Method for interpenetrating construction positioning of oversized steel column

Abstract

The invention provides a super-large steel column insertion construction positioning method, and belongs to the technical field of airport constructions.The method includes the steps that three-dimensional laser scanning is conducted on a super-large steel column to obtain an initial geometrical shape, a multi-field coupling finite element model is established to calculate a comprehensive deformation predicted value, and a reverse compensation function is used for determining the hoisting pre-bias amount; a hoisting path optimization game model is adopted to determine hoisting speed control parameters and measurement section arrangement spacing, real-time three-dimensional coordinate acquisition is performed on each measurement section positioning reference point in the hoisting process, spatial position deviation calculation is performed on a segmented actual measurement coordinate set and a design coordinate set, and hoisting equipment is dynamically adjusted; and when all measurement sections meet the positioning precision threshold value and the angle precision threshold value, steel column fixing and welding are completed, and the technical problem that multi-field coupling deformation is difficult to accurately predict and compensate in real time in the hoisting and positioning process of the ultra-large steel column is solved.

Description

Technical Field

[0001] This invention belongs to the field of airport construction technology, and more specifically, it relates to a method for positioning and interlocking ultra-large steel columns during construction. Background Technology

[0002] In the construction of super high-rise buildings and large-span structures, the hoisting and positioning of ultra-large steel columns exceeding 30 meters in length is a critical process. Traditional steel column positioning methods mainly rely on empirical formulas to estimate self-weight deformation, use static measurement methods for position verification, and manually adjust the hoisting equipment to achieve positioning correction. This type of method has certain applicability in conventional steel column construction and has been widely used in engineering fields such as industrial plants and bridge towers. However, during the hoisting of ultra-large steel columns, the steel column is simultaneously subjected to the coupled effects of self-weight load, ambient temperature changes, and welding thermal stress. Traditional methods only consider a single deformation factor and cannot accurately predict the comprehensive deformation under multi-field coupling, resulting in a large error in the pre-offset calculation. At the same time, traditional static measurement methods are difficult to capture real-time position deviations during the dynamic hoisting process, and cannot achieve continuous dynamic correction, resulting in insufficient final positioning accuracy. In other words, existing technologies suffer from the technical problem of accurately predicting and compensating for multi-field coupled deformation during the hoisting and positioning of ultra-large steel columns. Summary of the Invention

[0003] In view of this, the present invention provides a method for positioning ultra-large steel columns during interlocking construction, which can solve the technical problem in the prior art that it is difficult to accurately predict and compensate for multi-field coupled deformation during the hoisting and positioning of ultra-large steel columns.

[0004] This invention is implemented as follows: A method for positioning ultra-large steel columns during interlocking construction includes: performing three-dimensional laser scanning on the ultra-large steel column to obtain its initial geometric shape data; marking positioning reference points every 5m along the axial direction on the column surface; establishing a multi-field coupled finite element model of the ultra-large steel column; inputting the column's self-weight load, ambient temperature parameters, and weld joint positions; calculating the column's self-weight elastic deformation, temperature deformation, and welding shrinkage deformation under hoisting conditions to obtain a comprehensive deformation prediction value; inputting the comprehensive deformation prediction value, the measured stiffness coefficient of the steel column, and the ambient temperature difference into a reverse compensation function to calculate the hoisting pre-offset; establishing a hoisting path optimization game model; calculating the hoisting speed control parameters through the upper-level model; and using the lower-level model... The spacing of the measurement sections is calculated. During the hoisting of the super-large steel column, multiple measurement sections are arranged along the axial direction of the steel column according to the spacing of the measurement sections. The positioning reference points of each measurement section are collected in real time to obtain the actual measured coordinate set of the steel column segment. The spatial position deviation of the steel column segment measured coordinate set and the design coordinate set is calculated. When the spatial position deviation of any measurement section is greater than the deviation judgment threshold, the horizontal displacement and vertical lifting amount of the hoisting equipment are adjusted according to the deviation direction and deviation amount to achieve dynamic correction. When the spatial position deviation of all measurement sections is less than the positioning accuracy threshold and the axial angle deviation of adjacent measurement sections is less than the angle accuracy threshold, the super-large steel column is fixed and welded to complete the positioning of the super-large steel column through construction.

[0005] The multi-field coupled finite element model is a simulation model of steel column deformation that considers the interaction of self-weight field, temperature field and welding stress field.

[0006] In establishing the multi-field coupled finite element model, the steel column is divided into element meshes with a length of no more than 1m. Gravity acceleration load, temperature gradient load and welding line heat source load are applied to each element mesh, and the displacement field distribution of each node of the steel column is obtained through iterative calculation.

[0007] Wherein, the self-weight elastic deformation is the vector sum of the longitudinal tensile deformation and the transverse bending deformation of the steel column under the action of gravity; the temperature deformation is the linear deformation caused by thermal expansion and contraction due to the difference between the ambient temperature and the manufacturing temperature of the steel column; and the welding shrinkage deformation is the local deformation of the steel column caused by the cooling shrinkage of the weld area during the welding process.

[0008] The measured stiffness coefficient of the steel column is the ratio of the bending stiffness per unit length of the steel column to the theoretical bending stiffness, as measured by a small-scale model experiment.

[0009] In determining the measured stiffness coefficient of the steel column, a scaled-down model of the steel column with a geometric scale of 1:10 is made. A distributed load is applied to the scaled-down model to measure the deflection value of the scaled-down model. The deflection value is compared with the theoretical deflection value to obtain the stiffness correction coefficient. The stiffness correction coefficient is substituted into the numerical analysis model for full-size simulation verification. The weighted average of the two calculation results is used as the measured stiffness coefficient of the steel column.

[0010] The reverse compensation function is used to calculate the spatial position pre-offset amount required during hoisting based on the steel column deformation prediction parameters. The inputs include the comprehensive deformation prediction value, the measured stiffness coefficient of the steel column, and the ambient temperature difference. The output is the hoisting pre-offset amount.

[0011] The deviation judgment threshold, positioning accuracy threshold, and angle accuracy threshold are all determined by combining small-scale model experiments with numerical analysis.

[0012] Among them, when determining the deviation judgment threshold, a hoisting simulation experiment was carried out on the 1:10 scale model. A data jitter model was established based on physical constraints. The data jitter model was used to simulate the random fluctuations caused by environmental vibration, the accuracy of measuring instruments and human operation errors in the measurement data of the scale model.

[0013] In the data jitter model, the scaling factor is considered. The scaling factor is the amplification or attenuation coefficient of the data jitter characteristics by the scaling ratio. The scaling factor is applied to the jitter amplitude range to obtain the jitter amplitude after scaling correction.

[0014] The jitter data will form a jitter data distribution cloud after multiple measurements. Statistical analysis is performed on the jitter data distribution cloud to calculate the center coordinates, dispersion, and skewness of the jitter data distribution cloud. The center coordinates are used as the representative value of the experimental measurement, and the dispersion is multiplied by the confidence coefficient to obtain the uncertainty range of the experimental measurement.

[0015] Among them, an improved two-parameter Gaussian kernel function filter is applied to the jitter data distribution cloud. The improved two-parameter Gaussian kernel function is used to perform weighted smoothing of the measurement data and suppress the influence of abnormal fluctuation points.

[0016] Specifically, the experimental deviation threshold processed by the improved two-parameter Gaussian kernel function and the simulation deviation threshold processed by the improved two-parameter Gaussian kernel function are fused together for calculation. The fusion calculation adopts the Bayesian weighted fusion method, and the fusion result is used as the deviation judgment threshold.

[0017] The hoisting path optimization game model includes an upper-level model and a lower-level model. The upper-level model takes the stability of the steel column posture as the optimization objective, while the lower-level model takes the positioning accuracy as the optimization objective. The two models are coupled through hoisting speed control parameters.

[0018] The spatial position deviation is the Euclidean distance between the measured three-dimensional coordinates and the designed three-dimensional coordinates of the measurement section positioning reference point, and the dynamic correction is the process of continuously fine-tuning the hoisting equipment based on real-time measurement data.

[0019] The deviation of the axial angle is the absolute value of the difference between the angle between the axial direction vectors of two adjacent measurement sections and the theoretical angle. First, the axial direction vector is fitted by the coordinates of the positioning reference points of the adjacent measurement sections. Then, the angle is calculated using the vector dot product formula. Finally, it is compared with the theoretical angle in the design drawings.

[0020] This invention establishes a multi-field coupled finite element model considering the interaction of self-weight field, temperature field, and welding stress field. Combining the measured stiffness coefficient of the steel column and the ambient temperature difference, a reverse compensation function is used to calculate the pre-offset during hoisting, achieving accurate prediction of multi-field coupled deformation. By establishing a hoisting path optimization game model consisting of an upper-level model targeting the steel column's attitude stability and a lower-level model targeting positioning accuracy, optimal hoisting speed control parameters and measurement section spacing are determined. During hoisting, real-time three-dimensional coordinate acquisition of each positioning reference point is performed according to the optimized measurement section spacing. When the spatial position deviation exceeds a threshold, the hoisting equipment is immediately adjusted for dynamic correction. This solves the shortcomings of traditional methods that only consider a single deformation factor, leading to large prediction errors, and the inability of static measurement methods to correct in real time. In summary, this invention solves the technical problem mentioned in the background art of the difficulty in accurately predicting and real-time compensating for multi-field coupled deformation during the hoisting and positioning of ultra-large steel columns. Attached Figure Description

[0021] Figure 1 This is a flowchart of the method of the present invention.

[0022] Figure 2 This is a schematic diagram of an ultra-large steel column structure.

[0023] Figure 3 This is a deformation cloud diagram of the finite element model of the steel column.

[0024] Figure 4 This diagram illustrates the relationship between the posture stability and positioning accuracy of the steel column during the hoisting process. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0026] like Figure 1 The diagram shows a flowchart of a method for positioning and interlocking ultra-large steel columns provided by this invention. This method includes the following steps:

[0027] S01. Perform three-dimensional laser scanning on the super-large steel column with a length exceeding 30m to obtain the initial geometric shape data of the steel column, and mark the positioning reference points every 5m along the axial direction on the surface of the steel column.

[0028] S02. Establish a multi-field coupled finite element model of the super-large steel column, input the self-weight load of the steel column, the ambient temperature parameters, and the location of the welding nodes, calculate the self-weight elastic deformation, temperature deformation and welding shrinkage deformation of the steel column under hoisting conditions, and obtain the comprehensive deformation prediction value.

[0029] S03. Input the comprehensive deformation prediction value, the measured stiffness coefficient of the steel column, and the ambient temperature difference into the reverse compensation function to calculate the hoisting pre-offset amount. The hoisting pre-offset amount is the spatial offset distance and angle of the steel column relative to the design position during hoisting.

[0030] S04. Establish a hoisting path optimization game model, which includes an upper-level model with the steel column attitude stability as the objective and a lower-level model with the positioning accuracy as the objective. The hoisting speed control parameters are calculated through the upper-level model, and the measurement section layout spacing is calculated through the lower-level model.

[0031] S05. During the hoisting of the super-large steel column, multiple measurement sections are arranged along the axial direction of the steel column according to the spacing of the measurement sections. The positioning reference points of each measurement section are collected in real time in three dimensions to obtain the measured coordinate set of the steel column segments.

[0032] S06. Calculate the spatial position deviation between the measured coordinate set and the design coordinate set of the steel column segments. When the spatial position deviation of any measured section is greater than the deviation judgment threshold, adjust the horizontal displacement and vertical lifting of the hoisting equipment according to the deviation direction and deviation amount to achieve dynamic correction.

[0033] S07. When the spatial position deviation of all measurement sections is less than the positioning accuracy threshold and the axial angle deviation of adjacent measurement sections is less than the angle accuracy threshold, fix the super-large steel column and weld it to complete the super-large steel column interlacing construction positioning.

[0034] The multi-field coupled finite element model is a simulation model of steel column deformation considering the interaction of self-weight field, temperature field, and welding stress field. When establishing the multi-field coupled finite element model, the steel column is divided into element meshes with a length not exceeding 1m. Gravitational acceleration load, temperature gradient load, and welding line heat source load are applied to each element mesh, and the displacement field distribution of each node of the steel column is obtained through iterative calculation. The self-weight elastic deformation is the vector sum of the longitudinal tensile deformation and transverse bending deformation of the steel column under gravity. The temperature deformation is the linear deformation caused by thermal expansion and contraction due to the difference between the ambient temperature and the manufacturing temperature of the steel column. The welding shrinkage deformation is the local deformation of the steel column caused by the cooling shrinkage of the weld area during the welding process.

[0035] The measured stiffness coefficient of the steel column is the ratio of the bending stiffness per unit length of the steel column to the theoretical bending stiffness, obtained through small-scale model experiments. In determining the measured stiffness coefficient, a 1:10 scale model of the steel column is constructed. A distributed load is applied to the scaled model, and the deflection value is measured. The deflection value is compared with the theoretical deflection value to obtain a stiffness correction coefficient. This stiffness correction coefficient is then substituted into a numerical analysis model for full-scale simulation verification. The weighted average of the two calculations is taken as the measured stiffness coefficient of the steel column. In the weighted average, the weighting coefficient for the small-scale model experimental results is 0.6, and the weighting coefficient for the numerical analysis results is 0.4.

[0036] The reverse compensation function is used to calculate the required spatial offset during hoisting based on the predicted deformation parameters of the steel column. The inputs include the comprehensive deformation prediction value, the measured stiffness coefficient of the steel column, and the ambient temperature difference. The output is the hoisting pre-offset. The reverse compensation function is expressed as follows: the hoisting pre-offset divided by the comprehensive deformation prediction value equals the measured stiffness coefficient of the steel column multiplied by the ambient temperature difference divided by the standard temperature difference, plus a reference correction coefficient. The standard temperature difference is taken as 20℃, and the reference correction coefficient is taken as 1.0. The ambient temperature difference is the difference between the measured temperature at the construction site and the ambient temperature during the manufacture of the steel column.

[0037] The deviation judgment threshold, the positioning accuracy threshold, and the angle accuracy threshold are all determined through a combination of small-scale model experiments and numerical analysis. When determining the deviation judgment threshold, a hoisting simulation experiment is first conducted on a 1:10 scale model. During the experiment, a data jitter model is established based on physical constraints, including the material elastic modulus constraint, geometric similarity constraint, and gravity field scaling constraint of the scale model. The data jitter model is used to simulate the random fluctuations in the scale model's measurement data caused by environmental vibration, instrument accuracy, and human error. The amplitude range of the data jitter is determined by the standard deviation of multiple repeated measurements on the scale model. When establishing the data jitter model, the original measurement data sequence of the scaled model is input into the jitter amplitude calculation function. The jitter amplitude calculation function is used to calculate the fluctuation range of the measurement data. The input includes the original measurement data sequence, the number of repeated measurements, and the accuracy level of the measuring instrument. The output is the jitter amplitude range. The jitter amplitude calculation function is expressed as the jitter amplitude range divided by the standard jitter amplitude equals the standard deviation of the original measurement data sequence divided by the mean of the original measurement data sequence multiplied by the square root of the number of repeated measurements divided by the square root of the standard number of measurements divided by the accuracy level of the measuring instrument divided by the standard accuracy level. The standard jitter amplitude is 0.1 mm, the standard number of measurements is 10, the standard accuracy level is 0.01 mm, and the number of repeated measurements ∈ [8, 15].

[0038] The data jitter model considers a scaling factor, which is the amplification or attenuation coefficient of the scaling ratio on the data jitter characteristics. The scaling factor is calculated by multiplying the reciprocal of the scaling ratio by a material property correction coefficient, and then by a boundary condition similarity coefficient. The material property correction coefficient is the ratio of the material density of the scaled model to the material density of the full-size steel column. The boundary condition similarity coefficient is determined by comparing the support stiffness ratio of the scaled model and the full-size steel column. The scaling factor is applied to the jitter amplitude range to obtain the scaled-corrected jitter amplitude, which is equal to the jitter amplitude range multiplied by the scaling factor. The scaling factor has a value range of [8.5, 11.5].

[0039] After multiple measurements, the data jitter forms a jitter data distribution cloud, which is the set of all measured data points distributed in three-dimensional coordinate space. Statistical analysis is performed on this jitter data distribution cloud to calculate its center coordinates, dispersion, and skewness. The center coordinates are the arithmetic mean of the coordinates of all data points; the dispersion is the standard deviation of the distances from all data points to the center coordinates; and the skewness is the degree of asymmetry in the data point distribution relative to the center coordinates. Using the center coordinates as a representative value of the experimental measurement, the dispersion is multiplied by a confidence coefficient to obtain the uncertainty range of the experimental measurement. The confidence coefficient is set to 1.96, corresponding to a 95% confidence level.

[0040] Record the final positioning error of the steel column corresponding to different deviation values, establish a deviation-error relationship curve, and fit the deviation-error relationship curve to obtain the experimental deviation threshold. Then, establish a full-size numerical simulation model, input the same boundary conditions as the experiment into the full-size numerical simulation model, and introduce a numerical jitter disturbance term into the numerical calculation. The numerical jitter disturbance term is the jitter amplitude obtained from the scaled-down model experiment after proportional correction and magnified 10 times according to the scaled-down ratio. The numerical jitter disturbance term is superimposed into the displacement calculation result of the numerical simulation to simulate the measurement uncertainty of the full-size steel column in the actual construction environment. 1000 random jitter simulation calculations are performed using the Monte Carlo method to obtain the probability distribution interval of the simulation deviation threshold.

[0041] To improve the reliability of threshold acquisition, the jittery data distribution cloud is first subjected to an improved two-parameter Gaussian kernel function filter. The improved two-parameter Gaussian kernel function is used to weight and smooth the measurement data and suppress the influence of abnormal fluctuation points. The inputs include the distance from the data point to the center coordinate, the dispersion deviation of the data point, and the kernel function bandwidth parameter. The output is the kernel function weight value of the data point. The improved two-parameter Gaussian kernel function is expressed as follows: the kernel function weight value of the data point divided by the baseline weight value is equal to the negative half power of the natural constant multiplied by the distance from the data point to the center coordinate divided by the square of the kernel function bandwidth parameter, plus the exponent of the dispersion deviation of the data point divided by the square of the dispersion adjustment parameter. The baseline weight value is 1.0. The kernel function bandwidth parameter is determined based on the overall dispersion of the jitter data distribution cloud, and is equal to the dispersion of the jitter data distribution cloud multiplied by the bandwidth adjustment coefficient, which is 1.5. The dispersion adjustment parameter is equal to the dispersion of the jitter data distribution cloud multiplied by 0.8. The dispersion deviation of the data point is the absolute value of the difference between the distance from a single data point to the center coordinate and the dispersion of the jitter data distribution cloud.

[0042] When applying the improved two-parameter Gaussian kernel function, the kernel function weight value is calculated for each data point in the jittery data distribution cloud. All data points are then weighted and averaged according to their kernel function weight values ​​to obtain the filtered center value. This filtered center value replaces the original center coordinates as the final representative value of the experimental measurement. Compared to the traditional single-parameter Gaussian kernel function, the improved two-parameter Gaussian kernel function adds a dispersion deviation constraint term. This dispersion deviation constraint term is used to identify and reduce the weight of abnormal measurement points that deviate from the main body of the data distribution. These abnormal measurement points are data points with large deviations caused by sudden environmental interference or instantaneous instrument errors. Through the filtering process of the improved two-parameter Gaussian kernel function, the measurement error caused by random disturbances in the experimental data can be reduced by 30% to 45%, thus reducing the dispersion of the experimental deviation threshold and improving the repeatability of threshold acquisition.

[0043] When processing the numerical simulation results, the improved two-parameter Gaussian kernel function is used to estimate the kernel density of the simulation deviation threshold probability distribution obtained from 1000 Monte Carlo simulations. This kernel density estimation is used to extract a continuous probability density function from the discrete simulation results. During kernel density estimation, each simulation result is used as the center of the kernel function, and its contribution to the overall probability density is calculated using the improved two-parameter Gaussian kernel function. All 1000 kernel functions are superimposed to obtain a smooth probability density curve for the simulation deviation threshold. The peak position of the probability density is extracted from this smooth probability density curve as the optimal estimate of the simulation deviation threshold, and a 95% confidence interval is extracted as the uncertainty range of the simulation deviation threshold.

[0044] After kernel function filtering, the experimental deviation threshold processed by the improved two-parameter Gaussian kernel function and the simulation deviation threshold processed by the improved two-parameter Gaussian kernel function are fused together. The fusion calculation employs a Bayesian weighted fusion method, which dynamically determines the weight coefficients based on the confidence level of the experimental data and the convergence accuracy of the simulation data. The fusion calculation is performed by multiplying the experimental deviation threshold processed by the improved two-parameter Gaussian kernel function by the experimental data weight coefficient, and then adding the simulation deviation threshold processed by the improved two-parameter Gaussian kernel function by multiplying by the simulation data weight coefficient. The sum of the experimental data weight coefficient and the simulation data weight coefficient equals 1.0. The experimental data weight coefficient is set to 0.55, and the simulation data weight coefficient is set to 0.45. The fusion result serves as the deviation judgment threshold. By introducing the improved two-parameter Gaussian kernel function for filtering and kernel density estimation, the standard deviation of the deviation judgment threshold is reduced by 40% compared to the method without kernel function processing, and the consistency of threshold determination is improved by 35%. This reduces the fluctuation range of the threshold under different batches of experiments and different simulation conditions to 60% of the original method, significantly improving the stability of threshold acquisition and the repeatability of engineering applications. To further improve the stability of threshold acquisition, a sensitivity analysis is performed on the deviation judgment threshold. The support stiffness, material parameters, and ambient temperature of the scaled model are changed, and the change range of the deviation judgment threshold is observed. When the change range of the deviation judgment threshold is less than 5%, it is confirmed that the deviation judgment threshold has sufficient robustness. The value range of the deviation judgment threshold is [18mm, 22mm].

[0045] The positioning accuracy threshold is determined using the same method. When determining the positioning accuracy threshold, the dispersion weighting coefficient of the jitter data distribution cloud is adjusted to 0.65, and the number of Monte Carlo simulations in the numerical simulation is increased to 1500 to improve the accuracy of the threshold acquisition. The value range of the positioning accuracy threshold is [12mm, 16mm]. The angle accuracy threshold is determined using the same method. When determining the angle accuracy threshold, an angle measurement error amplification factor is added to the calculation of the scale adjustment influence factor. The angle measurement error amplification factor is the 1.2 power of the scaling ratio to compensate for the relatively low angle measurement accuracy in the small-scale model. The value range of the angle accuracy threshold is [0.25°, 0.35°].

[0046] The hoisting path optimization game model includes an upper-level model and a lower-level model. The upper-level model optimizes the steel column's attitude stability, while the lower-level model optimizes positioning accuracy. The two models are coupled through hoisting speed control parameters. The objective function of the upper-level model is used to calculate the attitude stability index of the steel column during hoisting. The inputs include hoisting speed control parameters, steel column length, and steel column section moment of inertia. The output is the attitude stability index. The objective function of the upper-level model is expressed as: the attitude stability index divided by the baseline stability index equals the hoisting speed control parameter divided by the standard hoisting speed multiplied by the steel column length divided by the square root of the standard steel column length, multiplied by the steel column section moment of inertia divided by the cube root of the standard section moment of inertia. The baseline stability index is set to 1.0, the standard hoisting speed to 0.5 m / min, the standard steel column length to 30 m, and the standard section moment of inertia to... The optimization objective of the upper-level model is to maximize the attitude stability index. The constraints of the upper-level model are that the hoisting speed control parameter ∈ [0.3m / min, 0.8m / min] and the attitude stability index ≥ 0.85.

[0047] The objective function of the lower-level model is used to calculate the relationship between the positioning accuracy evaluation index and the measurement section layout spacing. The inputs include the measurement section layout spacing, hoisting speed control parameters, and the measured stiffness coefficient of the steel column. The output is the positioning accuracy evaluation index. The objective function of the lower-level model is expressed as: the positioning accuracy evaluation index divided by the baseline positioning accuracy equals the measurement section layout spacing divided by the standard measurement spacing multiplied by the hoisting speed control parameters divided by the negative half power of the standard hoisting speed, multiplied by the measured stiffness coefficient of the steel column. Here, the baseline positioning accuracy is taken as 1.0, the standard measurement spacing as 5m, and the standard hoisting speed as 0.5m / min. The optimization objective of the lower-level model is to maximize the positioning accuracy evaluation index. The constraints of the lower-level model are: the measurement section layout spacing ∈ [3m, 7m] and the positioning accuracy evaluation index ≥ 0.90.

[0048] The hoisting speed control parameter is a coupling term between the upper-level model and the lower-level model, and it simultaneously affects both the attitude stability index and the positioning accuracy evaluation index. When solving the hoisting path optimization game model, the upper-level model is first solved with the hoisting speed control parameter fixed to obtain the optimal attitude stability index. Then, the hoisting speed control parameter is substituted into the lower-level model to obtain the optimal measurement section spacing. The positioning accuracy evaluation index corresponding to the measurement section spacing is then fed back to the upper-level model. Iterative calculations continue until both the upper-level and lower-level models simultaneously satisfy their respective constraints and the objective function value no longer changes. The convergence criterion for the iterative calculation is that the change in attitude stability index between two adjacent iterations is less than 0.01 and the change in positioning accuracy evaluation index is less than 0.01.

[0049] The segmented measurement method involves dividing the massive steel column axially into multiple measurement sections and independently measuring the coordinates of each section. When using this method, the bottom measurement section is adjusted first during the initial stage of steel column hoisting, the middle measurement section is corrected during the middle stage, and the top measurement section is finally checked during the later stage, gradually approaching the target position.

[0050] The spatial position deviation is the Euclidean distance between the measured three-dimensional coordinates and the designed three-dimensional coordinates of the positioning reference point of the measurement section. The dynamic correction is a process of continuously fine-tuning the hoisting equipment based on real-time measurement data. The frequency of the dynamic correction is to perform a correction action immediately after the coordinate acquisition of each measurement section is completed.

[0051] The axial angle deviation is the absolute value of the difference between the angle between the axial direction vectors of two adjacent measurement sections and the theoretical angle. To calculate the axial angle deviation, the axial direction vector is first fitted using the coordinates of the positioning reference points of adjacent measurement sections. Then, the angle is calculated using the vector dot product formula. Finally, it is compared with the theoretical angle in the design drawings.

[0052] The attitude stability index is a parameter characterizing the steel column's ability to resist swaying and torsion during hoisting; a larger value indicates a more stable attitude. The positioning accuracy evaluation index is a comprehensive evaluation parameter characterizing the final positioning error of the steel column; a larger value indicates higher positioning accuracy.

[0053] The specific implementation methods of the above steps are described in detail below.

[0054] The specific implementation of step S01 is as follows: First, a 3D laser scanner is used to perform a full-surface scan of the super-large steel column with a length exceeding 30m. During the scan, the distance between the laser scanner and the surface of the steel column is maintained between 2m and 5m to ensure scanning accuracy. The point cloud density is set to no less than 10,000 data points per square meter. The original point cloud is then processed by point cloud data processing software to perform noise reduction, registration, and stitching to establish a complete 3D geometric model of the steel column and extract the initial geometric shape data of the steel column. At the same time, positioning reference points are marked every 5m along the axial direction on the surface of the steel column using a laser marker or mechanical dotting method. The marking depth of the positioning reference points is controlled between 0.5mm and 1.0mm to ensure that they will not disappear due to wear during the subsequent hoisting process. The purpose of this step is to obtain the true geometric shape of the steel column and establish a unified measurement reference system to provide initial data support for subsequent deformation prediction and position correction.

[0055] The specific implementation of step S02 is to establish a multi-field coupled numerical model of the ultra-large steel column based on the finite element analysis principle. This model simultaneously considers the interaction of three physical fields: self-weight field, temperature field, and welding stress field. The steel column is discretized into hexahedral or tetrahedral elements with a length not exceeding 1m using finite element mesh generation technology. Gravitational acceleration load is applied to each element node to simulate the self-weight effect of the steel column. Temperature gradient load is applied to simulate the thermal expansion or contraction effect caused by changes in ambient temperature. Welding line heat source load is applied to simulate the local heating and cooling contraction effect during the welding process. The displacement field distribution of each node of the steel column is obtained by multi-field coupled calculation through a nonlinear iterative solver. The self-weight elastic deformation includes the vector synthesis result of two components: longitudinal tensile deformation and transverse bending deformation. The temperature deformation is calculated based on the linear expansion coefficient of the material and the difference between the ambient temperature and the manufacturing temperature. The welding shrinkage deformation is obtained by simulating the local shrinkage displacement of the weld area through welding thermal cycle simulation. The three deformation quantities are superimposed to obtain the comprehensive deformation prediction value. The purpose of this step is to quantitatively predict the spatial deformation of the steel column under hoisting conditions caused by the combined action of multiple physical factors, and to provide a theoretical basis for formulating compensation strategies.

[0056] The specific implementation of step S03 involves inputting the comprehensive deformation prediction value, the measured stiffness coefficient of the steel column, and the ambient temperature difference into the reverse compensation function for calculation. The measured stiffness coefficient of the steel column is obtained by creating a scaled-down model with a geometric scale of 1:10 and applying distributed loads to measure the deflection value. The measured deflection is compared with the theoretical deflection to obtain the stiffness correction coefficient. This coefficient is then substituted into the full-size numerical model for simulation verification. The two results are weighted and averaged with weighting coefficients of 0.6 and 0.4 to obtain the final measured stiffness coefficient of the steel column. The reverse compensation function is established based on the deformation compensation principle, and its input parameters include the comprehensive deformation prediction value, the measured stiffness coefficient of the steel column, and the ambient temperature difference. The function combines the predicted deformation value, the measured stiffness coefficient of the steel column, and the ambient temperature difference, and outputs the pre-offset value for hoisting. This function counteracts the actual deformation during hoisting by applying the predicted deformation value in reverse to the initial position of the steel column. The standard temperature difference value in the function is taken as 20℃ as the reference value, and the reference correction coefficient is taken as 1.0 as the calibration constant. The calculated pre-offset value for hoisting represents the spatial distance and angle that the steel column needs to be offset from the design position in advance during hoisting. The purpose of this step is to formulate an accurate position compensation scheme based on the deformation prediction results so that the steel column can accurately reach the design position after deformation.

[0057] The specific implementation of step S04 involves establishing a two-layer game optimization model to coordinate the two mutually constraining optimization objectives of attitude stability and positioning accuracy during the hoisting process. The upper-layer model takes attitude stability as the optimization objective, and its objective function input includes hoisting speed control parameters, steel column length, and steel column section moment of inertia. The output is an attitude stability index, which comprehensively reflects the steel column's ability to resist swaying and torsion during hoisting. The baseline stability index is set to 1.0, the standard hoisting speed is 0.5 m / min, the standard steel column length is 30 m, and the standard section moment of inertia is... The upper-level model's constraints require a hoisting speed between 0.3 m / min and 0.8 m / min and an attitude stability index of no less than 0.85. The lower-level model optimizes positioning accuracy. Its objective function inputs include the spacing of the measurement sections, hoisting speed control parameters, and the measured stiffness coefficient of the steel columns. The output is a positioning accuracy evaluation index. The baseline positioning accuracy is set to 1.0, and the standard measurement spacing is 5 m. The lower-level model's constraints require a measurement section spacing between 3 m and 7 m and an positioning accuracy evaluation index of no less than 0.90. The two models are compared through hoisting... The speed control parameters are coupled together. During the solution process, the upper model is first solved with a fixed hoisting speed to obtain the optimal attitude stability. Then, this speed is substituted into the lower model to obtain the optimal spacing of the measurement sections. The positioning accuracy evaluation index is then fed back to the upper model for iterative calculation. When the change in attitude stability index between two adjacent iterations is less than 0.01 and the change in positioning accuracy evaluation index is less than 0.01, convergence is determined. The purpose of this step is to optimize the arrangement of the measurement sections while ensuring the stability of the steel column hoisting process, so as to achieve a balance between hoisting safety and positioning accuracy.

[0058] The specific implementation of step S05 involves setting multiple measurement sections along the axial direction of the steel column during the hoisting process, based on the measurement section spacing calculated in step S04. Each measurement section corresponds to a positioning reference point. High-precision measuring equipment such as a total station or laser tracker is used to collect the positioning reference points of each measurement section in real time in three dimensions. The measurement frequency is set to collect coordinate data at least once per minute for each measurement section. All the collected measurement section coordinate data are organized in axial position order to form a segmented measured coordinate set of the steel column. This coordinate set contains the real-time position information of each measurement section in three-dimensional space. During the measurement process, the measuring equipment needs to be calibrated periodically to eliminate instrument system errors. The purpose of this step is to obtain the spatial attitude information of the steel column in real time during the hoisting process through segmented measurement, providing accurate measurement data for subsequent deviation calculation and dynamic correction.

[0059] The specific implementation of step S06 involves comparing the measured coordinate set of the steel column segments with the design coordinate set point by point to calculate the spatial position deviation. The spatial position deviation is calculated using the Euclidean distance formula to determine the straight-line distance between the measured coordinates and the design coordinates. The deviation judgment threshold is determined by combining scaled model experiments and numerical analysis. First, a hoisting simulation experiment is conducted on a 1:10 scaled model, and a data jitter model is established to simulate the random fluctuations caused by environmental vibration, instrument accuracy, and human operation errors in the measurement data. The data jitter amplitude is determined by the standard deviation of multiple repeated measurements. A scaling adjustment influence factor is introduced to convert the jitter characteristics of the scaled model into the jitter characteristics of the full-size steel column. The scaling adjustment influence factor is calculated by multiplying the reciprocal of the scaling ratio by the material property correction coefficient and then by the boundary condition similarity coefficient, with a value range of 8.5 to 11.5. Multiple measurement data form a jitter data distribution cloud. This distribution cloud is filtered using an improved two-parameter Gaussian kernel function. Compared with the traditional single-parameter Gaussian kernel function, this kernel function adds a constraint term for dispersion deviation to identify and reduce the weight of abnormal measurement points. The kernel function bandwidth parameter is determined based on the overall dispersion of the jitter data distribution cloud. A bandwidth adjustment factor of 1.5 is used to determine the filtering process, which can reduce the measurement error by 30% to 45%. Then, a full-scale numerical simulation model is established, and a numerical jitter disturbance term is introduced. This disturbance term is the jitter amplitude obtained from the scaled-down model experiment after proportional correction, magnified 10 times. The probability distribution of the simulation deviation threshold is obtained by performing 1000 random jitter simulations using the Monte Carlo method. An improved two-parameter Gaussian kernel function is used to estimate the kernel density of the simulation results and extract a smooth probability density curve. The experimental deviation threshold processed by the kernel function is then compared with the simulation deviation threshold according to... A deviation judgment threshold is obtained by Bayesian weighted fusion using weight coefficients of 0.55 and 0.45. The value of this threshold ranges from 18mm to 22mm. When the spatial position deviation of any measurement section is greater than the deviation judgment threshold, the horizontal displacement and vertical lifting amount that the hoisting equipment needs to adjust are calculated according to the direction and magnitude of the deviation. The micro-position adjustment is achieved by controlling the hydraulic system or motor drive system of the hoisting equipment. The purpose of this step is to ensure that the steel column is always kept within the allowable deviation range during the hoisting process through real-time monitoring and dynamic correction, so as to avoid the cumulative error leading to the final positioning inaccuracy.

[0060] The specific implementation of step S07 involves continuously monitoring the spatial position deviation of all measurement sections and the axial angle deviation of adjacent measurement sections during the hoisting process. The positioning accuracy threshold is determined using the same method as the deviation judgment threshold, but the dispersion weighting coefficient of the jitter data distribution cloud is adjusted to 0.65, and the number of Monte Carlo simulations is increased to 1500 to improve the accuracy of threshold acquisition. The positioning accuracy threshold ranges from 12mm to 16mm. The axial angle deviation is calculated by fitting the axial direction vector to the positioning reference point coordinates of adjacent measurement sections, and using the vector dot product formula to calculate the absolute value of the difference between the actual angle and the theoretical angle. In the process of determining the angle accuracy threshold, the angle measurement is added to the calculation of the proportional adjustment influence factor. The measurement error amplification factor is calculated as 1.2 times the scale ratio to compensate for the relatively low angle measurement accuracy in small-scale models. The angle accuracy threshold ranges from 0.25° to 0.35°. When the spatial position deviation of all measurement sections is less than the positioning accuracy threshold and the axial angle deviation of adjacent measurement sections is less than the angle accuracy threshold, the steel column fixing procedure is initiated. The steel column is locked in its current position using temporary support structures or clamps. Then, welding is performed to firmly connect the steel column to the installed components, completing the insertion construction positioning of the ultra-large steel column. The purpose of this step is to perform final fixing after confirming that the position and posture of the steel column meet the design requirements, ensuring the installation quality and structural safety of the ultra-large steel column.

[0061] Specifically, the principle of this invention is as follows: The solution to this technical problem lies in the organic combination of multi-field coupling analysis and dynamic feedback control. First, by dividing the steel column into a grid of no more than 1 meter, and simultaneously applying gravitational acceleration load, temperature gradient load, and welding line heat source load to each unit, iterative calculations are used to obtain the displacement field distribution of each node. This reveals the coupling law of self-weight elastic deformation, temperature deformation, and welding shrinkage deformation at the physical mechanism level, ensuring that the comprehensive deformation prediction value truly reflects the actual deformation state of the steel column under complex load conditions. Second, through a reverse compensation function, the comprehensive deformation prediction value, measured stiffness coefficient, and ambient temperature difference are converted into a hoisting pre-offset, realizing a quantitative mapping relationship from deformation prediction to position compensation. Third, through the coupling optimization of the upper and lower layer game model, the optimal measurement section spacing is determined while ensuring attitude stability, providing a scientific basis for real-time data acquisition. Finally, by calculating the spatial position deviation between the segmented measured coordinate set and the design coordinate set and immediately adjusting the hoisting equipment, a closed-loop control system of measurement-judgment-correction is formed, which transforms the positioning process from open-loop prediction to closed-loop dynamic control, thus enabling the precise positioning of ultra-large steel columns.

[0062] The following provides a specific embodiment 1 of the present invention, and the specific implementation of each step in this embodiment 1 is described in detail below.

[0063] The specific implementation of step S01 is as follows: First, a three-dimensional laser scanner is used to perform a full-surface scan of the super-large steel column with a length exceeding 30m. During the scan, the distance between the laser scanner and the surface of the steel column is maintained between 2m and 5m to ensure scanning accuracy. The point cloud density is set to no less than 10,000 data points per square meter. The original point cloud is then processed by point cloud data processing software to perform noise reduction, registration, and stitching to establish a complete three-dimensional geometric model of the steel column and extract the initial geometric shape data of the steel column. At the same time, positioning reference points are marked every 5m along the axial direction on the surface of the steel column using a laser marker or mechanical dotting method. The marking depth of the positioning reference points is controlled between 0.5mm and 1.0mm to ensure that they will not disappear due to wear during the subsequent hoisting process.

[0064] The specific implementation of step S02 is to establish a multi-field coupled numerical model of the ultra-large steel column based on the finite element analysis principle. This model simultaneously considers the interaction of three physical fields: self-weight field, temperature field, and welding stress field. The steel column is discretized into hexahedral or tetrahedral elements with a length not exceeding 1m using finite element mesh generation technology. Gravitational acceleration loads are applied to each element node to simulate the self-weight effect of the steel column; temperature gradient loads are applied to simulate the thermal expansion or contraction effect caused by changes in ambient temperature; and welding line heat source loads are applied to simulate the local heating and cooling contraction effects during the welding process. The displacement field distribution of each node of the steel column is obtained through multi-field coupled calculation using a nonlinear iterative solver, and the comprehensive deformation prediction value is then calculated. The calculation formula is expressed as follows:

[0065] ;

[0066] In the formula, The value is the predicted value of the overall deformation, in mm; The elastic deformation under its own weight, in mm, is obtained by calculating the vector sum of the longitudinal tensile deformation and the transverse bending deformation of the steel column under gravity through finite element analysis. The value is the temperature deformation, expressed in mm, and is obtained by linear thermal expansion and contraction caused by the difference between the ambient temperature and the steel column manufacturing temperature through finite element analysis. The value represents the welding shrinkage deformation, expressed in mm. It is obtained by calculating the local deformation of the steel column caused by cooling shrinkage in the weld area during the welding process using finite element analysis.

[0067] The specific implementation of step S03 involves inputting the comprehensive deformation prediction value, the measured stiffness coefficient of the steel column, and the ambient temperature difference into the reverse compensation function for calculation. The measured stiffness coefficient of the steel column... By constructing a scaled-down model with a geometric scale of 1:10 and applying distributed loads to measure deflection values, the measured deflection is compared with the theoretical deflection to obtain a stiffness correction coefficient. This coefficient is then substituted into a full-scale numerical model for simulation verification. The two results are weighted and averaged using weighting coefficients of 0.6 and 0.4, respectively. The calculation formula is as follows:

[0068] ;

[0069] In the formula, The measured stiffness coefficient of the steel column is dimensionless. The measured deflection value of the scaled model is in mm. It was obtained by using a displacement sensor after applying a concentrated load at the mid-span of the scaled model. The deflection value is the theoretical value of the scaled model, in mm, calculated based on the beam deflection theory of mechanics of materials. The deflection values ​​are full-scale numerical simulation values ​​in mm, calculated using finite element software under the same load conditions. This represents the full-size theoretical deflection value, in mm, calculated based on beam deflection theory in mechanics of materials. The formula for the reverse compensation function is as follows:

[0070] ;

[0071] In the formula, This refers to the pre-offset adjustment for hoisting, expressed in mm. The value is the predicted value of the overall deformation, in mm; The measured stiffness coefficient of the steel column is dimensionless. This is the difference in ambient temperature, expressed in °C, which is equal to the difference between the actual temperature measured at the construction site and the ambient temperature during the manufacture of the steel column. This is the standard temperature difference, in °C, with an empirical value of 20 °C. This is the baseline correction factor, dimensionless, and defaults to 1.0.

[0072] The specific implementation of step S04 is to establish a two-layer game optimization model to coordinate the two mutually restrictive optimization objectives of attitude stability and positioning accuracy during the hoisting process. The objective function formula of the upper-layer model is expressed as follows:

[0073] ;

[0074] In the formula, This is a dimensionless index for attitude stability. The baseline stability index is dimensionless and has an empirical value of 1.0. These are hoisting speed control parameters, in m / min. The standard hoisting speed is expressed in m / min, with an empirical value of 0.5 m / min. This refers to the length of the steel column, in meters (m). The standard steel column length is in meters (m), with an empirical value of 30m. The moment of inertia of the steel column section is expressed in units of 1000 m / s. ; The moment of inertia of the standard cross section, in units of experience value The constraints of the upper-level model are: m / min and The objective function of the lower-level model is expressed as follows:

[0075] ;

[0076] In the formula, It is a dimensionless indicator for evaluating positioning accuracy. The baseline positioning accuracy is dimensionless, and its empirical value is 1.0. The measurement unit is meters (m). The standard measurement interval is in meters (m), with an empirical value of 5 meters. These are hoisting speed control parameters, in m / min. The standard hoisting speed is expressed in m / min, with an empirical value of 0.5 m / min. The measured stiffness coefficient of the steel column is dimensionless. The constraints of the lower-level model are as follows: m and The solution process involves first fixing the hoisting speed to obtain the optimal attitude stability of the upper model, then substituting this speed into the lower model to obtain the optimal spacing of the measurement sections, and finally feeding the positioning accuracy evaluation index back to the upper model for iterative calculation. Convergence is determined when the change in attitude stability index between two adjacent iterations is less than 0.01 and the change in positioning accuracy evaluation index is less than 0.01.

[0077] The specific implementation of step S05 is as follows: during the hoisting of the super-large steel column, multiple measurement sections are set along the axial direction of the steel column according to the measurement section layout spacing calculated in step S04. Each measurement section position corresponds to a positioning reference point. High-precision measurement equipment such as a total station or laser tracker is used to collect the positioning reference points of each measurement section in real time in three-dimensional coordinates. The measurement frequency is set to collect coordinate data at least once per minute for each measurement section. All the collected measurement section coordinate data are organized in axial position order to form a segmented measured coordinate set of the steel column.

[0078] The specific implementation of step S06 involves comparing the measured coordinate set of the steel column segments with the design coordinate set point by point to calculate the spatial position deviation. The calculation formula is expressed as follows:

[0079] ;

[0080] In the formula, Spatial position deviation, in mm; , , The measured three-dimensional coordinates of the reference point for measuring the cross-section are in mm. , , The design three-dimensional coordinates of the reference points for measuring the cross-section, in mm. Deviation judgment threshold. The method combines scaled-down model experiments with numerical analysis. First, a hoisting simulation experiment was conducted on a 1:10 scaled-down model to establish a data jitter model. The formula for calculating the jitter amplitude is as follows:

[0081] ;

[0082] In the formula, The range of jitter amplitude is expressed in mm. The standard jitter amplitude is expressed in mm, with an empirical value of 0.1 mm. The standard deviation of the original data sequence, in mm, was calculated by repeatedly measuring the position data of the scaled model. The mean of the original data sequence, in mm, was calculated by repeatedly measuring the position data of the scaled model. The number of repeated measurements is dimensionless and ranges from 8 to 15. The standard number of measurements is dimensionless, and the empirical value is 10. The accuracy class of the measuring instrument is expressed in mm. This is the standard accuracy class, in mm, with an empirical value of 0.01 mm. (Scale adjustment influence factor) The calculation formula is expressed as follows:

[0083] ;

[0084] In the formula, The proportional adjustment factor is dimensionless and ranges from 8.5 to 11.5. The scale is dimensionless and has a value of 10. This is the material density of the scaled-down model, in units of... ; This refers to the density of the full-size steel column material, in units of... ; The stiffness of the scaled-down model is expressed in N / m. The stiffness of the full-size steel column support is expressed in N / m. The vibration amplitude is adjusted proportionally. The calculation formula is expressed as follows:

[0085] ;

[0086] In the formula, This is the jitter amplitude after proportional correction, in mm; The range of jitter amplitude is expressed in mm. The influence factor for proportional adjustment is dimensionless. Statistical analysis is performed on the jitter data distribution cloud, and the center coordinates of the jitter data distribution cloud are determined. The calculation formula is expressed as follows:

[0087] ;

[0088] In the formula, The center coordinates of the jitter data distribution cloud are given by , and is a three-dimensional vector containing x, y, and z coordinates, in mm. The total number of measurement data points, dimensionless; Let be the three-dimensional coordinate vector of the j-th measurement data point, containing x, y, and z coordinates, in mm. The dispersion of the jitter data distribution cloud. The calculation formula is expressed as follows:

[0089] ;

[0090] In the formula, The dispersion of jitter data distribution in the cloud, in mm; The total number of measurement data points, dimensionless; Let be the three-dimensional coordinate vector of the j-th measurement data point, which is a three-dimensional vector containing x, y, and z coordinates, in mm; The center coordinates of the jitter data distribution cloud are given by , and is a three-dimensional vector containing x, y, and z coordinates, in mm. This represents the Euclidean distance from the j-th data point to the center coordinates, in mm. The dispersion deviation of the data points. The calculation formula is expressed as follows:

[0091] ;

[0092] In the formula, represents the dispersion deviation of the j-th data point, in mm; Let be the three-dimensional coordinate vector of the j-th measurement data point, which is a three-dimensional vector containing x, y, and z coordinates, in mm; The center coordinates of the jitter data distribution cloud are given by , and is a three-dimensional vector containing x, y, and z coordinates, in mm. The distance from the j-th data point to the center coordinates is expressed in mm. The dispersion of the jittery data distribution cloud is expressed in mm. An improved two-parameter Gaussian kernel function is used. The formula is expressed as follows:

[0093] ;

[0094] In the formula, Let be the kernel function weight value for the j-th data point, which is dimensionless; The baseline weight value is dimensionless and has an empirical value of 1.0. Let be the three-dimensional coordinate vector of the j-th measurement data point, which is a three-dimensional vector containing x, y, and z coordinates, in mm; The center coordinates of the jitter data distribution cloud are given by , and is a three-dimensional vector containing x, y, and z coordinates, in mm. The distance from the j-th data point to the center coordinates is expressed in mm. This is the kernel function bandwidth parameter, in mm; represents the dispersion deviation of the j-th data point, in mm; This is the dispersion adjustment parameter, in mm. Kernel function bandwidth parameter. The calculation formula is expressed as follows:

[0095] ;

[0096] In the formula, This is the kernel function bandwidth parameter, in mm; The dispersion of jitter data distribution in the cloud, in mm; This is the bandwidth adjustment factor, dimensionless, with an empirical value of 1.5. Dispersion adjustment parameter. The calculation formula is expressed as follows:

[0097] ;

[0098] In the formula, This is a dispersion adjustment parameter, in mm. The dispersion of the jittery data distribution cloud, in mm. The center value after kernel filtering. The calculation formula is expressed as follows:

[0099] ;

[0100] In the formula, The center value after kernel filtering is a three-dimensional vector containing x, y, and z coordinates, in mm. Let be the kernel function weight value for the j-th data point, which is dimensionless; Let be the three-dimensional coordinate vector of the j-th measurement data point, which is a three-dimensional vector containing x, y, and z coordinates, in mm; The total number of measurement data points is dimensionless. A numerical jitter disturbance term is introduced when establishing the full-scale numerical simulation model. The calculation formula is expressed as follows:

[0101] ;

[0102] In the formula, This is the numerical jitter / disturbance term, in mm; This is the jitter amplitude after proportional correction, in mm; For scaling, dimensionless, and with a value of 10. Kernel density estimation is performed on the probability distribution of simulation bias thresholds obtained from 1000 Monte Carlo simulations, and the probability density function is smoothed. The calculation formula is expressed as follows:

[0103] ;

[0104] In the formula, The value of the smoothed probability density function is expressed in units of 1. ; The deviation threshold variable is in mm. The number of Monte Carlo simulations is dimensionless and takes a value of 1000. The kernel function weight value corresponding to the k-th simulation result is dimensionless. Bandwidth parameter for kernel density estimation, in mm. Bias judgment threshold. The fusion calculation formula is expressed as follows:

[0105] ;

[0106] In the formula, The deviation judgment threshold is in mm and ranges from 18 to 22 mm. The experimental bias threshold after processing with the improved two-parameter Gaussian kernel function is expressed in mm. The simulation deviation threshold after processing with the improved two-parameter Gaussian kernel function is expressed in mm. The weighting coefficient for the experimental data is dimensionless and has an empirical value of 0.55. The weighting coefficient for the simulation data is dimensionless and has an empirical value of 0.45. When the spatial position deviation of any measurement section exceeds the deviation judgment threshold, the horizontal displacement and vertical lifting amount that the hoisting equipment needs to adjust are calculated based on the direction and magnitude of the deviation.

[0107] The specific implementation of step S07 involves continuously monitoring the spatial position deviation of all measurement sections and the axial angle deviation of adjacent measurement sections during the hoisting process, and setting a positioning accuracy threshold. The same method as the deviation judgment threshold is used, but the dispersion weighting coefficient of the jitter data distribution cloud is adjusted to 0.65, the number of Monte Carlo simulations is increased to 1500, and the positioning accuracy threshold ranges from 12 to 16 mm. Axis direction vector. The calculation formula is as follows: It is obtained by fitting the coordinates of adjacent positioning reference points.

[0108] ;

[0109] In the formula, Let be the axial direction vector of the i-th measurement section, and let be a three-dimensional unit vector containing x, y, and z components, which is dimensionless. Let be the three-dimensional coordinate vector of the i-th measurement section positioning reference point, which is a three-dimensional vector containing x, y, and z coordinates, in mm; Let be the three-dimensional coordinate vector of the (i+1)th measurement section positioning reference point, which is a three-dimensional vector containing x, y, and z coordinates, in mm; The Euclidean distance between the i-th and (i+1)-th measurement sections is in mm. Axis angle deviation. The calculation formula is expressed as follows:

[0110] ;

[0111] In the formula, This represents the deviation of the included angle of the axis, in degrees. Let be the axial direction vector of the i-th measurement section, and let be a three-dimensional unit vector containing x, y, and z components, which is dimensionless. Let be the axial direction vector of the (i+1)th measurement section, and let be a three-dimensional unit vector containing x, y, and z components, which is dimensionless. It is the dot product of two unit vectors, dimensionless; This is the theoretical included angle, in degrees. Angle accuracy threshold. In the process of determining the proportion adjustment influence factor, an angle measurement error amplification factor is added. The calculation formula is expressed as follows:

[0112] ;

[0113] In the formula, This is the angle measurement error amplification factor, which is dimensionless. The scale is dimensionless and has a value of 10. The angle accuracy threshold ranges from 0.25 degrees to 0.35 degrees. When the spatial position deviation of all measured sections is less than the positioning accuracy threshold and the axial angle deviation of adjacent measured sections is less than the angle accuracy threshold, the steel column fixing procedure is initiated. The steel column is locked in its current position using a temporary support structure or clamps. Then, welding is performed to firmly connect the steel column to the installed components, completing the insertion construction positioning of the ultra-large steel column.

[0114] To better understand and implement this invention, a specific application scenario, Example 2, is provided below: The technical team first conducted a detailed initial state inspection of one of the steel columns, numbered C-03. After the steel column was welded and assembled at the manufacturing plant, technicians used a 3D laser scanner to perform a comprehensive scan of the steel column, achieving a point cloud density of 5000 data points per square meter. The scanning process was conducted in a temperature-controlled workshop, with the ambient temperature controlled at 22°C and the relative humidity at 45%. After denoising and registration processing, the point cloud data acquired by the scan was used to establish an initial geometric digital model of the steel column. Figure 2 As shown, technicians marked eight positioning reference points every 5 meters along the axial direction on the surface of the steel column. These reference points were made using laser engraving, with a diameter of 8 mm and a depth of 0.5 mm, to ensure they will not wear away during subsequent transportation and hoisting. The steel column adopts a box-section design with cross-sectional dimensions of [missing information]. The main lifting point is set at 12m from the bottom of the column, the secondary lifting point is set at 8m from the top of the column, and the three main welds are located at 6m, 18m and 30m from the bottom of the column, respectively.

[0115] To accurately predict the deformation of the steel column during hoisting, the technical team established a multi-field coupled finite element model. The model was built using the ANSYS software platform, dividing the 36m long steel column into 0.8m long element meshes, totaling 45 element nodes. The model was input with a steel column self-weight load of 58t and a steel density of... elastic modulus The Poisson's ratio is 0.3. Considering that the steel column will be hoisted on-site, and the actual measured temperature on-site is 34℃, a 12℃ difference from the manufacturing temperature of 22℃, the technicians input temperature field parameters into the model, and the coefficient of linear expansion of the steel was set to [value missing]. The steel column has three main weld seams during manufacturing, located at distances of 6m, 18m, and 30m from the column base, respectively, each weld seam being 4.8m long. Technicians applied welding line heat source loads to these three weld seam locations in the finite element model, setting the welding heat input power to [value missing]. The welding speed is Cooldown time: 180 seconds.

[0116] Model calculations show that the longitudinal tensile deformation of the steel column under its own weight in the hoisted state is 18mm, mainly concentrated in the upper half of the column; the maximum lateral bending deformation occurs in the middle of the column, at 35mm, in a horizontal eastward direction. Temperature deformation calculations indicate that a 12℃ temperature difference will cause a 5.2mm thermal expansion elongation in the steel column. Welding shrinkage deformation analysis shows that the cooling shrinkage of the three welds will cause local shrinkage of the steel column at the weld locations, with a lateral shrinkage of 2.8mm and a longitudinal shrinkage of 1.5mm at the welds. Figure 3 As shown, considering the three deformation factors, the technical team obtained the comprehensive deformation prediction value of the steel column. The finite element model was discretized with 45 element nodes. The deformation cloud map clearly shows the overall deformation characteristics of the steel column under hoisting conditions. The maximum deformation occurs in the upper middle part of the column, and the weld area shows obvious local shrinkage effect. The comprehensive deformation prediction value is shown in Table 1.

[0117] Table 1. Predicted values ​​of comprehensive deformation of steel columns

[0118] As shown in Table 1, after vector superposition of various deformations, the predicted comprehensive deformation of the steel column under hoisting conditions is 42mm, mainly manifested as eastward shift of the middle of the column and overall longitudinal elongation.

[0119] To improve the accuracy of deformation prediction, the technical team created a 1:10 scale model of the steel column in the laboratory. The scaled-down model uses the same Q345B steel as the actual steel column, with the box-section reduced to [specific dimensions]. The model is 3.6m long. Technicians applied a uniformly distributed load to the scaled-down model, the magnitude of which was converted according to similarity theory. Under load, the measured deflection of the scaled-down model was 4.2 mm, while the theoretically calculated deflection was 3.8 mm. The ratio of the measured deflection to the theoretical deflection was 1.105, which was used as a stiffness correction factor. The technical team then established a full-scale numerical analysis model, substituting the stiffness correction factor of 1.105 into the model for simulation calculations. The simulation results showed that the deflection of the steel column was 38 mm. Using the weighted average method, the weighting factor for the experimental results of the small-scale model was 0.6, and the weighting factor for the numerical analysis results was 0.4, resulting in a calculated measured stiffness factor of 1.063 for the steel column.

[0120] Based on the predicted comprehensive deformation of 42mm, the measured stiffness coefficient of the steel column of 1.063, and the ambient temperature difference of 12℃, the technical team calculated the pre-offset for hoisting using a reverse compensation function. In the calculation process, the standard temperature difference was set at 20℃, and the baseline correction factor was set at 1.0. The calculated pre-offset for hoisting was 48mm. This means that during the hoisting of the steel column, its initial position needs to be offset westward by 48mm relative to the design position to compensate for the eastward deformation that occurs during hoisting. Simultaneously, the bottom elevation of the steel column needs to be lowered by 5.2mm to compensate for longitudinal elongation caused by temperature.

[0121] Before determining the hoisting control parameters, the technical team needed to establish deviation thresholds through a combination of small-scale model experiments and numerical analysis. The technicians conducted a hoisting simulation experiment on a 1:10 scale model in an outdoor setting at an ambient temperature of 28°C and a wind speed of [missing information]. During the experiment, technicians used a total station to repeatedly measure the positioning reference points of the scaled model 15 times, with each measurement spaced 5 minutes apart. The standard deviation of the original measurement data sequence was 0.085 mm, and the mean was 3600.235 mm. The total station's accuracy class was 0.01 mm. Based on the jitter amplitude calculation function, and substituting the number of repeated measurements (15), the jitter amplitude range was calculated to be 0.124 mm.

[0122] The technical team further considered the impact of scaling adjustments. The material density of the scaled model is the same as that of the full-size steel column, and the material property correction factor is 1.0. The scaled model uses hinged supports with a support stiffness of [missing value]. The support stiffness of the full-size steel column is The boundary condition similarity coefficient is 1.0. The scaling adjustment influence factor is calculated to be 10.0. Applying the scaling adjustment influence factor to the jitter amplitude range, the jitter amplitude after scaling correction is 1.24mm.

[0123] In 15 repeated measurements, technicians obtained the three-dimensional coordinates of 45 data points, which constituted a jitter data distribution cloud. For example... Figure 3 As shown, the jitter data distribution cloud exhibits an ellipsoidal distribution in space, with its major axis aligned with the measurement line of sight. The technical team conducted statistical analysis on the jitter data distribution cloud, calculating the center coordinates as (3600.235, 1200.018, 2400.142), in mm. The dispersion calculation result is 0.089 mm, indicating that the data points are relatively concentrated. Skewness analysis shows that the data distribution is slightly biased towards the measuring instrument, with a skewness value of 0.15. Using the center coordinates as a representative value for the experimental measurement, the technicians multiplied the dispersion of 0.089 mm by a confidence coefficient of 1.96, obtaining an uncertainty range of 0.174 mm for the experimental measurement.

[0124] The technical team set different deviation values ​​in the experiment: 10mm, 15mm, 20mm, 25mm, and 30mm, and observed the final positioning error of the steel column corresponding to each deviation value. The experimental results are shown in Table 2.

[0125] Table 2 Relationship between Deviation Value and Positioning Error

[0126] As shown in Table 2, when the deviation value is 20mm, the final positioning error is 13.8mm, which is within an acceptable range. When the deviation value increases to 25mm, the positioning error surges to 18.6mm, exceeding the design allowable range. The technical team fitted the deviation-error relationship curve and obtained an experimental deviation threshold of 21mm.

[0127] The technical team then established a full-scale numerical simulation model, inputting the same boundary conditions as the experiment, including the location of the hoisting point, support stiffness, and ambient temperature. In the numerical calculations, the technicians introduced a numerical jitter disturbance term. This disturbance term, after proportional correction, resulted in a jitter amplitude of 1.24 mm, which was then scaled up 10 times to 12.4 mm. The team used the Monte Carlo method to perform 1000 random jitter simulations, randomly adding the numerical jitter disturbance term to the displacement calculation results in each calculation. The results of the 1000 simulations showed that the probability distribution of the simulation deviation threshold exhibited a normal distribution, with a mean of 19.8 mm, a standard deviation of 1.5 mm, and a 95% confidence interval of [16.8 mm, 22.8 mm].

[0128] To improve the reliability of threshold acquisition, the technical team implemented an improved two-parameter Gaussian kernel function filter on the jitter data distribution cloud. The kernel function bandwidth parameter was determined based on the dispersion of the jitter data distribution cloud (0.089 mm), and was calculated to be 0.134 mm. The dispersion adjustment parameter was calculated to be 0.071 mm. The technicians calculated the kernel function weight value for each data point in the jitter data distribution cloud. For example, for the first data point, its distance to the center coordinates is 0.052 mm, and the dispersion deviation is 0.037 mm, resulting in a kernel function weight value of 0.892. A weighted average of all 45 data points was calculated according to their respective kernel function weight values, yielding a kernel-filtered center value of (3600.228, 1200.015, 2400.138), in mm. Compared to the original center coordinates, the coordinate offset of the kernel-filtered center value is less than 0.01 mm, indicating a relatively stable data distribution.

[0129] The technical team used the same improved two-parameter Gaussian kernel function to estimate the kernel density of 1000 Monte Carlo simulations. Each simulation result was used as the center of the kernel function, and its contribution to the overall probability density was calculated. The sum of the 1000 kernel functions yielded a smooth probability density curve for the simulation deviation threshold. The curve peaked at 19.5 mm, and the team used this peak position as the optimal estimate of the simulation deviation threshold. The 95% confidence interval was extracted to be [17.2 mm, 21.8 mm].

[0130] The technical team performed a Bayesian weighted fusion calculation on the experimental deviation threshold of 21 mm and the simulation deviation threshold of 19.5 mm, processed using an improved two-parameter Gaussian kernel function. The weighting coefficient for the experimental data was set to 0.55, and the weighting coefficient for the simulation data was set to 0.45. The fusion calculation yielded a deviation judgment threshold of 20.3 mm. The technical team further conducted a sensitivity analysis on the deviation judgment threshold, changing the support stiffness of the scaled model from... Change to The elastic modulus of the material is from Change to The ambient temperature changed from 28℃ to 32℃. Sensitivity analysis showed that the deviation judgment threshold changed by 3.8%, which is less than the 5% requirement, confirming that the deviation judgment threshold has sufficient robustness. The final deviation judgment threshold was determined to be 20mm.

[0131] Using the same method, the technical team determined the positioning accuracy threshold. In determining the positioning accuracy threshold, the dispersion weighting coefficient of the jitter data distribution cloud was adjusted to 0.65, and the number of Monte Carlo simulations was increased to 1500. After improving the two-parameter Gaussian kernel function processing and Bayesian weighted fusion, a positioning accuracy threshold of 14mm was obtained. In determining the angle accuracy threshold, an angle measurement error amplification factor was added to the calculation of the scaling factor. This factor is 10 to the power of 1.2, with a calculated value of 15.85. After the same processing flow, an angle accuracy threshold of 0.30° was obtained.

[0132] The technical team established a game theory model for optimizing the hoisting path to determine the optimal hoisting speed control parameters and the spacing of the measurement sections. The upper-level model uses the steel column's attitude stability as the optimization objective, with input parameters including the steel column length of 36m and the moment of inertia of the steel column section. The technicians set the initial hoisting speed control parameters as follows: The values ​​were substituted into the objective function of the upper-level model for calculation. The calculated attitude stability index was 0.88, satisfying the constraint condition of not less than 0.85. The lower-level model optimized positioning accuracy, with input parameters including the measured stiffness coefficient of the steel column of 1.063. Technicians set the initial measurement section spacing to 5m and substituted it into the objective function of the lower-level model for calculation. The calculated positioning accuracy evaluation index was 0.92, satisfying the constraint condition of not less than 0.90. Figure 4 As shown, the technical team solved the game theory model through iterative calculations. The figure clearly illustrates the dual impact of the hoisting speed control parameters on both the attitude stability index and the positioning accuracy evaluation index. The blue curve represents the trend of the attitude stability index with hoisting speed, and the red curve represents the trend of the positioning accuracy evaluation index with hoisting speed. The two constraint lines correspond to the threshold requirements of 0.85 and 0.90, respectively. The green feasible region shows the hoisting speed range that satisfies both constraints. In the first iteration, the hoisting speed control parameters... With the upper-level model fixed, the optimal attitude stability index is obtained as 0.88. Substituting this hoisting speed control parameter into the lower-level model, the optimal measurement section spacing is found to be 4.8m, corresponding to a positioning accuracy evaluation index of 0.93. The positioning accuracy evaluation index of 0.93 is fed back to the upper-level model, and the hoisting speed control parameter is adjusted accordingly. In the second iteration, the hoisting speed control parameters were fixed. Solving the upper-level model yielded an attitude stability index of 0.89, while solving the lower-level model yielded a measurement section spacing of 4.7m and a positioning accuracy evaluation index of 0.94. After eight iterations, the change in attitude stability index between adjacent iterations was 0.005, and the change in positioning accuracy evaluation index was 0.008, both less than the convergence criterion of 0.01. The final hoisting speed control parameters were determined as follows: The spacing between the measurement sections is 4.6m.

[0133] Before the formal hoisting of steel column C-03, the technical team conducted detailed construction preparations. Two 800t crawler cranes were used for the hoisting, with the main hoisting point set 12m from the bottom of the steel column and the auxiliary hoisting point set 8m from the top. High-strength steel wire ropes with a diameter of 52mm and a breaking strength of [missing information - likely a specific value] were used for the hoisting. Five measurement benchmarks were set up at the construction site. The coordinates of the benchmarks adopted the unified coordinate system of the construction site, and the coordinate accuracy was achieved through GPS-RTK technology, with a horizontal accuracy of ±3mm and an elevation accuracy of ±5mm.

[0134] Before the hoisting began, technicians arranged eight measurement sections along the axial direction on the steel column at 4.6m intervals. The first measurement section was located at the bottom of the steel column, and the eighth measurement section was located at the top of the steel column. Each measurement section included previously marked positioning reference points. Technicians used a total station to collect the three-dimensional coordinates of each positioning reference point at the initial position of the steel column, establishing a segmented measured coordinate set for the steel column, as shown in Table 3.

[0135] Table 3 Measured coordinates of the initial position of the steel column in segments

[0136] As shown in Table 3, the coordinates of each measurement section at the initial position of the steel column have been collected, in mm. The technical team compared the measured coordinates with the design coordinates. The design coordinates of the first measurement section in the design coordinate set are (12500.000, 8300.000, 1500.000), in mm. The spatial position deviation of the first measurement section was calculated: the deviation in the X direction is 0.015 mm, the deviation in the Y direction is 0.022 mm, the deviation in the Z direction is 0.008 mm, and the Euclidean distance deviation is 0.028 mm, which is far less than the deviation judgment threshold of 20 mm, and no adjustment is required.

[0137] The hoisting operation officially began, with the crane... The steel column was slowly raised. When the column was 1.2m off the ground, technicians collected the first coordinates of the first measurement section. The measured coordinates were (12500.068, 8300.085, 1501.208), in mm. Compared with the design coordinates (12500.000, 8300.000, 1501.200), the spatial position deviation was 0.109mm, which was less than the deviation threshold, so the column continued to be raised.

[0138] When the steel column was raised to 5.5m above the ground, technicians collected coordinates at the second measurement section. The measured coordinates were (12500.145, 8304.732, 1509.805), in mm. The design coordinates were (12500.000, 8304.600, 1509.800), with a spatial position deviation of 0.188mm, less than the deviation threshold. Technicians continued to monitor the third measurement section. The measured coordinates were (12500.328, 8309.455, 1514.398), and the design coordinates were (12500.000, 8309.200, 1514.400), with a spatial position deviation of 0.514mm, still within the allowable range.

[0139] When the steel column was raised to 12m above the ground, the central section of the column entered the key monitoring phase. Technicians collected coordinates for the fourth measurement section: measured coordinates (12501.850, 8314.268, 1519.025), designed coordinates (12500.000, 8313.800, 1519.000), with a spatial deviation of 1.91mm. Coordinates for the fifth measurement section were collected: measured coordinates (12503.520, 8318.985, 1523.618), designed coordinates (12500.000, 8318.400, 1523.600), with a spatial deviation of 3.58mm. Technicians observed a significant eastward shift in the central section of the steel column, but the deviation remained within the 20mm threshold, thus not triggering dynamic correction.

[0140] When the steel column was raised to 18m above the ground, technicians collected coordinates again at the fifth measurement section. The measured coordinates were (12506.850, 8319.652, 1528.835), while the designed coordinates were (12500.000, 8318.400, 1528.800), with a spatial position deviation of 6.94mm. Coordinates were collected at the sixth measurement section. The measured coordinates were (12510.250, 8324.385, 1533.428), while the designed coordinates were (12500.000, 8323.000, 1533.400), with a spatial position deviation of 10.41mm. The measured coordinates of the seventh measurement section are (12514.680, 8329.205, 1538.058), while the designed coordinates are (12500.000, 8327.600, 1538.000), with a spatial deviation of 14.91 mm. Technicians noted that the deviation is increasing rapidly, but has not yet exceeded the deviation threshold of 20 mm.

[0141] When the steel column was raised to 24m above the ground, technicians collected coordinates for the sixth measurement section. The measured coordinates were (12518.250, 8325.185, 1538.652), while the designed coordinates were (12500.000, 8323.000, 1538.600), with a spatial position deviation of 18.48mm. Coordinates for the seventh measurement section were collected. The measured coordinates were (12523.580, 8330.025, 1543.285), while the designed coordinates were (12500.000, 8327.600, 1543.200), with a spatial position deviation of 23.71mm. The spatial position deviation of the seventh measurement section exceeded the deviation threshold of 20mm, triggering the dynamic correction procedure. Technicians immediately calculated the direction and amount of the deviation, which was mainly an eastward shift of 23.71mm. Based on the deviation, technicians directed the main crane to move westward by 12mm and the auxiliary crane by 15mm. Simultaneously, they adjusted the vertical lifting of the steel column, lowering the top of the column by 8mm. After the adjustment, technicians again collected coordinates from the seventh measurement section. The measured coordinates were (12508.350, 8328.285, 1542.458), reducing the spatial position deviation to 8.56mm, which is less than the deviation threshold.

[0142] The steel column was raised to 30m above the ground. Technicians collected coordinate data at the eighth measurement section. The measured coordinates were (12510.850, 8332.985, 1547.825), and the design coordinates were (12500.000, 8332.200, 1547.800). The spatial deviation was 10.98mm, within the allowable range. Technicians performed a final check on all measurement sections, and the measurement results are shown in Table 4.

[0143] Table 4 Measured Coordinates of the Final Position of the Steel Columns in Sections

[0144] As shown in Table 4, the spatial position deviation of all measured sections is less than the positioning accuracy threshold of 14 mm, meeting the positioning accuracy requirements. The technical team further checked the axial angle deviation of adjacent measured sections. The axial direction vector was fitted using the coordinates of the positioning reference points of the first and second measured sections, resulting in vectors (0.067, 4.590, 4.597), which were normalized to (0.010, 0.707, 0.707). The normalized axial direction vectors fitted using the second and third measured sections were (0.024, 0.707, 0.707). The angle between the two vectors was calculated using the vector dot product formula to be 0.18°, less than the angle accuracy threshold of 0.30°. The axial angle deviation of the remaining adjacent measured sections was checked sequentially, and the calculated results were 0.22°, 0.25°, 0.28°, 0.26°, 0.23°, and 0.20°, all less than the angle accuracy threshold of 0.30°.

[0145] After all accuracy indicators met the requirements, the technical team fixed the position of the steel column and began welding connections. Welding was performed using... Gas shielded welding was used, with a welding current of 280A, a welding voltage of 32V, and a welding wire diameter of 1.2mm. The welding process was carried out in three layers, with a cooling time of no less than 30 minutes after each layer. After welding, technicians checked the position of the steel column again to confirm that the change in position of the steel column during the welding process was less than 2mm, meeting the construction specifications. The positioning of steel column C-03 was completed.

[0146] The technical team used the same method to complete the hoisting and positioning of the remaining seven extra-large steel columns. Throughout the construction process, the average positioning accuracy of the eight steel columns was 9.5 mm, with a maximum deviation of 10.58 mm, far superior to the 30 mm deviation level of traditional methods. The average deviation of the axial angle between adjacent measurement sections was 0.23°, with a maximum of 0.28°, ensuring the verticality requirements of the steel columns. A total of 15 dynamic corrections were performed during construction, averaging 1.875 corrections per steel column, significantly improving correction efficiency.

[0147] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for positioning a super large steel column during construction, characterized in that, The method comprises the following steps: Three-dimensional laser scanning is performed on the super-large steel column to obtain initial geometric shape data of the steel column, and positioning reference points are marked on the surface of the steel column at an interval of 5m along the axial direction, a multi-field coupling finite element model of the super-large steel column is established, the self-weight load, environmental temperature parameters and welding joint positions of the steel column are input, the self-weight elastic deformation, temperature deformation and welding shrinkage deformation of the steel column in the hoisting state are calculated, the comprehensive deformation prediction value is obtained, the comprehensive deformation prediction value, the measured stiffness coefficient of the steel column and the environmental temperature difference value are input into a reverse compensation function to calculate the hoisting pre-offset amount, a hoisting path optimization game model is established, the hoisting speed control parameter is calculated by the upper model, the measurement section arrangement interval is calculated by the lower model, a plurality of measurement sections are arranged along the axial direction of the steel column at the measurement section arrangement interval during the hoisting process of the super-large steel column, the real-time three-dimensional coordinate of the positioning reference point of each measurement section is collected to obtain the segmented measured coordinate set of the steel column, the spatial position deviation calculation is performed on the segmented measured coordinate set and the design coordinate set, when the spatial position deviation of any measurement section is greater than the deviation judgment threshold, the horizontal displacement and the vertical lifting amount of the hoisting equipment are adjusted according to the deviation direction and the deviation amount to realize dynamic correction, when the spatial position deviation of all measurement sections is less than the positioning precision threshold and the axial angle deviation of adjacent measurement sections is less than the angle precision threshold, the super-large steel column is fixed and welded to complete the positioning of the super-large steel column.

2. The method of claim 1, wherein, The multi-field coupling finite element model is a steel column deformation simulation model considering the interaction of the self-weight field, the temperature field and the welding stress field.

3. The method of claim 2, wherein, When the multi-field coupling finite element model is established, the steel column is divided into unit grids with a length of not greater than 1m, the gravity acceleration load, the temperature gradient load and the welding line heat source load are applied to each unit grid, and the displacement field distribution of each node of the steel column is obtained through iterative calculation.

4. The method of claim 3, wherein, The self-weight elastic deformation is the vector sum of the longitudinal tensile deformation and the transverse bending deformation of the steel column under the action of gravity, the temperature deformation is the linear deformation caused by the thermal expansion and contraction due to the difference between the environmental temperature and the manufacturing temperature of the steel column, and the welding shrinkage deformation is the local deformation of the steel column caused by the cooling shrinkage of the weld area during the welding process.

5. The method of claim 4, wherein, The measured stiffness coefficient of the steel column is the ratio of the unit length bending stiffness of the steel column to the theoretical bending stiffness, which is measured through a small-scale model experiment.

6. The method of claim 5, wherein, When the measured stiffness coefficient of the steel column is measured, a scaled-down model with a geometric scale of 1:10 is made, the deflection value of the scaled-down model is measured by applying a distributed load to the scaled-down model, the stiffness correction coefficient is obtained by comparing the deflection value with the theoretical deflection value, the stiffness correction coefficient is substituted into the numerical analysis model for full-size simulation verification, and the weighted average value of the two calculation results is taken as the measured stiffness coefficient of the steel column.

7. The method of claim 6, wherein, The reverse compensation function is used to calculate the spatial position pre-offset amount required during hoisting according to the steel column deformation prediction parameters, the input includes the comprehensive deformation prediction value, the measured stiffness coefficient of the steel column and the environmental temperature difference value, and the output is the hoisting pre-offset amount.

8. The method of claim 7, wherein, The deviation judgment threshold, the positioning precision threshold and the angle precision threshold are determined by combining a small-scale model experiment with numerical analysis.

9. The method of claim 8, wherein, In determining the deviation decision threshold, the hoisting simulation experiment is carried out on the 1:10 scale model, and the data jitter model is established according to the physical constraint condition, and the data jitter model is used for simulating the random fluctuation of the measurement data of the scale model influenced by the environmental vibration, the measurement instrument precision and the human operation error.

10. The method of claim 9, wherein, The proportional adjustment influence factor is considered in the data jitter model, the proportional adjustment influence factor is the amplification or attenuation coefficient of the scale ratio to the data jitter characteristics, and the proportional adjustment influence factor is applied to the jitter amplitude range to obtain the proportional correction jitter amplitude.

Images