Virtual pre-assembly method of bridge tower reinforcement parts

Abstract

The application provides a virtual pre-assembly method for a bridge tower steel reinforcement part, adopts a BIM tool to perform data analysis and modeling on profile point coordinate of each section of the steel reinforcement part, forms theoretical coordinate point data, considers monitoring pre-deviation value of a current actual construction stage and influence factors of concrete pouring deformation, forms a section theoretical data model, builds a measured data model, collects steel reinforcement part forming precision information through a total station, a three-dimensional laser and various image visual technologies, imports the information into a digital system to build a real-time data model, compares and analyzes the theoretical model and the real-time model in real time, realizes deviation visualization, predicts size data of a next section of the steel reinforcement part through a special algorithm, guides a forming process of the next section of the steel reinforcement part, realizes dynamic adjustment of production data, and develops a set of virtual pre-assembly software for the tower column steel reinforcement part, integrates all the above functions, and realizes digital virtual construction.

Description

Technical Field

[0001] This invention relates to the field of virtual pre-assembly of steel reinforcement components, and in particular to a method for virtual pre-assembly of steel reinforcement components for bridge towers. Background Technology

[0002] For complex steel components, pre-assembly is often necessary to ensure accurate and smooth on-site installation. This allows for timely correction and reduction of cumulative errors should deviations occur. Pre-assembly of physical components not only consumes significant manpower, materials, transportation, and site resources, but also places higher demands on project schedule management. Pre-assembly operations require large areas of land, involve cumbersome inspection processes, are time-consuming and costly, and have relatively low inspection accuracy, while also posing certain safety hazards. With the development of measurement and computer technologies, three-dimensional pre-assembly technology, represented by digitalization and virtualization, has become the future direction. Summary of the Invention

[0003] The main objective of this invention is to provide a virtual pre-assembly method for bridge tower steel reinforcement components. By applying digital construction and virtual pre-assembly technology for steel reinforcement components, it achieves data-driven quality, visualized deviations, and standardized construction.

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a virtual pre-assembly method for bridge tower steel reinforcement components, the method comprising:

[0005] S1. Use BIM tools to analyze and model the coordinates of the outline points of the upper, middle and lower sections of each segment of the steel reinforcement component, form theoretical coordinate point data, and consider the monitoring pre-deviation value and the influence of concrete pouring deformation in the current actual construction stage to form a segment theoretical data model.

[0006] S2. Build a measured data model. Collect steel reinforcement component forming accuracy information through a variety of technologies such as total station, 3D laser and image vision, and import it into the digital system to build a real-time data model.

[0007] S3. Real-time comparative analysis: By comparing and analyzing the theoretical model and the real-time model, deviations are visualized. Furthermore, a dedicated algorithm is used to predict the size data of the next segment of steel reinforcement components, guiding the forming process of the next segment of steel reinforcement components and enabling dynamic adjustment of production data.

[0008] S4. Develop a set of virtual pre-assembly software for tower column steel reinforcement components, integrating all the above functions to realize digital virtual construction.

[0009] In the preferred scheme, the acquisition of theoretical point data includes:

[0010] Monitoring pre-offset values: Considering the tower limb profile, the pre-offset values ​​of the control points are obtained through simulation analysis using MIDAS / CIVIL;

[0011] Concrete pouring deformation: Deformation values ​​are obtained by measuring the coordinates of control points before and after pouring using a total station;

[0012] Deformation of steel reinforcement components during hoisting: The deformation value of the steel reinforcement components after hoisting was obtained by simulation analysis using MIDAS / CIVIL;

[0013] Adjustment value: Used to guide the construction of the next segment of steel reinforcement components.

[0014] In the preferred scheme, the measured data includes:

[0015] The measured coordinates of eight control points at the top and bottom openings, used to control the overall alignment and orientation of the reinforcing steel components, were obtained using a total station.

[0016] The coordinates of the main reinforcing bars used to control the splicing of steel reinforcement components are obtained through a scanner or drone;

[0017] After acquiring data using a total station, scanner, and drone, the data is processed to extract the measured coordinates of the eight control points controlling the steel reinforcement components and the coordinates of the main reinforcement bars at the joints of the steel reinforcement components, thus obtaining the measured data.

[0018] In the preferred scheme, the specific steps for collecting measured data are as follows:

[0019] A1. Standardize the measured coordinate data of the eight control points obtained by the total station and the coordinate data of the main reinforcement obtained by the scanner or UAV. For the data obtained by the total station, the coordinates of the eight control points are as follows: Calculate the mean of each coordinate component. , , Then, each coordinate is standardized, and the standardized coordinates are... , , , .

[0020] For the main reinforcement coordinate data acquired by scanners or drones, the coordinate of a single main reinforcement is: Calculate its distance to the origin. Then, standardization is performed, and the standardized coordinates are... ;

[0021] A2. Calculate the weights of total station data and scanner / UAV data in the fusion process. The weight of the total station data is: The weight of scanner / drone data is ,and ;

[0022] The formula for calculating the weights is: , ,in , , Adjust the parameters based on experience or actual data.

[0023] A3. Merge the standardized total station data and scanner / UAV data. The resulting coordinate data for a single main reinforcement bar is: Perform a fusion operation on the coordinates of each main reinforcement bar separately;

[0024] A4. Reconstruct the fused data to obtain the final measured data used to control the reinforcing steel components. For the data of the eight control points, the reconstructed coordinates are as follows: , , , For the main reinforcement coordinates, the restored coordinates are: .

[0025] These restored coordinate data are extracted as the final measured data for subsequent analysis and processing.

[0026] In the preferred embodiment, the specific steps in step S3 are as follows:

[0027] S31, Theoretical Model The coordinates of the control points are The coordinates of the corresponding control points in the real-time model are , , Number of control points;

[0028] Calculate the spatial deviation vector for each control point ;

[0029] Extract the magnitude of the deviation vector As a characteristic of the magnitude of the deviation;

[0030] The cosine of the angle between the deviation vector and the coordinate axis is calculated as the directional feature; Taking the shaft as an example, Similarly, we can obtain the following: Cosine value of the included angle and with Cosine value of the included angle ;

[0031] Among them, the deviation vector formula This directly reflects the spatial difference between theoretical points and real-time points, and serves as the basis for subsequent analysis of biases;

[0032] Deviation modulus formula Based on the Euclidean distance formula, it is used to quantify the magnitude of the deviation;

[0033] The formula for direction cosine is as follows: Used to describe the direction of the deviation vector relative to the coordinate axes, providing a basis for subsequent analysis that considers directional factors;

[0034] S32. Define the deviation visualization function ;set up ,in , , , To adjust the parameters;

[0035] according to The value maps the deviation to a visual range of color or brightness.

[0036] S33. Collect deviation characteristic data of multiple historical segments. , , Number of historical segments;

[0037] The next segment of steel reinforcement component size prediction vector is The next segment size vector in the theoretical model is ;

[0038] Building a predictive model:

[0039] First, the deviation characteristic data is normalized; let the normalized deviation modulus be... The normalized direction cosine values ​​are respectively , , ;

[0040] The prediction model then takes the following form:

[0041] ;

[0042] ;

[0043] ;

[0044] in , , These are weighting coefficients. , , These are bias terms; these weighting coefficients and bias terms are learned by minimizing the prediction error function.

[0045] The deviation feature data normalization formula maps the deviation magnitude and direction cosine value to the [0, 1] interval, which facilitates subsequent model training and weight calculation;

[0046] The prediction model formula predicts the size of the next segment by taking into account the magnitude and direction of historical deviation feature data;

[0047] Among them, the weighting coefficient , , and bias terms , , Adjustments are made by learning from historical data to adapt to different deviation patterns and the characteristics of steel reinforcement components.

[0048] In the preferred scheme, after comparing and analyzing the theoretical model and the real-time model, the measured point and theoretical point matching algorithm EOPA was first adopted, and then the virtual assembly alignment algorithm GPA was adopted to develop virtual pre-assembly software and carry out virtual matching analysis of single components and multiple components.

[0049] In the preferred scheme, the virtual assembly enhanced matching algorithm steps are as follows:

[0050] B1. Establishing the objective function, i.e., EOPA algorithm optimization: Let the theoretical point be... The actual measurement point is , , Number of control points;

[0051] Define a new distance metric function This function takes into account the relationship between the coordinate difference and the sum of squares of the coordinates themselves.

[0052] Establish the optimization objective function ,in To adjust the parameters;

[0053] Distance metric function By applying a special normalization process to the coordinate difference, the magnitude information of the coordinates themselves is taken into account, making the distance measurement more reasonable and comprehensive. This special form helps to more accurately measure the difference between theoretical points and measured points under coordinates of different scales.

[0054] Optimize objective function Built upon the exponential function and distance metric function, when The smaller the value, the larger the value of the exponential function. The smaller the value, the better the matching effect; adjust the parameter. The sensitivity of this matching degree can be controlled;

[0055] B2. Optimization of Virtual Assembly Alignment GPA Algorithm: The original control point matrix is... ( The target point matrix is ( );

[0056] Calculate the weighted covariance matrix ,in , and They are and The fundamental mean matrix, The points are calculated based on the distance metric function mentioned above. to the mean distance, To adjust the parameters;

[0057] For the weighted covariance matrix Perform singular value decomposition The optimal rotation matrix is ​​obtained. ;

[0058] Calculate the optimal translation vector ;

[0059] Weighted covariance matrix formula By introducing weights The contribution of control points in covariance calculation is adjusted based on their distance from the mean, making the algorithm pay more attention to control points that are far from the mean and improving the accuracy of location calculation.

[0060] Singular Value Decomposition Formula Used to find the optimal rotation matrix This matrix decomposition method obtains the optimal rotation relationship;

[0061] Optimal translation vector formula Used to calculate the amount of translation that makes the control point move from its original position to the target position;

[0062] B3. The optimal rotation matrix is ​​obtained from the above calculations. Translation vector Based on this, we further consider the consistency of direction between the measured point and the theoretical point; let the direction vectors of the measured point and the theoretical point be respectively... and ,in , , Similarly, for theoretical points;

[0063] Calculation of directional consistency index This indicates the degree of similarity in direction between adjacent points between measured and theoretical points;

[0064] For translation vector Adjustments are made, and the adjusted translation vector is... ,in These are weighting coefficients. It is the mean of the directional consistency index;

[0065] Finally, the translation matrix is ​​obtained. Translation matrix ,in It is the adjusted translation vector The resulting translation matrix (which can be represented as a single translation matrix) (The homogeneous coordinate matrix form extends the three-dimensional translation vector to homogeneous coordinate space).

[0066] Directional Consistency Index Formula By calculating the ratio of the vector dot product to the vector magnitude, the similarity of the directions between the measured point and the theoretical point in adjacent points can be obtained, providing a basis for adjusting the translation vector;

[0067] Translation vector adjustment formula The translation vector is adjusted based on the differences in the directional consistency index, so that the translated control points better conform to the requirements of the theoretical points in terms of direction; this is achieved by introducing weighting coefficients. and mean It allows for flexible control over the degree and direction of adjustment.

[0068] In the preferred scheme, based on the above results, multi-condition comparative analysis, component accuracy analysis, and virtual assembly accuracy analysis are carried out.

[0069] In the preferred scheme, C1 and multi-condition comparative analysis:

[0070] Define the working condition similarity function Used to measure two working conditions and The degree of similarity between them; ,in To adjust the parameters;

[0071] The working conditions are classified according to the value of the working condition similarity function; when At that time, the working condition was considered to be and Similar; when At that time, the working condition was considered to be and There are certain similarities; when At that time, the working condition was considered to be and The differences are significant;

[0072] C2. Calculate component precision indicators , ,in To adjust the parameters;

[0073] Evaluate component precision based on component precision index values;

[0074] when At that time, it was believed that the parts had high precision; when At that time, the precision of the parts was considered to be moderate; when At that time, it was believed that the parts had low precision.

[0075] The component accuracy index formula is obtained by exponentially weighting the deviations of each control point under each working condition, and then averaging the sums to obtain a comprehensive accuracy index; adjustment parameters The weight of the deviation magnitude in the accuracy index calculation can be controlled;

[0076] C3. The coordinates of the control points after virtual assembly are: Calculate the virtual assembly deviation vector ;

[0077] Extract the magnitude of the virtual assembly deviation vector As a characteristic of the magnitude of virtual assembly deviation;

[0078] Calculate virtual assembly accuracy index ,in To adjust the parameters;

[0079] The virtual assembly accuracy is evaluated based on the value of the virtual assembly accuracy index; when At the time, it was believed that virtual assembly had high precision; when At that time, the virtual assembly accuracy was considered to be moderate; when At that time, it was believed that virtual assembly had low precision;

[0080] Virtual assembly deviation vector formula This reflects the deviation between the control points after virtual assembly and the original coordinates, and is the basis for subsequent analysis of the accuracy of virtual assembly;

[0081] Virtual assembly deviation module formula Based on the Euclidean distance formula, it is used to quantify the magnitude of virtual assembly deviation.

[0082] This invention provides a virtual pre-assembly method for bridge tower steel reinforcement components. Compared to the pre-assembly of physical components, it avoids the consumption of significant manpower, material resources, transportation, and site resources, thus reducing the requirements for project progress management. It eliminates the need for large sites, simplifies the testing process, and may shorten measurement time and reduce testing costs.

[0083] By employing various advanced technologies and algorithms, using BIM tools for modeling, and total stations for data collection, and considering multiple influencing factors, higher accuracy is achieved. This enables the generation of accurate theoretical and measured data models for real-time comparative analysis, with deviations visualized, facilitating timely problem detection and correction.

[0084] The tower column steel reinforcement component structure realizes multiple functions, including automatic calculation of virtual assembly data, automatic assembly of measured data, automatic correction of various related factors of virtual assembly, error exceeding limit display, data report output, and free measurement and display of virtual assembly data.

[0085] By using specialized algorithms to predict the dimensions of the next steel reinforcement segment, the forming process is guided, enabling dynamic adjustments to production data and ensuring construction quality and installation accuracy. On-site assembly problems are identified early, allowing for timely rectification of deviations and ensuring smooth installation. Attached Figure Description

[0086] The present invention will be further described below with reference to the accompanying drawings and embodiments:

[0087] Figure 1 This is a flowchart of the digital pre-assembly technology for steel reinforcement components of this invention;

[0088] Figure 2 This is a diagram showing the simulation analysis results of Midas / civil from this invention;

[0089] Figure 3 This is a diagram showing the simulation analysis results of Midas / civil from this invention;

[0090] Figure 4 This is a single-component accuracy analysis diagram of the present invention;

[0091] Figure 5 This is a virtual assembly analysis diagram of multiple components according to the present invention;

[0092] Figure 6 This is a statistical analysis chart of the error data of this invention; Detailed Implementation

[0093] Example 1

[0094] like Figures 1-6 As shown, a virtual pre-assembly method for bridge tower steel reinforcement components includes:

[0095] S1. Use BIM tools to analyze and model the coordinates of the outline points of the upper, middle and lower sections of each segment of the steel reinforcement component, form theoretical coordinate point data, and consider the monitoring pre-deviation value and the influence of concrete pouring deformation in the current actual construction stage to form a segment theoretical data model.

[0096] S2. Build a measured data model. Collect steel reinforcement component forming accuracy information through a variety of technologies such as total station, 3D laser and image vision, and import it into the digital system to build a real-time data model.

[0097] S3. Real-time comparative analysis: By comparing and analyzing the theoretical model and the real-time model, deviations are visualized. Furthermore, a dedicated algorithm is used to predict the size data of the next segment of steel reinforcement components, guiding the forming process of the next segment of steel reinforcement components and enabling dynamic adjustment of production data.

[0098] S4. Develop a set of virtual pre-assembly software for tower column steel reinforcement components, integrating all the above functions to realize digital virtual construction.

[0099] In the preferred scheme, the acquisition of theoretical point data includes:

[0100] Monitoring pre-offset values: Considering the tower limb profile, the pre-offset values ​​of the control points are obtained through simulation analysis using MIDAS / CIVIL; for example... Figure 2 As shown.

[0101] Concrete pouring deformation: Deformation values ​​are obtained by measuring the coordinates of control points before and after pouring using a total station;

[0102] Deformation of reinforcing steel components during hoisting: The deformation values ​​of the reinforcing steel components after hoisting were obtained through simulation analysis using MIDAS / CIVIL; for example... Figure 3 As shown.

[0103] Adjustment value: Used to guide the construction of the next segment of steel reinforcement components.

[0104] Example 2

[0105] Further explanation in conjunction with Example 1, such as Figure 1 The structure shown, and the measured data include:

[0106] The measured coordinates of eight control points at the top and bottom openings, used to control the overall alignment and orientation of the reinforcing steel components, were obtained using a total station.

[0107] The coordinates of the main reinforcing bars used to control the splicing of steel reinforcement components are obtained through a scanner or drone;

[0108] After acquiring data using a total station, scanner, and drone, the data is processed to extract the measured coordinates of the eight control points controlling the steel reinforcement components and the coordinates of the main reinforcement bars at the joints of the steel reinforcement components, thus obtaining the measured data.

[0109] In the preferred scheme, the specific steps for collecting measured data are as follows:

[0110] A1. Standardize the measured coordinate data of the eight control points obtained by the total station and the coordinate data of the main reinforcement obtained by the scanner or UAV. For the data obtained by the total station, the coordinates of the eight control points are as follows: Calculate the mean of each coordinate component. , , Then, each coordinate is standardized, and the standardized coordinates are... , , , .

[0111] For the main reinforcement coordinate data acquired by scanners or drones, the coordinate of a single main reinforcement is: Calculate its distance to the origin. Then, standardization is performed, and the standardized coordinates are... ;

[0112] Mean formula This is used to centralize coordinate data, making it more comparable and stable. This eliminates systematic biases caused by different measurement locations.

[0113] Standardized formula Based on statistical principles, it transforms data to a relatively uniform scale, facilitating subsequent calculations and analysis.

[0114] A2. Calculate the weights of total station data and scanner / UAV data in the fusion process. The weight of the total station data is: The weight of scanner / drone data is ,and ;

[0115] The formula for calculating the weights is: , ,in , , Adjust the parameters based on experience or actual data.

[0116] The weighting formula is based on a non-linear mapping relationship, using an exponential function to dynamically allocate weights according to a certain "information content" of the total station data (measured here by the sum of the absolute values ​​of the coordinates). When the coordinates of the total station data change significantly (i.e., the information content is large), The larger the value, the more weight is given to the total station data; conversely, It will be relatively large. Adjust the parameters. , , The sensitivity of this mapping can be controlled.

[0117] A3. Merge the standardized total station data and scanner / UAV data. The resulting coordinate data for a single main reinforcement bar is: Perform a fusion operation on the coordinates of each main reinforcement bar separately;

[0118] A4. Reconstruct the fused data to obtain the final measured data used to control the reinforcing steel components. For the data of the eight control points, the reconstructed coordinates are as follows: , , , For the main reinforcement coordinates, the restored coordinates are: .

[0119] These restored coordinate data are extracted as the final measured data for subsequent analysis and processing.

[0120] Example 3

[0121] Further explanation in conjunction with Example 1, such as Figure 1 The structure shown, the specific steps in step S3 are as follows:

[0122] S31, Theoretical Model The coordinates of the control points are The coordinates of the corresponding control points in the real-time model are , , Number of control points;

[0123] Calculate the spatial deviation vector for each control point ;

[0124] Extract the magnitude of the deviation vector As a characteristic of the magnitude of the deviation;

[0125] The cosine of the angle between the deviation vector and the coordinate axis is calculated as the directional feature; Taking the shaft as an example, Similarly, we can obtain the following: Cosine value of the included angle and with Cosine value of the included angle ;

[0126] Among them, the deviation vector formula This directly reflects the spatial difference between theoretical points and real-time points, and serves as the basis for subsequent analysis of biases;

[0127] Deviation modulus formula Based on the Euclidean distance formula, it is used to quantify the magnitude of the deviation;

[0128] The formula for direction cosine is as follows: Used to describe the direction of the deviation vector relative to the coordinate axes, providing a basis for subsequent analysis that considers directional factors;

[0129] S32. Define the deviation visualization function ;set up ,in , , , To adjust the parameters;

[0130] according to The value maps the deviation to a range of colors or brightness in the visualization; when When the deviation is small, green indicates a smaller deviation; when When the deviation is moderate, yellow indicates a suitable deviation; when... When the deviation is large, red indicates a significant deviation.

[0131] The deviation visualization function formula maps the deviation vector to a value in the range [0, 1] using a non-linear function that incorporates factors of deviation magnitude and direction, facilitating visual classification based on a set threshold. Adjusting parameters... , , , The weight of the magnitude and direction of the deviation in the visualization mapping can be controlled.

[0132] S33. Collect deviation characteristic data of multiple historical segments. , , Number of historical segments;

[0133] The next segment of steel reinforcement component size prediction vector is The next segment size vector in the theoretical model is ;

[0134] Building a predictive model:

[0135] First, the deviation characteristic data is normalized; let the normalized deviation modulus be... The normalized direction cosine values ​​are respectively , , ;

[0136] The prediction model then takes the following form:

[0137] ;

[0138] ;

[0139] ;

[0140] in , , These are weighting coefficients. , , These are bias terms; these weighting coefficients and bias terms are learned by minimizing the prediction error function.

[0141] The deviation feature data normalization formula maps the deviation magnitude and direction cosine value to the [0, 1] interval, which facilitates subsequent model training and weight calculation;

[0142] The prediction model formula predicts the size of the next segment by taking into account the magnitude and direction of historical deviation feature data;

[0143] Among them, the weighting coefficient , , and bias terms , , Adjustments are made by learning from historical data to adapt to different deviation patterns and the characteristics of steel reinforcement components.

[0144] Example 3

[0145] Further explanation in conjunction with Example 1, such as Figure 1 Based on the structure shown, after comparing and analyzing the theoretical model and the real-time model, the measured point and theoretical point matching algorithm EOPA was first adopted, and then the virtual assembly alignment algorithm GPA was adopted to develop virtual pre-assembly software and carry out virtual matching analysis of single components and multiple components.

[0146] In the preferred scheme, the virtual assembly enhanced matching algorithm steps are as follows:

[0147] B1. Establishing the objective function, i.e., EOPA algorithm optimization: Let the theoretical point be... The actual measurement point is , , Number of control points;

[0148] Define a new distance metric function This function takes into account the relationship between the coordinate difference and the sum of squares of the coordinates themselves.

[0149] Establish the optimization objective function ,in To adjust the parameters;

[0150] Distance metric function By applying a special normalization process to the coordinate difference, the magnitude information of the coordinates themselves is taken into account, making the distance measurement more reasonable and comprehensive. This special form helps to more accurately measure the difference between theoretical points and measured points under coordinates of different scales.

[0151] Optimize objective function Built upon the exponential function and distance metric function, when The smaller the value, the larger the value of the exponential function. The smaller the value, the better the matching effect; adjust the parameter. The sensitivity of this matching degree can be controlled;

[0152] B2. Optimization of Virtual Assembly Alignment GPA Algorithm: The original control point matrix is... ( The target point matrix is ( );

[0153] Calculate the weighted covariance matrix ,in , and They are and The fundamental mean matrix, The points are calculated based on the distance metric function mentioned above. to the mean distance, To adjust the parameters;

[0154] For the weighted covariance matrix Perform singular value decomposition The optimal rotation matrix is ​​obtained. ;

[0155] Calculate the optimal translation vector ;

[0156] Weighted covariance matrix formula By introducing weights The contribution of control points in covariance calculation is adjusted based on their distance from the mean, making the algorithm pay more attention to control points that are far from the mean and improving the accuracy of location calculation.

[0157] Singular Value Decomposition Formula Used to find the optimal rotation matrix This matrix decomposition method obtains the optimal rotation relationship;

[0158] Optimal translation vector formula Used to calculate the amount of translation that makes the control point move from its original position to the target position;

[0159] B3. The optimal rotation matrix is ​​obtained from the above calculations. Translation vector Based on this, we further consider the consistency of direction between the measured point and the theoretical point; let the direction vectors of the measured point and the theoretical point be respectively... and ,in , , Similarly, for theoretical points;

[0160] Calculation of directional consistency index This indicates the degree of similarity in direction between adjacent points between measured and theoretical points;

[0161] For translation vector Adjustments are made, and the adjusted translation vector is... ,in These are weighting coefficients. It is the mean of the directional consistency index;

[0162] Finally, the translation matrix is ​​obtained. Translation matrix ,in It is the adjusted translation vector The resulting translation matrix can be represented as a (The homogeneous coordinate matrix form extends the three-dimensional translation vector to homogeneous coordinate space).

[0163] Directional Consistency Index Formula By calculating the ratio of the vector dot product to the vector magnitude, the similarity of the directions between the measured point and the theoretical point in adjacent points can be obtained, providing a basis for adjusting the translation vector;

[0164] Translation vector adjustment formula The translation vector is adjusted based on the differences in the directional consistency index, so that the translated control points better conform to the requirements of the theoretical points in terms of direction; this is achieved by introducing weighting coefficients. and mean It allows for flexible control over the degree and direction of adjustment.

[0165] In the preferred scheme, based on the above results, multi-condition comparative analysis, component accuracy analysis, and virtual assembly accuracy analysis are carried out.

[0166] C1. Multi-condition comparative analysis:

[0167] Define the working condition similarity function Used to measure two working conditions and The degree of similarity between them; ,in To adjust the parameters;

[0168] The working conditions are classified according to the value of the working condition similarity function; when At that time, the working condition was considered to be and Similar; when At that time, the working condition was considered to be and There are certain similarities; when At that time, the working condition was considered to be and The differences are significant;

[0169] C2. Calculate component precision indicators , ,in To adjust the parameters;

[0170] Evaluate component precision based on component precision index values;

[0171] when At that time, it was believed that the parts had high precision; when At that time, the precision of the parts was considered to be moderate; when At that time, it was believed that the parts had low precision.

[0172] The component accuracy index formula is obtained by exponentially weighting the deviations of each control point under each working condition, and then averaging the sums to obtain a comprehensive accuracy index; adjustment parameters The weight of the deviation magnitude in the accuracy index calculation can be controlled;

[0173] C3. The coordinates of the control points after virtual assembly are: Calculate the virtual assembly deviation vector ;

[0174] Extract the magnitude of the virtual assembly deviation vector As a characteristic of the magnitude of virtual assembly deviation;

[0175] Calculate virtual assembly accuracy index ,in To adjust the parameters;

[0176] The virtual assembly accuracy is evaluated based on the value of the virtual assembly accuracy index; when At the time, it was believed that virtual assembly had high precision; when At that time, the virtual assembly accuracy was considered to be moderate; when At that time, it was believed that virtual assembly had low precision;

[0177] Virtual assembly deviation vector formula This reflects the deviation between the control points after virtual assembly and the original coordinates, and is the basis for subsequent analysis of the accuracy of virtual assembly;

[0178] Virtual assembly deviation module formula Based on the Euclidean distance formula, it is used to quantify the magnitude of virtual assembly deviation.

[0179] The above embodiments are merely preferred technical solutions of the present invention and should not be considered as limitations on the present invention. The scope of protection of the present invention should be limited to the technical solutions described in the claims, including equivalent substitutions of the technical features described in the claims. That is, equivalent substitutions and improvements within this scope are also within the scope of protection of the present invention.

Claims

1. A method for virtual pre-erection of a bridge tower reinforcement part, characterized in that: The method comprises: S1, using BIM tools to analyze and model the profile point coordinates of each segment of the steel part, forming theoretical coordinate point data, and considering the current actual construction stage monitoring pre-offset value, concrete pouring deformation influence factor, forming segment theoretical data model; S2, build a real-time data model by collecting steel part forming precision information through total station, three-dimensional laser and image vision technology, and import it into the digital system to build a real-time data model; S3, real-time comparative analysis, by comparing and analyzing the theoretical model and real-time model, realizing deviation visualization, and predicting the size data of the next segment of steel part through special algorithm, guiding the forming process of the next segment of steel part, realizing dynamic adjustment of production data; Calculate the spatial deviation vector of each control point, extract the length of the deviation vector as the deviation size feature, and calculate the cosine value of the angle between the deviation vector and the coordinate axis as the direction feature; Defining a bias visualization function , according to the value of the bias to a color or luminance range of the visualization; S33: Constructing a prediction model, collecting multiple sets of historical segment deviation feature data , , is the number of historical segments The next segment steel reinforcement part size prediction vector is , the next segment size vector under the theoretical model is ; The prediction model is as follows: First, the deviation feature data is normalized. Let the normalized deviation module be , the normalized direction cosine value be , , ; The deviation feature data normalization formula maps the deviation length and direction cosine value to the [0, 1] interval, which is convenient for subsequent model training and weight calculation; ； ； ； wherein , , are weight coefficients, , , are bias terms; these weight coefficients and bias terms are learned by minimizing a prediction error function; The prediction model formula predicts the size of the next segment by considering the size and direction of the historical deviation feature data; S4, develop a set of tower column steel part virtual pre-assembly software, integrate all the above functions, and realize digital virtual construction. wherein the weight coefficient , , and the bias term , , is adjusted by learning historical data to adapt to different bias patterns and characteristics of the reinforcing parts; Theoretical point data acquisition includes:

2. The method for virtual pre-assembly of a bridge tower reinforcement part according to claim 1, characterized in that: Monitoring pre-offset value: considering the tower limb line type, the pre-offset value of the control point is obtained by simulation analysis through MIDAS / CIVIL; Concrete pouring deformation: the deformation value is obtained by measuring the coordinates of the control points before and after pouring through the total station; Steel part hoisting deformation: the deformation value of the steel part after hoisting is obtained by simulation analysis using MIDAS / CIVIL; Adjustment value: used to guide the construction of the next segment of steel part. The data includes:

3. The virtual pre-assembly method for bridge tower steel reinforcement components according to claim 1, characterized in that: actual measurement... The measured coordinates of the eight control points for controlling the overall line shape and attitude of the steel part are obtained by total station; The coordinates of the main reinforcement for controlling the butt joint of the steel part are obtained by scanner or unmanned aerial vehicle; After the total station, scanner and unmanned aerial vehicle acquire data, the measured data is obtained by processing the measured coordinates of the eight control points for controlling the steel part and the coordinates of the main reinforcement for butt joint of the steel part. The specific steps for data acquisition are as follows:

4. The method for virtual pre-assembling of bridge tower reinforcement parts according to claim 3, characterized in that the measured data Extract these restored coordinate data as the final measured data for subsequent analysis and processing. A1, the eight control point measured coordinate data obtained by the total station and the main reinforcement coordinate data obtained by the scanner or the unmanned aerial vehicle are respectively standardized, for the data obtained by the total station, the eight control point coordinates are , the mean value of each coordinate component is calculated , , , then each coordinate is standardized, and the standardized coordinate is , , , ; For the coordinate data of main reinforcement obtained by scanner or drone, the single main reinforcement coordinate is , the distance from the coordinate origin is calculated , and then normalized, and the normalized coordinate is ; A2, calculate the weight of total station data and scanner / drone data in fusion, the weight of total station data is , the weight of scanner / drone data is , and ; The formula for calculating the weight is , where , , is an adjustment parameter, adjusted according to experience or actual data; A3, fuse the standardized total station data and the scanner / drone data, and the coordinate data of a single main reinforcement after fusion is respectively fuse each main reinforcement coordinate; A4, the fused data is restored to obtain the final measured data for controlling the steel part, and the coordinates of the eight control points after restoration are , , , , for the main reinforcement coordinates, the coordinates after restoration are ; The specific steps in step S3 are as follows:

5. The method for virtual pre-assembly of bridge tower reinforcement parts according to claim 1, characterized in that: Calculate the cosine value of the angle between the deviation vector and the coordinate axis as the direction feature; S31, the coordinates of the control point in the theoretical model are , , , , is the number of control points; calculating a spatial deviation vector for each control point ; extracting the magnitude of the bias vector as a bias size feature; After comparing and analyzing the theoretical model and real-time model, the virtual pre-assembly software is developed by adopting the measured point and theoretical point matching algorithm EOPA and the virtual assembly alignment algorithm GPA, and single component and multi-component virtual matching analysis is carried out. With axis as an example, , similarly, the cosine value of the angle between the axis and the axis and the cosine value of the angle between the axis and the axis ;​​ Wherein, the bias vector formula Directly reflects the theoretical point and real-time point in the position difference in space, is the basis for subsequent analysis of deviation; Deviation module length formula Based on the Euclidean distance formula, for quantifying the size of the deviation; The direction cosine formula is as follows The direction cosine formula is as follows The direction cosine formula is as follows S32, defining a deviation visualization function ; set wherein , , , is an adjustment parameter; According to the values will be mapped to a color or luminance range of the visualization.

6. The method for virtual pre-assembling of a bridge tower reinforcement part according to claim 5, characterized in that: The steps of the virtual assembly enhanced matching algorithm are as follows:

7. The method for virtual pre-assembling of a bridge tower reinforcement part according to claim 6, characterized in that: Based on the above results, multi-condition comparative analysis, part precision analysis and virtual assembly precision analysis are carried out. B1, optimization of the objective function is established, that is, EOPA algorithm optimization: set the theoretical point as , the measured point as , , is the number of control points; A new distance metric function is defined which takes into account the relationship between the coordinate difference and the square of the coordinate itself. establishing an optimization objective function wherein is a tuning parameter; Distance metric function By a special normalization processing on the coordinate difference, the size information of the coordinate itself is considered, making the distance metric more reasonable and comprehensive; this special form helps to more accurately measure the difference between the theoretical point and the measured point under different scales of coordinates; Optimization objective function Based on the exponential function and distance metric function construction, when The smaller, the greater the value of the exponential function, The smaller, the better the matching effect; adjust the parameter Can control the sensitivity of this matching degree; B2, virtual assembly alignment GPA algorithm optimization: original control point matrix is ( ), target point matrix is ( ); Computing a weighted covariance matrix where , and are the fundamental mean matrices of and respectively, is the distance of a point to the mean computed based on the distance metric function described above, is a tuning parameter; For the weighted covariance matrix Perform singular value decomposition The optimal rotation matrix is ​​obtained. ; Computing optimal translation vector ; Weighted covariance matrix formula By introducing weights The distance of the control points from the mean is used to adjust their contribution to the covariance calculation, making the algorithm more sensitive to control points that are far from the mean, improving the accuracy of the position calculation; singular value decomposition formula for finding the optimal rotation matrix the matrix decomposition method to obtain the optimal rotation relationship; Optimal translation vector formula for calculating the amount of translation of the control points from the original position to the target position; B3、in the above calculation of the optimal rotation matrix and translation vector , further consider the direction consistency of the measured points and the theoretical points; let the measured point and the theoretical point direction vectors be and , wherein , , , and the same for the theoretical point; Computing a directional consistency index , which represents the similarity of the directions of the measured points and the theoretical points between adjacent points; to the translation vector is adjusted, the adjusted translation vector wherein is a weight coefficient, is the mean of the direction consistency indicator; The translation matrix is finally obtained and the translation matrix where is a translation matrix (which can be represented as a 4x4 homogeneous coordinate matrix that extends the three-dimensional translation vector to the homogeneous coordinate space) composed of the adjusted translation vector ​​ Direction consistency index formula By calculating the ratio of the vector dot product and the vector length, the similarity of the measured point and the theoretical point in the direction between adjacent points is obtained, which provides a basis for adjusting the translation vector. Translation vector adjustment formula The translation vector is adjusted according to the difference of the direction consistency index, so that the control points after translation are more consistent with the theoretical points in direction; by introducing weight coefficients and mean value , the degree and direction of adjustment can be flexibly controlled.

8. The method for virtual pre-assembling of a bridge tower reinforcement part according to claim 7, characterized in that:

9. The virtual pre-assembly method of the bridge tower steel part according to claim 8, characterized in that: C1, multi-condition comparative analysis: ​ defining a working condition similarity function for measuring the degree of similarity between two working conditions and ; wherein is an adjustment parameter; The working conditions are classified according to the value of the working condition similarity function; when At that time, the working condition was considered to be and Similar; when At that time, the working condition was considered to be and There are certain similarities; when At that time, the working condition was considered to be and The differences are significant; C2, a component precision index , wherein is a tuning parameter; The part accuracy is evaluated based on the value of the part accuracy index. When the part precision is high; when the part precision is medium; and when the part precision is low. The part precision index formula is obtained by exponentially weighting and summing the deviation of each control point under each working condition, and then averaging to obtain a comprehensive precision index; the adjustment parameters The weight of the deviation size in the precision index calculation can be controlled; C3, the coordinates of the virtual assembly control points are , the virtual assembly deviation vector is calculated ; extracting a length of the virtual assembly deviation vector as a virtual assembly deviation size feature Computing virtual assembly precision indicators wherein is a tuning parameter; The virtual assembly accuracy is evaluated based on the value of the virtual assembly accuracy index; when At the time, it was believed that virtual assembly had high precision; when At that time, the virtual assembly accuracy was considered to be moderate; when At that time, it was believed that virtual assembly had low precision; Virtual assembly deviation vector formula Reflects the deviation of control points after virtual assembly and original coordinates, which is the basis for subsequent analysis of virtual assembly accuracy; Formula for calculating the length of a virtual assembly deviation Based on the Euclidean distance formula, for quantifying the size of a virtual assembly deviation.

Images