multi-source sensing matrix is ​​used to collect the foundation displacement field data and foundation tilt angle data in real time, and the deviation between the foundation displacement field data and the predicted settlement field distribution is calculated to obtain the compensation correction amount. Step 6: Based on the compensation correction amount, perform closed-loop iterative adjustment of the compensation force at each compensation point in the hybrid compensation execution array until the deviation between the foundation displacement field data and the predicted settlement field distribution converges to a preset allowable threshold. Step 7: After the tunnel boring machine passes through the preset influence area, the compensation force of the hybrid compensation execution array is gradually released according to the preset unloading gradient, and the basic rebound amount is monitored to evaluate the compensation effect.

2. The active compensation method for shield tunneling under shallow foundation buildings according to claim 1, characterized in that, Step one includes: Multiple fiber optic grating sensors are installed at preset intervals around the perimeter of the outer wall foundation of the shallow foundation building. Each fiber optic grating sensor is buried in the soil layer below the bottom of the foundation along the longitudinal direction of the foundation to collect vertical displacement data at various locations of the foundation as ground displacement field data. Inclination sensors are installed at each corner of the shallow foundation building. Each inclination sensor is fixed to the upper surface of the foundation to collect the tilt angle data at each corner of the foundation as the foundation tilt angle data. The data collected by each fiber Bragg grating sensor and each tilt sensor are aggregated to the data receiving terminal through data transmission cables to form a multi-source sensing matrix.

3. The active compensation method for shield tunneling under shallow foundation buildings according to claim 1, characterized in that, Step two includes: The real-time thrust parameters, surface pressure parameters, and tunneling rate parameters during the shield tunneling process are obtained, as well as the physical and mechanical parameters of the strata obtained from the exploration, including soil compression modulus, Poisson's ratio, internal friction angle, and cohesion. Using thrust parameters, surface pressure parameters, and tunneling rate parameters during shield tunneling as dynamic input variables and geological physical and mechanical parameters as static geological parameters, a stratum-foundation coupled settlement prediction model is constructed. The stratum-foundation coupled settlement prediction model is a hybrid prediction model based on the fusion of finite element numerical simulation and long short-term memory neural network. The dynamic input variables and static geological parameters are input into the strata-foundation coupled settlement prediction model, and the settlement prediction values ​​of each preset spatial grid node within the future preset time window are output to form the predicted settlement field distribution.

4. The active compensation method for shield tunneling under shallow foundation buildings according to claim 1, characterized in that, Step three includes: For each preset spatial grid node in the predicted settlement field distribution, the settlement difference between adjacent preset spatial grid nodes is calculated, and the ratio of the settlement difference to the horizontal distance between adjacent preset spatial grid nodes is used as the differential settlement gradient. The required compensation force at each compensation point is calculated based on the product of the differential settlement gradient and the foundation stiffness. The direction of the compensation force is opposite to the direction of the differential settlement gradient, so as to generate a counteracting torque to counteract the differential settlement. Based on the tunnel boring machine's (TBM) excavation rate and the distance from the TBM's current position to each compensation point, the time it takes for the TBM to reach the corresponding stratum position at each compensation point is calculated, and this time is used as the application sequence.

5. The active compensation method for shield tunneling under shallow foundation buildings according to claim 1, characterized in that, Step four includes: At each compensation point below the foundation of a shallow foundation house, a hydraulic jack and a shape memory alloy actuator are installed. The hydraulic jack and the shape memory alloy actuator are arranged side by side and connected to the bottom of the foundation. The magnitude of the compensation force is decomposed into a basic compensation force component and a fine adjustment compensation force component, wherein the basic compensation force component is provided by a hydraulic jack and the fine adjustment compensation force component is provided by a shape memory alloy actuator. According to the application sequence, the hydraulic jack is first activated to apply the basic compensation force component, and then the shape memory alloy actuator applies the fine adjustment compensation force component. The resultant force of the basic compensation force component and the fine adjustment compensation force component is equal to the magnitude of the compensation force.

6. The active compensation method for shield tunneling under shallow foundation buildings according to claim 1, characterized in that, Step five includes: During the application of the compensation force, the foundation displacement field data and foundation tilt angle data are collected in real time using a multi-source sensing matrix at a preset sampling frequency. The foundation displacement field data are mapped to each preset spatial grid node in the predicted settlement field distribution to obtain the actual settlement value of each preset spatial grid node. For each preset spatial grid node, the difference between the actual settlement value and the predicted settlement value of the corresponding node in the predicted settlement field distribution is calculated, and the difference is used as the settlement deviation value. The compensation correction amount at each compensation point is calculated based on the settlement deviation value and the preset proportional coefficient. When the settlement deviation value is greater than zero, the compensation correction amount is positive to increase the compensation force, and when the settlement deviation value is less than zero, the compensation correction amount is negative to reduce the compensation force.

7. The active compensation method for shield tunneling under shallow foundation buildings according to claim 1, characterized in that, Step six includes: The compensation correction amount is added to the current compensation force magnitude of each compensation point to update the compensation force magnitude of each compensation point, and the hybrid compensation execution array is controlled to reapply the compensation force according to the updated compensation force magnitude. During the application of the updated compensation force, the foundation displacement field data is re-acquired using a multi-source sensing matrix, and the deviation between the foundation displacement field data and the predicted settlement field distribution is calculated to obtain the updated settlement deviation value. Determine whether the absolute value of the updated settlement deviation is less than or equal to the preset allowable threshold. If so, stop the iterative adjustment. If not, return to the step of adding the compensation correction amount to the current compensation force magnitude of each compensation point until the absolute value of the updated settlement deviation value is less than or equal to the preset allowable threshold.

8. The active compensation method for shield tunneling under shallow foundation buildings according to claim 1, characterized in that, Step seven includes: Based on the relative distance between the tunnel boring machine's location and the boundary of the preset influence zone, determine whether the tunnel boring machine has passed through the preset influence zone; After the tunnel boring machine passes through the preset influence area, the compensation force of the hybrid compensation execution array is gradually released in stages according to the preset unloading gradient. In each unloading stage, the compensation force is reduced by a preset proportion at a preset time interval. During each unloading phase, the base rebound amount is continuously monitored using a multi-source sensing matrix, and the deviation between the base rebound amount and the expected elastic recovery amount is calculated. After all unloading stages are completed, the compensation effect is evaluated based on the deviation between the final base rebound amount and the expected elastic recovery amount. If the deviation is within the preset evaluation threshold, the compensation effect is deemed qualified. If the deviation exceeds the preset evaluation threshold, an abnormal compensation effect report will be generated.

9. An active compensation system for shield tunneling under shallow foundation buildings, characterized in that, include: The sensing module is used to deploy a multi-source sensing matrix composed of fiber optic grating sensors and tilt sensors around the shallow foundation houses in the preset influence area in front of the shield tunneling, so as to collect the foundation displacement field data and foundation tilt angle data of the shallow foundation houses in real time. The prediction module is used to input the thrust parameters, surface pressure parameters, tunneling rate parameters, and stratum physical and mechanical parameters during the shield tunneling process into a pre-established stratum-foundation coupled settlement prediction model, and output the predicted settlement field distribution within a preset future time window. The calculation module is used to calculate the differential settlement gradient at each compensation point based on the predicted settlement field distribution, and to determine the required compensation force and application sequence at each compensation point based on the differential settlement gradient. An execution module is used to activate a hybrid compensation execution array deployed below the foundation of the shallow foundation house. The hybrid compensation execution array includes hydraulic jacks and shape memory alloy actuators, and applies compensation force to each compensation point according to the magnitude of the compensation force and the application sequence. The feedback module is used to collect foundation displacement field data and foundation tilt angle data in real time using the multi-source sensing matrix during the application of compensation force, and to calculate the deviation between the foundation displacement field data and the predicted settlement field distribution to obtain the compensation correction amount. The adjustment module is used to perform closed-loop iterative adjustment of the compensation force of each compensation point in the hybrid compensation execution array based on the compensation correction amount, until the deviation between the foundation displacement field data and the predicted settlement field distribution converges to a preset allowable threshold. The unloading evaluation module is used to gradually release the compensation force of the hybrid compensation execution array according to the preset unloading gradient after the tunnel boring machine passes through the preset influence area, and monitor the basic rebound amount to evaluate the compensation effect.

Images