Building bolt assembly quality control method and system based on BIM digital twinning

Abstract

The invention provides a building bolt assembly quality control method and system based on BIM digital twinning, and belongs to the field of building construction. According to the technical scheme, the method comprises the following steps: S1, establishing a corresponding relation between a physical bolt and a BIM model bolt object; s2, acquiring axial force data and environment temperature data of the physical bolt; s3, constructing an initial digital twinborn body of the physical bolt, and generating a calibrated digital twinborn body; s4, calculating a pre-tightening force uniformity index, stress and deformation by using the calibrated digital twin; and S5, generating a disposal instruction and verifying a disposal effect. The method has the beneficial effects that a data foundation is laid by bolt-level mapping registration and temperature compensation perception, the authenticity of the model is ensured through online calibration of the digital twinborn body, double-domain evaluation is realized through unified calculation of the pre-tightening force uniformity, key stress and deformation, torque and a sequence instruction are automatically generated, and finally a quality control closed loop is opened through retest verification, so that the accuracy of the model is improved. Therefore, the accuracy, the reliability and the performability are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of building construction, and particularly relates to a building bolt assembly quality control method and system based on BIM digital twinning. BACKGROUND

[0002] With the popularization of fabricated buildings and steel structure engineering, high-strength bolt connection has become the mainstream assembly method of key load-bearing nodes. In the technical aspect, building information modeling has been widely used in three-dimensional design and construction collaborative management of components; while in the field of quality control, torque method, angle method or ultrasonic length measurement are commonly used for indirect or direct axial force monitoring. At the same time, finite element analysis, as a mature technology, is mainly used for stress performance checking in the design stage.

[0003] In recent years, Internet of Things sensing technology and digital twinning concept have been introduced into construction process monitoring, and the industry has begun to try to compare field sensing data with virtual simulation results, aiming to improve the visualization and traceability of assembly quality. However, this preliminary combination of "sensing-simulation" is still in the exploratory stage in terms of engineering-level integration, and it is difficult to meet the industry's high-level needs for "fine-grained at the bolt level, online throughout the process, and closed-loop quality control".

[0004] Specifically, the existing technical solutions mainly have the following four obvious deficiencies: Firstly, quality control relies on sampling inspection, and its single criterion based on torque or angle cannot truly reflect the spatial dispersion and uniformity of pretightening force in the group bolt system, and cannot provide a reliable quantitative index for the overall stress performance of the node.

[0005] Secondly, BIM models focus on geometry and static attribute management, while there is a lack of effective online calibration mechanism between field sensing data and background mechanical models. This leads to a serious disconnection between the theoretical assumptions in the design stage (such as contact conditions, friction coefficients, and connection stiffness) and the actual state on the construction site, making the offline simulation results unable to truly reflect the actual stress and deformation of the structure. In addition, the coordinate registration from the field to the model and the binding of physical objects lack standardized processes, making it difficult for massive sensing data to form a one-to-one mapping relationship with specific bolt objects in the BIM model. At the same time, environmental factors such as temperature are not systematically compensated, and data noise and temperature drift directly affect the accuracy of axial force evaluation.

[0006] Finally, and most importantly, existing systems mostly stay at the level of "data visualization and abnormal alarm", lacking the ability to automatically convert diagnostic conclusions into specific and executable field disposal instructions (such as target axial force / torque, optimized tightening sequence), and even less able to form an "decision-execution-verification" engineering closed loop through retesting mechanisms.

[0007] These limitations collectively lead to a core contradiction: most existing digital twin applications are "static displays" or "post-event records," rather than a "dynamic intelligent system" capable of guiding construction online and self-verifying its intervention effects. Summary of the Invention

[0008] The purpose of this invention is to provide a BIM-based digital twin-based method and system for quality control of building bolt assembly. This system establishes a data foundation through bolt-level mapping registration and temperature compensation sensing, ensures model authenticity through online calibration of digital twins, achieves dual-domain evaluation by uniformly calculating preload uniformity, key stresses and deformations, automatically generates torque and sequence commands, and finally verifies the quality control loop through retesting. This results in significant improvements in accuracy, reliability and executability.

[0009] This invention is achieved through the following measures: A method for quality control of building bolt assembly based on BIM digital twins, characterized by the following steps: Establish a one-to-one mapping relationship between physical bolts and BIM model bolt objects, and register the field space coordinate system of the physical bolts with the model space coordinate system of the BIM model bolt objects; Collect axial force data of the physical bolts and simultaneously collect ambient temperature data; Based on the geometric and attribute information of the bolt object in the BIM model, an initial digital twin of the physical bolt is constructed. Then, using the axial force data and the ambient temperature data, the mechanical parameters of the initial digital twin are calibrated through a parameter identification algorithm to generate a calibrated digital twin. Using the calibrated digital twin, the preload uniformity index of the physical bolt is calculated, and the stress and deformation in the component area connected to the physical bolt are calculated. Based on the preload uniformity index and the calculation results of stress and deformation, a treatment command is generated for the physical bolt. After executing the treatment command, the axial force data and ambient temperature data of the physical bolt are collected again to verify the treatment effect.

[0010] The invention also has the following specific features: Establishing a one-to-one mapping relationship between physical bolts and BIM model bolt objects includes: obtaining unique identification information attached to the physical bolts, writing the unique identification information into the attribute information of the BIM model bolt objects, and verifying the consistency between the specifications, connection type, and hole number of the physical bolts corresponding to the unique identification information and the preset design parameters in the BIM model bolt objects. When the verification passes, the one-to-one mapping relationship is confirmed and recorded.

[0011] The registration of the physical bolt's on-site spatial coordinate system with the BIM model bolt object's model spatial coordinate system includes: The coordinate data of multiple control points on site in the site spatial coordinate system are obtained by measuring equipment, and the coordinate data of the corresponding control points in the model spatial coordinate system are extracted from the BIM model. Based on two sets of coordinate data, a seven-parameter transformation model is used to calculate the transformation parameters between coordinate systems, and the adjustment optimization is performed using the least squares method. Registration accuracy is verified using checkpoints independent of control points. When the coordinate transformation deviation of the checkpoint does not exceed the preset deviation threshold, the registration is confirmed to be qualified. The qualified transformation parameters are then applied to all subsequent coordinate transformation operations to ensure the consistency of spatial data.

[0012] The process of collecting axial force data of the physical bolt and simultaneously collecting ambient temperature data includes: collecting axial force measurement signals through a preload sensing shim set on the physical bolt, and collecting ambient temperature signals through a temperature sensor arranged around the physical bolt. The acquired axial force measurement signal and ambient temperature signal are preprocessed. The preprocessing includes data filtering and outlier removal of the axial force measurement signal and ambient temperature signal, respectively, to obtain processed axial force data and ambient temperature data.

[0013] Constructing an initial digital twin of the physical bolt based on the geometric and attribute information of the bolt object in the BIM model includes: establishing a parametric geometric model of the physical bolt based on the geometric information of the bolt object in the BIM model, wherein the parametric geometric model includes the main structural feature dimensions of the bolt; Simultaneously, the material's mechanical properties, including elastic modulus and Poisson's ratio, are assigned based on the aforementioned attribute information; Boundary constraints are set according to the design connection relationship of the physical bolts in the structure, including the contact relationship and constraint type of the connection parts; An initial load based on the design preload value is applied to create an initial digital twin containing complete information on geometry, materials, boundaries, and loads.

[0014] Using the axial force data and the ambient temperature data, the mechanical parameters of the initial digital twin are calibrated through a parameter identification algorithm to generate a calibrated digital twin, including: Using the processed axial force data as an observation benchmark, a comparison relationship is established between the output response of the initial digital twin and the measured axial force data. The key mechanical parameters of the initial digital twin are calculated by inversion using a parameter identification algorithm. The key mechanical parameters include connection stiffness and friction coefficient. The ambient temperature data is introduced during the parameter inversion process for compensation and correction to eliminate the impact of temperature changes on parameter identification. The key mechanical parameters are adjusted through an iterative optimization process, so that the output response of the initial digital twin gradually converges with the measured axial force data. When the preset convergence condition is met, the parameter calibration is completed, and the calibrated digital twin is generated.

[0015] Using the calibrated digital twin, the preload uniformity index of the physical bolt is calculated, and the stress and deformation in the component area connected to the physical bolt are calculated, including: performing mechanical simulation analysis based on the calibrated digital twin, extracting the simulated axial force values ​​of each physical bolt at the same connection node, and calculating the preload uniformity index based on the simulated axial force values. Mechanical simulation analysis is performed based on the calibrated digital twin to extract the stress distribution in the component area connected to the physical bolt and obtain the stress values ​​at key locations. Mechanical simulation analysis is performed based on the calibrated digital twin to extract the displacement distribution of the component area connected to the physical bolts and obtain deformation data at key locations.

[0016] Based on the calculation results of the preload uniformity index and the stress and deformation, the treatment instructions for the physical bolts are generated by: comprehensively analyzing and determining the physical bolt objects that need to be treated and the corresponding treatment measures based on the comparison results of the preload uniformity index and the preset threshold, the comparison results of the stress values ​​at the key locations and the material allowable values, and the comparison results of the deformation data at the key locations and the design allowable values, and generating treatment instructions that include specific treatment objects and treatment measure types.

[0017] After executing the treatment command, the axial force data and ambient temperature data of the physical bolts are collected again to verify the treatment effect, including: After executing the disposal instruction, steps S2 to S4 are re-executed to obtain a new preload uniformity index, new stress values ​​at key locations, and new deformation data at key locations. The new calculation results are compared and analyzed with the corresponding thresholds. The disposal effect is verified based on the comparison and analysis results. When all evaluation parameters meet the requirements, the quality control process is completed. Otherwise, a new disposal instruction is generated based on the new comparison and analysis results to continue optimizing the process.

[0018] A system employing the aforementioned BIM-based digital twin-based building bolt assembly quality control method is characterized by comprising: Mapping and Registration Module: Establishes a one-to-one mapping relationship between physical bolts and BIM model bolt objects, and registers the field space coordinate system of the physical bolts with the model space coordinate system of the BIM model bolt objects; Data acquisition and preprocessing module: Acquires axial force data of the physical bolts and simultaneously acquires ambient temperature data; Digital twin construction and parameter calibration module: Based on the geometric and attribute information of the bolt object in the BIM model, an initial digital twin of the physical bolt is constructed, and the mechanical parameters of the initial digital twin are calibrated using the axial force data and the ambient temperature data through a parameter identification algorithm to generate a calibrated digital twin; Evaluation module: Using the calibrated digital twin, calculate the preload uniformity index of the physical bolt, and calculate the stress and deformation in the component area connected to the physical bolt; Treatment and verification module: Based on the preload uniformity index and the calculation results of stress and deformation, a treatment instruction is generated for the physical bolt. After executing the treatment instruction, the axial force data and ambient temperature data of the physical bolt are collected again to verify the treatment effect.

[0019] The beneficial effects of this invention are as follows: This invention lays the data foundation with bolt-level mapping registration and temperature compensation sensing, ensures the model's authenticity through online calibration of the digital twin, and achieves dual-domain evaluation by uniformly calculating preload uniformity, key stresses and deformations, and automatically generates torque and sequence commands. Finally, the quality control closed loop is established through retesting and verification, thereby achieving significant improvements in accuracy, reliability and executability. Attached Figure Description

[0020] Figure 1 The present invention provides an overall flowchart of a method and system for quality control of building bolt assembly based on BIM digital twin. Detailed Implementation

[0021] To clearly illustrate the technical features of this solution, the following detailed implementation method will be used to explain the solution.

[0022] Example 1 See Figure 1 A method for quality control of building bolt assembly based on BIM digital twins includes the following steps: Step S1: Establish a one-to-one mapping relationship between physical bolts and BIM model bolt objects, and register the on-site spatial coordinate system of the physical bolts with the model spatial coordinate system of the BIM model bolt objects, including: Obtain the unique identification information attached to the physical bolt, write the unique identification information into the attribute information of the bolt object in the BIM model, and verify the consistency between the specification parameters, connection type and hole number of the physical bolt corresponding to the unique identification information and the preset design parameters in the bolt object in the BIM model. When the verification is successful, confirm the establishment of a one-to-one mapping relationship and record it. The coordinate data of multiple control points on site in the site spatial coordinate system are obtained by measuring equipment, and the coordinate data of the corresponding control points in the model spatial coordinate system are extracted from the BIM model. Based on two sets of coordinate data, a seven-parameter transformation model is used to calculate the transformation parameters between coordinate systems, and the adjustment optimization is performed using the least squares method. Registration accuracy is verified using checkpoints independent of control points. When the coordinate transformation deviation of the checkpoint does not exceed the preset deviation threshold, the registration is confirmed to be qualified. The qualified transformation parameters are then applied to all subsequent coordinate transformation operations to ensure the consistency of spatial data.

[0023] Specifically, step S1 includes: I. One-to-one mapping and consistency verification Generate a unique identifier for each physical bolt and write it into the attribute information of the corresponding bolt object in the BIM model; After obtaining unique identification information by scanning a code or near-field reading, unique indexes are established in the quality control database for both the "unique identification information" and the "BIM model bolt object identifier". Uniqueness verification is performed first. If there is a conflict, the mapping is rejected and the reason for the conflict is recorded. After the uniqueness is passed, the "specification parameters, connection type, and hole position number" of the physical bolt are compared with the preset design parameters in the BIM model item by item. If any one is inconsistent, the process is stopped. Only when all are consistent is the mapping confirmed and a mapping record is generated (including unique identification information, BIM model bolt object identifier, specification parameters, connection type, hole position number, operator identifier, equipment identifier, timestamp, and mapping status). The mapping state adopts a finite state machine of "valid / pending review / invalid": passing the consistency check means it is "valid", the trigger condition (see Parts 4 and 5) occurs and it enters "pending review", and the review fails or a conflict occurs and it turns into "invalid".

[0024] II. Control Network Acquisition and Seven-Parameter Registration Select at least six control points with reasonable spatial distribution and different elevations, and obtain their coordinate data in the field spatial coordinate system and the model spatial coordinate system; perform robust pre-screening on paired coordinates (e.g., 3σ or RANSAC to remove outliers), then establish a seven-parameter similarity transformation and solve it using the least squares method: ; ; in: The model coordinates are a three-dimensional column vector; Scale factor; For Euler angles , , The three-dimensional rotation matrix formed; T is a three-dimensional column vector representing the on-site coordinates; T is a three-dimensional translation column vector. Let the objective function be the sum of squared residuals; For summation operators; For control point indexing; This represents the total number of control points. It is a 2-norm operator.

[0025] III. External Inspection at Checkpoints and Threshold Determination External accuracy verification is performed by selecting checkpoints independent of control points; preferably, registration is deemed qualified when the root mean square (RMS) residual of the checkpoint conversion deviation is not greater than 5 mm and the maximum residual is not greater than 8 mm. Under qualified conditions, the "Coordinate Transformation Parameter Version Number (TransformVersion)" is fixed and associated with the mapping record to form a traceable reference; if unqualified, control points are added or the solution is re-measured until the threshold is met.

[0026] IV. Versioning Application and Spatial Consistency Verification Use the latest TransformVersion for all site-to-model coordinate transformations, and perform a hole position spatial consistency check before each data write to bolt objects in the BIM model; define plane deviation and elevation deviation as follows: ; ; in: This refers to the deviation of the hole position plane. The difference in coordinates in the X direction; E represents the Y-axis coordinate difference; E represents the borehole elevation deviation. The Z-direction coordinate difference is used. Preferably, when D≤5mm and E≤5mm, the verification is considered successful. Otherwise, the mapping status is marked as "pending verification", the writing is blocked, and a prompt is given to retest or update the model hole position before comparison.

[0027] V. Triggering Update and Rollback Strategies To adapt to scenarios such as BIM geometry or attribute updates during construction, control network reconstruction, replacement of surveying equipment, or sudden changes in ambient temperature (e.g., temperature difference greater than 15 ℃), the system automatically triggers rapid re-inspection: Extract at least two checkpoints to verify the validity of the TransformVersion; if the estimated residual will exceed the threshold, force reregistration and generate a new TransformVersion; if the new version fails the external check, it can be rolled back to the previous valid version without interrupting the data flow and without changing the reference relationship and timestamp of the existing mapping records, while retaining the failed version and error code for auditing.

[0028] Step S2: Collect axial force data of the physical bolt and simultaneously collect ambient temperature data, including: collecting axial force measurement signals through a preload sensing shim set on the physical bolt, and collecting ambient temperature signals through a temperature sensor placed around the physical bolt. The acquired axial force measurement signal and ambient temperature signal are preprocessed. The preprocessing includes data filtering and outlier removal for the axial force measurement signal and ambient temperature signal, respectively, to obtain the processed axial force data and ambient temperature data.

[0029] Specifically, step S2 includes: After step S1 is completed and the mapping status is "valid", and the TransformVersion is obtained, the system starts the acquisition session according to the unique identification information in the mapping record. All data frames uniformly use UTC millisecond-level timestamps and carry unique identification information and TransformVersion to ensure consistent writing with the bolt objects in the BIM model in the future. Preferably, the preload sensing shim adopts a shim-type force sensor and is provided with a through hole for the bolt to pass through. During the bolt tightening process, a signal corresponding to the preload is output. The temperature sensor can be a patch type or a digital temperature probe. Each connection node is provided with at least one temperature point. Dense bolt groups can be arranged in such a way that 4 to 8 bolts share one temperature point.

[0030] (1) Calibration conversion and time unification The original measurements of the sensing pads were converted into initial values ​​of axial force, and the two types of sequences were mapped to a unified time grid:

[0031]

[0032] in: For the first Initial axial force values ​​(in N) at each sampling time. Zero-bias coefficient (unit: N) The coefficient is a linear coefficient (unit N / count). The coefficient is a quadratic coefficient (unit N / count²). For the first The raw measurement values ​​at each sampling time (in units of sensor count or voltage). For discrete sampling index (dimensionless), coefficients Obtained through laboratory calibration; For the first Timestamps for a unified time grid (in seconds or milliseconds, using UTO reference); The timestamp of the session start (in seconds or milliseconds) is used for data collection. The fixed time interval (in seconds or milliseconds) between adjacent sampling points.

[0033] (2) Cascaded filtering and anomaly removal First, perform three-point median filtering to suppress shot noise: ; in: This is the axial force value after median filtering (in N); It is a median operator (dimensionless); For the first Initial values ​​of axial force (in N) at each sampling time; For the first Initial axial force values ​​(in N) at each sampling time; For the first Initial values ​​of axial force (in N) at each sampling time; For discrete sampling indexes (dimensionless); A first-order IIR low-pass filter is then applied to suppress high-frequency vibrations. ; in: The axial force value after the first-order low-pass filter (in N); The low-pass smoothing factor (dimensionless, satisfying...) ); For the first The first-order low-pass axial force value (in N) at each sampling time; This is the axial force value after median filtering (in N); for Complementary weighting coefficients (dimensionless).

[0034] (3) Hampel robust anomaly identification and replacement; In Calculate the median and the absolute deviation of the median over a sliding window centered on the median, and identify outliers accordingly:

[0035]

[0036] in: For window The calculated median of the first-order low-pass axial force (in N); It is a median operator (dimensionless); For index The first-order low-pass axial force value at the location (in N); This is the index for discrete sampling within the window (dimensionless); For the first A set of sliding window indices centered on each sampling point (dimensionless); This represents the median absolute deviation (in N) for the corresponding window. It is an absolute value operator (dimensionless).

[0037] When the following conditions are met: hour; Mark the point as an anomaly and use Alternatively, linear interpolation of adjacent valid points may be used, where: This is the temperature-adaptive anomaly threshold coefficient (dimensionless); This represents the magnitude of the deviation from the median at the current point (in N). This represents the absolute deviation of the center within the corresponding window (in N).

[0038] (4) Temperature adaptive threshold Considering the impact of temperature fluctuations on sensor sensitivity, the following adaptive coefficient is used to correct the abnormal threshold, and the temperature sequence is processed and aligned to the same procedure as the axial force sequence. :

[0039] in: This is the temperature-adaptive anomaly threshold coefficient (dimensionless); The baseline anomaly threshold coefficient (dimensionless); This is the temperature sensitivity coefficient (dimensionless); For the first Temperature values ​​(in °C) for a unified time grid; For window The calculated average temperature (in °C); It is an absolute value operator (dimensionless).

[0040] (5) Data consistency gating and packaged writing The following ordered triplet will only be written to the bolt object in the BIM model if the mapping status of the corresponding unique identifier information is "valid" and the TranstormVersion is the latest version:

[0041] in: For the first An ordered triplet record of a unified time grid (dimensionless set representation); In the first Axial force (in N) processed at a unified time grid after median filtering, first-order IIR low-pass filtering, Hampel anomaly removal, and linear interpolation when necessary; In the first Temperature values ​​(in °C) processed at a unified time grid after median filtering, first-order IIR low-pass filtering, Hampel anomaly removal, and linear interpolation when necessary; For the first A unified time grid timestamp (in seconds or milliseconds, using UTC reference).

[0042] Step S3: Based on the geometric and attribute information of the bolt objects in the BIM model, construct an initial digital twin of the physical bolt. Then, using axial force data and ambient temperature data, calibrate the mechanical parameters of the initial digital twin using a parameter identification algorithm to generate the calibrated digital twin, which includes: A parametric geometric model of the physical bolt is established based on the geometric information of the bolt object in the BIM model. The parametric geometric model includes the main structural feature dimensions of the bolt. At the same time, the mechanical properties of the material are assigned based on the attribute information, including the elastic modulus and Poisson's ratio parameters; Set boundary constraints based on the design connection relationship of physical bolts in the structure, including the contact relationship and constraint type of the connection parts; An initial load based on the design preload value is applied to create an initial digital twin containing complete information on geometry, materials, boundaries, and loads; Using the processed axial force data as the observation benchmark, a comparison relationship between the output response of the initial digital twin and the measured axial force data is established. The key mechanical parameters of the initial digital twin are calculated by inverting the parameter identification algorithm. The key mechanical parameters include connection stiffness and friction coefficient. During the parameter inversion process, ambient temperature data is introduced for compensation and correction to eliminate the impact of temperature changes on parameter identification. By adjusting key mechanical parameters through an iterative optimization process, the output response of the initial digital twin gradually converges with the measured axial force data. When the preset convergence condition is met, parameter calibration is completed, and a calibrated digital twin is generated.

[0043] Step S3 specifically includes: Prerequisites and units: This step is entered only if the mapping status of step S1 is "valid" and the TransformVersion is up-to-date; Standardized units: force (N), length (mm), stress (The MPa output from the BIM / solver is set to 1 MPa = 1) (Conversion), temperature (°C), time (ms); then the normal / tangential equivalent stiffness is denoted as (Initial value recorded) The remaining symbols will be explained one by one in the "Single Explanation" section after each formula.

[0044] Initial parameter vector Based on BIM attributes (nominal dimensions, connection type), material manuals (elastic modulus, surface treatment friction coefficient), and empirical / experimental values ​​of typical connection surfaces; upper and lower bounds of parameters , , , , , Determined based on the physical feasible domain and engineering experience (or on-site small-scale testing), and recorded in project configuration for reuse and traceability.

[0045] I. Constructing the "Initial Digital Twin" Based on the geometric and attribute information of the bolt objects in the BIM model, parametric geometry, material constitutive model, contact and constraint are established, and the design preload is applied to obtain the initial parameters. The twin solver interface indexes at time step. Back Based on the positive field quantity, the "initial predicted axial force" is calculated:

[0046] in: For the first Initial predicted axial force for a unified time grid; The initial normal equivalent stiffness; The soft regularity kurtosis is dimensionless. The natural logarithm is dimensionless; The exponent is dimensionless; For normal stress, To contact the opening and closing reference stress, It is the normal equivalent flexure / extension measure; The baseline friction coefficient is dimensionless; The amplitude of the temperature effect is dimensionless; It is a dimensionless hyperbolic tangent. The rate is affected by temperature; Pretreatment temperature, For reference temperature, The initial tangential equivalent stiffness; This is the tangential equivalent slip. The discrete sampling index is dimensionless.

[0047] II. Calibration: "Generating a calibrated digital twin" 1. Temperature compensation (subtracting sensor temperature drift from observations)

[0048] in: Axial force after temperature compensation; To pre-treat axial force; β is the linear temperature sensitivity; β is the second-order temperature sensitivity; degree Pretreatment temperature; For reference temperature; 2. Predicted axial force during the calibration phase

[0049] in: For parameters Predicted axial force below; The units for each component of the parameter set to be calibrated are the same as above.

[0050] 3. Standardized Residuals and Objective Function

[0051]

[0052] in: The standardized residuals are dimensionless; Axial force after temperature compensation; To predict axial force; The baseline standard deviation; This is the coefficient of temperature heteroscedasticity; Pretreatment temperature; For reference temperature, The absolute value operator is dimensionless; The overall objective is dimensionless; The Tikhonov regular weight is dimensionless; The second norm is dimensionless; The initial parameter vector unit is the same as above; The total number of samples is dimensionless.

[0053] 4. Parameter Iterative Update

[0054] in: For the first The parameter vector after the next update; For the first Secondary parameter vector; J is in to about The Jacobian matrix of its first degree Behavior ; Let be a diagonal weight matrix, and its first... The diagonal elements are This diagonal element corresponds to the time sampling index k. The damping factor is dimensionless; The identity matrix is ​​dimensionless; e is the residual column vector whose first... Components ; The iterative counting index is dimensionless.

[0055] 5. Convergence Criterion and Parameter Boundary

[0056] in: The residual norm 2; The residual threshold; The parameter step size is dimensionless or normalized to component units; The parameter step size threshold is dimensionless; For logic and dimensionless; Apply physical boundaries during optimization: , , , , , .

[0057] 6. Versioned Hammer Drop and Output Once the above equation converges, take... For the "calibrated digital twin" parameters, generate the parameter version number ParamVersion and write it back along with TransformVersion and the bolt's unique identifier; Recalculate the axial force of each bolt within the node and the stress / deformation field of the component region, and provide the results to step S4 for the calculation of the preload uniformity index and stress / deformation.

[0058] Step S4: Using the calibrated digital twin, calculate the preload uniformity index of the physical bolts, and calculate the stress and deformation in the component area connected to the physical bolts, including: Mechanical simulation analysis is performed based on the calibrated digital twin. The simulated axial force values ​​of each physical bolt at the same connection node are extracted, and the preload uniformity index is calculated based on the simulated axial force values. Mechanical simulation analysis is performed based on the calibrated digital twin to extract the stress distribution in the component area connected to the physical bolts and obtain the stress values ​​at key locations. Mechanical simulation analysis is performed based on the calibrated digital twin to extract the displacement distribution of the component area connected to the physical bolts and obtain deformation data at key locations.

[0059] Specifically, step S4 includes: In step S3, the parameter set of the "calibrated digital twin" is obtained. Next, in this step, under the same coordinate system as step S1 and the same working condition label as step S2, the same version (ParamVersion and TransformVersion correspond to each other) of contact-friction-stiffness parameters is used to perform static / contact nonlinear solutions. This yields the simulated axial force of each physical bolt within the same connection node, as well as the stress and displacement fields of the component region connected to that node. Based on this, the preload uniformity index is calculated. Stress at key locations Deformation at key locations .

[0060] (a) Uniformity index of bolt preload at the same node After solving for the target connection node, read the simulated axial force sequence of n bolts at that node. Construct a dimensionless homogeneity index using the inverse mapping of the coefficient of variation:

[0061] in: The preload uniformity index is dimensionless and ranges from (0,1], and CV is the dimensionless coefficient of variation of axial force samples. The standard deviation of the sample is 1. The sample mean. For the first Simulates axial force on a single bolt. The bolt number index is dimensionless. The number of bolts at the same node is dimensionless and , To make the summation operator dimensionless, The square root operator is dimensionless. by Solving for the results Then, calculate according to the above formula. , CV, finally yielding the preload uniformity index of the node. (Dimensionless).

[0062] (ii) Stress at critical locations in the component area In the same simulation step, the principal stress field of the connecting component region is extracted, and the extreme values ​​are obtained in the region of interest using the "maximum principal stress method":

[0063] in: The maximum principal stress at the critical location, For the operator to reach its maximum value, it must be dimensionless. It is a spatial position vector. The point set for the region of interest of the connecting components is dimensionless (the region is defined by BIM constraints and typically includes bolt hole flare, plate edge band, and stiffening zone). For position The first principal stress is located at; Work process: The stress field obtained below is constrained to The internal structure is scanned, and the extreme values ​​are obtained according to the above formula to obtain the stress at the critical location of the node component. .

[0064] (iii) Deformation at key locations in the component area To determine the degree of fit and potential opening tendency of the connecting surfaces, the extreme values ​​of the normal displacement are extracted from the set of contact surfaces:

[0065] in: This represents the normal deformation at the critical location. For the operator to reach its maximum value, it must be dimensionless. The absolute value operator is dimensionless. It is a spatial position vector. The set of contact surfaces is dimensionless (a set of patches defined by BIM contact pairs). For position Displacement vector at point, The vector dot product operator is dimensionless. For position The unit normal vector at the contact surface is dimensionless. exist Extract the displacement field and project it onto the normal. Then, take the extreme value according to the above formula to obtain the deformation at the critical location of the node. .

[0066] Step S5: Based on the calculation results of the preload uniformity index and stress and deformation, generate a treatment command for the physical bolt. After executing the treatment command, re-collect the axial force data and ambient temperature data of the physical bolt to verify the treatment effect, including: Based on the comparison results of the preload uniformity index and the preset threshold, the comparison results of the stress values ​​at key locations and the allowable values ​​of the material, and the comparison results of the deformation data at key locations and the allowable values ​​of the design, the physical bolt objects that need to be treated and the corresponding treatment measures are determined through comprehensive analysis, and treatment instructions containing specific treatment objects and treatment measures are generated. After executing the disposal instruction, steps S2 to S4 are repeated to obtain a new preload uniformity index, new stress values ​​at key locations, and new deformation data at key locations. The new calculation results are compared and analyzed with the corresponding thresholds. The disposal effect is verified based on the comparison and analysis results. When all evaluation parameters meet the requirements, the quality control process is completed. Otherwise, a new disposal instruction is generated based on the new comparison and analysis results to continue optimizing the process.

[0067] Specifically, step S5 includes: In step S4, the preload uniformity index of the same connection node has been obtained. Stress at key locations Deformation at key locations Then, this step uses the "calibrated digital twin" parameter version output in step S3. Using this as the sole model basis, and under the premise of unifying the coordinates with step S1 and maintaining consistency with the data caliber of step S2, the following steps are completed: "Threshold determination - object identification and sorting - target axial force and torque command - sequence and execution constraints". The process of "retesting and acceptance" is closed-loop, and the results of the entire process are versioned and written back for traceability.

[0068] I. Threshold Determination and Decision Triggering Let the project threshold set be set. If and only if both conditions are met , , If the condition is met, it is marked as "passed"; otherwise, it proceeds to the processing flow. The threshold is recorded together with ParamVersion / TransformVersion to ensure alignment across batches.

[0069] II. Identification of Disposal Targets and Priority Scoring To obtain the reference axial force that resists outliers, we first define:

[0070] in: The reference axial force at this node is... The median operator is dimensionless. For the first Simulates axial force with a single bolt. The bolt number index is dimensionless. The number of bolts at this node is dimensionless.

[0071] The construction prioritizes and scores considerations for "axial force deviation, proximity to high stress, and sensitivity to openings":

[0072] in: For the first The priority scoring for bolt handling is dimensionless. The weight of axial force deviation is dimensionless. The stress proximity weight is dimensionless. The opening-sensitive weight is dimensionless. The absolute value operator is dimensionless. For the first bolt and The normalized proximity coefficient of the key point is dimensionless (linearly mapped from BIM distance to [0,1]). For the first bolt pair The normalized sensitivity coefficient of the contact surface is dimensionless (mapped from the local normal stiffness and the hole edge geometry to [0,1]), and the meanings of the other symbols are the same as before.

[0073] Weight constraints: and Project default .according to Bolts with large deviations and those more likely to affect high stress / openings are selected for treatment in descending order of priority for re-tightening or loosening.

[0074] III. Generation of Target Axial Force and Torque Commands The target axial force adopts a "conservative approximation equalization" strategy:

[0075] in: For the first The target axial force of each bolt. To make real numbers Cut off in interval The operators within are dimensionless. The approximation coefficients are dimensionless and take values ​​in the range (0,1], with a default value of 0.3-0.5. To design the lower / upper limits of the allowable axial force, The meaning is the same as before.

[0076] Convert the target axial force into a torque command (considering the combined effects of temperature-related friction and thread / bearing surface):

[0077] in: For the first The target torque of each bolt The empirical coefficient for torque and axial force is infinite (determined by thread type and bearing surface conditions). For the first Nominal diameter of the bolt, The dimensionless weighting for friction correction is represented by the outline. For the first The equivalent friction coefficient of a bolt is dimensionless.

[0078] Equivalent friction coefficient calculation formula:

[0079] Symbol Explanation (Single Segment): For the first The equivalent friction coefficient of a bolt is dimensionless. The calibrated baseline friction coefficient is dimensionless. The amplitude of the temperature effect is dimensionless. The hyperbolic tangent operator is dimensionless. For the rate affected by temperature, For the first Ambient temperature at the bolt location, For reference temperature, The bolt number index is dimensionless.

[0080] IV. Tightening Sequence and Execution Constraints A "diagonal alternating circumferential diffusion" path is adopted to suppress the peak value caused by redistribution: Symmetrically paired and tightened with reference to the geometric center, then pushed outward along the ring band; if If the limit is exceeded, first increase the torque in small steps at the point symmetrical to the opening direction; if Approaching the allowable value, the stress is reduced in the high-stress adjacent region. (Smaller step size) to avoid abrupt changes in stiffness; the processing card specifies BoltUID, tool model, and allowable deviation.

[0081] V. Retesting and Closed-Loop Acceptance (Rerunning S2 to S4) Immediately after treatment, retest according to step S2, and recalculate according to the caliber of steps S3 / S4. , , , .

[0082] Acceptance testing was completed using logical tables:

[0083] in: This is for closed-loop acceptance logic determination. The preload uniformity index after retesting. The stress at key locations after retesting, This refers to the deformation at key locations after retesting. For logic AND operators.

[0084] To avoid ineffective and repeated adjustments, a maximum number of processing rounds and a stop-loss convergence criterion are defined:

[0085] in: For the first With the Wheel uniformity change, For key stress changes, For key deformation changes, For their respective minimum improvement thresholds, To handle the round index, The maximum number of disposal rounds allowed. This is a logical OR operator. If the stop-loss condition is met, the current batch ends and a diagnostic report is output.

[0086] Example 2 A system employing a BIM-based digital twin-based method for quality control of building bolt assembly includes: Mapping and Registration Module: Establishes a one-to-one mapping relationship between physical bolts and bolt objects in the BIM model, and registers the on-site spatial coordinate system of the physical bolts with the model spatial coordinate system of the bolt objects in the BIM model; Data acquisition and preprocessing module: Acquires axial force data of physical bolts and simultaneously acquires ambient temperature data; Digital Twin Construction and Parameter Calibration Module: Based on the geometric and attribute information of bolt objects in the BIM model, an initial digital twin of the physical bolt is constructed. Using axial force data and ambient temperature data, the mechanical parameters of the initial digital twin are calibrated through a parameter identification algorithm to generate a calibrated digital twin. Evaluation module: Using the calibrated digital twin, calculate the preload uniformity index of the physical bolts and calculate the stress and deformation in the component area connected to the physical bolts; Treatment and verification module: Based on the calculation results of the preload uniformity index and stress and deformation, a treatment command is generated for the physical bolt. After the treatment command is executed, the axial force data and ambient temperature data of the physical bolt are collected again to verify the treatment effect.

[0087] The technical features of this invention not described can be implemented by or using existing technology, and will not be repeated here. Of course, the above description is not a limitation of this invention, and this invention is not limited to the examples above. Any changes, modifications, additions or substitutions made by those skilled in the art within the scope of this invention should also be within the protection scope of this invention.

Claims

1. A method for quality control of building bolt assembly based on BIM digital twin, characterized in that, Includes the following steps: Establish a one-to-one mapping relationship between physical bolts and BIM model bolt objects, and register the field space coordinate system of the physical bolts with the model space coordinate system of the BIM model bolt objects; Collect axial force data of the physical bolts and simultaneously collect ambient temperature data; Based on the geometric and attribute information of the bolt object in the BIM model, an initial digital twin of the physical bolt is constructed. Then, using the axial force data and the ambient temperature data, the mechanical parameters of the initial digital twin are calibrated through a parameter identification algorithm to generate a calibrated digital twin. Using the calibrated digital twin, the preload uniformity index of the physical bolt is calculated, and the stress and deformation in the component area connected to the physical bolt are calculated. Based on the preload uniformity index and the calculation results of stress and deformation, a treatment command is generated for the physical bolt. After executing the treatment command, the axial force data and ambient temperature data of the physical bolt are collected again to verify the treatment effect.

2. The method for quality control of building bolt assembly based on BIM digital twins according to claim 1, characterized in that, Establishing a one-to-one mapping relationship between physical bolts and BIM model bolt objects includes: obtaining unique identification information attached to the physical bolts, writing the unique identification information into the attribute information of the BIM model bolt objects, and verifying the consistency between the specifications, connection type, and hole number of the physical bolts corresponding to the unique identification information and the preset design parameters in the BIM model bolt objects. When the verification passes, the one-to-one mapping relationship is confirmed and recorded.

3. The method for quality control of building bolt assembly based on BIM digital twins according to claim 2, characterized in that, The registration of the physical bolt's on-site spatial coordinate system with the BIM model bolt object's model spatial coordinate system includes: The coordinate data of multiple control points on site in the site spatial coordinate system are obtained by measuring equipment, and the coordinate data of the corresponding control points in the model spatial coordinate system are extracted from the BIM model. Based on two sets of coordinate data, a seven-parameter transformation model is used to calculate the transformation parameters between coordinate systems, and the adjustment optimization is performed using the least squares method. Registration accuracy is verified using checkpoints independent of control points. When the coordinate transformation deviation of the checkpoint does not exceed the preset deviation threshold, the registration is confirmed to be qualified. The qualified transformation parameters are then applied to all subsequent coordinate transformation operations to ensure the consistency of spatial data.

4. The method for quality control of building bolt assembly based on BIM digital twins according to claim 3, characterized in that, The process of collecting axial force data of the physical bolt and simultaneously collecting ambient temperature data includes: collecting axial force measurement signals through a preload sensing shim set on the physical bolt, and collecting ambient temperature signals through a temperature sensor arranged around the physical bolt. The acquired axial force measurement signal and ambient temperature signal are preprocessed. The preprocessing includes data filtering and outlier removal of the axial force measurement signal and ambient temperature signal, respectively, to obtain processed axial force data and ambient temperature data.

5. The method for quality control of building bolt assembly based on BIM digital twins according to claim 4, characterized in that, Constructing an initial digital twin of the physical bolt based on the geometric and attribute information of the bolt object in the BIM model includes: establishing a parametric geometric model of the physical bolt based on the geometric information of the bolt object in the BIM model, wherein the parametric geometric model includes the main structural feature dimensions of the bolt; Simultaneously, the material's mechanical properties, including elastic modulus and Poisson's ratio, are assigned based on the aforementioned attribute information; Boundary constraints are set according to the design connection relationship of the physical bolts in the structure, including the contact relationship and constraint type of the connection parts; An initial load based on the design preload value is applied to create an initial digital twin containing complete information on geometry, materials, boundaries, and loads.

6. The method for quality control of building bolt assembly based on BIM digital twins according to claim 5, characterized in that, Using the axial force data and the ambient temperature data, the mechanical parameters of the initial digital twin are calibrated through a parameter identification algorithm to generate a calibrated digital twin, including: Using the processed axial force data as an observation benchmark, a comparison relationship is established between the output response of the initial digital twin and the measured axial force data. The key mechanical parameters of the initial digital twin are calculated by inversion using a parameter identification algorithm. The key mechanical parameters include connection stiffness and friction coefficient. The ambient temperature data is introduced during the parameter inversion process for compensation and correction to eliminate the impact of temperature changes on parameter identification. The key mechanical parameters are adjusted through an iterative optimization process, so that the output response of the initial digital twin gradually converges with the measured axial force data. When the preset convergence condition is met, the parameter calibration is completed, and the calibrated digital twin is generated.

7. The method for quality control of building bolt assembly based on BIM digital twins according to claim 6, characterized in that, Using the calibrated digital twin, the preload uniformity index of the physical bolt is calculated, and the stress and deformation in the component area connected to the physical bolt are calculated, including: performing mechanical simulation analysis based on the calibrated digital twin, extracting the simulated axial force values ​​of each physical bolt at the same connection node, and calculating the preload uniformity index based on the simulated axial force values. Mechanical simulation analysis is performed based on the calibrated digital twin to extract the stress distribution in the component area connected to the physical bolt and obtain the stress values ​​at key locations. Mechanical simulation analysis is performed based on the calibrated digital twin to extract the displacement distribution of the component area connected to the physical bolts and obtain deformation data at key locations.

8. The method for quality control of building bolt assembly based on BIM digital twins according to claim 7, characterized in that, Based on the calculation results of the preload uniformity index and the stress and deformation, the treatment instructions for the physical bolts are generated by: comprehensively analyzing and determining the physical bolt objects that need to be treated and the corresponding treatment measures based on the comparison results of the preload uniformity index and the preset threshold, the comparison results of the stress values ​​at the key locations and the material allowable values, and the comparison results of the deformation data at the key locations and the design allowable values, and generating treatment instructions that include specific treatment objects and treatment measure types.

9. The method for quality control of building bolt assembly based on BIM digital twins according to claim 8, characterized in that, After executing the treatment command, the axial force data and ambient temperature data of the physical bolts are collected again to verify the treatment effect, including: After executing the disposal instruction, steps S2 to S4 are re-executed to obtain a new preload uniformity index, new stress values ​​at key locations, and new deformation data at key locations. The new calculation results are compared and analyzed with the corresponding thresholds. The disposal effect is verified based on the comparison and analysis results. When all evaluation parameters meet the requirements, the quality control process is completed. Otherwise, a new disposal instruction is generated based on the new comparison and analysis results to continue optimizing the process.

10. A system employing the BIM digital twin-based building bolt assembly quality control method as described in any one of claims 1-9, characterized in that, include: Mapping and Registration Module: Establishes a one-to-one mapping relationship between physical bolts and BIM model bolt objects, and registers the field space coordinate system of the physical bolts with the model space coordinate system of the BIM model bolt objects; Data acquisition and preprocessing module: Acquires axial force data of the physical bolts and simultaneously acquires ambient temperature data; Digital twin construction and parameter calibration module: Based on the geometric and attribute information of the bolt object in the BIM model, an initial digital twin of the physical bolt is constructed, and the mechanical parameters of the initial digital twin are calibrated using the axial force data and the ambient temperature data through a parameter identification algorithm to generate a calibrated digital twin; Evaluation module: Using the calibrated digital twin, calculate the preload uniformity index of the physical bolt, and calculate the stress and deformation in the component area connected to the physical bolt; Treatment and verification module: Based on the preload uniformity index and the calculation results of stress and deformation, a treatment instruction is generated for the physical bolt. After executing the treatment instruction, the axial force data and ambient temperature data of the physical bolt are collected again to verify the treatment effect.

Images