A multi-source data fusion house foundation settlement early warning method and system

Abstract

The present application relates to the technical field of building health monitoring, and proposes a kind of house foundation settlement early warning method and system of multi-source data fusion, house foundation settlement early warning method includes obtaining pit support structure internal force response, surrounding soil seepage field and surface three-dimensional deformation etc.Multiple source monitoring data;Based on geological survey data, construct stratum structure model, map each kind of monitoring data to model layer position node according to spatial coordinates, form the stratum node state matrix driven by multi-source data;According to the settlement correlation of node under the coupling of multi-source data, construct the settlement propagation atlas of support-soil-groundwater three-field coupling, utilize edge weight to represent the transmission intensity of settlement influence between adjacent nodes;Finally, based on the settlement contribution degree of node and the safety degree of support structure, determine the risk level of house and trigger graded early warning.The present application realizes the depth fusion of monitoring data and stratum space, can accurately depict settlement evolution path, significantly improves the accuracy of house foundation settlement risk prediction.

Description

Technical Field

[0001] This invention relates to the field of building health monitoring technology, and in particular to a method and system for early warning of building foundation settlement through multi-source data fusion. Background Technology

[0002] The root causes of building foundation settlement originate during the foundation and substructure construction phases. However, existing detection technologies primarily focus on observing apparent settlement and identifying structural damage after the main building structure is completed. Monitoring methods rely mainly on single physical quantities, failing to synchronously correlate the internal force response of the support structure during the foundation pit construction period, the evolution of the seepage field in the surrounding soil, and the three-dimensional deformation of the ground surface. Because ground disturbances caused by foundation pit excavation have significant spatiotemporal transmission characteristics, and once construction-period settlement solidifies, it is difficult to reverse during the operational phase, traditional methods lack effective modeling of the coupling mechanism between the support, soil, and groundwater fields. This results in semantic gaps between multi-source heterogeneous monitoring data and the spatial structure of the geological strata, making it difficult to trace the causes of settlement anomalies and causing early warning responses to lag behind risk evolution. Especially under complex geological conditions, existing technologies cannot characterize the propagation path and intensity of settlement between strata nodes, hindering the transition from "cause localization" to "path prediction," thus limiting the improvement of building foundation quality from passive post-event detection to proactive pre-event early warning capabilities. Summary of the Invention

[0003] To address the aforementioned shortcomings, the present invention aims to propose a multi-source data fusion method and system for early warning of building foundation settlement. By constructing a geological structure model and a three-field coupled settlement propagation map, the method achieves deep semantic fusion of multi-source monitoring data at stratigraphic nodes, accurately depicts the evolution path and contribution of settlement risk, and improves the predictive ability and early warning accuracy of building quality monitoring.

[0004] To achieve this objective, the present invention adopts the following technical solution: A method for early warning of building foundation settlement through multi-source data fusion includes: Acquire the internal force response data of the foundation pit support structure of the target building, the seepage field monitoring data of the soil surrounding the foundation pit, and the three-dimensional deformation monitoring data of the ground surface surrounding the foundation pit; Based on the geological survey data of the foundation pit, a stratigraphic structure model is constructed. The internal force response data, the seepage field monitoring data, and the three-dimensional deformation monitoring data are mapped to the corresponding layer nodes of the stratigraphic structure model according to the stratigraphic spatial coordinates, forming a stratigraphic node state matrix driven by multi-source data. Based on the settlement correlation of each layer node in the stratigraphic node state matrix under the coupling effect of multi-source data, a settlement propagation map of the three fields of support-soil-groundwater coupling is constructed. The edge weight of the settlement propagation map represents the intensity of settlement influence transmission between adjacent layer nodes. Based on the settlement contribution of each node in the settlement propagation map and the safety of the support structure, the risk level of the building is determined and a graded early warning is triggered.

[0005] Preferably, the acquisition of internal force response data of the target building foundation pit support structure, seepage field monitoring data of the soil surrounding the foundation pit, and three-dimensional deformation monitoring data of the ground surface surrounding the foundation pit includes: Sensors are deployed on the foundation pit support components, retaining walls and anchor cable stress points to collect axial force, bending moment and prestress data of the support structure as internal force response data. The pore water pressure and groundwater level gradient within the influence range of the foundation pit excavation are obtained by using a piezometer array and a water level monitoring well as the seepage field monitoring data. The spatial displacement vector of the ground surface around the foundation pit within a preset sampling period is obtained using a three-dimensional laser scanning device or a GNSS monitoring array, and is used as the three-dimensional deformation monitoring data.

[0006] Preferably, constructing a stratigraphic structure model based on the geological survey data of the foundation pit includes: Extract borehole columnar sections from the geological survey data, obtain the physical and mechanical parameters of each soil layer, and use spatial interpolation algorithms to reconstruct the three-dimensional envelope surfaces of the interfaces of each soil layer in the foundation pit to divide the three-dimensional stratigraphic space. Based on the thickness distribution of each stratum in the three-dimensional stratigraphic space, rasterization is performed according to a preset step size to establish multiple stratigraphic nodes. The physical and mechanical parameters are mapped to the corresponding stratigraphic nodes to obtain the first stratigraphic structure model.

[0007] Preferably, mapping the internal force response data to the corresponding stratigraphic nodes of the stratigraphic structure model according to the stratigraphic spatial coordinates includes: Obtain the spatial coordinates of the internal force monitoring points corresponding to the internal force response data. The internal force monitoring points include axial force monitoring points arranged on the supporting components, bending moment monitoring points of the retaining wall, and prestress monitoring points of the anchor cable stress points. Calculate the three-dimensional spatial distance between each internal force monitoring point and each layer node in the first geological structure model; Based on the three-dimensional spatial distance, and following a weighted allocation strategy based on a spatial attenuation function, the internal force response data is mapped to one or more nearest layer nodes, which are then used as the internal force response components of the layer nodes. The weighted allocation strategy includes: calculating the mapping weight using an exponential decay function based on the distance between the internal force monitoring point and the layer node, and allocating the internal force response data of the internal force monitoring point to the adjacent layer node according to the mapping weight, thereby obtaining the force constraint attributes of each layer node.

[0008] Preferably, mapping the seepage field monitoring data to the corresponding stratigraphic nodes of the stratigraphic structure model according to the stratigraphic spatial coordinates includes: Obtain the spatial coordinates of the seepage monitoring points corresponding to the seepage field monitoring data, wherein the seepage monitoring points include the locations of the piezometer array and the water level monitoring wells; Extract the set of stratigraphic nodes in the first stratigraphic structure model that are located within the stratigraphic space of each seepage monitoring point; Based on the spatial coordinates of the seepage monitoring points and the spatial coordinates of each layer node, the seepage field monitoring data is distributed to each layer node in the layer node set using the inverse distance weighted interpolation algorithm to obtain the seepage field components of each layer node. The inverse distance weighted interpolation algorithm includes: for the j-th layer node, its seepage field component is obtained by weighted averaging of the monitoring values ​​of each seepage monitoring point within a preset neighborhood of the j-th layer node, and the weight is inversely proportional to the spatial distance from the seepage monitoring point to the j-th layer node, thereby obtaining the seepage driving attribute of each layer node.

[0009] Preferably, mapping the three-dimensional deformation monitoring data to the corresponding stratigraphic nodes of the stratigraphic structure model according to stratigraphic spatial coordinates includes: Obtain the horizontal spatial coordinates of the surface monitoring points corresponding to the three-dimensional deformation monitoring data; Based on the horizontal spatial coordinates, a target stratigraphic node is determined on the horizontal projection plane of the first stratigraphic structure model. The target stratigraphic node is the stratigraphic node whose horizontal coordinates are closest to the surface monitoring point and located in the uppermost stratigraphic space. The vertical and horizontal displacement components in the three-dimensional deformation monitoring data are vertically projected onto the target layer node to serve as the surface deformation components of the target layer node. If the same target stratum node corresponds to multiple surface monitoring points, the average value of the three-dimensional deformation monitoring data of the multiple surface monitoring points or the weighted average value according to the monitoring accuracy is used as the surface deformation component of the current target stratum node, and the displacement boundary attributes of each stratum node are obtained.

[0010] Preferably, based on the settlement correlation of each layer node in the stratigraphic node state matrix under the coupling effect of multi-source data, the settlement propagation map of the three fields coupled by support, soil, and groundwater is constructed, including: The spatial adjacency relationship of each stratigraphic node in the stratigraphic structure model is determined. The spatial adjacency relationship includes the horizontal adjacency relationship between adjacent stratigraphic nodes within the same stratigraphic space and the vertical adjacency relationship between stratigraphic nodes between adjacent stratigraphic spaces. For each pair of stratigraphic nodes with spatial adjacency, the multi-source state vectors of the source node and the target node are extracted based on the stratigraphic node state matrix, and the coupling difference between the source node and the target node in terms of force constraint attributes, seepage driving attributes and displacement boundary attributes is calculated. Based on the coupling difference, the ratio of the permeability coefficients of the strata to which the source node and the target node belong, the stiffness coupling coefficient of the support structure, and the spatial distance between the source node and the target node, the settlement propagation probability from the source node to the target node is calculated and used as the edge weight of the settlement propagation map. The subsidence propagation probability is compared with a preset propagation threshold. If the subsidence propagation probability is greater than the preset propagation threshold, a directed propagation edge is established between the source node and the target node. The direction of the directed propagation edge is from the subsidence impact source to the subsidence impact target. Traverse all pairs of layer nodes with spatial adjacency, and construct the settlement propagation map of the three-field coupling of support, soil and groundwater based on the directed propagation edge and its corresponding edge weight.

[0011] Preferably, comparing the settlement propagation probability with a preset propagation threshold includes: Set a first propagation threshold and a second propagation threshold, wherein the first propagation threshold is less than the second propagation threshold; If the subsidence propagation probability is not less than the second propagation threshold, then the stratigraphic node is determined. To the hierarchical node There exists a strong subsidence propagation path, and the directed propagation edge is marked as a strong propagation edge; If the subsidence propagation probability is not less than the first propagation threshold and less than the second propagation threshold, then the stratigraphic node is determined. To the hierarchical node There exists a moderate settlement propagation path, and the directed propagation edge is marked as a moderate propagation edge; If the subsidence propagation probability is less than the first propagation threshold, then the stratigraphic node is determined. To the hierarchical node There is no effective settlement propagation path between them, and no directed propagation edge is established.

[0012] Preferably, the first propagation threshold and the second propagation threshold satisfy the following relationship: ; ; in, Indicates the first propagation threshold. Indicates the second propagation threshold. Indicates the basic threshold. This indicates the current depth of the foundation pit excavation. Indicates the reference excavation depth. This indicates the equivalent stiffness of the enclosure wall. Indicates reference stiffness. and Indicates the depth correction factor. This represents the stiffness correction factor.

[0013] A multi-source data fusion-based early warning system for building foundation settlement includes: The data acquisition module is used to acquire the internal force response data of the foundation pit support structure of the target building, the seepage field monitoring data of the soil around the foundation pit, and the three-dimensional deformation monitoring data of the ground surface around the foundation pit. The data fusion module is used to construct a stratigraphic structure model based on the geological survey data of the foundation pit, and to map the internal force response data, the seepage field monitoring data and the three-dimensional deformation monitoring data to the corresponding layer nodes of the stratigraphic structure model according to the stratigraphic spatial coordinates, forming a stratigraphic node state matrix driven by multi-source data. The map construction module is used to construct a settlement propagation map of the support-soil-groundwater three-field coupling based on the settlement correlation of each layer node in the stratigraphic node state matrix under the coupling effect of multi-source data. The edge weight of the settlement propagation map represents the intensity of settlement influence transmission between adjacent layer nodes. The risk warning module is used to determine the risk level of the building and trigger a graded warning based on the settlement contribution of each node in the settlement propagation map and the safety of the support structure.

[0014] One of the above technical solutions has the following advantages or beneficial effects: This invention expands the data dimensions of building quality inspection from a single apparent displacement to a collaborative perception level of structural stress, groundwater seepage, and surface three-dimensional deformation by simultaneously collecting three types of heterogeneous monitoring data: internal forces of the foundation pit support structure, seepage field of the surrounding soil, and three-dimensional deformation of the ground surface. This overcomes the blind spots in tracing the causes caused by fragmented data sources in traditional monitoring methods. Subsequently, a stratigraphic structure model is constructed based on geological survey data, mapping the multi-source heterogeneous data to corresponding stratigraphic nodes according to stratigraphic spatial coordinates. This gives the discrete monitoring data stratigraphic semantic attribution and spatial consistency constraints, solving the semantic discontinuity problem between heterogeneous data and geological structure. Based on the settlement relationship of stratigraphic nodes under the coupling effect of multi-source data... A settlement propagation map coupling the support, soil, and groundwater fields was constructed using a correlation model. The transmission intensity of settlement influence between adjacent strata nodes was explicitly quantified using edge weights, transforming the originally implicit settlement evolution process into an interpretable propagation path. This achieved a leap from single-point anomaly alarm to chain-like propagation prediction. Finally, risk classification was performed by comprehensively evaluating the settlement contribution of strata nodes and the safety of the support structure. The trend of stratum deformation propagation was linked with the bearing redundancy of the support system for judgment. This enabled the early warning mechanism to take into account both soil disturbance evolution and support structure response, significantly improving the physical interpretability and engineering relevance of early warning for building foundation quality, and effectively supporting proactive prevention and control of building safety during the construction period. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0016] Figure 1 This is a flowchart of a multi-source data fusion method for early warning of building foundation settlement provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of the multi-source data fusion early warning system for building foundation settlement provided in an embodiment of the present invention. Detailed Implementation

[0017] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0018] In this invention, the terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0019] A multi-source data fusion method for early warning of building foundation settlement is applied to the foundation and substructure construction phase of building construction, specifically during earthwork excavation and underground structure construction after the completion of the foundation pit support structure. During this phase, the redistribution of internal forces in the support structure caused by foundation pit excavation, the evolution of the seepage field in the surrounding soil, and the three-dimensional deformation of the ground surface directly determine the final bearing capacity and long-term settlement stability of the building foundation. Therefore, this invention, through the fusion analysis and settlement propagation inference of multi-source monitoring data during the foundation pit construction period, achieves early source location and dynamic hierarchical early warning of building foundation settlement risks, representing a crucial preliminary step in the quality control of the entire building lifecycle. Figure 1 As shown, a preferred embodiment of the present invention includes the following steps: S1: Obtain the internal force response data of the foundation pit support structure of the target building, the seepage field monitoring data of the soil around the foundation pit, and the three-dimensional deformation monitoring data of the ground surface around the foundation pit; It should be noted that internal force response data refers to the internal mechanical state data of the foundation pit support structure during excavation due to soil pressure release, changes in support prestress, and external loads. This includes axial force of support components, bending moment of retaining wall, and prestress of anchor cables. This data is acquired through strain gauges, axial force gauges, or fiber optic sensors deployed at key stress points of the support structure. Its function is to reflect the degree of mechanical disturbance caused by foundation pit excavation to the support system from the perspective of structural stress. Seepage field monitoring data refers to the spatial distribution of groundwater seepage in the soil surrounding the foundation pit. The data includes pore water pressure and groundwater level gradient, acquired through piezometer arrays and water level monitoring wells. Its purpose is to reveal the driving effect of seepage field redistribution caused by foundation pit dewatering or excavation on the effective stress and deformation characteristics of the soil from a groundwater dynamics perspective. Three-dimensional deformation monitoring data refers to the displacement vector data of the ground surface around the foundation pit in three-dimensional space, including vertical settlement and horizontal displacement. This data is collected through three-dimensional laser scanning or GNSS monitoring arrays at a preset sampling period. Its purpose is to capture the surface response caused by foundation pit excavation from a geometric deformation perspective. For example, in a foundation pit project in a soft soil area, a rebar gauge is installed in the middle of the retaining wall to collect the wall bending moment, a piezometer is installed three meters outside the pit to monitor pore water pressure changes, and a GNSS receiver is installed on the road beside the pit to obtain the surface displacement vector. These three types of data together constitute the multi-source data foundation for monitoring the foundation quality of buildings during the foundation pit construction period.

[0020] Understandably, step S1 establishes a multi-dimensional perception system covering structural stress, groundwater seepage, and surface deformation by simultaneously collecting three types of heterogeneous monitoring data: the internal force response of the foundation pit support structure, the seepage field of the surrounding soil, and the three-dimensional deformation of the ground surface. This achieves a leap from monitoring a single physical quantity to coordinating the perception of multiple fields, thereby providing complete data input including mechanical driving, seepage driving, and deformation boundary conditions for subsequent geological structure modeling and settlement propagation inference, avoiding blind spots in tracing the causes of settlement caused by fragmented data sources.

[0021] S2: Based on the geological survey data of the foundation pit, construct a stratigraphic structure model, and map the internal force response data, the seepage field monitoring data, and the three-dimensional deformation monitoring data to the corresponding layer nodes of the stratigraphic structure model according to the stratigraphic spatial coordinates to form a multi-source data-driven stratigraphic node state matrix. It should be noted that geological survey data refers to the data set reflecting the stratum distribution and physical and mechanical properties of soil and rock masses at the foundation pit site, obtained through borehole sampling, in-situ testing, and indoor geotechnical tests. This includes borehole columnar sections, soil layer thickness, permeability coefficient, compression modulus, and internal friction angle. Geological survey data forms the spatial geometric and physical property basis for constructing the stratigraphic structure model. The stratigraphic structure model refers to a digital model reflecting the three-dimensional stratum distribution of the foundation pit site, constructed based on geological survey data through spatial interpolation and rasterization. Internally, it is divided into stratum nodes according to a preset step size, and each node carries the physical and mechanical parameters of its respective stratum. A stratum node is a data-carrying unit discretely arranged in the stratigraphic structure model according to spatial coordinates. Its spatial coordinates are jointly determined by the spatial range of the stratum and the rasterization step size, and are used to mount multi-source monitoring data. The stratigraphic node state matrix is ​​a structured data matrix composed of stratum nodes as rows and multi-source data attributes as columns. Each element represents the quantitative state of the corresponding node in a specific attribute. For example, after determining the existence of four strata—miscellaneous fill, silty clay, fine sand, and gravel—based on the borehole columnar section, the three-dimensional envelope surface of each stratum interface is constructed using the Kriging interpolation algorithm. Stratification nodes are then set up in the silty clay layer at a step size of two meters horizontally and one meter vertically. The coordinates of each node are correlated with the compression modulus of the soil layer to form an initial stratum structure model.

[0022] Understandably, step S2 reconstructs geological exploration data into a three-dimensional stratigraphic structure model containing stratigraphic nodes, and maps multi-source heterogeneous monitoring data to corresponding stratigraphic nodes according to spatial coordinates. This enables discrete monitoring data to obtain stratigraphic semantic attribution and a unified spatial representation framework, achieving semantic fusion between heterogeneous data and geological structures. This provides a spatially consistent node-based data foundation for the subsequent construction of subsidence propagation maps, avoiding fusion obstacles caused by differences in coordinate systems and semantic systems among multi-source data.

[0023] For example, geological surveys of a foundation pit site revealed that the strata, from top to bottom, consisted of miscellaneous fill, silty clay, fine sand, and gravel. Using the soil layer boundary depth data from the borehole columnar section, a three-dimensional envelope surface of the top and bottom plates of each stratum was constructed using Kriging interpolation, dividing the site into four three-dimensional stratigraphic spaces. Within the silty clay layer, the stratum nodes N1 to N12 were established at horizontal intervals of two meters and vertical intervals of one meter. Each node was associated with the layer's compression modulus and permeability coefficient. Bending moment data collected by the rebar gauge in the middle of the retaining wall was weighted and distributed to the two nodes (N3 and N4) according to the spatial distance between the rebar gauge and these nodes using an exponential decay function, forming the stress constraint attribute. Pore water pressure data from the piezometer array three meters outside the pit was distributed to nodes N3 to N8 using inverse distance weighted interpolation, forming the seepage drive attribute. Surface settlement data from the GNSS monitoring array was vertically projected to the uppermost nodes (N1 and N2), forming the displacement boundary attribute. The above attributes are arranged in node order to form a six-row, three-column formation node state matrix. Each row in the matrix corresponds to a stratum node, and the columns are force constraint attributes, seepage driving attributes, and displacement boundary attributes, respectively.

[0024] S3: Based on the settlement correlation of each layer node in the state matrix of the strata under the coupling effect of multi-source data, construct a settlement propagation map of the three fields of support-soil-groundwater coupling. The edge weight of the settlement propagation map represents the intensity of settlement influence transmission between adjacent layer nodes. It should be noted that settlement correlation refers to the causal relationship between stratigraphic nodes caused by changes in the stress of the supporting structure, changes in the soil seepage field, and the transmission of surface deformation. It is simultaneously constrained by the physical properties of the strata and their spatial location. Three-field coupling refers to the interaction and energy exchange mechanism between the stress field of the supporting structure, the soil seepage field, and the surface deformation field. Among them, the redistribution of internal forces in the supporting structure causes changes in the stress state of the soil, changes in the seepage field lead to a reduction in effective stress, and surface deformation provides displacement boundary conditions. Settlement propagation map refers to a directed graph structure with stratigraphic nodes as vertices and settlement propagation probability as edge weights. It quantifies the transmission intensity of settlement influence between adjacent nodes through edge weights and characterizes the directionality of settlement propagation through the direction of the edges. For example, when the effective stress of a node in a silty clay layer decreases due to increased pore water pressure, the settlement trend of that node is transmitted to the adjacent silty sand layer nodes through soil stress diffusion. At the same time, under the constraint of the stiffness of the support structure, it affects the stress state of the adjacent nodes of the retaining wall. The above-mentioned multi-source coupling effect is explicitly expressed in the graph through directed edges and edge weights.

[0025] Understandably, step S3 identifies the spatial adjacency relationships between stratigraphic nodes based on the stratigraphic node state matrix, and calculates the subsidence propagation probability by comprehensively considering the coupling differences of nodes in multi-source attributes, stratigraphic permeability characteristics, support stiffness constraints, and spatial distance attenuation effects. This transforms the implicit subsidence evolution process into an explicit directed propagation path and quantified edge weights, achieving a leap from single-point anomaly identification to chain propagation prediction, thereby providing complete topological information including propagation path, propagation intensity, and propagation direction for subsequent risk classification.

[0026] S4: Based on the settlement contribution of each layer node in the settlement propagation map and the safety of the support structure, determine the building risk level and trigger a graded early warning.

[0027] It should be noted that settlement contribution refers to the ability of a stratum node to transmit settlement trends to adjacent nodes as a settlement influence source in the settlement propagation map, as well as the cumulative effect of its own deformation on the overall risk situation. It is determined by the strength of the incoming propagation paths received by the node and the node's own displacement boundary properties. The safety of the support structure refers to the redundancy of the support system relative to the design bearing capacity under the current internal force response state, which is reflected by the ratio of the actual internal force of the support components to the allowable internal force. The building risk level refers to the classification of the building foundation quality and safety status based on the settlement propagation trend and the support bearing capacity, which links the evolution trend of stratum deformation with the bearing redundancy of the support structure. For example, if a stratum node receives multiple strong propagation paths in the settlement propagation map and its force constraint properties show that the axial force of the adjacent support is close to the design value, then the settlement contribution of this node is high and the support safety is low, and it is comprehensively judged as a high-risk level.

[0028] Understandably, step S4 comprehensively evaluates the settlement contribution of the stratum nodes and the safety of the support structure, and links the intensity of stratum deformation propagation with the bearing redundancy of the support system to make a joint judgment. This allows the risk classification to take into account both the soil disturbance evolution trend and the mechanical response of the support structure, thus realizing the transformation from a single threshold alarm to a multi-field coupled risk classification. This provides a closed-loop early warning output for the safety of building foundations during the construction period, including risk location, risk level and response strategy.

[0029] Specifically, all incoming propagation paths ending at nodes at each level are extracted from the settlement propagation map. The edge weights of each path are summed to obtain the cumulative path risk intensity of each node. The settlement contribution is calculated by combining the node's own displacement boundary attributes. The allowable design internal force and actual monitored internal force values ​​of the adjacent support components at each level are obtained to calculate the safety of the support structure. The settlement contribution and safety of each node at each level are mapped to a two-dimensional risk assessment space. According to the preset classification rules, the nodes are divided into different risk levels. Nodes with high settlement contribution and low safety are classified as high-risk, while nodes with low settlement contribution and high safety are classified as low-risk. The corresponding graded early warning response is triggered according to the risk level. High-risk levels trigger alarms and push emergency response suggestions, medium-risk levels start encrypted monitoring, and low-risk levels maintain regular recording.

[0030] Preferably, the acquisition of internal force response data of the target building foundation pit support structure, seepage field monitoring data of the soil surrounding the foundation pit, and three-dimensional deformation monitoring data of the ground surface surrounding the foundation pit includes: Sensors are deployed on the foundation pit support components, retaining walls and anchor cable stress points to collect axial force, bending moment and prestress data of the support structure as internal force response data. The pore water pressure and groundwater level gradient within the influence range of the foundation pit excavation are obtained by using a piezometer array and a water level monitoring well as the seepage field monitoring data. The spatial displacement vector of the ground surface around the foundation pit within a preset sampling period is obtained using a three-dimensional laser scanning device or a GNSS monitoring array, and is used as the three-dimensional deformation monitoring data.

[0031] It should be noted that internal force response data refers to the quantitative values ​​of the internal mechanical response of the foundation pit support structure under the action of earth pressure and external loads. These include the axial force of the support members, the bending moment of the retaining wall, and the prestress of the anchor cables. The axial force refers to the tensile or compressive force along the axis of the support member, which is collected by strain gauges or vibrating wire axial force gauges installed on the surface of the support member. It is used to characterize the axial stress state of the support system during the foundation pit excavation process. The bending moment refers to the internal moment of the retaining wall under the action of lateral earth pressure. It is collected by attaching a steel bar gauge to the main reinforcement of the wall or embedding a concrete strain gauge in the middle of the wall. It is used to reflect the bending performance and deformation trend of the retaining wall. The prestress refers to the initial tension force held by the anchor cable or soil nail after locking. It is collected by an anchor cable gauge or a through-type pressure sensor at the anchoring end. It is used to monitor the preload loss and load increase of the anchoring system. Seepage field monitoring data refers to the spatial distribution quantification of groundwater seepage state in the soil surrounding the foundation pit, including pore water pressure and groundwater level gradient. Pore water pressure refers to the sum of hydrostatic pressure and excess hydrostatic pressure borne by water in the soil pores, automatically acquired by an array of piezometers buried at different depths in the soil at a preset frequency, used to reveal changes in effective stress in the soil caused by foundation pit dewatering or rainfall. Groundwater level gradient refers to the ratio of the water level difference between adjacent monitoring points to the horizontal distance, acquired by float-type or pressure-type water level gauges in water level monitoring wells, used to determine the direction and velocity trend of groundwater seepage. Three-dimensional deformation monitoring data refers to the quantification of displacement vectors of the ground surface and structures surrounding the foundation pit in three-dimensional space, including vertical settlement and horizontal displacement. Spatial displacement vector refers to the change in three-dimensional coordinates of the monitoring point relative to its initial position, acquired by a three-dimensional laser scanning device using point cloud registration or a GNSS monitoring array using carrier phase differential sampling at a preset sampling period, used to capture the spatial morphology and temporal evolution of surface deformation. The above three types of data describe the disturbance mechanism of the foundation pit construction on the building foundation from different physical dimensions, and together they constitute the data foundation for multi-source perception.

[0032] Understandably, step S2, by clarifying the specific objects of collection, sensor placement, and data acquisition methods for internal force response data, seepage field monitoring data, and three-dimensional deformation monitoring data, transforms the abstract multi-source data acquisition operation in step S1 into a feasible engineering implementation plan. This achieves a refinement from a general data concept to specific monitoring points, physical quantity types, and acquisition technologies, thereby providing standardized data input with clear spatial coordinates and physical semantics for subsequent data mapping and model construction. This avoids data loss and spatial misalignment problems caused by inconsistent data acquisition standards or improper sensor placement.

[0033] Specifically, during the deployment of the monitoring network, strain gauges were installed at the mid-span sections and end nodes of the first and second reinforced concrete supports, reinforcement gauges were installed on the main reinforcement bars facing the soil at the middle and corners of the diaphragm wall, and anchor gauges were installed at the anchoring ends of the anchor cables. All internal force sensors were connected to the data acquisition instrument via shielded cables, with a sampling frequency of once per hour. Within the influence boundary of the excavation pit, a piezometer array was arranged in a quincunx pattern with a plane spacing of five meters and a depth interval of two meters. A water level monitoring well was set at the midpoint of the long side of the pit and at each corner of the pit, with a pressure water level gauge installed in the well, and a sampling frequency of once every two hours. GNSS static monitoring points were set up on the road surface around the pit, the top of the existing building foundations, and the ground markings of underground pipelines. A three-star eight-frequency receiver was used and connected to the base station, with a sampling frequency of once per day. At the same time, a three-dimensional laser scanner was set up next to the main road at the edge of the pit to acquire surface point cloud data once a week. The aforementioned sensor network performs timestamp alignment and outlier removal at the edge computing gateway. When a sensor experiences a short-term communication interruption, linear interpolation is used to fill in the missing data, ensuring that the three types of data are synchronously transmitted to the data processing center under a unified time reference.

[0034] Preferably, constructing a stratigraphic structure model based on the geological survey data of the foundation pit includes: Extract borehole columnar sections from the geological survey data, obtain the physical and mechanical parameters of each soil layer, and use spatial interpolation algorithms to reconstruct the three-dimensional envelope surfaces of the interfaces of each soil layer in the foundation pit to divide the three-dimensional stratigraphic space. Based on the thickness distribution of each stratum in the three-dimensional stratigraphic space, rasterization is performed according to a preset step size to establish multiple stratigraphic nodes. The physical and mechanical parameters are mapped to the corresponding stratigraphic nodes to obtain the first stratigraphic structure model.

[0035] It should be noted that a borehole columnar section refers to a columnar profile obtained through mechanical or manual drilling during geological exploration, recording the soil layer type, thickness, and burial depth according to depth. It provides discrete sample points for spatial interpolation by extracting the depth coordinates of the interfaces between different strata. Physical and mechanical parameters refer to quantitative indicators characterizing the engineering properties of soil layers obtained through indoor geotechnical tests or in-situ tests, including permeability coefficient, compression modulus, and internal friction angle. The permeability coefficient refers to the ease with which water flows through the pores of the soil, determined through constant head or variable head permeability tests, and is used to characterize the permeability of the soil. The compression modulus is the ratio of vertical stress to vertical strain in soil under lateral confinement conditions, determined through consolidation tests, and is used to characterize the compressibility of the soil. The internal friction angle is the angle between the shear strength envelope of the soil and the normal stress axis, determined through direct shear tests or triaxial shear tests, and is used to characterize the shear strength characteristics of the soil. Spatial interpolation algorithms refer to algorithms that use mathematical methods to calculate the elevation of stratigraphic interfaces in un-drilled areas based on the depth data of stratigraphic interfaces at discrete borehole points. These include inverse distance weighting, Kriging interpolation, or radial basis function methods, and are used to extend discrete borehole data into continuous three-dimensional surfaces. A three-dimensional envelope surface refers to a continuous three-dimensional surface generated through spatial interpolation that encloses the spatial extent of each stratigraphic region, including the top and bottom surfaces of each stratum, used to delineate the three-dimensional stratigraphic space of the foundation pit site. Rasterization is the process of discretizing the continuous three-dimensional stratigraphic space into a regular grid according to a preset step size. The preset step size refers to the horizontal and vertical spacing of the stratigraphic nodes, set according to monitoring accuracy requirements and computational resource constraints, used to control node density. A stratigraphic node is a data-carrying unit established after rasterization, capable of carrying three-dimensional spatial coordinates and stratigraphic attributes, used for subsequent spatial mapping of multi-source monitoring data.

[0036] Understandably, by extracting stratigraphic information and physical and mechanical parameters from borehole columnar sections, spatial interpolation algorithms are used to reconstruct the three-dimensional envelope surface of stratigraphic interfaces to divide the three-dimensional stratigraphic space. Within each stratigraphic space, rasterization is performed according to a preset step size to establish stratigraphic nodes. This achieves the transformation from discrete geological exploration data to a continuous digital stratigraphic model, expanding the point-like geological information originally based on boreholes into a spatial continuum covering the entire foundation pit site. This provides a node-based framework with spatial coordinates and physical attributes for subsequent multi-source data mapping, avoiding fusion barriers caused by different spatial representation methods between geological data and monitoring data.

[0037] Specifically, columnar section data from all boreholes within the site are extracted from the geological survey report, and the borehole depth coordinates at each stratigraphic interface are recorded. Kriging interpolation is used to spatially interpolate the elevation data of the top and bottom strata of each site, generating three-dimensional envelope surfaces for each stratigraphic interface, thus dividing the foundation pit site into several three-dimensional stratigraphic spaces. Based on the thickness distribution of the stratigraphic spaces and the monitoring accuracy requirements, a horizontal step size of three meters and a vertical step size of one meter are set. Stratigraphic nodes are evenly distributed within each stratigraphic space according to these step sizes, and each node is assigned three-dimensional spatial coordinates and the physical and mechanical parameters of its respective stratum. This includes writing the permeability coefficient, compression modulus, and internal friction angle as inherent attributes of the node into the model database, resulting in the first stratigraphic structure model. During the rasterization process, thin interlayers with a thickness less than the vertical step size are merged into adjacent main strata or marked as special nodes, ensuring that the stratigraphic structure model reflects both geological stratification characteristics and meets the requirements for computational node density.

[0038] For example, geological surveys of a foundation pit site revealed that the strata, from top to bottom, consisted of miscellaneous fill, silty clay, fine sand, and gravel. After extracting columnar section data from four boreholes and obtaining the depth coordinates of each stratum boundary, the three-dimensional envelope surfaces of the top and bottom plates of each stratum were reconstructed using the Kriging interpolation algorithm, dividing the site into four three-dimensional stratigraphic spaces. Within the silty clay layer, a raster was created with a horizontal spacing of three meters and a vertical spacing of one meter to establish stratigraphic nodes. Each node was associated with the permeability coefficient, compression modulus, and internal friction angle of that layer, forming the first stratigraphic structure model.

[0039] Preferably, mapping the internal force response data to the corresponding stratigraphic nodes of the stratigraphic structure model according to the stratigraphic spatial coordinates includes: Obtain the spatial coordinates of the internal force monitoring points corresponding to the internal force response data. The internal force monitoring points include axial force monitoring points arranged on the supporting components, bending moment monitoring points of the retaining wall, and prestress monitoring points of the anchor cable stress points. Calculate the three-dimensional spatial distance between each internal force monitoring point and each layer node in the first geological structure model; Based on the three-dimensional spatial distance, and following a weighted allocation strategy based on a spatial attenuation function, the internal force response data is mapped to one or more nearest layer nodes, which are then used as the internal force response components of the layer nodes. The weighted allocation strategy includes: calculating the mapping weight using an exponential decay function based on the distance between the internal force monitoring point and the layer node, and allocating the internal force response data of the internal force monitoring point to the adjacent layer node according to the mapping weight, thereby obtaining the force constraint attributes of each layer node.

[0040] It should be noted that internal force monitoring points refer to physical locations set up at key stress-bearing parts of the support structure to collect internal force response data. These include axial force monitoring points, bending moment monitoring points, and prestressing monitoring points. Axial force monitoring points are located at the mid-span or end sections of the supporting members; bending moment monitoring points are located at the main reinforcement positions in the middle or at corners of the retaining wall; and prestressing monitoring points are located at the anchoring ends of anchor cables or soil nails. The spatial coordinates of these monitoring points are obtained through total station or RTK measurements to establish the correspondence between internal force data and spatial locations. Three-dimensional spatial distance refers to the Euclidean distance between the internal force monitoring point and the floor node in the three-dimensional coordinate system. It is obtained by calculating the square root of the sum of the squares of the coordinate differences between the two points in the X, Y, and Z directions, and is used to quantify the spatial distance between the monitoring point and the node. The spatial attenuation function is a mathematical function describing the attenuation of internal force response data with increasing spatial distance. It includes exponential attenuation functions or inverse distance-squared functions. The exponential attenuation function expresses the attenuation law in natural exponential form, and the attenuation rate is controlled by setting an attenuation coefficient. It reflects the fact that the influence of the internal forces of the support structure on the surrounding soil weakens with increasing distance. The weighted allocation strategy is a data allocation method that assigns different mapping weights to different layer nodes based on spatial distance. Nodes closer to each other receive higher mapping weights, ensuring that internal force data is preferentially mapped to the spatially most closely related nodes. Mapping weight refers to the proportion of data allocated from a specific internal force monitoring point to a specific layer node. It is calculated using the spatial attenuation function and normalized, and is used to control the data allocation ratio when multiple nodes share data. Internal force response components are quantitative data received by a layer node reflecting the stress state of its adjacent support structure, including axial force components, bending moment components, and prestress components, used to characterize the support constraint strength at that node location. Force constraint attributes refer to the comprehensive attribute labels that the layer nodes bear in subsequent analysis, characterizing the degree of mechanical constraint of the support structure on them. They are obtained by standardizing the internal force response components and are used to participate in the subsequent coupling difference calculation.

[0041] Understandably, by acquiring the spatial coordinates of the internal force monitoring points and calculating their three-dimensional spatial distance from the stratum nodes, the internal force response data is mapped to the adjacent stratum nodes according to a weighted allocation strategy based on the spatial attenuation function. This transforms the internal force data, which was originally attached to the discrete monitoring points of the support structure, into force constraint attributes distributed across various nodes in the stratum space. This expands the monitoring from local structural monitoring to the distribution of the stratum space field, thereby providing the support mechanical state input covering the stratum space for subsequent three-field coupling analysis. This avoids the problem that internal force data cannot participate in the stratum node state assessment due to being limited to the monitoring point location.

[0042] Specifically, the three-dimensional spatial coordinates of all internal force monitoring points are obtained, including the axial force monitoring point located at the center of the mid-span section of the supporting member, the bending moment monitoring point located on the surface of the main reinforcement in the middle of the retaining wall, and the prestressing monitoring point located at the center of the anchorage end of the anchor cable. These coordinates are obtained through construction layout records or as-built measurement results. For each internal force monitoring point, all layer nodes in the first stratum structure model are traversed, and the coordinate differences between the monitoring point and each node in the X, Y, and Z directions are calculated. The three-dimensional spatial distance is obtained using the Euclidean distance formula. The attenuation coefficient is set to 15% per meter. The initial mapping weights of nodes at each layer are calculated using an exponential decay function. The weight values ​​decrease exponentially with increasing distance, and nodes located outside the five-meter range are assigned weights approaching zero. The initial mapping weights are then normalized so that the sum of the weights of all nodes corresponding to the same monitoring point is one. According to the normalized mapping weights, the axial force, bending moment, and prestress data of the internal force monitoring points are distributed to the nodes at each adjacent layer to obtain the internal force response components of each node. The internal force response components are sorted by node number to form the force constraint attributes characterizing the support constraint strength of each node.

[0043] Preferably, mapping the seepage field monitoring data to the corresponding stratigraphic nodes of the stratigraphic structure model according to the stratigraphic spatial coordinates includes: Obtain the spatial coordinates of the seepage monitoring points corresponding to the seepage field monitoring data, wherein the seepage monitoring points include the locations of the piezometer array and the water level monitoring wells; Extract the set of stratigraphic nodes in the first stratigraphic structure model that are located within the stratigraphic space of each seepage monitoring point; Based on the spatial coordinates of the seepage monitoring points and the spatial coordinates of each layer node, the seepage field monitoring data is distributed to each layer node in the layer node set using the inverse distance weighted interpolation algorithm to obtain the seepage field components of each layer node. The inverse distance weighted interpolation algorithm includes: for the j-th layer node, its seepage field component is obtained by weighted averaging of the monitoring values ​​of each seepage monitoring point within a preset neighborhood of the j-th layer node, and the weight is inversely proportional to the spatial distance from the seepage monitoring point to the j-th layer node, thereby obtaining the seepage driving attribute of each layer node.

[0044] It should be noted that seepage monitoring points refer to physical locations within the soil surrounding the foundation pit used to collect seepage field monitoring data. These include piezometer array locations and water level monitoring well locations. Piezometer array locations refer to the locations of pore water pressure sensors arranged in a planar grid or quincunx pattern at different depths within the soil surrounding the pit. Water level monitoring wells refer to vertical shaft structures drilled around the pit to observe groundwater levels. The spatial coordinates of these monitoring points are determined through surveying and setting out, and are used to establish the positional correlation between seepage data and the geological space. The stratigraphic node set refers to a subset of all stratigraphic nodes located within the spatial range of a particular seepage monitoring point. It is determined by determining whether the spatial coordinates of the nodes fall between the top and bottom elevations of that stratigraphic space, and is used to limit the spatial range for seepage data interpolation. The inverse distance weighted interpolation algorithm is a spatial interpolation method that uses the inverse ratio of the spatial distance between the point to be interpolated and known monitoring points as weights. The inverse ratio of weight to spatial distance means that the closer the monitoring point, the greater its influence on the node, and its higher weight value, reflecting the continuous spatial variation of the seepage field. The seepage field components refer to the quantitative data obtained by interpolation of the stratigraphic nodes, characterizing the pore water pressure and water level at their location. These include pore water pressure values ​​and water level gradient values, reflecting the groundwater seepage driving state at the nodes. The seepage driving attributes refer to the comprehensive attribute labels that the stratigraphic nodes bear in subsequent analyses, characterizing the degree of groundwater seepage driving their settlement. These attributes are obtained by standardizing the seepage field components and are used in subsequent three-field coupling analyses.

[0045] Understandably, by obtaining the spatial coordinates of seepage monitoring points, extracting the set of stratum nodes within the corresponding stratum space, and using the inverse distance weighted interpolation algorithm to distribute the seepage field monitoring data to each stratum node, the seepage data that was originally discretely distributed in the piezometer array and monitoring wells is transformed into a continuous seepage field distribution covering each node in the stratum space. This achieves the expansion from point monitoring to spatial field distribution, thereby providing spatially continuous groundwater seepage state input for subsequent three-field coupling analysis and avoiding the problem that seepage data cannot accurately characterize the differences in seepage driving within the stratum due to the sparseness of monitoring points.

[0046] Specifically, the three-dimensional spatial coordinates of all seepage monitoring points are obtained. The piezometer array is deployed around the pit at a 5-meter plane spacing and a 2-meter depth interval. Water level monitoring wells are located at the pit corners and the midpoint of the long side. These coordinates are obtained through construction survey records. For each seepage monitoring point, all stratum nodes in the first stratigraphic structure model are traversed. It is determined whether the spatial coordinates of the node fall within the top and bottom elevation range of the stratum space where the monitoring point is located. Nodes that meet the condition are included in the stratum node set of that monitoring point. For the j-th stratum node in the stratum node set, its preset neighborhood range is determined to be a horizontal radius of 5 meters and a vertical thickness of... A one-meter-long cylindrical space is used to extract all seepage monitoring points falling within its neighborhood. The three-dimensional spatial distance from each seepage monitoring point to the j-th layer node is calculated, and the inverse ratio of the distance is used as the initial weight. A minimum distance threshold is set for monitoring points with a distance of less than 0.5 meters to avoid infinite weights. After normalizing the initial weights, the pore water pressure value and water level gradient value of each monitoring point are weighted and averaged to obtain the seepage field component of the j-th layer node. After traversing all nodes in the layer node set, the seepage field components of each node are sorted by node number to form the seepage driving attribute characterizing the groundwater seepage driving state of each node.

[0047] Preferably, mapping the three-dimensional deformation monitoring data to the corresponding stratigraphic nodes of the stratigraphic structure model according to stratigraphic spatial coordinates includes: Obtain the horizontal spatial coordinates of the surface monitoring points corresponding to the three-dimensional deformation monitoring data; Based on the horizontal spatial coordinates, a target stratigraphic node is determined on the horizontal projection plane of the first stratigraphic structure model. The target stratigraphic node is the stratigraphic node whose horizontal coordinates are closest to the surface monitoring point and located in the uppermost stratigraphic space. The vertical and horizontal displacement components in the three-dimensional deformation monitoring data are vertically projected onto the target layer node to serve as the surface deformation components of the target layer node. If the same target stratum node corresponds to multiple surface monitoring points, the average value of the three-dimensional deformation monitoring data of the multiple surface monitoring points or the weighted average value according to the monitoring accuracy is used as the surface deformation component of the current target stratum node, and the displacement boundary attributes of each stratum node are obtained.

[0048] It should be noted that surface monitoring points refer to physical locations on the surface surrounding the foundation pit used to collect three-dimensional deformation monitoring data. These include monitoring positions on road surfaces, the tops of building foundations, and ground markers for pipelines. Their horizontal spatial coordinates refer to the two-dimensional coordinates of the monitoring points on a horizontal plane, obtained through measurements using a GNSS receiver or total station. These coordinates are used to determine the location of the surface monitoring data on the horizontal projection plane of the geological structure model. The horizontal projection plane is a two-dimensional plane obtained by projecting the three-dimensional geological structure model vertically onto a horizontal plane. It is used to establish the horizontal positional correspondence between surface monitoring points and underground stratum nodes. Target stratum nodes are stratum nodes located near the horizontal coordinates of the surface monitoring points and within the uppermost stratum. These nodes are determined by calculating the distances between the surface monitoring points and each stratum node on the horizontal plane and selecting the uppermost node. They are used to receive the vertical projection of the surface deformation monitoring data. The vertical displacement component refers to the settlement or uplift of the surface monitoring points in the vertical direction. It is obtained through GNSS elevation measurements or three-dimensional laser scanning point cloud elevation difference analysis and is used to characterize the deformation of the surface in the direction of gravity. Horizontal displacement components refer to the movement vectors of surface monitoring points on the horizontal plane, obtained through GNSS planar coordinate difference or laser scanning point cloud planar registration, and are used to characterize the deformation of the surface in the horizontal direction. Surface deformation components refer to the quantitative data received by the stratigraphic node reflecting the deformation state at its corresponding surface location, including vertical and horizontal displacement components, and are used to characterize the displacement boundary conditions at that node location. Displacement boundary attributes refer to the comprehensive attribute labels carried by the stratigraphic node in subsequent analysis, characterizing the degree of constraint of surface deformation on its settlement. These are obtained by standardizing the surface deformation components and are used in subsequent three-field coupled analysis.

[0049] Understandably, by acquiring the horizontal spatial coordinates of surface monitoring points, the target stratigraphic node located in the uppermost stratigraphic space is determined on the horizontal projection plane of the first stratigraphic structure model. The vertical and horizontal displacement components in the three-dimensional deformation monitoring data are then vertically projected onto the target stratigraphic node, transforming the deformation data originally distributed at discrete surface monitoring points into the displacement boundary attributes of the uppermost stratigraphic node. This achieves the transformation from surface monitoring to stratigraphic node boundary conditions, thereby providing boundary inputs that reflect surface deformation constraints for subsequent three-field coupling analysis. This avoids the problem that surface deformation data cannot participate in the assessment of subsidence propagation within the stratigraphy because it only represents the surface state.

[0050] Specifically, the horizontal spatial coordinates of all surface monitoring points are obtained, including the planar coordinates of GNSS monitoring points and the control point coordinates of 3D laser scanning stations; the 3D coordinates of all stratigraphic nodes in the first stratigraphic structure model are projected vertically onto the horizontal plane to establish a node distribution map on the horizontal projection plane; for each surface monitoring point, its horizontal distance to each stratigraphic node on the horizontal projection plane is calculated, the node with the smallest horizontal distance is selected, and it is determined whether the node is located in the uppermost stratigraphic space. If the node is located in the lower stratigraphic space, the search continues upward until the nearest node in the uppermost stratigraphic space is found, and the node is confirmed as the uppermost stratigraphic node. The target stratigraphic node for each surface monitoring point is designated; the vertical and horizontal displacement components in the 3D deformation monitoring data are directly assigned to the corresponding target stratigraphic node as the surface deformation components of that node; when multiple surface monitoring points correspond to the same target stratigraphic node, the arithmetic mean of the vertical and horizontal displacement components of each monitoring point is taken, or a weighted average is calculated based on the monitoring accuracy, with higher-accuracy monitoring points assigned greater weight, and the averaged displacement data is used as the surface deformation component of the current target stratigraphic node; the surface deformation components are sorted by node number to form displacement boundary attributes characterizing the displacement boundary constraints of each node.

[0051] Preferably, based on the settlement correlation of each layer node in the stratigraphic node state matrix under the coupling effect of multi-source data, the settlement propagation map of the three fields coupled by support, soil, and groundwater is constructed, including: The spatial adjacency relationship of each stratigraphic node in the stratigraphic structure model is determined. The spatial adjacency relationship includes the horizontal adjacency relationship between adjacent stratigraphic nodes within the same stratigraphic space and the vertical adjacency relationship between stratigraphic nodes between adjacent stratigraphic spaces. For each pair of stratigraphic nodes with spatial adjacency, the multi-source state vectors of the source node and the target node are extracted based on the stratigraphic node state matrix, and the coupling difference between the source node and the target node in terms of force constraint attributes, seepage driving attributes and displacement boundary attributes is calculated. Based on the coupling difference, the ratio of the permeability coefficients of the strata to which the source node and the target node belong, the stiffness coupling coefficient of the support structure, and the spatial distance between the source node and the target node, the settlement propagation probability from the source node to the target node is calculated and used as the edge weight of the settlement propagation map. The subsidence propagation probability is compared with a preset propagation threshold. If the subsidence propagation probability is greater than the preset propagation threshold, a directed propagation edge is established between the source node and the target node. The direction of the directed propagation edge is from the subsidence impact source to the subsidence impact target. Traverse all pairs of layer nodes with spatial adjacency, and construct the settlement propagation map of the three-field coupling of support, soil and groundwater based on the directed propagation edge and its corresponding edge weight.

[0052] It should be noted that spatial adjacency refers to the connection between stratigraphic nodes in three-dimensional space due to their proximity. This includes horizontal and vertical adjacency. Horizontal adjacency refers to the connection between stratigraphic nodes adjacent in the horizontal direction within the same stratigraphic space, determined by whether the horizontal spacing between nodes equals a preset step size, and is used to characterize the lateral transfer of settlement within the same soil layer. Vertical adjacency refers to the connection between stratigraphic nodes adjacent in the vertical direction between adjacent stratigraphic spaces, determined by whether the vertical spacing between nodes equals a preset step size and whether they are located on opposite sides of the adjacent stratigraphic interface, and is used to characterize the vertical transfer of settlement at the soil interface. The multi-source state vector refers to the attribute vector borne by the stratigraphic node, composed of force constraint attributes, seepage driving attributes, and displacement boundary attributes, used to comprehensively characterize the current state of the node under the coupling of three fields. The coupling difference degree refers to the quantified difference between the source node and the target node in the multi-source state vector. It is obtained by calculating the degree of deviation between the two nodes in terms of force constraint attributes, seepage driving attributes, and displacement boundary attributes. It reflects the inconsistency between the two nodes in the three-field coupling state. The greater the difference, the stronger the settlement driving gradient between the two nodes. The settlement propagation probability refers to the quantified value of the possibility of settlement influence being transmitted from the source node to the target node. It is calculated by combining the coupling difference degree, the ratio of stratum permeability coefficient, the stiffness coupling coefficient of the support structure, and the spatial distance attenuation effect. It is used to quantify the intensity of settlement influence transmission between nodes. The edge weight refers to the weight value of the directed propagation edge in the settlement propagation spectrum. It is obtained by assigning the settlement propagation probability and is used to characterize the intensity of settlement influence transmission between adjacent layer nodes. The preset propagation threshold is the probability critical value used to determine whether to establish an effective propagation edge. It is set through engineering experience or historical data statistics and is used to screen reliable settlement propagation paths. A directional propagation edge is a propagation path with a clear direction, pointing from the source of settlement impact to the target of settlement impact. It is used to characterize the unidirectionality of settlement propagation and reflects the transmission trend of settlement from the strong driving point to the weak driving point.

[0053] Understandably, by determining the spatial adjacency of stratigraphic nodes, extracting multi-source state vectors of source and target nodes and calculating coupling difference, and comprehensively considering stratum permeability characteristics, support stiffness constraints, and spatial distance attenuation effects to calculate the subsidence propagation probability, the propagation probability is compared with a preset propagation threshold to establish directed propagation edges. By traversing all adjacent node pairs, a subsidence propagation map is constructed, making the subsidence correlation originally implicit in the stratum node state matrix explicit into a graph structure with directionality and weight. This achieves a leap from node state description to propagation path reasoning, thereby providing complete topological information including propagation path, propagation intensity, and propagation direction for subsequent risk classification, avoiding the problem of being unable to assess the evolution trend of subsidence in the spatial network by relying solely on the state of the nodes themselves.

[0054] Specifically, all stratigraphic nodes in the first stratigraphic structure model are traversed. For each node, its horizontally adjacent nodes within the same stratigraphic space with a preset step size are searched to establish horizontal adjacency. Simultaneously, its vertically adjacent nodes within adjacent stratigraphic spaces with a preset step size are searched to establish vertical adjacency. All spatially adjacent node pairs are summarized. For each pair of spatially adjacent nodes, the multi-source state vectors of the source node and the target node are extracted from the stratigraphic node state matrix. The Euclidean distance between the two nodes in terms of force constraint attributes, the difference between pore water pressure and seepage velocity vectors in terms of seepage driving attributes, and the difference between vertical and horizontal displacement vectors in terms of displacement boundary attributes are calculated. These differences are then fused into a comprehensive coupling difference degree according to a preset weight. The source node and target node are obtained. The permeability coefficient and compressibility modulus of the stratum to which the point belongs are used to calculate the ratio of stratum permeability coefficients. The equivalent stiffness and monitored axial force of the adjacent support structure between the two nodes are obtained, and the stiffness coupling coefficient of the support structure is calculated. Combined with the three-dimensional spatial distance between the two nodes, the settlement propagation probability is calculated using a weighted summation formula. A preset propagation threshold of 0.5 is set, and the settlement propagation probability is compared with this threshold. If the propagation probability is greater than the threshold, a directed propagation edge is established between the source node and the target node, pointing from the settlement influence source to the settlement influence target, and the propagation probability is assigned to this edge as the edge weight. If the propagation probability is not greater than the threshold, the node pair is skipped. After traversing all spatially adjacent node pairs, all directed propagation edges and their edge weights are summarized to construct a settlement propagation map with the stratum nodes as vertices and the directed propagation edges as edges.

[0055] Preferably, comparing the settlement propagation probability with a preset propagation threshold includes: Set a first propagation threshold and a second propagation threshold, wherein the first propagation threshold is less than the second propagation threshold; If the subsidence propagation probability is not less than the second propagation threshold, then the stratigraphic node is determined. To the hierarchical node There exists a strong subsidence propagation path, and the directed propagation edge is marked as a strong propagation edge; If the subsidence propagation probability is not less than the first propagation threshold and less than the second propagation threshold, then the stratigraphic node is determined. To the hierarchical node There exists a moderate settlement propagation path, and the directed propagation edge is marked as a moderate propagation edge; If the subsidence propagation probability is less than the first propagation threshold, then the stratigraphic node is determined. To the hierarchical node There is no effective settlement propagation path between them, and no directed propagation edge is established.

[0056] It should be noted that the first propagation threshold is a lower critical value used to distinguish between effective and ineffective propagation paths. Its value is less than the second propagation threshold and is used to screen node pairs with at least a moderate probability of settlement propagation. The second propagation threshold is an upper critical value used to distinguish between strong and moderate propagation paths and is used to identify node pairs with highly active settlement propagation. A strong settlement propagation path refers to a node connection with a settlement propagation probability reaching or exceeding the second propagation threshold, indicating an extremely high intensity of settlement impact transmission between the two nodes, reflecting a significant settlement-driving effect of the source node on the target node. A strong propagation edge refers to a directed propagation edge marked as a strong settlement propagation path in the settlement propagation map, usually displayed with a striking color or thick line width, used to highlight high-risk propagation channels in the map. A moderate settlement propagation path refers to a node connection with a settlement propagation probability between the first and second propagation thresholds, indicating a moderate intensity of settlement impact transmission between the two nodes, reflecting a certain degree of settlement driving effect of the source node on the target node, but not yet reaching a high-risk level. Intermediate propagation edges refer to directed propagation edges marked as intermediate settlement propagation paths in the settlement propagation map, used to identify general risk propagation channels in the map. Effective settlement propagation paths refer to the connections between nodes where the settlement propagation probability reaches or exceeds the first propagation threshold, including strong and intermediate settlement propagation paths, used to preserve settlement transmission relationships with actual physical significance in the map.

[0057] Understandably, by setting two levels of propagation thresholds—the first and second—the probability of subsidence propagation is divided into three intervals: strong propagation, moderate propagation, and ineffective propagation. This transforms the originally continuous propagation probability values ​​into risk classification labels with engineering semantics, realizing the mapping from probability values ​​to risk levels. This allows for the differentiation of propagation channels with different risk levels in the subsidence propagation map, providing a direct map identification basis for the differentiated formulation of subsequent risk response strategies. It also avoids the problem of all propagation edges being treated equally, causing high-risk channels to be submerged among ordinary channels.

[0058] Specifically, based on engineering experience and historical monitoring data, a first propagation threshold of 0.5 and a second propagation threshold of 0.8 are set, ensuring that the first propagation threshold is strictly less than the second propagation threshold. All layer node pairs with directed propagation edges in the settlement propagation map are traversed to obtain the settlement propagation probability of each edge. For a propagation edge from layer node i to layer node j, if its settlement propagation probability is not less than 0.8, the propagation edge is marked as a strong propagation edge and displayed as a thick red line in the map visualization. Simultaneously, this path is recorded as a strong settlement propagation path. If its settlement propagation probability is less than 0.8, the path is marked as a strong propagation edge. If the propagation probability is not less than 0.5 and less than 0.8, the propagation edge is marked as a medium propagation edge and displayed as an orange midline in the graph visualization. At the same time, the path is recorded as a medium settlement propagation path. If its settlement propagation probability is less than 0.5, the propagation edge is removed from the graph, and it is determined that there is no effective settlement propagation path between layer node i and layer node j. After traversal, a graded settlement propagation graph composed of strong propagation edges and medium propagation edges is obtained, where strong propagation edges identify high-risk settlement transmission channels and medium propagation edges identify general-risk settlement transmission channels.

[0059] Preferably, the first propagation threshold and the second propagation threshold satisfy the following relationship: ; ; in, Indicates the first propagation threshold. Indicates the second propagation threshold. Indicates the basic threshold. This indicates the current depth of the foundation pit excavation. Indicates the reference excavation depth. This indicates the equivalent stiffness of the enclosure wall. Indicates reference stiffness. and Indicates the depth correction factor. This represents the stiffness correction factor.

[0060] It should be noted that the basic threshold refers to a benchmark value determined based on historical monitoring data statistics or engineering experience, reflecting the critical state of settlement propagation probability under general geological conditions, and is used as the starting point for calculating the propagation threshold. The current excavation depth of the foundation pit refers to the vertical depth of the foundation pit project at the current moment, obtained through on-site measurements or construction log records, and is used to reflect the current degree of ground disturbance caused by the excavation. The reference excavation depth refers to a benchmark depth value set based on regional engineering experience or specification recommendations, used to dimensionlessly represent the current excavation depth, and to avoid the incomparability of thresholds caused by differences in absolute depth between foundation pits of different sizes. The equivalent stiffness of the retaining wall refers to the comprehensive characterization value of the bending stiffness of the retaining wall under lateral earth pressure, obtained through calculations using the elastic modulus of the retaining wall material, the moment of inertia of the section, and the support spacing, and is used to reflect the constraint capacity of the support structure on ground deformation. The reference stiffness refers to a benchmark stiffness value calibrated according to typical support structure forms, used to dimensionlessly represent the equivalent stiffness of the retaining wall, and to avoid the incomparability of thresholds caused by differences in stiffness between different support forms. The depth correction factor is a parameter used to adjust the foundation threshold to reflect the impact of excavation depth on the activation degree of settlement propagation. It includes a first depth correction factor and a second depth correction factor. The first depth correction factor is used to calculate the first propagation threshold, and the second depth correction factor is used to calculate the second propagation threshold. The second depth correction factor is greater than the first depth correction factor, reflecting that increased depth has a stronger activation effect on high-risk propagation paths than on general risk paths. The stiffness correction factor is a parameter used to adjust the foundation threshold to reflect the impact of retaining wall stiffness on the inhibition degree of settlement propagation. In the calculation of the second propagation threshold, it reflects the physical fact that higher support stiffness results in a more stringent determination of high-risk propagation paths.

[0061] Understandably, by establishing a functional relationship between the basic threshold and the excavation depth and the stiffness of the retaining wall, the preset propagation threshold can be dynamically adjusted according to the current construction status of the foundation pit and the characteristics of the support structure. This achieves the transformation from a static fixed threshold to a dynamic adaptive threshold, thereby enabling foundation pit projects with different excavation depths and different support forms to adopt a unified risk assessment benchmark. This avoids the problem of over-warning in shallow foundation pits or insufficient warning in deep foundation pits caused by the use of fixed thresholds, and improves the applicability and accuracy of settlement propagation maps under different engineering conditions.

[0062] Specifically, monitoring data from historical foundation pit projects in the region were collected, and the distribution characteristics of settlement propagation probability under general geological conditions were statistically analyzed to determine a foundation threshold of 0.35. The current foundation pit excavation depth was obtained through on-site measurements, for example, 12m. A reference excavation depth of 10m was set according to regional regulations. The equivalent stiffness of the retaining wall was calculated using design parameters: a retaining wall thickness of 0.8m, a concrete elastic modulus of 30GPa, and a horizontal support spacing of 6m. The calculated equivalent stiffness of the retaining wall was... The reference stiffness is determined based on the typical diaphragm wall support structure. Based on engineering experience, the first depth correction coefficient is set to 0.10, the second depth correction coefficient is set to 0.25, and the stiffness correction coefficient is set to 0.15. Substituting the above parameters into the formula, the first propagation threshold is calculated to be 0.47, and the second propagation threshold is calculated to be 0.89. The calculated first propagation threshold of 0.47 and the second propagation threshold of 0.89 are applied to the comparison of settlement propagation probability to realize the dynamic setting of the threshold.

[0063] A multi-source data fusion-based early warning system for building foundation settlement, such as Figure 2 As shown, it includes: The data acquisition module is used to acquire the internal force response data of the foundation pit support structure of the target building, the seepage field monitoring data of the soil around the foundation pit, and the three-dimensional deformation monitoring data of the ground surface around the foundation pit. The data fusion module is used to construct a stratigraphic structure model based on the geological survey data of the foundation pit, and to map the internal force response data, the seepage field monitoring data and the three-dimensional deformation monitoring data to the corresponding layer nodes of the stratigraphic structure model according to the stratigraphic spatial coordinates, forming a stratigraphic node state matrix driven by multi-source data. The map construction module is used to construct a settlement propagation map of the support-soil-groundwater three-field coupling based on the settlement correlation of each layer node in the stratigraphic node state matrix under the coupling effect of multi-source data. The edge weight of the settlement propagation map represents the intensity of settlement influence transmission between adjacent layer nodes. The risk warning module is used to determine the risk level of the building and trigger a graded warning based on the settlement contribution of each node in the settlement propagation map and the safety of the support structure.

[0064] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0065] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for early warning of building foundation settlement through multi-source data fusion, characterized in that, include: Acquire the internal force response data of the foundation pit support structure of the target building, the seepage field monitoring data of the soil surrounding the foundation pit, and the three-dimensional deformation monitoring data of the ground surface surrounding the foundation pit; Based on the geological survey data of the foundation pit, a stratigraphic structure model is constructed. The internal force response data, the seepage field monitoring data, and the three-dimensional deformation monitoring data are mapped to the corresponding layer nodes of the stratigraphic structure model according to the stratigraphic spatial coordinates, forming a stratigraphic node state matrix driven by multi-source data. Based on the settlement correlation of each layer node in the stratigraphic node state matrix under the coupling effect of multi-source data, a settlement propagation map of the three fields of support-soil-groundwater coupling is constructed. The edge weight of the settlement propagation map represents the intensity of settlement influence transmission between adjacent layer nodes. Based on the settlement contribution of each node in the settlement propagation map and the safety of the support structure, the risk level of the building is determined and a graded early warning is triggered.

2. The method for early warning of building foundation settlement based on multi-source data fusion according to claim 1, characterized in that, The acquisition of internal force response data of the target building foundation pit support structure, seepage field monitoring data of the soil surrounding the foundation pit, and three-dimensional deformation monitoring data of the ground surface surrounding the foundation pit includes: Sensors are deployed on the foundation pit support components, retaining walls and anchor cable stress points to collect axial force, bending moment and prestress data of the support structure as internal force response data. The pore water pressure and groundwater level gradient within the influence range of the foundation pit excavation are obtained by using a piezometer array and a water level monitoring well as the seepage field monitoring data. The spatial displacement vector of the ground surface around the foundation pit within a preset sampling period is obtained using a three-dimensional laser scanning device or a GNSS monitoring array, and is used as the three-dimensional deformation monitoring data.

3. The method for early warning of building foundation settlement based on multi-source data fusion according to claim 2, characterized in that, The geological survey data of the foundation pit is used to construct a stratigraphic structure model, which includes: Extract borehole columnar sections from the geological survey data, obtain the physical and mechanical parameters of each soil layer, and use spatial interpolation algorithms to reconstruct the three-dimensional envelope surfaces of the interfaces of each soil layer in the foundation pit to divide the three-dimensional stratigraphic space. Based on the thickness distribution of each stratum in the three-dimensional stratigraphic space, rasterization is performed according to a preset step size to establish multiple stratigraphic nodes. The physical and mechanical parameters are mapped to the corresponding stratigraphic nodes to obtain the first stratigraphic structure model.

4. The method for early warning of building foundation settlement based on multi-source data fusion according to claim 3, characterized in that, Mapping the internal force response data to the corresponding stratigraphic nodes of the stratigraphic structure model according to the stratigraphic spatial coordinates includes: Obtain the spatial coordinates of the internal force monitoring points corresponding to the internal force response data. The internal force monitoring points include axial force monitoring points arranged on the supporting components, bending moment monitoring points of the retaining wall, and prestress monitoring points of the anchor cable stress points. Calculate the three-dimensional spatial distance between each internal force monitoring point and each layer node in the first geological structure model; Based on the three-dimensional spatial distance, and following a weighted allocation strategy based on a spatial attenuation function, the internal force response data is mapped to one or more nearest layer nodes, which are then used as the internal force response components of the layer nodes. The weighted allocation strategy includes: calculating the mapping weight using an exponential decay function based on the distance between the internal force monitoring point and the layer node, and allocating the internal force response data of the internal force monitoring point to the adjacent layer node according to the mapping weight, thereby obtaining the force constraint attributes of each layer node.

5. The method for early warning of building foundation settlement based on multi-source data fusion according to claim 3, characterized in that, Mapping the seepage field monitoring data to the corresponding stratigraphic nodes of the stratigraphic structure model according to the stratigraphic spatial coordinates includes: Obtain the spatial coordinates of the seepage monitoring points corresponding to the seepage field monitoring data, wherein the seepage monitoring points include the locations of the piezometer array and the water level monitoring wells; Extract the set of stratigraphic nodes in the first stratigraphic structure model that are located within the stratigraphic space of each seepage monitoring point; Based on the spatial coordinates of the seepage monitoring points and the spatial coordinates of each layer node, the seepage field monitoring data is distributed to each layer node in the layer node set using the inverse distance weighted interpolation algorithm to obtain the seepage field components of each layer node. The inverse distance weighted interpolation algorithm includes: for the j-th layer node, its seepage field component is obtained by weighted averaging of the monitoring values ​​of each seepage monitoring point within a preset neighborhood of the j-th layer node, and the weight is inversely proportional to the spatial distance from the seepage monitoring point to the j-th layer node, thereby obtaining the seepage driving attribute of each layer node.

6. The method for early warning of building foundation settlement based on multi-source data fusion according to claim 3, characterized in that, Mapping the three-dimensional deformation monitoring data to the corresponding stratigraphic nodes of the stratigraphic structure model according to stratigraphic spatial coordinates includes: Obtain the horizontal spatial coordinates of the surface monitoring points corresponding to the three-dimensional deformation monitoring data; Based on the horizontal spatial coordinates, a target stratigraphic node is determined on the horizontal projection plane of the first stratigraphic structure model. The target stratigraphic node is the stratigraphic node whose horizontal coordinates are closest to the surface monitoring point and located in the uppermost stratigraphic space. The vertical and horizontal displacement components in the three-dimensional deformation monitoring data are vertically projected onto the target layer node to serve as the surface deformation components of the target layer node. If the same target stratum node corresponds to multiple surface monitoring points, the average value of the three-dimensional deformation monitoring data of the multiple surface monitoring points or the weighted average value according to the monitoring accuracy is used as the surface deformation component of the current target stratum node, and the displacement boundary attributes of each stratum node are obtained.

7. The method for early warning of building foundation settlement based on multi-source data fusion according to claim 1, characterized in that, Based on the settlement correlation of each layer node in the stratigraphic node state matrix under the coupling effect of multi-source data, a settlement propagation map of the three fields coupled by support, soil and groundwater is constructed, including: The spatial adjacency relationship of each stratigraphic node in the stratigraphic structure model is determined. The spatial adjacency relationship includes the horizontal adjacency relationship between adjacent stratigraphic nodes within the same stratigraphic space and the vertical adjacency relationship between stratigraphic nodes between adjacent stratigraphic spaces. For each pair of stratigraphic nodes with spatial adjacency, the multi-source state vectors of the source node and the target node are extracted based on the stratigraphic node state matrix, and the coupling difference between the source node and the target node in terms of force constraint attributes, seepage driving attributes and displacement boundary attributes is calculated. Based on the coupling difference, the ratio of the permeability coefficients of the strata to which the source node and the target node belong, the stiffness coupling coefficient of the support structure, and the spatial distance between the source node and the target node, the settlement propagation probability from the source node to the target node is calculated and used as the edge weight of the settlement propagation map. The subsidence propagation probability is compared with a preset propagation threshold. If the subsidence propagation probability is greater than the preset propagation threshold, a directed propagation edge is established between the source node and the target node. The direction of the directed propagation edge is from the subsidence impact source to the subsidence impact target. Traverse all pairs of layer nodes with spatial adjacency, and construct the settlement propagation map of the three-field coupling of support, soil and groundwater based on the directed propagation edge and its corresponding edge weight.

8. The method for early warning of building foundation settlement based on multi-source data fusion according to claim 7, characterized in that, Comparing the settlement propagation probability with a preset propagation threshold includes: Set a first propagation threshold and a second propagation threshold, wherein the first propagation threshold is less than the second propagation threshold; If the subsidence propagation probability is not less than the second propagation threshold, then the stratigraphic node is determined. To the hierarchical node There exists a strong subsidence propagation path, and the directed propagation edge is marked as a strong propagation edge; If the subsidence propagation probability is not less than the first propagation threshold and less than the second propagation threshold, then the stratigraphic node is determined. To the hierarchical node There exists a moderate settlement propagation path, and the directed propagation edge is marked as a moderate propagation edge; If the subsidence propagation probability is less than the first propagation threshold, then the stratigraphic node is determined. To the hierarchical node There is no effective settlement propagation path between them, and no directed propagation edge is established.

9. The method for early warning of building foundation settlement based on multi-source data fusion according to claim 8, characterized in that, The first propagation threshold and the second propagation threshold satisfy the following relationship: ； ； in, Indicates the first propagation threshold. This represents the second propagation threshold. Indicates the basic threshold. This indicates the current depth of the foundation pit excavation. Indicates the reference excavation depth. This indicates the equivalent stiffness of the enclosure wall. Indicates reference stiffness. and Indicates the depth correction factor. This represents the stiffness correction factor.

10. A multi-source data fusion-based early warning system for building foundation settlement, characterized in that, include: The data acquisition module is used to acquire the internal force response data of the foundation pit support structure of the target building, the seepage field monitoring data of the soil around the foundation pit, and the three-dimensional deformation monitoring data of the ground surface around the foundation pit. The data fusion module is used to construct a stratigraphic structure model based on the geological survey data of the foundation pit, and to map the internal force response data, the seepage field monitoring data and the three-dimensional deformation monitoring data to the corresponding layer nodes of the stratigraphic structure model according to the stratigraphic spatial coordinates, forming a stratigraphic node state matrix driven by multi-source data. The map construction module is used to construct a settlement propagation map of the support-soil-groundwater three-field coupling based on the settlement correlation of each layer node in the stratigraphic node state matrix under the coupling effect of multi-source data. The edge weight of the settlement propagation map represents the intensity of settlement influence transmission between adjacent layer nodes. The risk warning module is used to determine the risk level of the building and trigger a graded warning based on the settlement contribution of each node in the settlement propagation map and the safety of the support structure.

Images