A BIM-based foundation pit deformation monitoring method

Abstract

The application provides a BIM-based foundation pit deformation monitoring method, and belongs to the field of building information processing. The method comprises the following steps: obtaining a BIM model of a target foundation pit, introducing a spatial topological relationship into the BIM model, performing semantic enhancement processing according to the introduction result, and constructing a deformation analysis model; performing multi-source monitoring interaction of the target foundation pit, establishing a multi-source monitoring data set, and mapping the multi-source monitoring data set to the deformation analysis model according to the spatial position; performing spatial reconstruction processing on the mapped multi-source monitoring data, and converting the multi-source monitoring data into a continuous three-dimensional deformation field; extracting a deformation characteristic parameter representing a response characteristic of the foundation pit structure, performing deformation mode recognition, and outputting a discrimination result; taking the discrimination result as a first input, taking the continuous three-dimensional deformation field as a second input, performing time sequence evolution trend prediction of the deformation of the target foundation pit, establishing a prediction monitoring result, and improving the comprehensiveness and timeliness of the foundation pit deformation monitoring.

Description

Technical Field

[0001] This invention relates to the field of building information processing, and in particular to a BIM-based method for monitoring foundation pit deformation. Background Technology

[0002] During the construction of foundation pits, the support structure and surrounding soil undergo continuous deformation under the excavation and unloading action. The purpose of deformation monitoring is to acquire key data such as the displacement of the support structure and the settlement of the surrounding ground surface in real time, providing a basis for safety control and scheme adjustments during construction. With the widespread application of Building Information Modeling (BIM) technology in the engineering field, using BIM models to carry foundation pit geometric and construction information has become an industry trend. However, existing foundation pit deformation monitoring methods still have significant shortcomings.

[0003] Currently, foundation pit deformation monitoring mainly relies on discrete measuring points deployed at specific locations. Displacement and settlement data are collected using instruments such as total stations, inclinometers, and levels. Then, manual or simple programs are used to perform threshold comparisons and trend judgments on the data from each measuring point. This approach has the following problems: First, the data from each measuring point are independent, lacking consideration of the mechanical coupling relationship between the support structure and the surrounding soil, making it difficult to reflect the true deformation state of the foundation pit from the overall structural perspective, resulting in biased monitoring results. Second, the lack of deformation information between discrete measuring points prevents the formation of a continuous three-dimensional deformation field, making it difficult to detect localized abnormal deformation outside the coverage area of ​​the measuring points in a timely manner. Furthermore, existing methods rely heavily on engineering experience and manual analysis for deformation pattern identification, resulting in low automation and insufficient predictive ability for deformation development trends. Warnings are often triggered only after deformation has already significantly developed, posing a risk of delayed warnings. Although some studies have attempted to introduce BIM models into the foundation pit monitoring process, most of them remain at the level of three-dimensional visualization. They have failed to fully utilize the structural topological relationships and component mechanical properties contained in the BIM model, and to deeply integrate them with the monitoring data to achieve continuous deformation field reconstruction and automatic identification of deformation modes under structure-soil coupling conditions. The information carrying capacity advantage of BIM models has not been effectively utilized in deformation analysis and early warning prediction. Summary of the Invention

[0004] This invention addresses the technical problem in existing foundation pit deformation monitoring technologies that rely on discrete measuring point data, making it difficult to combine the structure-soil coupling relationship of the BIM model to achieve continuous three-dimensional deformation field reconstruction and automatic identification of deformation modes, resulting in one-sided monitoring results and delayed early warnings. The invention provides a BIM-based foundation pit deformation monitoring method to solve this problem.

[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0006] This invention provides a BIM-based method for monitoring foundation pit deformation, comprising: acquiring a BIM model of the target foundation pit, and introducing soil layering parameters, support structure mechanical parameters, and spatial topological relationships between components into the BIM model; performing semantic enhancement processing based on the introduced results to construct a deformation analysis model including structure-soil coupling relationships; performing multi-source monitoring interaction on the target foundation pit to establish a multi-source monitoring dataset, the multi-source monitoring dataset including displacement monitoring data and settlement monitoring data, and mapping the multi-source monitoring dataset to the component units or soil units corresponding to the deformation analysis model according to spatial location; performing spatial reconstruction processing on the mapped multi-source monitoring data based on the structural topological relationships and mechanical constraints of the deformation analysis model to convert discrete monitoring data into a continuous three-dimensional deformation field that satisfies structural boundary constraints; extracting deformation feature parameters characterizing the structural response features of the foundation pit based on the continuous three-dimensional deformation field, performing deformation pattern recognition using the deformation analysis model, and outputting a discrimination result; using the discrimination result as the first input and the continuous three-dimensional deformation field as the second input to perform temporal evolution trend prediction of the target foundation pit deformation and establish a predictive monitoring result.

[0007] Optionally, semantic enhancement processing is performed based on the introduced results to construct a deformation analysis model including the structure-soil coupling relationship. This includes: defining the support structure component units and soil sub-region units divided by soil layers in the BIM model as network nodes, and assigning corresponding geometric attributes and mechanical parameters to each node; based on the contact relationship and spatial adjacency relationship between the support structure component units and soil sub-region units, establishing coupling edges representing the mechanical action relationship between corresponding nodes, and assigning edge weight parameters reflecting the action intensity to the coupling edges; based on the network nodes and coupling edges, introducing support structure constraints and soil mechanical response constraints to constrain and limit the deformation transmission path between nodes, and constructing a structure-soil coupling topology network; performing unified semantic annotation on each node based on the structure-soil coupling topology network, so that monitoring data propagates along the coupling edges between nodes and forms a constrained deformation response relationship, and constructing a deformation analysis model based on the structure-soil coupling topology network.

[0008] Optionally, deformation feature parameters characterizing the structural response of the foundation pit are extracted based on a continuous three-dimensional deformation field. Deformation pattern recognition is performed using a deformation analysis model, and the discrimination result is output. This includes: extracting the displacement vector, settlement, and displacement gradient and curvature distribution of each node according to the spatial distribution of nodes in the structure-soil coupled topology network based on the continuous three-dimensional deformation field, and establishing node-level deformation feature parameters; calculating the deformation difference between adjacent nodes based on the connection relationship of the coupling edge, and obtaining deformation transfer feature parameters along the direction of the coupling edge to characterize the deformation coordination and transfer intensity between the structure and soil; mapping the node-level deformation feature parameters and the deformation transfer feature parameters to the structure-soil coupled topology network, performing constrained feature propagation and aggregation processing along the coupling edge to form a topological correlation feature vector reflecting the overall structural response; and performing classification recognition using the matching relationship between the topological correlation feature vector and the preset deformation pattern discrimination rule, and outputting the discrimination result.

[0009] Optionally, the preset deformation mode discrimination rule includes a first topological correlation feature vector matching discrimination rule for characterizing the bending deformation of the cantilever. The construction of the first topological correlation feature vector matching discrimination rule includes: extracting the horizontal displacement components of the nodes based on the node sequence of the retaining structure along the depth direction and establishing a displacement gradient distribution sequence in which the displacement gradient changes with depth; calculating the sign consistency of the displacement change rate between nodes based on the displacement gradient distribution sequence and constructing a monotonicity constraint condition characterizing the monotonic change of displacement along the depth direction; calculating the displacement curvature distribution parameters based on the node sequence and constructing a curvature constraint condition characterizing the degree of bending deformation of the retaining structure; and combining the displacement gradient distribution sequence, the monotonicity constraint condition, and the curvature constraint condition to form a first topological correlation feature vector matching discrimination rule for matching with the topological correlation feature vector.

[0010] Optionally, the preset deformation mode discrimination rule includes a second topological correlation feature vector matching discrimination rule for characterizing the constraint failure of the retaining structure. The construction of the second topological correlation feature vector matching discrimination rule includes: extracting the displacement difference between the nodes of the supporting member and the nodes of the adjacent retaining structure to establish the deformation difference distribution characteristics across the support position; calculating the deformation transfer characteristic parameters between the nodes on both sides of the supporting node based on the coupling edge to construct deformation continuity constraint conditions characterizing the deformation continuity change; performing abrupt change analysis on the deformation difference distribution characteristics to construct abrupt change constraint conditions characterizing the local deformation discontinuity; and combining the deformation difference distribution characteristics, deformation continuity constraint conditions, and abrupt change constraint conditions to form a second topological correlation feature vector matching discrimination rule for matching with the topological correlation feature vector.

[0011] Optionally, the preset deformation mode discrimination rule includes a third topological correlation feature vector matching discrimination rule for characterizing soil fluid diffusion. The construction of the third topological correlation feature vector matching discrimination rule includes: extracting the displacement vectors of nodes in the soil sub-region and performing directional consistency analysis to establish the displacement direction distribution characteristics within the region; calculating the spatial dispersion of displacement differences based on the adjacency relationship between nodes to construct a dispersion constraint condition characterizing the deformation diffusion range; performing diffusion analysis on the deformation transmission characteristics of soil nodes along the coupling edge to construct a diffusion constraint condition characterizing the multi-directional diffusion of deformation; and combining the displacement direction distribution characteristics, dispersion constraint condition, and diffusion constraint condition to form a third topological correlation feature vector matching discrimination rule for matching with topological correlation feature vectors.

[0012] Optionally, the classification and recognition are performed by using the matching relationship between the topological correlation feature vector and the preset deformation pattern discrimination rule. This further includes: comparing the topological correlation feature vector and the preset deformation pattern discrimination rule by matching degree, and outputting the discrimination result according to the matching degree.

[0013] Optionally, based on the structural topology and mechanical constraints of the deformation analysis model, spatial reconstruction processing is performed on the mapped multi-source monitoring data to convert the discrete monitoring data into a continuous three-dimensional deformation field that satisfies the structural boundary constraints. This includes: using the multi-source monitoring data mapped to each node in the structure-soil coupled topology network as the initial deformation state quantity, and defining nodes without monitoring data as nodes to be estimated; establishing a weighted correlation relationship for deformation transmission between nodes based on the connection relationship and edge weight parameters of the coupling edges, and transmitting the deformation state of known nodes to adjacent nodes along the coupling edges to form an initial deformation estimation distribution; introducing support structure constraints and soil mechanical response constraints during the deformation transmission process to constrain the deformation transmission values ​​and directions between nodes; and performing iterative balancing processing of the initial deformation estimation distribution to establish a continuous three-dimensional deformation field.

[0014] Optionally, before mapping the multi-source monitoring dataset to the deformation analysis model, time synchronization and outlier removal processing of the multi-source monitoring dataset are performed.

[0015] Optionally, the multi-source monitoring dataset is subjected to self-verification of data sampling, the self-verification result is output, an early warning is issued based on the self-verification result, and resampling processing is performed.

[0016] The beneficial effects of this invention are:

[0017] First, a BIM model of the target foundation pit is acquired, and soil layering parameters, support structure mechanical parameters, and spatial topological relationships between components are introduced into the BIM model. Semantic enhancement processing is performed based on the introduced results to construct a deformation analysis model that includes the structure-soil coupling relationship. This transforms the BIM model from merely carrying geometric information into one capable of describing the mechanical interaction between the support structure and the soil, providing a structured computational foundation for subsequent deformation analysis. Next, multi-source monitoring interaction of the target foundation pit is executed, establishing a multi-source monitoring dataset including displacement and settlement monitoring data. This dataset is mapped to the corresponding component or soil elements in the deformation analysis model according to their spatial location, ensuring that the various monitoring data no longer exist as isolated values ​​but are associated with specific structural locations in the model, achieving spatial binding between monitoring data and model components. Then, based on the structural topological relationships and mechanical constraints of the deformation analysis model, spatial reconstruction processing is performed on the mapped multi-source monitoring data, converting discrete monitoring data into a continuous three-dimensional deformation field that satisfies structural boundary constraints. This compensates for the information gaps between discrete measuring points, restoring the continuous deformation distribution of the entire foundation pit under structural mechanical constraints. Furthermore, deformation characteristic parameters characterizing the structural response of the foundation pit are extracted based on the continuous three-dimensional deformation field. Deformation pattern recognition is then performed using a deformation analysis model, outputting discrimination results to automatically classify and identify the current deformation state of the foundation pit, replacing the traditional manual analysis method that relies on engineering experience. Subsequently, the discrimination results are used as the first input, and the continuous three-dimensional deformation field as the second input to predict the temporal evolution trend of the target foundation pit deformation, establishing predictive monitoring results. This enables the monitoring system to have a forward-looking judgment capability on deformation development trends, providing a time window for early warning decisions.

[0018] The above technical solution introduces soil layering parameters, support structure mechanical parameters, and spatial topological relationships based on the BIM model to construct a structure-soil coupled deformation analysis model. Multi-source discrete monitoring data are reconstructed into a continuous three-dimensional deformation field under structural topological and mechanical constraints. On this basis, automatic identification of deformation modes and prediction of temporal evolution trends are realized. This solves the technical problem in the existing technology that foundation pit deformation monitoring relies on discrete measuring point data, making it difficult to realize continuous deformation field reconstruction and automatic identification of deformation modes, resulting in one-sided monitoring results and delayed early warning. The solution achieves the technical effect of improving the comprehensiveness of foundation pit deformation monitoring and the timeliness of early warning. Attached Figure Description

[0019] Figure 1 A flowchart illustrating a BIM-based foundation pit deformation monitoring method provided by this invention;

[0020] Figure 2 This is a schematic diagram of the process for establishing a continuous three-dimensional deformation field provided by the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0023] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0024] Example 1, as Figure 1 As shown, this embodiment of the invention provides a BIM-based method for monitoring foundation pit deformation, including:

[0025] S1. Obtain the BIM model of the target foundation pit, and introduce soil layering parameters, support structure mechanical parameters and spatial topological relationships between components into the BIM model. Perform semantic enhancement processing based on the introduction results to construct a deformation analysis model including the structure-soil coupling relationship.

[0026] Specifically, the first step is to obtain the BIM model of the target foundation pit. This BIM model is a three-dimensional digital model built based on Building Information Modeling (BIM) technology, containing the retaining structure, support system, support structure components such as columns and piles, as well as the geometric and spatial location information of the surrounding soil. The BIM model can be created using modeling software such as Revit and Tekla, or it can be directly obtained from existing engineering design stage models. Its data format can be the IFC standard format or the native format of each modeling platform.

[0027] After obtaining the BIM model of the target foundation pit, soil layering parameters, support structure mechanical parameters, and spatial topological relationships between components are introduced into the BIM model.

[0028] Regarding the introduction of soil stratification parameters, the distribution depth, layer thickness, and stratigraphic sequence of strata around and at the bottom of the foundation pit are determined based on the borehole columnar section and stratigraphic profile information in the engineering geological survey report. In the BIM model, corresponding three-dimensional solid elements are established according to the actual burial depth and distribution range of each stratum, serving as soil units for each soil layer. The geometric boundary of each soil unit is consistent with the stratigraphic interface determined in the survey report. After establishing the soil units, corresponding physical and mechanical parameters are attached to each soil unit using custom attribute fields, specifically including the natural unit weight, compression modulus, internal friction angle, cohesion, Poisson's ratio, and permeability coefficient of the soil layer. The parameter values ​​are derived from the indoor test results or in-situ test results of the corresponding soil layer in the survey report. For each soil layer within the influence range of the foundation pit excavation, the state attributes of each soil unit under different excavation conditions also need to be marked according to the excavation depth and soil layer distribution relationship, i.e., whether the soil unit is in the excavated and removed state or the retained loaded state, so that the subsequent deformation analysis model can identify the effective soil layer distribution under each condition.

[0029] Regarding the introduction of mechanical parameters for the support structure, the BIM model already includes three-dimensional geometric models of support structure components such as the retaining structure, internal bracing, and columns / piles. Each support structure component in the BIM model corresponds to a component element. Based on this, mechanical parameters are assigned to each component element using custom attribute fields. For the component elements corresponding to the retaining structure, parameters such as material elastic modulus, moment of inertia, equivalent bending stiffness, wall thickness, and embedment depth are assigned. The equivalent bending stiffness is calculated based on the cross-sectional dimensions and material properties of the retaining structure. For the component elements corresponding to the internal bracing, parameters such as axial stiffness, preloaded axial force, bracing spacing, and bracing elevation are assigned. For the component elements corresponding to the columns / piles, parameters such as vertical bearing stiffness, pile diameter, and pile length are assigned. The values ​​of these mechanical parameters are determined according to the foundation pit support design scheme. The foundation pit support design scheme refers to the scheme formed before the construction of the foundation pit project, based on the excavation depth, site geological conditions, and surrounding environmental protection requirements, after comprehensively designing the structural form, component dimensions, material selection, and stress calculations of the support structure.

[0030] Regarding the introduction of spatial topological relationships between components, spatial connections and adjacency relationships are established between component units and between component units and soil units in the BIM model. Specifically, all component units and soil units in the BIM model are traversed, and spatial adjacency is determined based on their geometric positions. When the surface of a component unit corresponding to the retaining structure is in contact with the geometric boundary of a soil unit, a contact adjacency relationship is established between the component unit and the soil unit, and the area and normal direction of the contact surface are recorded. When the end node of a component unit corresponding to an internal support is located within the geometric range of a component unit corresponding to the retaining structure, a node connection relationship is established between the internal support component unit and the corresponding retaining structure component unit, and the coordinates of the connection position and the connection method are recorded. When adjacent component units on the same retaining structure are arranged sequentially along the depth or length direction in space, a structural connection relationship is established between adjacent component units, and the spacing between adjacent component units is recorded. When the surface of the pile body of the component unit corresponding to the column pile is in contact with the geometric boundary of a soil unit, a contact adjacency relationship is established between the column pile component unit and the soil unit, and the area of ​​the contact surface on the pile side is recorded. When two adjacent soil units share a stratum interface, an inter-layer adjacency relationship is established between them. Through the above traversal and judgment, a topological relationship table is formed that records the spatial connection relationships between all component units and soil units. This topological relationship table stores the adjacency, contact, and connection relationships between each unit in the form of node pairs and connection types.

[0031] After the soil layering parameters, support structure mechanical parameters, and spatial topological relationships are introduced, the BIM model expands from simply carrying geometric information to becoming an information carrier that simultaneously carries mechanical properties and structural relationships. Based on this, semantic enhancement processing is performed according to the introduced results, so that each component unit and soil unit in the model not only has geometric position attributes but also carries mechanical parameter attributes and topological connection attributes with adjacent units. The semantics of each unit in the model expands from a simple geometric object to an analytical entity with physical meaning and structural relationships. Through semantic enhancement processing, a deformation analysis model including the structure-soil coupling relationship is constructed. This deformation analysis model uses the semantically enhanced BIM model as its carrier, and its core lies in solidifying the mechanical interaction relationship between the support structure and the surrounding soil in the form of topological connections and mechanical constraints within the model structure. The structure-soil coupling relationship means that the deformation of the support structure is transmitted to the adjacent soil through the contact surface, causing changes in the stress and deformation state of the soil. Conversely, the changes in active and passive earth pressures generated by excavation unloading also act in the opposite direction on the support structure, affecting its deformation response. The deformation analysis model uses the aforementioned spatial topological relationship to structurally express this two-way mechanical interaction relationship, so that the monitoring data received in subsequent steps can be transmitted, reconstructed and analyzed within the framework of this model along the coupling path between the component and the soil, rather than being processed in isolation at each discrete measuring point.

[0032] Through the above processing, the BIM model is transformed from a three-dimensional model that only carries geometric information into a deformation analysis model that simultaneously includes the physical and mechanical properties of the soil, the mechanical properties of the support structure, and the coupling relationship between the two. This enables the model to carry monitoring data and perform deformation analysis under structure-soil coupling conditions.

[0033] S2. Perform multi-source monitoring interaction of the target foundation pit, establish a multi-source monitoring dataset, the multi-source monitoring dataset includes displacement monitoring data and settlement monitoring data, and map the multi-source monitoring dataset to the component unit or soil unit corresponding to the deformation analysis model according to the spatial location.

[0034] Specifically, based on the construction of the deformation analysis model, step S2 is executed to perform multi-source monitoring interaction on the target foundation pit, establish a multi-source monitoring dataset, and map the multi-source monitoring dataset to the deformation analysis model.

[0035] Multi-source monitoring interaction for the target foundation pit refers to the coordinated data collection of various monitoring items of the foundation pit by deploying various types of monitoring instruments at the foundation pit site in accordance with a unified collection plan and monitoring frequency. This ensures that the data obtained by different monitoring methods have a corresponding relationship in the time dimension, and can reflect the deformation status of various parts of the foundation pit at the same time or within the same collection period. The collected monitoring data are then uniformly collected to form a multi-source monitoring dataset.

[0036] This multi-source monitoring dataset includes displacement monitoring data and settlement monitoring data. Displacement monitoring data refers to the horizontal deformation data of the foundation pit retaining structure and surrounding soil acquired by monitoring instruments. Specifically, inclinometer tubes are pre-embedded in the soil inside or adjacent to the retaining structure. The inclinometer measures the inclination angle at each depth along the inclinometer tube from bottom to top. Based on the difference in inclination angle between adjacent measurement segments and the segment length, the horizontal displacement values ​​of each measuring point along the depth direction of the retaining structure relative to the bottom reference point of the tube are accumulated and calculated, forming the deep horizontal displacement data of the retaining structure. Similarly, inclinometer tubes are pre-embedded in the soil surrounding the foundation pit, and the same measurement method is used to acquire the horizontal displacement values ​​of each measuring point along the depth direction of the soil, forming the deep horizontal displacement data of the soil. Prisms or reflectors are pre-set at the top of the retaining structure as displacement monitoring markers. Coordinate observations of each marker point are performed from the reference point using a total station. The observed coordinates at each period are compared with the initial coordinates to obtain the horizontal displacement value of the top of the retaining structure, forming the horizontal displacement data of the top of the retaining structure. Displacement monitoring data comprises the deep horizontal displacement data of the retaining structure, the deep horizontal displacement data of the soil, and the horizontal displacement data of the top of the retaining structure. Settlement monitoring data refers to the vertical deformation data of the ground surface around the foundation pit and the internal structure of the foundation pit, obtained through elevation observation. Specifically, settlement monitoring points are set up at certain intervals along the edge of the foundation pit on the ground surface around the pit. The monitoring points are set up in the form of pre-embedded settlement markers or settlement plates. Starting from a benchmark point far away from the influence range of the foundation pit, elevation observations are carried out at each settlement monitoring point in sequence using a level or electronic level. The observed elevations at each period are compared with the initial elevation to obtain the cumulative settlement and settlement increment of each monitoring point, forming the surface settlement monitoring data. Settlement observation points are set up on the top of the columns inside the foundation pit, and the vertical displacement values ​​of the columns at each period are obtained using the same elevation observation method, forming the column pile settlement monitoring data. The surface settlement monitoring data and the column pile settlement monitoring data together constitute the settlement monitoring data. Each monitoring data point includes its corresponding monitoring point number, three-dimensional spatial coordinates, data acquisition time, and corresponding monitoring value. The three-dimensional spatial coordinates of the monitoring points are determined during the monitoring plan development stage and are calibrated using an engineering coordinate system consistent with the BIM model, ensuring that the spatial location of each monitoring point establishes a clear spatial correspondence with the component units or soil units in the deformation analysis model.

[0037] After establishing the multi-source monitoring dataset, it is mapped to the corresponding component or soil unit in the deformation analysis model according to their spatial location. During the mapping process, the three-dimensional spatial coordinates of the monitoring points carried by each monitoring data point in the multi-source monitoring dataset are read. The shortest distance between this coordinate point and the geometric boundaries of each component and soil unit is calculated in the deformation analysis model, and the monitoring point is assigned to the unit corresponding to the shortest distance. Based on this, the assignment results are further verified according to the physical meaning of the monitoring point and its associated monitoring project. For deep horizontal displacement data of the retaining structure and horizontal displacement data of the top of the retaining structure, it is verified whether they are mapped to the corresponding component unit of the retaining structure. If they are assigned to adjacent soil units due to coordinate deviation, they are corrected and mapped to the nearest retaining structure component unit according to the monitoring project attributes of the monitoring data. For deep horizontal displacement data of the soil, it is verified whether they are mapped to the corresponding soil unit. If they are assigned to adjacent retaining structure component units, they are corrected and mapped to the nearest soil unit. For surface settlement monitoring data, it is verified whether they are mapped to the corresponding soil unit. For the settlement monitoring data of the column piles, verify whether it is mapped to the corresponding component unit of the column pile. When the monitoring point is located at the boundary of two adjacent units and the shortest distance difference is within the preset tolerance range, use the physical attributes of the monitored item as the basis for assignment and map it to the unit with a more matching physical meaning.

[0038] After mapping, measured monitoring data corresponding to the spatial locations are attached to each component and soil element in the deformation analysis model, establishing a binding relationship between the discrete monitoring values ​​and the analysis units with mechanical properties and topological relationships in the model. Through the above processing, the independent data originally scattered in various monitoring instruments are transformed into monitoring information with clear spatial assignment and structural association on the deformation analysis model, enabling the monitoring data to be used for subsequent spatial reconstruction and deformation analysis based on the structure-soil coupling relationship in the deformation analysis model.

[0039] S3. Based on the structural topological relationships and mechanical constraints of the deformation analysis model, spatial reconstruction processing is performed on the mapped multi-source monitoring data to convert the discrete monitoring data into a continuous three-dimensional deformation field that satisfies the structural boundary constraints.

[0040] Specifically, after the mapping process in step S2, only the component units and soil units with monitoring instruments have actual monitoring data attached to them in the deformation analysis model, while a large number of units without monitoring points do not have deformation data. Through spatial reconstruction processing, the discrete monitoring data on the limited measuring points are extrapolated and extended to the unmonitored areas, so that all component units and soil units in the deformation analysis model obtain deformation state quantities, forming a continuous three-dimensional deformation field covering the entire foundation pit.

[0041] Spatial reconstruction processing is based on two aspects: structural topology and mechanical constraints in the deformation analysis model. Structural topology refers to the spatial connections and adjacencies established in step S1 between component units, soil units, and between component and soil units. This topology determines the transmission path of deformation information in the model; deformation data can only be transmitted between units with connections or adjacencies. Units without topological connections do not directly transmit deformation. Mechanical constraints refer to the physical constraints that must be met when deformation is transmitted between units, including support structure constraints and soil mechanical response constraints. Support structure constraints require that the deformation between adjacent component units remain continuous at the connection point, without abrupt changes inconsistent with the structural stiffness. Simultaneously, they require that the deformation of the retaining structure component units and adjacent soil units at the contact surface remain coordinated, meaning that the deformation amounts at the contact surface match. Soil mechanical response constraints require that the deformation transmission between soil units conforms to the physical and mechanical properties of the soil layer, and that the attenuation and diffusion patterns of deformation are compatible with parameters such as the soil layer's compression modulus and Poisson's ratio.

[0042] During the spatial reconstruction process, the measured monitoring data mapped to each unit is used as known constraints. Along the topological connection path of the deformation analysis model, the deformation state of the known measuring points is transferred and calculated to the adjacent unmonitored areas. During the transfer process, the above-mentioned support structure constraints and soil mechanical response constraints are applied, and the transfer results are adjusted. After repeated iterations, the overall deformation distribution converges to a state that simultaneously satisfies the measured data constraints and mechanical constraints, forming a continuous three-dimensional deformation field.

[0043] This continuous three-dimensional deformation field covers all component and soil elements in the deformation analysis model. It is no longer limited to the location of discrete monitoring points, but rather, under the combined effect of structural topological relationships and mechanical constraints, expands the discrete monitoring data into a continuous field reflecting the overall deformation distribution of the foundation pit. The deformation state quantities of each element in the continuous three-dimensional deformation field include both direct constraints from the measured monitoring data and indirect constraints from structural mechanical conditions, ensuring that the deformation distribution conforms to the mechanical response laws of the support structure and soil while satisfying the measured data.

[0044] S4. Based on the continuous three-dimensional deformation field, extract the deformation feature parameters that characterize the response features of the foundation pit structure, use the deformation analysis model to perform deformation pattern recognition, and output the discrimination results.

[0045] Specifically, the continuous three-dimensional deformation field provides complete deformation state quantities for all component units and soil units in the deformation analysis model. The purpose of step S4 is to extract feature parameters that can characterize the current structural response state of the foundation pit from the continuous deformation field, and to determine the current deformation mode of the foundation pit based on these feature parameters, so as to provide clear deformation state discrimination input for the prediction of time-series evolution trend.

[0046] Deformation characteristic parameters refer to quantitative parameters extracted from a continuous three-dimensional deformation field that reflect the degree of deformation and spatial distribution of deformation in various parts of the foundation pit. The extraction of deformation characteristic parameters focuses on each component element and soil element in the deformation analysis model. Basic deformation quantities such as displacement vector and settlement are extracted from the deformation state quantities of each element. Simultaneously, based on the spatial variation relationship of deformation state quantities between adjacent elements, parameters reflecting the spatial variation trend of deformation, such as displacement gradient and curvature distribution, are calculated. Furthermore, based on the coupling relationship between component elements and soil elements in the deformation analysis model, the deformation difference and deformation transfer intensity between adjacent elements are calculated to characterize the deformation coordination state between the support structure and the soil. These various parameters collectively constitute the deformation characteristic parameters, describing the structural response characteristics of the foundation pit at multiple levels, from the local deformation state at the element level to the deformation transfer and coordination relationship between elements.

[0047] After extracting deformation feature parameters, deformation pattern recognition is performed using a deformation analysis model. A deformation pattern refers to a typical deformation form that may occur in the foundation pit during construction. Different deformation patterns reflect the structural response characteristics of the foundation pit support system under different stress states, and each deformation pattern exhibits different characteristics in its deformation feature parameters. During deformation pattern recognition, the extracted deformation feature parameters are compared and matched with pre-established deformation pattern discrimination rules. Each discrimination rule defines the combination of conditions that the deformation feature parameters should satisfy under that pattern. By judging the degree of matching between the current deformation feature parameters and each discrimination rule, the current deformation pattern of the foundation pit is determined, and the discrimination result is output. The discrimination result includes the current deformation pattern type of the foundation pit and the corresponding degree of matching.

[0048] Through the above processing, the deformation information contained in the continuous three-dimensional deformation field is transformed into a qualitative judgment of the current deformation state of the foundation pit. This enables the monitoring system to not only obtain the deformation values ​​of various parts of the foundation pit, but also to automatically identify the deformation mode of the foundation pit, providing a structured state judgment basis for subsequent deformation trend prediction and early warning decision-making.

[0049] S5. Using the discrimination result as the first input and the continuous three-dimensional deformation field as the second input, perform time-series evolution trend prediction of the target foundation pit deformation and establish prediction monitoring results.

[0050] Specifically, by utilizing the output discrimination results and continuous three-dimensional deformation field, the development trend of foundation pit deformation in the future period can be predicted, enabling the monitoring system to expand from passively recording and judging the current deformation state to actively predicting the deformation evolution trend, providing a time window for safety early warning and scheme adjustment during the construction process.

[0051] Using the discrimination result as the first input means incorporating the current excavation pit deformation mode type and corresponding matching degree output from step S4 into the prediction process. The discrimination result provides information on the current deformation mode of the excavation pit. Different deformation modes have different evolutionary patterns and development trends. During the prediction process, constraints corresponding to the current deformation mode need to be applied to the prediction based on the discrimination result. Specifically, the deformation mode type in the discrimination result is used to determine the mode constraint parameters in the prediction process. These mode constraint parameters are a set of constraints predetermined based on the mechanical mechanism of the deformation mode, including the dominant direction constraint of deformation development under this mode, the typical trend constraint of deformation rate over time, and the main spatial expansion region constraint of deformation. The dominant direction constraint specifies the main directional component of the deformation increase of each unit under this deformation mode, such as whether horizontal displacement growth is dominant or vertical settlement growth is dominant. The deformation rate trend constraint specifies the typical characteristics of deformation rate change over time under this mode, i.e., whether the deformation rate tends to converge, remain uniform, or accelerate. The expansion region constraint specifies the spatial range where the deformation increment is mainly concentrated under this mode. The aforementioned mode constraint parameters are pre-established and stored based on the mechanical mechanisms and engineering experience of each deformation mode, and are retrieved during prediction according to the deformation mode type in the discrimination results. When the matching degree in the discrimination results indicates that the deformation characteristics of the foundation pit are similar to multiple deformation modes simultaneously, the mode constraint parameters of each relevant mode are retrieved separately. The matching degree value corresponding to each mode is used as a weighting coefficient to perform a weighted calculation on each group of mode constraint parameters, resulting in the fused mode constraint parameters, which are used for subsequent prediction.

[0052] Using a continuous three-dimensional deformation field as the second input means incorporating the complete deformation distribution state covering all component and soil elements, formed in step S3, into the prediction process. The continuous three-dimensional deformation field provides a complete picture of the spatial deformation distribution of the foundation pit at the current moment, including the deformation amount, deformation direction, and deformation gradient distribution between elements. In the time-series evolution trend prediction, the continuous three-dimensional deformation field serves as the deformation state benchmark at the current moment, providing the initial deformation state of each element for the prediction process. This allows the trend prediction to start from a complete spatial deformation distribution, rather than solely from the numerical values ​​of discrete measurement points.

[0053] After acquiring the first and second inputs, the temporal evolution trend prediction of the target foundation pit deformation is performed. First, using the current deformation state of each element in the continuous three-dimensional deformation field as a benchmark, and combining the deformation state of each element obtained through step S2 mapping and step S3 spatial reconstruction in each historical acquisition cycle, a deformation time series for each element is constructed. Temporal features are extracted from the deformation time series of each element, including the rate of change of deformation over time, the acceleration of change, the growth trend of cumulative deformation, and the evolution law of deformation differences between adjacent elements over time. Specifically, the rate of change is the ratio of the deformation increment to the time interval between two adjacent acquisition cycles; the acceleration of change is the ratio of the rate of change increment to the time interval between two adjacent acquisition cycles; and the growth trend of cumulative deformation is obtained by curve fitting of the cumulative deformation in each historical cycle. Then, the mode constraint parameters determined by the first input are applied to the temporal features of each element, and the deformation time series of each element is extrapolated. The specific method for trend extrapolation is as follows: A polynomial fitting method is used to establish a fitting function between deformation and time for the deformation time series of each unit. The fitting order is determined based on the variation characteristics of the historical deformation time series. When the historical series shows an approximately linear change, a first-order linear fitting is used; when the historical series shows a significant nonlinear acceleration or deceleration trend, a second- or third-order polynomial fitting is used. After establishing the fitting function, the time variable is extended to several future acquisition cycles. The initial predicted deformation of each unit at each prediction time is calculated using the fitting function. After the initial prediction is completed, model constraint parameters are applied to the initial prediction results. Specifically, the directional component of the initial predicted deformation increment of each unit is checked to see if it conforms to the dominant direction constraint. If not, the predicted deformation increment of that unit is projected and corrected towards the dominant direction. The trend of the predicted deformation rate of each unit is checked to see if it conforms to the deformation rate trend constraint. If the initial prediction shows that the deformation rate tends to converge while the model constraint indicates that it should be accelerating, the prediction rate is adjusted accordingly based on the model constraint. The spatial distribution of the predicted deformation increment of each unit is checked to see if it is concentrated within the range limited by the spatial extension region constraint. If the predicted deformation increment is too large in units outside the constraint range, it is attenuated and corrected. After mode constraint adjustment, the prediction results of each element still need to meet the limitations of structural topological relationships and mechanical constraints in the deformation analysis model. That is, the predicted deformation state of each element must still meet the structural continuity constraints, deformation compatibility constraints, and soil mechanical response constraints described in step S3 in space. For the predicted deformation of each element after mode constraint adjustment, spatial consistency verification is performed according to the same constraint check and iterative adjustment method as in step S3. When the predicted deformation value between adjacent elements does not meet the above mechanical constraints, it is adjusted until convergence is achieved to ensure the physical rationality of the prediction results in the spatial dimension.

[0054] Through the above predictive processing, the predicted deformation state quantities of each component unit and soil unit are obtained in several future acquisition cycles, forming the predictive monitoring results. The predictive monitoring results include the predicted displacement, predicted settlement, predicted deformation rate, and the evolution trend of the predicted deformation mode for each unit at each prediction time. Among them, the evolution trend of the predicted deformation mode refers to substituting the predicted deformation state quantities at each prediction time back into the deformation feature parameter extraction and deformation mode identification process in step S4 to obtain the deformation mode discrimination results corresponding to each prediction time. By comparing the changes in the deformation mode at the current time with those at each prediction time, it is determined whether the foundation pit deformation mode has a trend of transforming into a more unfavorable mode.

[0055] A deformation warning is issued when the predicted deformation of any unit in the predictive monitoring results exceeds the preset deformation warning threshold, or the predicted deformation rate exceeds the preset deformation rate warning threshold, or the predicted deformation mode shows a trend of transformation to a more unfavorable mode. The deformation warning threshold is determined based on the control values ​​of each monitoring item specified in the foundation pit support design scheme, and the deformation rate warning threshold is determined based on the allowable deformation rate of each monitoring item. These thresholds are set during the initialization phase according to the specific design parameters and specification requirements of the foundation pit project.

[0056] Through the above treatment, the foundation pit deformation monitoring system is equipped with the ability to make forward-looking judgments on the deformation development trend, and the monitoring work is upgraded from post-deformation recording and analysis of deformation to pre-determining and proactively warning of future deformation evolution trends.

[0057] Through the above steps, soil layering parameters, support structure mechanical parameters, and spatial topological relationships are introduced into the BIM model and semantic enhancement processing is performed to construct a deformation analysis model that includes structure-soil coupling relationship. On this basis, multi-source monitoring data are mapped to component elements and soil elements in the model according to spatial location. The discrete monitoring data are reconstructed into a continuous three-dimensional deformation field by using the structural topological relationship and mechanical constraints of the model. Then, deformation feature parameters are extracted from the continuous deformation field to perform automatic deformation mode identification. Combined with the discrimination results and the continuous three-dimensional deformation field, the temporal evolution trend is predicted to establish predictive monitoring results. This embodiment uses a deformation analysis model as its core carrier, unifying the acquisition, mapping, reconstruction, identification, and prediction of monitoring data within a structure-soil coupled analysis framework. This solves the problems in existing technologies where foundation pit deformation monitoring relies on discrete measuring point data, making it difficult to reconstruct continuous deformation fields and automatically identify deformation patterns, resulting in incomplete monitoring results and delayed early warnings. It expands foundation pit deformation monitoring from discrete measuring points to continuous deformation fields in space, upgrades the analysis from manual experience-based judgment to automatic pattern recognition, and upgrades the temporal aspect from post-event recording to pre-event trend prediction, thereby improving the comprehensiveness of foundation pit deformation monitoring and the timeliness of early warnings.

[0058] Furthermore, semantic enhancement processing is performed based on the introduced results to construct a deformation analysis model that includes the structure-soil coupling relationship, including:

[0059] S11. Define the support structure component units and the soil sub-region units divided by soil layers in the BIM model as network nodes, and assign corresponding geometric properties and mechanical parameters to each node.

[0060] S12. Based on the contact relationship and spatial adjacency relationship between the support structure component unit and the soil sub-region unit, establish coupling edges representing the mechanical action relationship between corresponding nodes, and assign edge weight parameters reflecting the action intensity to the coupling edges.

[0061] S13. Based on the network nodes and coupling edges, introduce support structure constraints and soil mechanical response constraints to constrain and limit the deformation transmission path between nodes, and construct a structure-soil coupled topology network.

[0062] S14. Based on the structure-soil coupled topology network, perform unified semantic annotation on each node, so that the monitoring data can propagate along the coupling edge between nodes and form a constrained deformation response relationship. Based on the structure-soil coupled topology network, construct a deformation analysis model.

[0063] In a preferred embodiment, firstly, the support structure component units and soil sub-region units divided according to soil layers in the BIM model are defined as network nodes, and each node is assigned corresponding geometric attributes and mechanical parameters. Specifically, all support structure component units and soil sub-region units divided according to soil layers in the BIM model after processing in step S1 are traversed, and each support structure component unit and each soil sub-region unit is defined as a network node. Here, the support structure component unit refers to the component unit corresponding to the retaining structure, internal bracing, column piles, etc., in the BIM model in step S1; the soil sub-region unit refers to each soil unit established according to the stratum distribution in step S1, and each soil unit participates in the construction of the coupled topology network as an independent sub-region according to its soil layer and spatial distribution location. After defining the network nodes, geometric attributes and mechanical parameters are assigned to each node. The geometric attributes include the spatial center coordinates, geometric dimensions, and spatial orientation of the unit corresponding to the node. The mechanical parameters are assigned based on the type of the unit corresponding to the node. For nodes corresponding to the retaining structure component unit, parameters such as elastic modulus, moment of inertia, and equivalent bending stiffness already loaded in step S1 are assigned. For nodes corresponding to the internal support component unit, parameters such as axial stiffness and preloaded axial force are assigned. For nodes corresponding to the column pile component unit, parameters such as vertical bearing stiffness are assigned. For nodes corresponding to the soil sub-region unit, parameters such as natural unit weight, compression modulus, internal friction angle, cohesion, and Poisson's ratio of the soil layer are assigned.

[0064] Then, based on the contact and spatial adjacency relationships between the support structure component units and the soil sub-region units, coupling edges representing the mechanical interaction relationships are established between corresponding nodes, and edge weight parameters reflecting the intensity of the interaction are assigned to the coupling edges. The establishment of coupling edges is based on the contact, adjacency, and connection relationships between each unit recorded in the topology table formed in step S1. When there is a contact adjacency relationship, inter-layer adjacency relationship, or node connection relationship between the units corresponding to two nodes in the topology table, a coupling edge is established between these two nodes. The coupling edge represents the mechanical interaction relationship between the units corresponding to the two nodes, that is, the deformation of one unit will affect the deformation of the other unit through the coupling edge. After establishing the coupling edges, an edge weight parameter is assigned to each coupling edge, which is used to quantify the intensity of the mechanical interaction represented by the coupling edge. The calculation method of the edge weight parameter varies depending on the type of the two nodes connected by the coupling edge.

[0065] For example, a support structure component unit includes an enclosure structure component unit, an internal bracing component unit, and a column / pile component unit. For a coupling edge between adjacent support structure component unit nodes on the same support structure, the edge weight parameter is determined based on the stiffness parameter of the support structure and the distance between the two nodes. For a coupling edge between adjacent enclosure structure component unit nodes, the edge weight parameter is equal to the equivalent bending stiffness of the enclosure structure divided by the distance between the two nodes. For a coupling edge between adjacent internal bracing component unit nodes, the edge weight parameter is equal to the axial stiffness of the internal bracing divided by the distance between the two nodes. For a coupling edge between adjacent column / pile component unit nodes, the edge weight parameter is equal to the vertical bearing stiffness of the column / pile divided by the distance between the two nodes.

[0066] For the coupling edges between unit nodes of different types of support structures, that is, the coupling edges between unit nodes of internal support components and unit nodes of enclosure structure components, the edge weight parameter is equal to the axial stiffness of the internal support divided by the distance between the unit nodes of the internal support components and the unit nodes of the enclosure structure components.

[0067] For the coupling edge between the retaining structure component node and the soil sub-region unit node, the edge weight parameter is determined based on the contact surface characteristics and soil mechanical parameters. For the coupling edge between the retaining structure component node and the soil sub-region unit node, the edge weight parameter is equal to the product of the contact surface area between the retaining structure component unit and the soil sub-region unit and the compressibility modulus of the corresponding soil layer of the soil sub-region unit, divided by the distance between the spatial centers of the two nodes. The physical meaning of this edge weight parameter is the ability to transmit deformation per unit distance through the contact surface. For the coupling edge between the column pile component node and the adjacent soil sub-region unit node, the edge weight parameter is equal to the product of the vertical bearing stiffness of the column pile and the pile-side contact area between the column pile and the soil sub-region unit, divided by the distance between the spatial centers of the two nodes.

[0068] For the coupling edge between adjacent soil sub-region unit nodes, the edge weight parameter is equal to the product of the area of ​​the shared interface of the two soil sub-region units and the arithmetic mean of the soil layer compression modulus of the two soil sub-region units, and then divided by the distance between the spatial centers of the two nodes.

[0069] Next, based on the network nodes and coupling edges, support structure constraints and soil mechanical response constraints are introduced to limit the deformation transmission path between nodes, constructing a structure-soil coupled topology network. The network nodes and coupling edges constitute the basic topological connection structure, but relying solely on edge weight parameters cannot fully reflect the mechanical constraint relationships in actual engineering. Therefore, constraints are introduced into the topology to limit the deformation transmission behavior between nodes. Support structure constraints include deformation continuity constraints and deformation compatibility constraints. Deformation continuity constraints are applied to the coupling edges between adjacent member element nodes on the same support structure, requiring that the deformation value transmitted along the coupling edge remains continuous at both nodes, without any abrupt changes inconsistent with the structural stiffness. Deformation continuity constraints are achieved by setting an upper limit threshold for the deformation difference. This threshold is equal to the distance between two adjacent member element nodes divided by the product of the member's equivalent bending stiffness and the constraint coefficient. The constraint coefficient is a calibration parameter related to the member type, determined according to the deformation characteristics of each member type. When the deformation difference between adjacent component unit nodes exceeds the upper limit threshold, the deformation transmission at that point is deemed to fail to meet the continuity constraint and requires adjustment during subsequent spatial reconstruction. Deformation compatibility constraints are applied to the coupling edge between the retaining structure component unit node and the adjacent soil sub-region unit node. This requires that the deformation value transmitted through this coupling edge meets the compatibility relationship at both nodes; that is, the difference between the lateral displacement of the retaining structure node and the deformation of the adjacent soil node in the normal direction of the contact surface does not exceed the allowable deviation range. The allowable deviation range is equal to the normal stress on the contact surface between the retaining structure and the soil divided by the compressive modulus of the corresponding soil layer in the soil sub-region unit. Its physical meaning is the amount of elastic deformation that the soil can produce under the current normal stress level at the contact surface. The normal stress is calculated based on the earth pressure at that depth. Soil mechanical response constraints are applied to the coupling edge between adjacent soil sub-region unit nodes, requiring that the deformation transmitted along this coupling edge conforms to the stress-strain relationship of the corresponding soil layer. Soil mechanical response constraints are implemented by setting a deformation transmission attenuation coefficient. First, the constraint modulus of the soil layer is calculated based on the compressive modulus and Poisson's ratio of the corresponding soil layer. The constraint modulus is calculated as follows: subtract Poisson's ratio from 1 to obtain the first intermediate value; add Poisson's ratio to 1 to obtain the second intermediate value; subtract twice Poisson's ratio from 1 to obtain the third intermediate value; multiply the second and third intermediate values ​​to obtain the denominator; multiply the compression modulus by the first intermediate value to obtain the numerator; and divide the numerator by the denominator to obtain the constraint modulus. Then, the attenuation coefficient is calculated by dividing the distance between two nodes by the constraint modulus to obtain the distance modulus ratio. The attenuation coefficient is equal to 1 divided by 1 plus the distance modulus ratio. When deformation is transferred from one soil node to an adjacent soil node, the amount of deformation after transfer is the amount of deformation before transfer multiplied by the attenuation coefficient. Soil layers with larger constraint moduli have an attenuation coefficient close to 1, indicating slower deformation attenuation. Soil layers with smaller constraint moduli have a smaller attenuation coefficient, indicating faster deformation attenuation.The aforementioned constraints are applied as parameters to the corresponding coupling edges, and together with the edge weight parameters, they define the mode and magnitude of deformation transmission on that coupling edge. By applying the corresponding constraints to all coupling edges, the network nodes and coupling edges together constitute a structure-soil coupled topology network.

[0070] Subsequently, based on the structure-soil coupled topology network, unified semantic annotation was applied to each node, enabling monitoring data to propagate along the coupling edges between nodes and form constrained deformation response relationships. A deformation analysis model was then constructed based on the structure-soil coupled topology network. Unified semantic annotation refers to assigning standardized semantic label information to each node in the coupled topology network. The semantic label information includes the node's element type identifier, its structural component identifier, depth location identifier, and its topological role identifier within the network. The element type identifier indicates whether the node corresponds to a retaining structure component, internal support component, column / pile component, or soil sub-region unit. The structural component identifier indicates the node's location within the overall structure of the foundation pit, such as a retaining structure on one side of the pit, a certain support, or a layer of soil below the pit bottom. The depth location identifier indicates the depth range of the node. The topological role identifier indicates the node's association characteristics with other nodes in the coupled topology network, such as whether the node is a boundary node, a support connection node, or a multi-layer soil interface node. Through unified semantic annotation, each node in the coupled topology network not only possesses geometric attributes and mechanical parameters but also a standardized semantic identifier. This enables monitoring data mapped to a node in subsequent steps to automatically identify the node's type, location, and topological role. Based on the connection relationships, edge weights, and constraints of the coupled edges, the data propagates along the coupled edges to adjacent nodes, forming a constrained deformation response relationship. Based on the semantically annotated structure-soil coupled topology network, a deformation analysis model is constructed. This model, with the coupled topology network as its core structure, carries node attributes, coupled edge connection relationships, edge weights, constraints, and semantic annotation information, providing a unified structured analysis framework for subsequent monitoring data mapping, spatial reconstruction, deformation pattern recognition, and trend prediction.

[0071] Through the above processing, the geometric information, mechanical parameters and spatial topological relationships in the BIM model are transformed into a structure-soil coupled topological network with network nodes, coupling edges, edge weight parameters, constraints and semantic annotations as elements. The deformation analysis model built on this basis has the ability to bear monitoring data and perform deformation transfer and constraint analysis under the coupling conditions of structure and soil.

[0072] Furthermore, deformation feature parameters characterizing the structural response of the foundation pit are extracted based on a continuous three-dimensional deformation field. Deformation pattern recognition is then performed using a deformation analysis model, and the discrimination results are output, including:

[0073] S41. Based on the continuous three-dimensional deformation field and the spatial distribution of nodes in the structure-soil coupled topology network, extract the displacement vector, settlement, displacement gradient and curvature distribution of each node, and establish node-level deformation characteristic parameters.

[0074] S42. Based on the connection relationship of the coupling edge, the deformation difference between adjacent nodes is calculated, and the deformation transfer characteristic parameters along the direction of the coupling edge are obtained to characterize the deformation coordination and transfer strength between the structure and the soil.

[0075] S43. Map the node-level deformation feature parameters and the deformation transfer feature parameters to the structure-soil coupled topology network, and perform constrained feature propagation and aggregation processing along the coupling edges to form a topological correlation feature vector that reflects the overall structural response.

[0076] S44. Utilize the matching relationship between the topological correlation feature vector and the preset deformation pattern discrimination rule to perform classification and recognition, and output the discrimination result.

[0077] In a preferred embodiment, firstly, based on the continuous three-dimensional deformation field and the spatial distribution of nodes in the structure-soil coupled topology network, the displacement vector, settlement, and displacement gradient and curvature distribution corresponding to each node are extracted to establish node-level deformation characteristic parameters. Specifically, each component element and soil element in the continuous three-dimensional deformation field already has deformation state quantities, and each node in the structure-soil coupled topology network corresponds one-to-one with the component element and soil element. Therefore, according to the spatial position of each node, the deformation state quantities of the corresponding element are read from the continuous three-dimensional deformation field. For each node, the extracted deformation characteristic parameters include displacement vector, settlement, displacement gradient, and curvature distribution. The displacement vector refers to the horizontal and vertical displacement components of the element corresponding to the node in three-dimensional space. The horizontal displacement component reflects the lateral deformation state of the element, and the vertical displacement component reflects the settlement or heave state of the element. The settlement refers to the vertical displacement value of the element corresponding to the node; a positive value downwards indicates settlement, and a negative value upwards indicates heave. After extracting the displacement vectors and settlement of each node, the displacement gradient and curvature distribution resulting from spatial changes are further calculated. The displacement gradient is calculated by selecting a sequence of adjacent nodes along a certain direction in the coupled topology network, calculating the difference in displacement components between two adjacent nodes, and dividing this difference by the distance between the two nodes. This yields the displacement gradient value in that direction, reflecting the rate of spatial change of deformation along that direction. The curvature distribution is calculated by taking the displacement values ​​of three consecutive adjacent nodes in the same node sequence, using the middle node as a reference, calculating the sum of the deviations between the displacement values ​​of the two adjacent nodes and the middle node, and dividing this sum by the square of the node distance. This yields the curvature value at that location, reflecting the degree of bending of deformation along that direction. The displacement vector, settlement, displacement gradient, and curvature distribution of each node together constitute the node-level deformation characteristic parameters of that node.

[0078] Then, based on the connection relationship of the coupling edges, the deformation difference between adjacent nodes is calculated to obtain the deformation transfer characteristic parameters along the coupling edge direction, so as to characterize the deformation compatibility and transfer intensity between the structure and soil. Specifically, all coupling edges in the structure-soil coupling topology network are traversed. For each coupling edge, the node-level deformation characteristic parameters of the nodes at both ends of the coupling edge are read, and the deformation difference between the two ends is calculated. The calculation of deformation difference includes displacement difference, settlement difference, and displacement gradient difference. Among them, displacement difference refers to the difference of each component of the displacement vector of the two ends of the node, reflecting the degree of inconsistency in deformation between the two ends of the coupling edge in various directions. Settlement difference refers to the difference of settlement of the two ends of the node, reflecting the degree of differential settlement between the two ends of the coupling edge. Displacement gradient difference refers to the difference of displacement gradient of the two ends of the node in the same direction, reflecting whether there is an abrupt change in the rate of change of deformation space at both ends of the coupling edge. Based on the calculation of deformation difference, deformation transfer characteristic parameters are further calculated. Deformation transfer characteristic parameters include deformation transfer ratio and deformation compatibility. The deformation transfer ratio is equal to the displacement value of the node at one end of the coupling edge divided by the displacement value of the node at the other end. It reflects the degree of attenuation or amplification of deformation as it is transmitted from one side to the other through the coupling edge. When the deformation transfer ratio is close to 1, it indicates that the deformation transfer is relatively sufficient. A large deviation from 1 indicates that there is significant attenuation or concentration of deformation at the coupling edge. The deformation compatibility is equal to 1 minus the absolute value of the displacement difference between the two nodes, divided by the larger of the absolute values ​​of the displacements of the two nodes. The deformation compatibility ranges from 0 to 1. A value close to 1 indicates that the deformation at both ends of the coupling edge is highly coordinated. A value close to 0 indicates that the deformation difference between the two ends is significant, and the deformation compatibility between the structure and the soil or between adjacent units is poor. The displacement difference, settlement difference, displacement gradient difference, deformation transfer ratio, and deformation compatibility of each coupling edge together constitute the deformation transfer characteristic parameters of the coupling edge.

[0079] Next, the node-level deformation feature parameters and deformation transfer feature parameters are mapped to the structure-soil coupled topology network. Constrained feature propagation and aggregation along the coupling edges are then performed to form a topologically related feature vector reflecting the overall structural response. Specifically, firstly, the obtained node-level deformation feature parameters of each node are assigned to the corresponding nodes in the coupled topology network, and the obtained deformation transfer feature parameters of each coupling edge are assigned to the corresponding coupling edge, so that each node and each coupling edge in the coupled topology network is attached with the corresponding feature parameters. Then, constrained feature propagation and aggregation along the coupling edges are performed. Feature propagation refers to each node transmitting its own node-level deformation feature parameters to adjacent nodes along the coupling edge. The transmission process is constrained by the edge weight parameters and constraints on the coupling edge. The specific propagation method is as follows: For each node, the node-level deformation feature parameters of all adjacent nodes connected to it through the coupling edge, as well as the deformation transfer feature parameters on the connecting coupling edge, are collected. Using the edge weight parameters of each coupling edge as weight coefficients, the collected adjacent node feature parameters are weighted and summed to obtain the propagated features received by the node from its adjacent nodes. During propagation, constraints on the coupling edges adjust the propagation features. When a feature parameter transmitted from an adjacent node does not meet the constraints on that coupling edge, the propagation feature is attenuated or corrected. Feature aggregation refers to merging the node-level deformation feature parameters of each node with the propagation features received from adjacent nodes to form the aggregated feature parameters of that node. The aggregation method involves weighting the node's own deformation feature parameters and the propagation features according to preset aggregation weights. The weight of the node's own features in the aggregation weights is determined based on whether the node is a measured monitoring point. Nodes corresponding to measured monitoring points have higher weights for their own features, while nodes corresponding to non-measured monitoring points have lower weights for their own features. For example, for nodes corresponding to measured monitoring points, the weight of their own features is 0.7, and the weight of the propagation features is 0.3; for nodes corresponding to non-measured monitoring points, the weight of their own features is 0.3, and the weight of the propagation features is 0.7. After feature propagation and aggregation, the aggregated feature parameters of each node no longer only reflect the local deformation state of the node itself, but also integrate the deformation information transmitted from adjacent nodes through coupling edges, thus possessing the ability to reflect the response features of local areas and even the overall structure. The aggregated feature parameters of all nodes in the coupled topology network are arranged and spliced ​​according to the node numbering order in the coupled topology network to form a unified topological correlation feature vector. This topological correlation feature vector comprehensively reflects the overall structural response characteristics of the foundation pit under the current deformation state.

[0080] Subsequently, classification and recognition are performed using the matching relationship between the topological correlation feature vector and the preset deformation mode discrimination rules, and the discrimination results are output. The preset deformation mode discrimination rules refer to a combination of discrimination conditions established in advance based on the mechanical characteristics of various typical deformation modes. Each deformation mode corresponds to a set of discrimination rules, which define the numerical and relational conditions that each feature component in the topological correlation feature vector should satisfy under that deformation mode. During classification and recognition, the formed topological correlation feature vector is compared with each preset deformation mode discrimination rule, the degree of matching between the topological correlation feature vector and each discrimination rule is calculated, and the deformation mode of the current foundation pit is determined based on the degree of matching, and the discrimination results are output. The discrimination results include the deformation mode type of the current foundation pit and the corresponding degree of matching.

[0081] Through the above steps, the structural response features were extracted layer by layer from the continuous three-dimensional deformation field, from node-level local deformation features to coupled edge-level deformation transmission features, and then to the overall structural response features that integrate topological correlation information. Based on the matching of these features with preset discrimination rules, the deformation mode was automatically identified, enabling the monitoring system to automatically classify and discriminate the current deformation state of the foundation pit.

[0082] Furthermore, the preset deformation mode discrimination rule includes a first topological correlation feature vector matching discrimination rule for characterizing the bending deformation of the cantilever. The construction of the first topological correlation feature vector matching discrimination rule includes:

[0083] S4411. Based on the node sequence of the enclosure structure along the depth direction, extract the horizontal displacement components of the nodes and establish a displacement gradient distribution sequence that varies with depth.

[0084] S4412. Based on the displacement gradient distribution sequence, calculate the sign consistency of the displacement change rate between nodes, and construct a monotonicity constraint condition characterizing the monotonic change of displacement along the depth direction.

[0085] S4413. Calculate the displacement curvature distribution parameters based on the node sequence, and construct curvature constraint conditions characterizing the degree of bending deformation of the enclosure structure;

[0086] S4414. The displacement gradient distribution sequence, monotonicity constraint, and curvature constraint are combined to form a first topological correlation feature vector matching discrimination rule for matching with the topological correlation feature vector.

[0087] In a preferred embodiment, the preset deformation mode discrimination rule includes a first topological correlation feature vector matching discrimination rule for characterizing cantilever bending deformation. Cantilever bending deformation refers to the bending deformation of the retaining structure in the pit under the action of earth pressure outside the pit, with the bottom of the wall as the fixed point and the top of the wall as the free end, when the retaining structure has not yet been internally supported or the internal support has not yet played a restraining role. Its typical characteristics are that the horizontal displacement of the retaining structure increases monotonically from the bottom to the top along the depth direction, and the displacement curve shows a bending shape that bulges into the pit.

[0088] The process of constructing the first topological association feature vector matching and discrimination rule is as follows:

[0089] First, based on the node sequence of the retaining structure along the depth direction, the horizontal displacement components of the nodes are extracted, and a displacement gradient distribution sequence showing the variation of displacement gradient with depth is established. Specifically, in the structure-soil coupled topology network, the retaining structure consists of a node sequence formed by multiple adjacent retaining structure component nodes arranged sequentially along the depth direction. The node sequence is arranged from the bottom node to the top node of the retaining structure, and the nodes are connected sequentially by coupling edges. The horizontal displacement components corresponding to each node in this node sequence are extracted from the topological association feature vector. The horizontal displacement components are taken as the displacement values ​​of the retaining structure in the direction into the pit. After extracting the horizontal displacement components of each node, the displacement gradient between two adjacent nodes is calculated, that is, the horizontal displacement component of the next node minus the horizontal displacement component of the current node, and then divided by the distance between the two nodes along the depth direction, to obtain the displacement gradient value at that location. The displacement gradient values ​​between each pair of adjacent nodes in the node sequence are arranged in order from bottom to top to form a displacement gradient distribution sequence. The displacement gradient distribution sequence reflects the rate of change of the horizontal displacement of the retaining structure along the depth direction and is the basic data for determining the bending deformation mode of the cantilever.

[0090] Then, based on the sign consistency of the displacement change rate between nodes calculated from the displacement gradient distribution sequence, a monotonicity constraint condition characterizing the monotonic change of displacement along the depth direction is constructed. Specifically, the sign of each displacement gradient value in the displacement gradient distribution sequence is determined one by one. Under the cantilever bending deformation mode, the horizontal displacement of the retaining structure monotonically increases from the bottom to the top. Therefore, all displacement gradient values ​​in the displacement gradient distribution sequence should be positive, i.e., the sign consistency is all positive. The sign consistency is calculated by counting the ratio of the number of positive displacement gradients to the total number of displacement gradients in the displacement gradient distribution sequence. When this ratio equals 1, it indicates that all displacement gradients are positive, and the displacement increases strictly monotonically along the depth direction. When this ratio is close to 1 but not equal to 1, it indicates that there are individual locations where the displacement gradient is negative or zero, and the monotonicity of the displacement deviates locally. Based on this ratio, a monotonicity constraint is constructed. The monotonicity constraint requires that the ratio of positive displacement gradients in the displacement gradient distribution sequence is not lower than a preset monotonicity threshold. This monotonicity threshold is set according to engineering experience, for example, 0.9, which means that no more than 10% of the displacement gradients are allowed to have non-positive local deviations.

[0091] Next, displacement curvature distribution parameters are calculated based on the node sequence to construct curvature constraints characterizing the degree of bending deformation of the enclosure structure. Specifically, in the node sequence along the depth direction of the enclosure structure, three consecutive adjacent nodes are selected, and the curvature value at that location is calculated according to the curvature distribution calculation method in step S41, using the middle node as the reference. The curvature values ​​at each location are calculated sequentially from bottom to top along the node sequence to form curvature distribution parameters. Curvature constraints are constructed based on curvature distribution parameters. The curvature constraints include two requirements: First, the curvature direction consistency requirement, that is, the ratio of the number of curvature values ​​with the same sign direction in the curvature distribution parameters to the total number of curvature values ​​is not less than a preset curvature sign consistency threshold, for example, 0.85. This requirement is used to determine whether the bending deformation of the enclosure structure is dominated by the same direction. Second, the curvature distribution trend requirement, that is, the node corresponding to the maximum curvature value in the curvature distribution parameters is located in the bottom 1 / 3 of the node sequence, and from the node where the maximum curvature value is located to the top, the ratio of the decrease in the number of curvature values ​​between adjacent nodes to the total number of adjacent node pairs is not less than a preset curvature decrease ratio threshold, for example, 0.8. This requirement is used to determine whether the spatial distribution of bending deformation conforms to the mechanical characteristics of cantilever bending.

[0092] Next, the displacement gradient distribution sequence, monotonicity constraint, and curvature constraint are combined to form the first topological correlation feature vector matching discrimination rule for matching with the topological correlation feature vector. Specifically, the first topological correlation feature vector matching discrimination rule consists of the following three conditions: The first condition is that the displacement gradient values ​​in the displacement gradient distribution sequence increase overall from bottom to top along the depth direction, that is, the ratio of the number of adjacent displacement gradient values ​​that show an increasing relationship to the total number of adjacent gradient pairs is not less than a preset gradient increase ratio threshold, for example, 0.8, reflecting that the deformation of the retaining structure accelerates when approaching the free end; the second condition is that the monotonicity constraint condition is met, that is, the ratio of positive values ​​in the displacement gradient distribution sequence is not less than the monotonicity threshold; the third condition is that the curvature constraint condition is met, that is, the curvature distribution parameters meet the requirement of consistency of curvature direction and the position of maximum curvature and the curvature decreasing trend conform to the cantilever bending characteristics. When performing deformation pattern recognition, the feature components corresponding to the sequence of building envelope nodes in the topological association feature vector are extracted, and the above three conditions are checked respectively. The degree of matching between the topological association feature vector and the first topological association feature vector matching discrimination rule is calculated based on the degree of satisfaction of each condition.

[0093] Through the above processing, a cantilever bending deformation discrimination rule based on the displacement gradient distribution, monotonicity and curvature distribution characteristics of the retaining structure along the depth direction was established, namely the first topological correlation feature vector matching discrimination rule, so that the deformation pattern recognition process can automatically determine whether there is a cantilever bending deformation mode based on the morphological characteristics of the retaining structure displacement curve.

[0094] Furthermore, the preset deformation mode discrimination rule includes a second topological correlation feature vector matching discrimination rule for characterizing the constraint failure of the support structure. The construction of the second topological correlation feature vector matching discrimination rule includes:

[0095] S4421. Extract the displacement difference between the support member node and the adjacent enclosure structure node, and establish the deformation difference distribution characteristics across the support location.

[0096] S4422. Based on the coupling edge, calculate the deformation transfer characteristic parameters between the nodes on both sides of the support node, and construct deformation continuity constraint conditions to characterize the change in deformation continuity.

[0097] S4423. Perform a mutation analysis on the deformation difference distribution characteristics to construct mutation constraint conditions characterizing local deformation discontinuity;

[0098] S4424. Combine the deformation difference distribution characteristics, deformation continuity constraints, and abrupt change constraints to form a second topological association feature vector matching discrimination rule for matching with topological association feature vectors.

[0099] In a preferred embodiment, the preset deformation mode discrimination rule includes a second topological correlation feature vector matching discrimination rule for characterizing the constraint failure of the support structure. The constraint failure of the support structure refers to the loss or significant weakening of the horizontal constraint effect of the internal supports on the retaining structure during construction due to reasons such as excessive axial force, failure of connection nodes, or instability of the support system. This results in the retaining structure losing effective horizontal support at the location of the internal supports, leading to localized abnormal deformation near that location. Its typical characteristics include an abnormally large increase in the displacement difference between the retaining structure and the internal supports at the support location, and a significant interruption in the continuity of deformation of the retaining structure on both sides of the support location.

[0100] The process of constructing the second topological association feature vector matching and discrimination rule is as follows:

[0101] First, the displacement differences between the supporting member nodes and adjacent retaining structure nodes are extracted to establish the deformation difference distribution characteristics across the supporting locations. Specifically, in the structure-soil coupled topology network, each internal supporting member node and retaining structure member node is connected through coupling edges. All internal supporting member nodes are traversed. For each internal supporting node, the displacement vector of that node and the displacement vector of the retaining structure node connected to it through the coupling edge are read, and the horizontal displacement difference between the two is calculated, which is the horizontal displacement component of the retaining structure node minus the horizontal displacement component of the internal supporting node. Under normal constraint conditions, the internal supports exert effective horizontal constraints on the retaining structure, resulting in a small horizontal displacement difference between the two. When the internal support constraints fail, the horizontal displacement of the retaining structure at that location increases while the internal supports cannot follow the constraints, leading to a significant increase in the horizontal displacement difference between the two. The horizontal displacement differences between each internal supporting node and its corresponding retaining structure node are arranged in order of the spatial location of the internal supports, forming the deformation difference distribution characteristics across the supporting locations. The deformation difference distribution characteristics reflect the deformation coordination state between the retaining structure and the internal supports at each internal support location.

[0102] Then, based on the coupling edge calculation, deformation transfer characteristic parameters between the nodes on both sides of the support node are calculated, and deformation continuity constraints characterizing the change in deformation continuity are constructed. Specifically, for each inner support node, the retaining structure node closest to the connection position of the inner support node is determined in the node sequence along the depth direction of the retaining structure, and one adjacent retaining structure node above and one below the inner support node are taken respectively, and the upper and lower nodes are taken as the nodes on both sides of the support node. The node-level deformation characteristic parameters of the upper and lower nodes are read, and the deformation transfer characteristic parameters between them are calculated, including the deformation transfer ratio and deformation coordination degree in step S42. Under normal constraint conditions, the inner support effectively plays a constraint role, the deformation of the retaining structure on both sides of the support position maintains a continuous transition, the deformation transfer ratio is close to 1, and the deformation coordination degree is high; when the inner support constraint fails, the support position becomes the boundary point of deformation, the deformation continuity on both sides is destroyed, the deformation transfer ratio deviates from 1, and the deformation coordination degree decreases. Based on the above analysis, deformation continuity constraints are constructed. The deformation continuity constraints require that when the deformation coordination degree between the nodes on both sides of the support node is lower than the preset deformation continuity threshold, the deformation continuity at that location is determined to have changed. The deformation continuity threshold is set according to engineering experience, for example, 0.6. That is, when the deformation coordination degree is lower than 0.6, the deformation continuity at that support location is determined to be abnormal.

[0103] Next, abrupt change analysis was performed on the deformation difference distribution characteristics to construct abrupt change constraint conditions characterizing local deformation discontinuities. Specifically, statistical analysis was conducted on the horizontal displacement differences at each internal support location within the deformation difference distribution characteristics, calculating the mean and standard deviation of all horizontal displacement differences. The abrupt change analysis included two aspects: First, single-point abrupt change judgment. When the horizontal displacement difference at a certain internal support location exceeded the mean plus a first preset multiple of the standard deviation, a deformation abrupt change was determined at that location. The first preset multiple was set based on engineering experience, for example, 2. Second, adjacent jump judgment. The absolute value of the difference between the horizontal displacement differences of two adjacent internal support locations was calculated, i.e., the absolute value of the difference between the horizontal displacement differences of the later internal support location and the difference between the horizontal displacement differences of the earlier internal support location. When this absolute value exceeded a second preset multiple of the standard deviation of all horizontal displacement differences, a discontinuous jump in deformation was determined between adjacent support locations. The second preset multiple was set based on engineering experience, for example, 1.5. Based on the above analysis, abrupt change constraint conditions were constructed. These constraints required that at least one internal support location within the deformation difference distribution characteristics satisfy either the single-point abrupt change judgment or the adjacent jump judgment.

[0104] Next, the deformation difference distribution characteristics, deformation continuity constraints, and abrupt change constraints are combined to form a second topological correlation feature vector matching discrimination rule for matching with the topological correlation feature vector. Specifically, the second topological correlation feature vector matching discrimination rule consists of the following three conditions: The first condition is that there is an inner support location with a significantly large horizontal displacement difference in the deformation difference distribution characteristics, that is, the horizontal displacement difference at least one inner support location exceeds a preset proportion of the average horizontal displacement difference of all inner support locations, for example, 1.5 times the average; the second condition is that the deformation continuity constraint condition is met, that is, the deformation coordination on both sides of the support node at the abnormal location is lower than the deformation continuity threshold; the third condition is that the abrupt change constraint condition is met, that is, the abnormal location meets the single-point abrupt change judgment or the adjacent jump judgment. When the above three conditions are simultaneously met at the same inner support location, it is determined that there is a support structure constraint failure at that location. When performing deformation pattern recognition, the feature components corresponding to each inner support node and its adjacent enclosure structure node in the topological association feature vector are extracted, and the above three conditions are checked respectively. The matching degree between the topological association feature vector and the second topological association feature vector matching discrimination rule is calculated based on the degree of satisfaction of each condition.

[0105] Through the above processing, a support structure constraint failure discrimination rule based on the deformation difference, deformation continuity and deformation abrupt change characteristics of the support position was established, namely the second topological correlation feature vector matching discrimination rule, which enables the deformation pattern recognition process to automatically determine whether there is a local abnormal deformation mode caused by the failure of internal support constraint.

[0106] Furthermore, the preset deformation mode discrimination rule includes a third topological correlation feature vector matching discrimination rule for characterizing soil fluid diffusion. The construction of the third topological correlation feature vector matching discrimination rule includes:

[0107] S4431. Extract the displacement vectors of the nodes in the soil sub-region and perform directional consistency analysis to establish the displacement direction distribution characteristics within the region;

[0108] S4432. Calculate the spatial dispersion of displacement difference based on the adjacency relationship between nodes, and construct dispersion constraints to characterize the deformation diffusion range.

[0109] S4433. Perform diffusion analysis on the deformation transfer characteristics of soil nodes along the coupling edge, and construct diffusion constraint conditions to characterize the multi-directional diffusion of deformation.

[0110] S4434. The displacement direction distribution characteristics, discreteness constraints, and diffusion constraints are combined to form a third topological association feature vector matching discrimination rule for matching with topological association feature vectors.

[0111] In a preferred embodiment, the preset deformation mode discrimination rule includes a third topological correlation feature vector matching discrimination rule for characterizing soil fluid diffusion. Soil fluid diffusion refers to the fluid deformation of soil under unloading during the excavation process when the soil at the bottom or side of the pit is in a weak and saturated state. The deformation is not limited to a fixed direction or a local area, but spreads outward in multiple directions. Its typical characteristics are that the displacement direction of each node in the soil area is dispersed rather than uniform, and the deformation is transmitted and extended outward along multiple coupled edge paths.

[0112] The process of constructing the third topological association feature vector matching and discrimination rule is as follows:

[0113] First, the displacement vectors of the soil sub-region nodes are extracted and directional consistency analysis is performed to establish the displacement direction distribution characteristics within the region. Specifically, in the structure-soil coupled topology network, nodes corresponding to all soil sub-region units are selected, and the displacement vectors of each soil node are extracted from the topological association feature vectors. For each soil node's displacement vector, its orientation angle is calculated; the orientation angle is the angle between the projection direction of the displacement vector in the horizontal plane and a preset reference direction. After extracting the orientation angles of all soil nodes, directional consistency analysis is performed. The method of directional consistency analysis is as follows: the average orientation angle of all soil node orientation angles is calculated, then the absolute value of the deviation between each node's orientation angle and the average orientation angle is calculated, and the ratio of the number of nodes with an absolute deviation value less than a preset directional deviation threshold to the total number of soil nodes is statistically analyzed. This ratio is defined as the directional consistency coefficient. The directional deviation threshold is set based on engineering experience, for example, 30 degrees. When the directional consistency coefficient is close to 1, it indicates that the displacement directions of each soil node are highly consistent, and the deformation is mainly overall translational, not a fluid diffusion characteristic. When the directional consistency coefficient is low, it indicates that the displacement directions of each soil node are dispersed, and the deformation exhibits characteristics of diffusion in multiple directions. The orientation angle and directional consistency coefficient of each soil node are used as the characteristics of displacement direction distribution within the region.

[0114] Then, based on the adjacency relationship between nodes, the spatial dispersion of displacement differences is calculated, and dispersion constraints characterizing the deformation diffusion range are constructed. Specifically, all coupling edges between adjacent soil sub-region unit nodes in the structure-soil coupled topology network are traversed. For each coupling edge, the difference vector between the displacement vectors of the two soil nodes is calculated, and the magnitude of this difference vector is taken as the displacement difference value on the coupling edge. The mean and standard deviation of the displacement difference values ​​on all coupling edges are calculated, and the standard deviation of the displacement difference values ​​is defined as the spatial dispersion. The spatial dispersion reflects the degree of dispersion of deformation differences between adjacent nodes within the soil region. The larger the spatial dispersion, the stronger the spatial non-uniformity of deformation and the wider the diffusion range. Furthermore, the ratio of the number of coupling edges with displacement difference values ​​exceeding the mean to the total number of coupling edges is calculated and defined as the diffusion coverage rate. The diffusion coverage rate reflects the spatial distribution breadth of deformation differences. Based on the above analysis, a dispersion constraint condition is constructed. The dispersion constraint condition includes two requirements: first, the spatial dispersion is not lower than a preset dispersion threshold, which is set according to engineering experience; second, the diffusion coverage is not lower than a preset coverage threshold, for example, 0.4, that is, at least 40% of the coupling edges have displacement differences exceeding the mean, indicating that the deformation differences are not concentrated in individual locations but are widely distributed.

[0115] Next, a diffusion analysis of the deformation transmission characteristics of the soil nodes is performed along the coupling edges to construct diffusion constraints characterizing the multi-directional diffusion of deformation. Specifically, for each soil sub-region unit node, all adjacent soil nodes connected to it through coupling edges are counted, and the deformation transmission characteristic parameters calculated in step S42 are read from each coupling edge. For each coupling edge of the node, it is determined whether the deformation transmission ratio is greater than a preset transmission activity threshold. A deformation transmission ratio greater than the transmission activity threshold indicates that there is effective transmission and diffusion of deformation along the direction of the coupling edge. The transmission activity threshold is set according to engineering experience, for example, 0.3. The number of coupling edges that satisfy the transmission activity judgment among the coupling edges of the node is counted and defined as the number of active transmission directions of the node. The number of active transmission directions reflects the diversity of the directions in which deformation diffuses from the node to the surrounding area. The more active transmission directions there are, the more deformation diffuses from the node to multiple directions simultaneously. Further, the ratio of the number of soil nodes with at least a preset multi-directional diffusion threshold to the total number of soil nodes is calculated, defined as the multi-directional diffusion rate. The multi-directional threshold is determined based on the total number of adjacent coupled edges of the node, for example, half of the total number of adjacent coupled edges. Based on the above analysis, diffusion constraints are constructed, requiring the multi-directional diffusion rate to be no less than a preset diffusion rate threshold, for example, 0.3, meaning that at least 30% of the soil nodes exhibit the characteristic of simultaneous diffusion deformation in multiple directions.

[0116] Next, the displacement direction distribution characteristics, dispersion constraints, and diffusion constraints are combined to form a third topological correlation feature vector matching discrimination rule for matching with the topological correlation feature vector. Specifically, the third topological correlation feature vector matching discrimination rule consists of the following three conditions: The first condition is that the directional consistency coefficient in the displacement direction distribution characteristics is lower than a preset directional consistency threshold, for example, 0.5, meaning that more than half of the soil nodes have a large deviation between their displacement directions and the average direction, indicating that the soil deformation direction is dispersed; the second condition is that the dispersion constraint condition is met, that is, the spatial dispersion is not lower than the dispersion threshold and the diffusion coverage is not lower than the coverage threshold; the third condition is that the diffusion constraint condition is met, that is, the multi-directional diffusion rate is not lower than the diffusion rate threshold. When the above three conditions are met simultaneously, it is determined that there is fluid plastic diffusion deformation in the soil region. When performing deformation pattern recognition, the feature components corresponding to all soil sub-region unit nodes in the topological correlation feature vector are extracted, and the above three conditions are checked respectively. The matching degree between the topological correlation feature vector and the third topological correlation feature vector matching discrimination rule is calculated based on the degree of satisfaction of each condition.

[0117] Through the above processing, a soil fluid diffusion discrimination rule based on soil displacement direction distribution, deformation space dispersion and multi-directional diffusion characteristics was established, namely the third topological correlation feature vector matching discrimination rule, which enables the deformation pattern recognition process to automatically determine whether there is a soil fluid diffusion deformation mode.

[0118] In a preferred embodiment, the classification and recognition are performed by utilizing the matching relationship between topological correlation feature vectors and preset deformation pattern discrimination rules, and further include:

[0119] S441. The topological association feature vector and the preset deformation mode discrimination rule are compared by matching degree, and the discrimination result is output according to the matching degree.

[0120] In a preferred embodiment, the topological correlation feature vector needs to be matched and compared with each preset deformation mode discrimination rule. The matching comparison is achieved by calculating the matching degree. The matching degree is a value between 0 and 1, used to quantify the degree of conformity between the topological correlation feature vector and a certain deformation mode discrimination rule. The closer the matching degree is to 1, the higher the degree of conformity between the current deformation state and the deformation mode. The closer the matching degree is to 0, the greater the difference between the current deformation state and the deformation mode.

[0121] Specifically, each deformation mode discrimination rule consists of a combination of several conditions. For each condition, the individual matching degree is calculated based on the satisfaction of the corresponding feature component in the topological association feature vector. When a condition is fully satisfied, the individual matching degree is 1; when a condition is not satisfied at all, the individual matching degree is 0; when a condition is partially satisfied, the individual matching degree is calculated based on the deviation between the actual calculated value and the threshold of the condition. The smaller the deviation, the closer the individual matching degree is to 1; the larger the deviation, the closer the individual matching degree is to 0. The specific calculation method for the individual matching degree is to take the absolute value of the difference between the actual calculated value and the threshold of the condition, and then divide it by the threshold of the condition to obtain the deviation ratio. The individual matching degree is equal to 1 minus the deviation ratio. When the deviation ratio is greater than 1, the individual matching degree is 0.

[0122] After obtaining the individual matching degree of each condition, the individual matching degrees of each condition under the same discrimination rule are weighted and summed to obtain the matching degree between the topological correlation feature vector and the discrimination rule. The weight of each condition is determined according to the importance of the condition in the corresponding deformation mode discrimination, and the sum of the weights of each condition is equal to 1. For example, for the first topological correlation feature vector matching discrimination rule, the weights of the displacement gradient increasing condition, the monotonicity constraint condition, and the curvature constraint condition can be set to 0.3, 0.3, and 0.4, respectively.

[0123] After calculating the matching degree between the topological correlation feature vector and each preset deformation mode discrimination rule, the matching degree value between the current deformation state and each deformation mode is obtained. The discrimination result is output based on the matching degree. The output method is to select the deformation mode with the highest matching degree as the deformation mode type of the current foundation pit, and use this maximum matching degree value as the matching degree in the discrimination result. When the maximum matching degree is lower than the preset minimum matching degree threshold, for example, 0.4, the current deformation state is determined not to belong to any preset deformation mode, and the discrimination result is output as an unidentified state. When the matching degree of two or more deformation modes exceeds the minimum matching degree threshold, the discrimination result includes each relevant deformation mode type and its corresponding matching degree, indicating that the current deformation state may simultaneously possess the characteristics of multiple deformation modes.

[0124] Through the above processing, the matching degree is used as a quantitative comparison basis to realize the systematic matching comparison between the topological association feature vector and each preset deformation pattern discrimination rule, so that the discrimination result of deformation pattern recognition has a clear numerical basis and avoids subjective judgment that relies on human experience.

[0125] Furthermore, such as Figure 2 As shown, based on the structural topological relationships and mechanical constraints of the deformation analysis model, spatial reconstruction processing is performed on the mapped multi-source monitoring data to convert the discrete monitoring data into a continuous three-dimensional deformation field that satisfies the structural boundary constraints, including:

[0126] S31. The multi-source monitoring data mapped to each node in the structure-soil coupled topology network is used as the initial deformation state quantity, and the nodes without monitoring data are defined as nodes to be estimated.

[0127] S32. Based on the connection relationship and edge weight parameters of the coupling edge, establish a weighted correlation relationship for deformation transmission between nodes, and transmit the deformation state of the known node to the adjacent node along the coupling edge to form an initial deformation estimation distribution.

[0128] S33. During the deformation transfer process, the support structure constraint and soil mechanical response constraint are introduced to constrain the deformation transfer values ​​and directions between nodes.

[0129] S34. Perform iterative balancing of the initial deformation estimation distribution to establish a continuous three-dimensional deformation field.

[0130] In a preferred embodiment, when establishing a continuous three-dimensional deformation field, firstly, the multi-source monitoring data mapped to each node in the structure-soil coupled topology network is used as the initial deformation state quantity, and nodes without monitoring data are defined as nodes to be estimated. Specifically, after the mapping process in step S2, some nodes in the structure-soil coupled topology network have been attached with measured monitoring data. These nodes are defined as known nodes, and their initial deformation state quantities are directly taken from the measured monitoring data values ​​mapped to that node, including the horizontal displacement value and settlement value of that node. For retaining structure component unit nodes, the initial deformation state quantity of the known node is taken from the deep horizontal displacement data of the retaining structure or the top horizontal displacement data of the retaining structure mapped to that node. For soil sub-region unit nodes, the initial deformation state quantity of the known node is taken from the deep horizontal displacement data of the soil or the surface settlement monitoring data mapped to that node. For column pile component unit nodes, the initial deformation state quantity of the known node is taken from the column pile settlement monitoring data mapped to that node. The remaining nodes in the network without attached measured monitoring data are defined as nodes to be estimated. The initial deformation state variables of the nodes to be estimated are null values, which need to be estimated through subsequent deformation propagation and iterative balancing processes. All nodes in the coupled topology network are traversed, and each node is marked as either a known node or a node to be estimated, establishing the distribution relationship between known nodes and nodes to be estimated.

[0131] Then, based on the connection relationship and edge weight parameters of the coupling edges, a weighted correlation relationship for deformation transmission between nodes is established. The deformation state of the known node is transmitted to the adjacent nodes along the coupling edges to form an initial deformation estimation distribution. Specifically, for each node to be estimated, all adjacent nodes directly connected to it through coupling edges are searched in the coupled topology network. Nodes with deformation state quantities are selected, including known nodes and nodes to be estimated that have completed deformation estimation in the current round. The deformation state quantities of each adjacent known or estimated node and the edge weight parameters of the connecting coupling edges are read, and a weighted correlation relationship is established using the edge weight parameters as weight coefficients. The initial deformation estimate of the node to be estimated is equal to the sum of the products of the deformation state quantities of each adjacent node and the corresponding edge weight parameters of the coupling edges, divided by the sum of the edge weight parameters of each coupling edge. That is, the deformation state quantities of adjacent nodes are weighted by the edge weight parameters. Adjacent nodes connected by coupling edges with larger edge weight parameters contribute more to the deformation estimation of the node to be estimated, while adjacent nodes connected by coupling edges with smaller edge weight parameters contribute less to the deformation estimation of the node to be estimated. The transmission process starts from the known node and extends outward layer by layer along the connection path of the coupling edge. First, the node to be estimated that is directly adjacent to the known node is estimated. Then, the node that has been estimated is used as the new transmission source to continue to transmit to the next layer of node to be estimated. This process continues until all nodes to be estimated in the coupled topology network have obtained the initial deformation estimate value, forming the initial deformation estimate distribution.

[0132] Next, during the deformation transfer process, constraints on the support structure and soil mechanical response are introduced to constrain the deformation transfer values ​​and directions between nodes. Specifically, during the deformation transfer process in step S32, after the initial deformation estimation of each node to be estimated is completed, it is necessary to check whether the estimated value meets the constraint conditions attached to its coupling edge. For coupling edges between adjacent component unit nodes on the same support structure, it is checked whether the difference between the estimated value of the node to be estimated and the deformation value of the adjacent known or estimated node exceeds the upper limit threshold of the deformation difference on the coupling edge. If it exceeds, the estimated value of the node to be estimated is adjusted in the direction of the deformation value of the adjacent node so that the difference is reduced to within the threshold range. For coupling edges between retaining structure component unit nodes and adjacent soil sub-region unit nodes, it is checked whether the difference between their deformation estimates in the direction of the contact surface normal exceeds the allowable deviation range on the coupling edge. If it exceeds, the deformation estimates of the two nodes are redistributed according to the stiffness ratio of their corresponding units. The adjustment range of the node estimate on the side with greater stiffness is smaller, and the adjustment range of the node estimate on the side with less stiffness is larger. For the coupling edge between adjacent soil sub-region unit nodes, the deformation value of the node before transfer is multiplied by the attenuation coefficient on the coupling edge to obtain the deformation value after transfer. If there is a deviation between the estimated value obtained by weighted averaging in step S32 and the attenuated deformation value, the estimated value is adjusted in the direction of the attenuated deformation value. Through the above constraint processing, the estimated values ​​of each node to be estimated during the deformation transfer process are subject to structural mechanical constraints while obtaining a weighted average estimate, ensuring the physical rationality of the estimated values.

[0133] Subsequently, an iterative balancing process is performed on the initial deformation estimation distribution to establish a continuous three-dimensional deformation field. Specifically, after steps S32 and S33, all nodes in the coupled topology network have obtained deformation state variables. The deformation state variables of known nodes are measured values, and the deformation state variables of nodes to be estimated are estimated values ​​after constraint adjustment, forming the initial deformation estimation distribution. However, since the deformation transmission process is carried out layer by layer, nodes estimated earlier may encounter situations where they do not meet the constraint conditions with their adjacent nodes after the subsequent nodes are estimated. Therefore, iterative balancing processing is required for the overall deformation distribution. The iterative balancing process involves traversing all coupling edges in the coupled topology network and checking whether the deformation state variables of the nodes at both ends of each coupling edge meet the constraint conditions on that coupling edge, including deformation continuity constraints, deformation compatibility constraints, and soil mechanical response constraints. For coupling edges that do not meet the constraint conditions, the deformation state variables of the nodes at both ends are adjusted according to the same adjustment method as in step S33. The deformation state variables of known nodes remain unchanged during the iterative adjustment process and participate in the iterative balancing as fixed constraints to ensure that the constraint effect of the measured monitoring data on the deformation distribution is not weakened. After completing one round of traversal and adjustment, the adjustment amount of the deformation state variables of each node to be estimated in this round is calculated. The maximum value of the adjustment amount of all nodes to be estimated is compared with a preset convergence threshold. When the maximum adjustment amount is less than the convergence threshold, the iteration is considered to have converged, and the iteration process ends. When the maximum adjustment amount is still greater than the convergence threshold, the adjusted deformation distribution is used as the input for a new round of iteration, and the traversal and adjustment process is repeated until the convergence condition is met. The convergence threshold is set according to the accuracy level of the monitoring data, for example, 0.1 mm. After the iteration converges, the deformation state variables of all nodes in the coupled topology network satisfy the constraint conditions on each coupling edge, forming a continuous three-dimensional deformation field.

[0134] Through the above steps, discrete monitoring data are reconstructed into a continuous three-dimensional deformation field covering all nodes under the combined effect of the topological connection relationship and mechanical constraints of the structure-soil coupled topological network, thus extending deformation analysis from the discrete measurement point level to the overall continuous distribution level.

[0135] Furthermore, before mapping the multi-source monitoring dataset to the deformation analysis model, time synchronization and outlier removal processing of the multi-source monitoring dataset are performed.

[0136] In a preferred embodiment, since the various monitoring data in the multi-source monitoring dataset originate from different types of monitoring instruments, the acquisition frequency and acquisition time of each instrument may differ. For example, there may be a minute-level time deviation between the acquisition time of the inclinometer and the total station, and the acquisition frequency of the level instrument may also differ from that of the inclinometer. If time synchronization processing is not performed and data acquired at different times are directly mapped to the same deformation analysis model, the monitoring data attached to each element in the model will not be comparable in time, affecting the accuracy of subsequent spatial reconstruction and deformation analysis.

[0137] During time synchronization, a unified time reference is first established. The earliest and latest times among the data collected by various monitoring instruments within the same acquisition cycle are used to determine the time window for that cycle. Within this time window, a unified reference time is selected, such as the midpoint of the window. Then, for data whose acquisition time differs from the reference time, linear interpolation is performed based on the monitoring values ​​of the monitoring point at two adjacent acquisition times to calculate the interpolated data at the reference time. The linear interpolation is calculated by using the monitoring values ​​at the previous and subsequent acquisition times as endpoints, and then calculating the interpolated data value corresponding to the reference time according to the time position ratio of the reference time between the two acquisition times. For monitoring data whose acquisition time is completely consistent with the reference time, the original data value is directly retained. After time synchronization, all monitoring data in the multi-source monitoring dataset are unified to the same reference time, ensuring consistency across all data types in the time dimension.

[0138] After time synchronization, outlier removal is performed on the multi-source monitoring dataset. Outliers refer to monitoring data that significantly deviate from normal deformation patterns due to instrument malfunction, human error, environmental interference, or other reasons. The specific outlier removal process involves reading monitoring data from each monitoring point over multiple consecutive acquisition periods to form a time series for that point. The mean and standard deviation of this time series are calculated. When the deviation between the monitoring value and the mean in a certain acquisition period exceeds a preset multiple of the standard deviation, the data is marked as a suspected outlier. The preset multiple is set according to the accuracy level of the monitoring data, for example, 3. For data marked as suspected outliers, trend verification is further performed by combining data from adjacent acquisition periods for that monitoring point. The trend of change between the monitoring values ​​of the previous and subsequent acquisition periods is calculated. If the suspected outlier changes in the opposite direction to the trend or the magnitude of the change far exceeds the expected trend, the data is confirmed as an outlier and removed. The data gaps after removal are filled using linear interpolation of data from adjacent acquisition periods for that monitoring point.

[0139] Through the above processing, the various types of data in the multi-source monitoring dataset are synchronized and aligned in the time dimension, and abnormal data introduced by non-deformation factors are eliminated, ensuring the consistency of monitoring data mapped to the deformation analysis model in time and the reliability of numerical values.

[0140] Furthermore, the multi-source monitoring dataset undergoes self-verification of data sampling, the self-verification result is output, an early warning is issued based on the self-verification result, and resampling processing is performed.

[0141] In a preferred embodiment, self-verification refers to automatically verifying and checking the sampling quality of the monitoring data after the multi-source monitoring dataset has completed time synchronization and outlier removal processing, and before it is mapped to the deformation analysis model, in order to determine whether the monitoring data of the current collection period meets the data quality requirements required for subsequent deformation analysis.

[0142] The specific method of self-verification is to verify the multi-source monitoring dataset of the current collection period from three aspects: data integrity, data stability, and data consistency.

[0143] Data integrity verification refers to checking whether all monitoring points in the current data collection period have successfully collected valid data. Specifically, the data integrity rate is defined as the ratio of the number of monitoring points that actually collected valid data in the current data collection period to the total number of monitoring points that should have collected data. When the data integrity rate is lower than a preset integrity rate threshold, for example, 0.9, it is determined that the data integrity of the current data collection period does not meet the requirements, and a self-verification result indicating insufficient data integrity is output.

[0144] Data stability verification refers to checking whether there are abnormal fluctuations in the data of each monitoring point in the current acquisition period. Specifically, for each monitoring point, the monitoring value of the current acquisition period is compared with the monitoring values ​​of that monitoring point in the most recent several acquisition periods. The deviation between the current value and the mean of the most recent periods is calculated, and then divided by the standard deviation of the most recent periods to obtain the standardized deviation value. When the standardized deviation value of a monitoring point exceeds a preset stability threshold, for example, 2.5, it is determined that the data of that monitoring point has abnormal fluctuations. The ratio of the number of monitoring points with abnormal fluctuations to the total number of monitoring points is defined as the abnormal volatility rate. When the abnormal volatility rate exceeds a preset volatility threshold, for example, 0.15, it is determined that the data stability of the current acquisition period does not meet the requirements, and the self-verification result of data stability abnormality is output.

[0145] Data consistency verification refers to checking for logical inconsistencies between different types of monitoring data in the current acquisition cycle. Specifically, data obtained from different monitoring methods at the same spatial location are cross-compared. For example, the inclinometer data on the deep horizontal displacement at a certain depth of the retaining structure and the total station data on the top displacement at that location should satisfy a consistency relationship in deformation trends; that is, when the top horizontal displacement increases, the corresponding deep horizontal displacement should also show an increasing trend. When the trends of change between different types of monitoring data show opposite directions, it is determined that there is a consistency inconsistency at that location. The number of locations with consistency inconsistencies is counted. When this number exceeds a preset threshold for inconsistent locations, for example, taking 10% of the total cross-comparison locations, it is determined that the data consistency of the current acquisition cycle does not meet the requirements, and a self-verification result indicating data consistency anomalies is output.

[0146] The self-verification results include verification conclusions and specific verification index values ​​for data integrity, data stability, and data consistency. Based on the self-verification results, an early warning is issued. If any of the three verification aspects fails to meet the requirements, a data quality early warning message is issued. The warning message includes the specific verification item that failed to meet the requirements, the number and location information of the abnormal monitoring point, and prompts monitoring management personnel to check the relevant monitoring instruments or monitoring points.

[0147] Simultaneously with issuing early warning information, resampling is performed. Resampling refers to targeted supplementary collection or data repair of monitoring data determined to be unsatisfactory during self-verification. For cases of insufficient data integrity, a re-collection instruction is issued to monitoring points with missing data, and the supplementary data is incorporated into the multi-source monitoring dataset after collection is complete. For cases of abnormal data stability, short-interval continuous collection is performed on monitoring points exhibiting abnormal fluctuations. Multiple collections are conducted on the monitoring point within a short period, and the average of the multiple collection results is used as the corrected data for that monitoring point, replacing the original abnormal data. For cases of abnormal data consistency, re-collection is performed simultaneously on all monitoring instruments involved at locations with inconsistencies. The re-collected data undergoes consistency verification again; if inconsistencies still exist, the data from the monitoring instrument with the higher accuracy level is used as the standard, and the data from the lower accuracy level is corrected.

[0148] After the resampling process is completed, the repaired multi-source monitoring dataset will be re-verified until all self-verification results meet the requirements before the subsequent mapping process is performed.

[0149] Through the above processing, an automated data quality verification and repair step is added before the monitoring data is mapped to the deformation analysis model. This ensures that the monitoring data entering the deformation analysis process meets the analysis requirements in terms of integrity, stability, and consistency, and reduces the risk of distortion of deformation analysis results due to data quality issues.

[0150] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0151] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0152] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0153] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0154] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0155] Although preferred embodiments of the invention have been described, those skilled in the art, once they have learned the basic inventive concept, can make other changes and modifications to these embodiments.

[0156] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A BIM-based method for monitoring foundation pit deformation, characterized in that, The method includes: Obtain the BIM model of the target foundation pit, and introduce soil layering parameters, support structure mechanical parameters and spatial topological relationships between components into the BIM model. Perform semantic enhancement processing based on the introduction results to construct a deformation analysis model including the structure-soil coupling relationship. Perform multi-source monitoring interaction of the target foundation pit, establish a multi-source monitoring dataset, the multi-source monitoring dataset includes displacement monitoring data and settlement monitoring data, and map the multi-source monitoring dataset to the component unit or soil unit corresponding to the deformation analysis model according to the spatial location; Based on the structural topological relationships and mechanical constraints of the deformation analysis model, spatial reconstruction processing is performed on the mapped multi-source monitoring data to convert the discrete monitoring data into a continuous three-dimensional deformation field that satisfies the structural boundary constraints. Deformation feature parameters characterizing the response characteristics of the foundation pit structure are extracted based on the continuous three-dimensional deformation field, deformation pattern recognition is performed using the deformation analysis model, and the discrimination results are output. Using the discrimination result as the first input and the continuous three-dimensional deformation field as the second input, the temporal evolution trend prediction of the target foundation pit deformation is performed, and the prediction monitoring results are established. Based on the introduced results, semantic enhancement processing is performed to construct a deformation analysis model that includes the structure-soil coupling relationship, including: The support structure component units and the soil sub-region units divided according to soil layers in the BIM model are defined as network nodes, and each node is assigned corresponding geometric properties and mechanical parameters. Based on the contact relationship and spatial adjacency relationship between the support structure component unit and the soil sub-region unit, a coupling edge representing the mechanical action relationship is established between the corresponding nodes, and the coupling edge is assigned a weight parameter reflecting the action intensity. Based on the network nodes and coupling edges, support structure constraints and soil mechanical response constraints are introduced to constrain and limit the deformation transmission path between nodes, thus constructing a structure-soil coupled topology network. Based on the structure-soil coupled topology network, each node is given a unified semantic label, which enables the monitoring data to propagate along the coupling edge between nodes and form a constrained deformation response relationship. Based on the structure-soil coupled topology network, a deformation analysis model is constructed.

2. The BIM-based method for monitoring foundation pit deformation as described in claim 1, characterized in that, Deformation feature parameters characterizing the response of the foundation pit structure are extracted based on a continuous three-dimensional deformation field. Deformation pattern recognition is performed using a deformation analysis model, and the discrimination results are output, including: Based on the continuous three-dimensional deformation field and the spatial distribution of nodes in the structure-soil coupled topology network, the displacement vector, settlement, displacement gradient and curvature distribution of each node are extracted, and node-level deformation characteristic parameters are established. Based on the connection relationship of the coupling edge, the deformation difference between adjacent nodes is calculated, and the deformation transfer characteristic parameters along the direction of the coupling edge are obtained to characterize the deformation compatibility and transfer strength between the structure and the soil. The node-level deformation feature parameters and the deformation transfer feature parameters are mapped to the structure-soil coupled topology network, and constrained feature propagation and aggregation processing along the coupling edges are performed to form a topological correlation feature vector that reflects the overall structural response. By utilizing the matching relationship between topological correlation feature vectors and preset deformation pattern discrimination rules, classification and recognition are performed, and the discrimination results are output.

3. The BIM-based method for monitoring foundation pit deformation as described in claim 2, characterized in that, The preset deformation mode discrimination rules include a first topological correlation feature vector matching discrimination rule for characterizing the bending deformation of a cantilever. The construction of the first topological correlation feature vector matching discrimination rule includes: Based on the node sequence along the depth direction of the enclosure structure, the horizontal displacement components of the nodes are extracted and a displacement gradient distribution sequence that varies with depth is established. Based on the displacement gradient distribution sequence, the sign consistency of the displacement change rate between nodes is calculated, and a monotonicity constraint condition characterizing the monotonic change of displacement along the depth direction is constructed. Based on the node sequence, the displacement curvature distribution parameters are calculated, and curvature constraint conditions characterizing the degree of bending deformation of the enclosure structure are constructed. The displacement gradient distribution sequence, monotonicity constraint, and curvature constraint are combined to form a first topological correlation feature vector matching discrimination rule for matching with topological correlation feature vectors.

4. The BIM-based method for monitoring foundation pit deformation as described in claim 3, characterized in that, The preset deformation mode discrimination rules include a second topological correlation feature vector matching discrimination rule for characterizing the constraint failure of the support structure. The construction of the second topological correlation feature vector matching discrimination rule includes: Extract the displacement difference between the nodes of the supporting components and the nodes of the adjacent enclosure structure, and establish the deformation difference distribution characteristics across the support locations; Based on the coupling edge, the deformation transmission characteristic parameters between the nodes on both sides of the support node are calculated, and deformation continuity constraint conditions characterizing the deformation continuity change are constructed. A mutation analysis was performed on the deformation difference distribution characteristics to construct mutation constraint conditions characterizing local deformation discontinuities; The deformation difference distribution characteristics, deformation continuity constraints, and abrupt change constraints are combined to form a second topological association feature vector matching discrimination rule for matching with topological association feature vectors.

5. The BIM-based method for monitoring foundation pit deformation as described in claim 3, characterized in that, The preset deformation mode discrimination rules include a third topological correlation feature vector matching discrimination rule for characterizing soil fluid diffusion. The construction of the third topological correlation feature vector matching discrimination rule includes: Extract the displacement vectors of nodes in the soil sub-region and perform directional consistency analysis to establish the displacement direction distribution characteristics within the region; The spatial discreteness of displacement difference is calculated based on the adjacency relationship between nodes, and discreteness constraints characterizing the deformation diffusion range are constructed. Diffusion analysis of deformation transfer characteristics of soil nodes along coupling edges is performed to construct diffusion constraints that characterize the multi-directional diffusion of deformation. The displacement direction distribution characteristics, discreteness constraints, and diffusion constraints are combined to form a third topological association feature vector matching discrimination rule for matching with topological association feature vectors.

6. The BIM-based method for monitoring foundation pit deformation as described in claim 3, characterized in that, The classification and recognition are performed by matching the topological correlation feature vectors with the preset deformation pattern discrimination rules, and also include: The topological association feature vector is compared with the preset deformation mode discrimination rule through matching degree, and the discrimination result is output according to the matching degree.

7. The BIM-based method for monitoring foundation pit deformation as described in claim 1, characterized in that, Based on the structural topological relationships and mechanical constraints of the deformation analysis model, spatial reconstruction processing is performed on the mapped multi-source monitoring data to convert the discrete monitoring data into a continuous three-dimensional deformation field that satisfies the structural boundary constraints, including: The multi-source monitoring data mapped to each node in the structure-soil coupled topology network are used as the initial deformation state variables, and the nodes without monitoring data are defined as nodes to be estimated. Based on the connection relationship and edge weight parameters of the coupling edge, a weighted correlation relationship for deformation transmission between nodes is established, and the deformation state of the known node is transmitted to the adjacent node along the coupling edge to form an initial deformation estimation distribution. During the deformation transfer process, support structure constraints and soil mechanical response constraints are introduced to constrain the deformation transfer values ​​and directions between nodes. Perform iterative balancing processing on the initial deformation estimation distribution to establish a continuous three-dimensional deformation field.

8. The BIM-based method for monitoring foundation pit deformation as described in claim 1, characterized in that, Before being mapped to the deformation analysis model, the multi-source monitoring dataset undergoes time synchronization and outlier removal processing.

9. The BIM-based method for monitoring foundation pit deformation as described in claim 8, characterized in that, The multi-source monitoring dataset is subjected to self-verification of data sampling, and the self-verification result is output. Based on the self-verification result, an early warning is issued, and resampling is performed.

Images