A comprehensive control method for the coaxiality of wind tunnel body loops

Abstract

This invention provides a comprehensive control method for the coaxiality of the wind tunnel loop, belonging to the field of wind tunnel construction technology. This invention calculates thermal deformation in real time by establishing a coupled temperature field and deformation field finite element model, and dynamically adjusts the density of measuring points using a three-layer multi-scale iterative leveling method to achieve efficient and accurate measurement. It uses the Jacobian mapping matrix and matrix pseudo-inverse operation to solve the optimal support point adjustment amount, and introduces a torque sensor and hydraulic jack in the corner section to form a closed-loop control system to dynamically balance the torsion angle. This solves the technical problem of the difficulty in achieving full-scale accurate coaxiality control of the wind tunnel loop under the coupling effect of multiple physical fields.

Description

Technical Field

[0001] This invention belongs to the field of wind tunnel construction technology, and specifically relates to a comprehensive control method for the coaxiality of the wind tunnel body loop. Background Technology

[0002] Coaxiality control of the wind tunnel loop is a key technology for ensuring airflow quality. Traditional control methods mainly employ segmented independent measurements and single-scale adjustments. Axial deviation is measured using a total station or laser tracker at fixed measurement point intervals, and adjustments are then made to each support point based on the measurement results. In current wind tunnel construction and maintenance projects, the loop length can reach hundreds of meters, and thermal deformation caused by temperature fields exists. Traditional methods often treat temperature effects as environmental interference factors, failing to couple the temperature field with the deformation field for analysis. This results in measurement data that cannot accurately reflect the true position and orientation of the tunnel. Traditional techniques use uniform measurement point intervals for full-loop measurements, failing to balance the contradiction between coarse adjustment efficiency and fine adjustment accuracy. Furthermore, they lack systematic optimization methods for handling measurement error propagation and cumulative error distribution, leading to mutual interference between adjustments of different tunnel segments and an increase in the number of iterations. Especially in corner areas, torsional deformation and axial deviation are coupled. Traditional methods lack targeted torsional control measures, and relying solely on axial adjustment is insufficient to solve the coaxiality deviation problem caused by torsion. In other words, existing technologies present a technical problem: it is difficult to achieve precise coaxiality control across all scales in the wind tunnel's internal circuitry under the coupling effect of multiple physical fields. Summary of the Invention

[0003] In view of this, the present invention provides a comprehensive control method for the coaxiality of wind tunnel circuits, which can solve the technical problem in the prior art that it is difficult to achieve full-scale accurate coaxiality control of wind tunnel circuits under the coupling effect of multiple physical fields.

[0004] This invention is implemented as follows: A comprehensive control method for the coaxiality of a wind tunnel's internal loop includes: establishing a three-dimensional solid finite element model based on the wind tunnel's internal loop design drawings; inputting geometric parameters and material property parameters of each segment of the tunnel into the model; setting an initial measurement point layout scheme for the coarse-scale layer; arranging a distributed temperature sensing network on the surface and inside the wind tunnel to collect temperature data from each measurement point in real time; inputting the temperature data as thermal load boundary conditions into the finite element model for coupled temperature field-deformation field analysis; and employing a multi-scale layered iterative leveling method to perform layered measurement and adjustment of the tunnel's internal loop. First, the overall axial deviation is measured at the coarse-scale layer, and the coarse adjustment amount at each support point is calculated, followed by coarse adjustment. Then, the measurement is performed in segments at the mesoscale level by increasing the density of measuring points. At the mesoscale level, the pose deviation of each section of the tunnel is identified, and a Jacobian mapping matrix is ​​established between the rigid body pose parameters of the tunnel and the adjustment of each support point. The adjustment of the support point is solved by matrix pseudo-inverse operation and the adjustment is executed. High-density measurements are performed on key areas at the fine-scale level. When the measurement point deviation exceeds the preset threshold, the measurement is returned to the mesoscale level for readjustment. When the deviation of all measuring points is less than the preset threshold, the measurement enters the torsion control of the corner section. Torque sensors and diagonally arranged hydraulic jacks are installed in the corner section of the wind tunnel to monitor the torsion angle data of the corner section in real time. When the torsion angle exceeds the allowable value, the reverse torque is applied by the hydraulic jacks for dynamic balance adjustment.

[0005] Specifically, the initial measurement point layout scheme for the coarse-scale layer involves setting measurement points on the tunnel loop according to the spacing between the coarse-scale layer measurement points. The spacing between the coarse-scale layer measurement points is calculated using a coarse-scale measurement point spacing function.

[0006] Specifically, the calculation of the coarse-scale measuring point spacing function involves multiplying the ratio of the tunnel diameter to the reference diameter by the ratio of the tunnel wall thickness to the reference wall thickness, and then multiplying by the reference measuring point spacing.

[0007] The reference diameter, reference wall thickness, and reference measuring point spacing are determined through a standard tunnel test. The standard tunnel test includes constructing a standard tunnel test section, arranging measuring point groups at different spacings on the standard tunnel test section, measuring the axial deviation data of each measuring point group using a three-dimensional laser scanner, calculating the measurement coverage and error propagation coefficient of each measuring point group, and selecting the maximum spacing with a measurement coverage greater than 90% and an error propagation coefficient less than 1.2 as the reference measuring point spacing.

[0008] Specifically, the temperature field-deformation field coupled analysis and calculation involves first solving the steady-state temperature field distribution to obtain the temperature values ​​of each node, then applying the temperature values ​​of each node as body loads to the structural field to calculate the thermal stress distribution and thermal deformation distribution, and extracting the displacement of key nodes on the tunnel axis.

[0009] The multi-scale hierarchical iterative leveling method includes three levels: coarse-scale, meso-scale, and fine-scale. The distance between measurement points in the meso-scale layer is calculated using the meso-scale measurement point distance function, and the distance between measurement points in the fine-scale layer is calculated using the fine-scale measurement point distance function. The adjustment amount for each layer is calculated using the weighted least squares fitting method.

[0010] Specifically, the calculation of the mesoscale measuring point spacing function involves multiplying the reference mesoscale spacing by the ratio of the measured maximum deviation value of the coarse-scale layer to the reference deviation value, and then multiplying it by the ratio of the reference stiffness parameter to the tunnel stiffness parameter.

[0011] Specifically, the calculation of the fine-scale measurement point spacing function involves multiplying the baseline fine-scale spacing by the ratio of the measured maximum deviation value of the mesoscale layer to the baseline fine-scale deviation value, and then multiplying it by the ratio of the baseline region length to the critical region length.

[0012] Specifically, the calculation of the coarse adjustment amount involves subtracting the measured axis coordinates of each measuring point in the coarse-scale layer from the design axis coordinates to obtain the deviation vector. The deviation vector is then subjected to global optimization adjustment calculation, and the accumulated error is evenly distributed to each section of the tunnel to obtain the vertical and horizontal adjustment amounts of each support point.

[0013] The global optimization adjustment calculation adopts an indirect adjustment model, which establishes a set of observation equations and a set of condition equations. The set of observation equations includes position observation equations, elevation observation equations and azimuth observation equations. The measurement errors and installation errors of each segment are used as observation values, and the optimal estimates of the unknown parameters are solved by the least squares criterion.

[0014] The Jacobian mapping matrix is ​​a matrix describing the sensitivity coefficient of each support point adjustment to the hole pose change, and the matrix elements are calculated by the finite difference method.

[0015] Specifically, the Jacobian mapping matrix is ​​established by applying a unit adjustment to each support point, recording the change in the hole's pose parameters, dividing the change in the hole's pose parameters by the unit adjustment to obtain the sensitivity coefficient, filling the sensitivity coefficient into the corresponding position in the Jacobian mapping matrix, and obtaining the complete Jacobian mapping matrix after traversing all support points.

[0016] The matrix pseudo-inverse operation uses the singular value decomposition method to handle the ill-conditioned problem of the matrix. A damping factor is set in the singular value decomposition process to avoid over-adjustment caused by excessive adjustment. The adjustment amount of each support point obtained by solving is the medium adjustment amount.

[0017] The preset thresholds include preset thresholds for the test section area and preset thresholds for the non-test section area. The preset threshold for the test section area is determined through test section accuracy tests, and the preset threshold for the non-test section area is twice the preset threshold for the test section area. When returning to the mesoscale layer, the measurement data of the fine-scale layer is retained for the next iteration.

[0018] The distributed temperature sensing network includes temperature sensors arranged on the outer surface of the cave and temperature sensors arranged on the inner surface of the cave. The spacing between the temperature sensors on the outer surface of the cave is calculated using an outer surface sensor spacing function, and the spacing between the temperature sensors on the inner surface of the cave is calculated using an inner surface sensor spacing function.

[0019] The magnitude of the reverse torque is determined by a set of torque calculation equations, which includes a torsion angle difference equation and a reverse torque equation. The torsion angle difference equation is used to calculate the torsion angle deviation based on the measured torsion angle and the allowable value, and the reverse torque equation is used to calculate the magnitude of the reverse torque based on the torsion angle deviation and the torsional stiffness of the corner segment.

[0020] This invention establishes a finite element model coupling the temperature field and the deformation field, and uses a distributed temperature sensing network to collect temperature data in real time as the boundary condition for the thermal load. It quantifies the thermal deformation caused by temperature and incorporates it into the calculation of the axis deviation, eliminating the interference of the temperature field on measurement accuracy and achieving accurate acquisition of the true pose of the tunnel. This invention designs a three-layer, multi-scale, iterative leveling method (coarse, medium, and fine), which dynamically adjusts the spacing between measuring points to adapt to the accuracy requirements of different adjustment stages. The coarse-scale layer quickly identifies the overall deviation trend and completes large-scale coarse adjustments. The medium-scale layer establishes a Jacobian mapping matrix to accurately correlate the pose of each segment with the adjustment amount of the support points. The fine-scale layer performs high-density measurements in key areas to ensure the accuracy of the test section meets the standards. This solves the problem of traditional single-scale methods struggling to balance efficiency and accuracy. Simultaneously, a closed-loop control system consisting of a torque sensor and a hydraulic jack is introduced in the corner section to monitor and dynamically balance the torsion angle in real time. In summary, this invention eliminates thermal deformation interference through temperature field deformation field coupling analysis, achieves efficient and accurate leveling through multi-scale layered iteration, and solves torsional coupling problems through torsional closed-loop control of corner segments. It solves the technical problem that it is difficult to achieve full-scale accurate coaxiality control of the wind tunnel circuit under the coupling effect of multiple physical fields. 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 the overall structure of the wind tunnel circuit in the embodiment.

[0023] Figure 3 This is a partial structural diagram of the multi-scale hierarchical measurement system in the embodiment.

[0024] Figure 4 This is a schematic diagram of the distributed temperature sensing network component structure in the embodiment.

[0025] Figure 5 This is a schematic diagram of the component structure of the torsion control system for the corner segment in the embodiment. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below.

[0027] like Figure 1 As shown, the present invention provides a comprehensive control method for the coaxiality of the wind tunnel loop, comprising:

[0028] S10. Establish a three-dimensional solid finite element model based on the wind tunnel circuit design drawings, input the geometric parameters and material property parameters of each section of the tunnel into the model, and set the initial measurement point layout scheme for the coarse-scale layer.

[0029] S20. A distributed temperature sensing network is arranged on the surface and inside the wind tunnel to collect temperature data at each measuring point in real time. The temperature data is used as the thermal load boundary condition and input into the finite element model to perform coupled analysis and calculation of temperature field and deformation field.

[0030] S30. The multi-scale layered iterative leveling method is used to measure and adjust the tunnel loop in layers. First, the overall axis deviation is measured and the coarse adjustment amount of each support point is calculated in the coarse-scale layer. After the coarse adjustment is performed, the measurement point density is increased in the medium-scale layer for segmented measurement.

[0031] S40. Identify the pose deviation of each section of the cave body at the mesoscale level, establish the Jacobian mapping matrix between the rigid body pose parameters of the cave body and the adjustment of each support point, solve the adjustment of the support point through matrix pseudo-inverse operation and perform the adjustment.

[0032] S50. Perform high-density measurements on key areas in the fine-scale layer. When the measurement point deviation exceeds the preset threshold, return to the mesoscale layer for readjustment. When the deviation of all measurement points is less than the preset threshold, enter the corner segment torsion control.

[0033] S60. Install torque sensors and diagonally arranged hydraulic jacks in the corner section of the wind tunnel to monitor the torsion angle data of the corner section in real time. When the torsion angle exceeds the allowable value, apply reverse torque through the hydraulic jacks to perform dynamic balance adjustment.

[0034] In step S10, the initial measurement point arrangement scheme for the coarse-scale layer involves setting measurement points on the tunnel loop according to the spacing between the coarse-scale layer measurement points. The spacing between the coarse-scale layer measurement points is calculated using a coarse-scale layer measurement point spacing function. This function is used to determine the spacing between the measurement points of the coarse-scale layer based on the tunnel diameter and tunnel wall thickness. The input includes the tunnel diameter and tunnel wall thickness, and the output is the spacing between the coarse-scale layer measurement points. The number of coarse-scale layer measurement points is the value obtained by dividing the entire loop length by the spacing between the coarse-scale layer measurement points and rounding it up.

[0035] The calculation process of the coarse-scale measuring point spacing function is to multiply the ratio of the tunnel diameter to the reference diameter by the ratio of the tunnel wall thickness to the reference wall thickness, and then multiply by the reference measuring point spacing. The reference diameter is 5 meters, the reference wall thickness is 0.02 meters, the reference measuring point spacing is 18 meters, and the value range of the coarse-scale layer measuring point spacing is 15 to 20 meters.

[0036] The reference diameter, reference wall thickness, and reference measuring point spacing are determined through a standard tunnel test. The standard tunnel test includes constructing a standard tunnel test section with a diameter of 5 meters, a wall thickness of 0.02 meters, and a length of 100 meters. Measuring point groups are arranged on the standard tunnel test section at intervals of 10 meters, 15 meters, 18 meters, 20 meters, and 25 meters, respectively. The axial deviation data of each measuring point group is measured using a three-dimensional laser scanner. The measurement coverage rate and error propagation coefficient of each measuring point group are calculated. The maximum spacing with a measurement coverage rate greater than 90% and an error propagation coefficient less than 1.2 is selected as the reference measuring point spacing. The measurement coverage rate is the reciprocal of the ratio of the maximum deviation in the area between measuring points to the deviation at the measuring point. The error propagation coefficient is the ratio of the cumulative error between adjacent measuring points to the single-point measurement error.

[0037] The temperature field-deformation field coupling analysis calculation in step S20 includes first solving the steady-state temperature field distribution to obtain the temperature values ​​of each node, then applying the temperature values ​​of each node as body loads to the structural field to calculate the thermal stress distribution and thermal deformation distribution, and extracting the displacement of key nodes on the tunnel axis.

[0038] In step S30, the multi-scale layered iterative leveling method includes three layers: a coarse-scale layer, a meso-scale layer, and a fine-scale layer. The spacing between measuring points in the meso-scale layer is calculated using a meso-scale measuring point spacing function. This function is used to determine the spacing of measuring points in the meso-scale layer based on the maximum measured deviation value of the coarse-scale layer and the stiffness parameters of the cavity. The input includes the maximum measured deviation value of the coarse-scale layer and the stiffness parameters of the cavity, and the output is the spacing between measuring points in the meso-scale layer. Similarly, the spacing between measuring points in the fine-scale layer is calculated using a fine-scale measuring point spacing function. This function is used to determine the spacing between measuring points in the fine-scale layer based on the maximum measured deviation value of the meso-scale layer and the length of the key region. The input includes the maximum measured deviation value of the meso-scale layer and the length of the key region, and the output is the spacing between measuring points in the fine-scale layer. The adjustment amount for each layer is calculated using a weighted least squares fitting method.

[0039] The calculation process for the mesoscale measuring point spacing function involves multiplying the reference mesoscale spacing by the ratio of the maximum measured deviation value of the coarse-scale layer to the reference deviation value, and then multiplying by the ratio of the reference stiffness parameter to the cavity stiffness parameter. The reference mesoscale spacing is 6.5 meters, the reference deviation value is 5 millimeters, and the reference stiffness parameter is... The mesoscale layer measurement point spacing ranges from 5 to 8 meters.

[0040] The calculation process of the fine-scale measuring point spacing function is as follows: multiply the reference fine-scale spacing by the ratio of the measured maximum deviation value of the mesoscale layer to the reference fine-scale deviation value, and then multiply by the ratio of the length of the reference area to the length of the key area. The reference fine-scale spacing is 1.5 meters, the reference fine-scale deviation value is 2 millimeters, the length of the reference area is 30 meters, and the value range of the fine-scale layer measuring point spacing is 1 to 2 meters.

[0041] Wherein, the tunnel stiffness parameter is the ratio of the tunnel section bending stiffness to the tunnel length, the tunnel section bending stiffness is the elastic modulus of the tunnel material multiplied by the tunnel section moment of inertia, and the critical region length is the sum of the length of the wind tunnel test section and the length of the connecting sections before and after the test section.

[0042] The reference mesoscale spacing, reference deviation value, reference stiffness parameter, reference fine-scale spacing, reference fine-scale deviation value, and reference region length are determined through multi-scale leveling tests. The multi-scale leveling tests involve artificially applying different deviations (2 mm, 5 mm, 8 mm, and 10 mm) to a standard tunnel test section. For each deviation, different measuring point spacings are used for measurement and leveling. The number of leveling iterations and the final accuracy are recorded. The measuring point spacing with less than 5 leveling iterations and a final accuracy of less than 0.5 mm is selected as the reference spacing. The relationship curve between the deviation value and the spacing is established through data fitting.

[0043] The method for calculating the coarse adjustment in step S30 is to obtain the deviation vector by subtracting the measured axis coordinates of each measuring point in the coarse-scale layer from the design axis coordinates, perform global optimization adjustment calculation on the deviation vector, and distribute the cumulative error evenly to each section of the tunnel to obtain the vertical and horizontal adjustment amounts of each support point.

[0044] The global optimization adjustment calculation adopts an indirect adjustment model, establishing a set of observation equations and a set of condition equations. The set of observation equations includes position observation equations, elevation observation equations, and azimuth observation equations. The position observation equations are used to calculate the plane position correction based on the measured plane coordinates and design plane coordinates of the measuring points. The inputs include the measured plane coordinates and design plane coordinates of the measuring points, and the output is the plane position correction. The elevation observation equations are used to calculate the elevation correction based on the measured elevation and design elevation of the measuring points. The inputs include the measured elevation and design elevation of the measuring points, and the output is the elevation correction. The azimuth observation equations are used to calculate the azimuth correction based on the measured azimuth angle and design azimuth angle of the line connecting adjacent measuring points. The inputs include the measured azimuth angle and design azimuth angle of the line connecting adjacent measuring points, and the output is the azimuth correction. The measurement error and installation error of each segment are used as observation values, and the optimal estimate of the unknown parameters is solved by the least squares criterion to achieve a reasonable distribution of errors in each segment.

[0045] In step S40, the Jacobi mapping matrix is ​​a matrix describing the sensitivity coefficient of each support point adjustment to the hole pose change. The number of rows in the matrix is ​​the number of hole pose degrees of freedom, the number of columns in the matrix is ​​the number of support points, and the matrix elements are calculated by the finite difference method.

[0046] The process of establishing the Jacobian mapping matrix includes applying a unit adjustment amount of 1 mm to each support point, recording the change in the hole's pose parameters, dividing the change in the hole's pose parameters by the unit adjustment amount to obtain a sensitivity coefficient, filling the sensitivity coefficient into the corresponding position of the Jacobian mapping matrix, and obtaining the complete Jacobian mapping matrix after traversing all support points.

[0047] In step S40, the matrix pseudo-inverse operation uses singular value decomposition to handle the ill-conditioned problem of the matrix. A damping factor is set during the singular value decomposition process. This damping factor is determined through a damping factor selection experiment, which includes leveling calculations on a standard tunnel test section using different damping factors. These different damping factors are 0.01, 0.02, 0.03, 0.04, and 0.05. The adjustment amplitude and leveling accuracy corresponding to each damping factor are recorded. A damping factor with an adjustment amplitude of less than 5 mm and a leveling accuracy of less than 1 mm is selected as the recommended value. The value range of the damping factor is 0.01 to 0.05 to avoid over-adjustment due to excessive adjustment. The adjustment amount obtained for each support point is the intermediate adjustment amount.

[0048] In step S50, the preset threshold is determined based on the wind tunnel coaxiality design requirements. The preset threshold includes a preset threshold for the test section area and a preset threshold for the non-test section area. The preset threshold for the test section area is determined through a test section accuracy test. The test section accuracy test includes airflow quality testing on a test section simulator, using different coaxiality deviations for comparative testing. The different coaxiality deviations are 0.3 mm, 0.5 mm, 0.8 mm, 1.0 mm, and 1.5 mm, respectively. The airflow velocity uniformity and turbulence are measured, and the maximum deviation with airflow velocity uniformity greater than 98% and turbulence less than 0.1% is selected as the preset threshold for the test section area. The value range of the preset threshold for the test section area is 0.5 to 1.0 mm. The preset threshold for the non-test section area is twice the preset threshold for the test section area, and the value range of the preset threshold for the non-test section area is 1.0 to 2.0 mm. When returning to the mesoscale layer, the measurement data of the fine-scale layer is retained for the next iteration.

[0049] In step S60, the corner section is the tunnel section in the wind tunnel loop where the airflow direction changes. The torque sensor is installed at the connection between the corner section tunnel and the support structure. The hydraulic jacks are arranged diagonally in the four quadrants of the corner section tunnel. The dynamic balance of torque is achieved through a closed-loop control system.

[0050] In step S60, the allowable value is the maximum allowable torsional angle of the corner section of the tunnel. The allowable value is determined by a torsion test of the corner section. The torsion test of the corner section includes applying different torque loads to the corner section simulation component. The different torque loads are 100 Nm, 200 Nm, 300 Nm, 400 Nm and 500 Nm, respectively. The corresponding torsional angle and stress distribution are measured. The torsional angle corresponding to the stress value being less than 80% of the allowable stress of the material is selected as the allowable value. The range of the allowable value is 0.02 to 0.05 degrees. When the measured torsional angle exceeds the allowable value, the control system automatically starts the hydraulic jack to apply the reverse torque.

[0051] The distributed temperature sensing network includes temperature sensors arranged on the outer surface of the cave and temperature sensors arranged on the inner surface of the cave. The spacing between the temperature sensors on the outer surface of the cave is calculated using an outer surface sensor spacing function. This function determines the spacing of the outer surface temperature sensors based on the outer surface area of ​​the cave and the rate of change of the ambient temperature. The inputs include the outer surface area of ​​the cave and the rate of change of the ambient temperature, and the output is the spacing of the temperature sensors on the outer surface of the cave. The spacing between the temperature sensors on the inner surface of the cave is calculated using an inner surface sensor spacing function. This function determines the spacing of the inner surface temperature sensors based on the inner surface area of ​​the cave and the temperature gradient inside the cave. The inputs include the inner surface area of ​​the cave and the temperature gradient inside the cave, and the output is the spacing of the temperature sensors on the inner surface of the cave. All temperature sensors are connected to a control computer through a data acquisition system.

[0052] The calculation process of the outer surface sensor spacing function is as follows: multiply the reference outer surface spacing by the ratio of the outer surface area of ​​the cavity to the reference outer surface area, and then divide by the ratio of the ambient temperature change rate to the reference temperature change rate. The reference outer surface spacing is 2.5 meters, the reference outer surface area is 500 square meters, the reference temperature change rate is 5°C per hour, and the value range of the outer surface temperature sensor spacing is 2 to 3 meters.

[0053] The calculation process of the inner surface sensor spacing function is as follows: multiply the reference inner surface spacing by the ratio of the inner surface area of ​​the cave to the reference inner surface area, and then divide by the ratio of the temperature gradient inside the cave to the reference inner temperature gradient. The reference inner surface spacing is 4 meters, the reference inner surface area is 480 square meters, the reference inner temperature gradient is 2℃ per meter, and the value range of the inner surface temperature sensor spacing is 3 to 5 meters.

[0054] The reference outer surface spacing, reference outer surface area, reference temperature change rate, reference inner surface spacing, reference inner surface area, and reference internal temperature gradient are determined through temperature field monitoring experiments. The temperature field monitoring experiments include arranging temperature sensor networks of different densities on a standard cave test section, namely, a 1-meter spacing network, a 2.5-meter spacing network, a 4-meter spacing network, and a 6-meter spacing network, continuously monitoring the temperature field distribution data for 24 hours, calculating the temperature field reconstruction error and the thermal deformation prediction error, and selecting the maximum spacing with a temperature field reconstruction error of less than 0.5℃ and a thermal deformation prediction error of less than 0.2 mm as the reference spacing.

[0055] The weighted least squares fitting method assigns different weight coefficients to the deviation values ​​of each measuring point. These weight coefficients include weight coefficients for measuring points in the test section area and weight coefficients for measuring points in the non-test section area. The weight coefficients for measuring points in the test section area are determined through weight coefficient experiments. These experiments include performing leveling calculations on the test section area and the non-test section area using different weight ratios: 1.2:1, 1.5:1, 1.8:1, 2.0:1, and 2.5:1. The accuracy of the test section and the overall leveling efficiency corresponding to each weight ratio are calculated. A weight ratio with an accuracy of less than 0.5 mm and an overall leveling efficiency greater than 85% is selected. The value range of the weight coefficients for measuring points in the test section area is 1.5 to 2.0, and the value range of the weight coefficients for measuring points in the non-test section area is 0.8 to 1.0. The optimal adjustment amount is solved by minimizing the weighted sum of squared residuals.

[0056] Among them, the key nodes are the nodes on the tunnel axis that have the greatest impact on coaxiality, including the center points of the connecting flanges of each tunnel section, the center points of the corner sections, and the center points of the inlet and outlet of the test section. The displacement of the key nodes includes the displacement component along the axis and the displacement component perpendicular to the axis.

[0057] The rigid body pose parameters include three translational degrees of freedom and three rotational degrees of freedom of the cavity. The three translational degrees of freedom are displacements along the three coordinate axes, and the three rotational degrees of freedom are rotation angles around the three coordinate axes. The rigid body pose parameters are obtained by measuring with a three-dimensional laser scanner or a total station.

[0058] The closed-loop control system includes a torque sensor, a controller, and a hydraulic jack. The torque sensor transmits the measured torsion angle signal to the controller. The controller calculates the magnitude of the reverse torque based on the difference between the measured torsion angle and the allowable value, and outputs a control signal to drive the hydraulic jack to move, thus forming a measurement-calculation-execution closed loop.

[0059] The magnitude of the reverse torque is determined by a set of torque calculation equations, which includes a torsion angle difference equation and a reverse torque equation. The torsion angle difference equation is used to calculate the torsion angle deviation based on the measured torsion angle and the allowable value. The input includes the measured torsion angle and the allowable value, and the output is the torsion angle deviation. The reverse torque equation is used to calculate the magnitude of the reverse torque based on the torsion angle deviation and the torsional stiffness of the corner segment. The input includes the torsion angle deviation and the torsional stiffness of the corner segment, and the output is the magnitude of the reverse torque. The torsional stiffness of the corner segment is the shear modulus of the corner segment material multiplied by the polar moment of inertia of the corner segment section and then divided by the length of the corner segment.

[0060] The specific implementation methods of the above steps are described in detail below.

[0061] The specific implementation of step S10 involves extracting the diameter, wall thickness, length, and material properties of each section of the wind tunnel from the geometric information in the wind tunnel loop design drawings. The extracted geometric and material properties are then imported into finite element analysis software to establish a three-dimensional solid element model. In this model, the mesh density and element type are defined, completing the initial construction of the finite element model. The spacing between coarse-scale measuring points is calculated using a coarse-scale measuring point spacing function. The input parameters of this function are the tunnel diameter and wall thickness, and the output parameter is the coarse-scale layer measuring point spacing. The function calculates the ratio between the actual tunnel parameters and the reference tunnel parameters based on the similarity principle. It divides the entire loop length by the calculated spacing between coarse-scale layer measuring points and rounds up to obtain the number of coarse-scale layer measuring points. The coarse-scale layer measuring points are then arranged on the tunnel loop axis according to the principle of equal spacing, thus completing the initial layout scheme of the coarse-scale layer measuring points. The reference range for the spacing between the coarse-scale layer measuring points is 15 to 20 meters. The establishment process uses the discretization principle to transform the continuous tunnel structure into a finite set of nodes and elements, providing a numerical calculation basis for subsequent mechanical analysis and deformation calculation.

[0062] The specific implementation of step S20 involves calculating the spacing between temperature sensors on the outer and inner surfaces of the cave based on the outer surface sensor spacing function and the inner surface sensor spacing function, respectively. The input parameters of the outer surface sensor spacing function are the outer surface area of ​​the cave and the rate of change of ambient temperature, and the output parameter is the spacing between the outer surface temperature sensors. The input parameters of the inner surface sensor spacing function are the inner surface area of ​​the cave and the internal temperature gradient of the cave, and the output parameter is the spacing between the inner surface temperature sensors. Temperature sensors are installed on the outer and inner surfaces of the cave according to the calculated spacing values ​​to form a distributed temperature sensing network. Real-time temperature data from each temperature sensor is collected at fixed time intervals through a data acquisition system, and the collected temperature data is mapped to a coordinate system according to spatial coordinate correspondence. Temperature field boundary conditions are formed at the node locations of the finite element model. A thermal analysis module is set up in the finite element analysis software, and the thermal expansion coefficient and thermal conductivity of the material are input. Steady-state thermal analysis calculations are performed to obtain the temperature distribution results at each node of the cavity. The temperature distribution results are then applied as a body load to the structural mechanics analysis module to solve for the spatial distribution of the thermal stress field and thermal deformation field. The three-dimensional displacement components of key nodes along the cavity axis are extracted. The temperature field-deformation field coupling analysis uses a sequential coupling method to decouple and solve the thermal and mechanical problems. The reference range for the spacing between the outer surface temperature sensors is 2 to 3 meters, and the reference range for the spacing between the inner surface temperature sensors is 3 to 5 meters. This coupling analysis method can accurately predict the deformation trend of the cavity caused by the temperature gradient, providing a theoretical basis for subsequent compensation and adjustment.

[0063] The specific implementation of step S30 involves using a 3D laser scanner or total station to measure the actual spatial coordinates of each measuring point at the pre-arranged locations on the coarse-scale layer. The measured coordinates are then compared point-by-point with the theoretical coordinates in the design drawings to calculate the deviation vector for each measuring point. This deviation vector includes deviation components along the three coordinate axes. An observation equation set for the indirect adjustment model is established, comprising position observation equations describing planar positional relationships, elevation observation equations describing elevation relationships, and azimuth observation equations describing directional relationships. The deviation vectors of each measuring point are input into the observation equation set as observed values. The least squares criterion is used to solve the observation equation set to obtain the optimal pose correction for each section of the tunnel. Based on the optimal pose correction, the values ​​of each support point are calculated. The required vertical and horizontal adjustments are called coarse adjustments. Coarse adjustments are made to the elevation and position of each support point. After the coarse adjustments are completed, the density of the measuring points is increased and the process moves to the mesoscale layer. In the mesoscale layer, new measuring points are arranged according to the spacing values ​​calculated by the mesoscale measuring point spacing function. The input parameters of the mesoscale measuring point spacing function are the maximum deviation value measured in the coarse-scale layer and the stiffness parameters of the tunnel body. The output parameter is the spacing of the measuring points in the mesoscale layer. Actual measurements are taken at the measuring point locations in the mesoscale layer, and the deviation data is recorded. The reference range for the spacing of the measuring points in the mesoscale layer is 5 to 8 meters. The multi-scale layering strategy adopts a progressive refinement approach from coarse to fine to reduce the workload of measurement and computational complexity. The least squares criterion can reasonably distribute the accumulated error to each section of the tunnel body to achieve global optimal adjustment.

[0064] The specific implementation of step S40 involves identifying the deviation of each segment of the tunnel relative to its designed pose at the mesoscale layer. This deviation includes displacement deviations in three translational degrees of freedom and angular deviations in three rotational degrees of freedom. A Jacobi mapping matrix is ​​established to describe the mapping relationship between the support point adjustment and the tunnel pose change. The Jacobi mapping matrix is ​​established using the finite difference method. A unit adjustment of 1 mm is applied to each support point, and the response change of the tunnel pose parameters is recorded. The response change is divided by the unit adjustment to obtain the sensitivity coefficient, which is then filled into the corresponding position in the matrix. After traversing all support points, a complete Jacobi mapping matrix is ​​formed. The number of rows in the matrix is ​​equal to the number of tunnel pose degrees of freedom (6), and the number of columns in the matrix is ​​equal to the number of support points. The method employs singular value decomposition (SVD) to perform pseudo-inverse operations on the Jacobian mapping matrix. A damping factor is introduced during SVD to address ill-conditioned matrix conditions. The damping factor has a reference range of 0.01 to 0.05. The difference between the measured pose deviation at the mesoscale layer and the target pose is calculated to obtain the pose error vector. The optimal adjustment combination for each support point is obtained by left-multiplying the pose error vector with the pseudo-inverse matrix; this is the mid-adjustment amount. The mid-adjustment operation completes the adjustment of the mesoscale layer. The Jacobian matrix method transforms the multi-point collaborative adjustment problem into a linear algebra problem. The SVD method effectively handles matrix rank deficiency and numerical instability. The introduction of the damping factor avoids oscillations caused by excessively large adjustment amounts.

[0065] The specific implementation of step S50 involves calculating the spacing of measurement points in the fine-scale layer based on the fine-scale measurement point spacing function. The input parameters of this function are the maximum measured deviation value of the mesoscale layer and the length of the critical region. The output parameter is the spacing of measurement points in the fine-scale layer. High-density measurement points are arranged in the critical region according to the calculated spacing. The critical region includes the test section region and the corner section region. High-precision measuring equipment is used to measure the spatial coordinates of each measurement point in the fine-scale layer. The deviation value between the measured coordinates and the design coordinates of each measurement point is calculated. The deviation value is compared with a preset threshold. The preset threshold reference range for the test section region is 0.5 to 1.0 mm, and the preset threshold reference range for the non-test section region is... The measurement range is 1.0 to 2.0 mm. When the deviation of a measurement point exceeds the preset threshold, all measurement data of the current fine-scale layer are retained and the system returns to the mesoscale layer. In the mesoscale layer, the Jacobian mapping matrix is ​​re-established and the adjustment amount is re-solved by combining the measurement data of the fine-scale layer. After the adjustment is performed, the system re-enters the fine-scale layer for measurement verification. The above iterative process is repeated until the deviation of all measurement points is less than the preset threshold. When all measurement points meet the accuracy requirements, the iteration is terminated and the system enters the corner segment torsion control step. The reference range of the measurement point spacing in the fine-scale layer is 1 to 2 meters. The iterative adjustment strategy is based on the closed-loop feedback control principle and gradually converges to the target accuracy. The threshold judgment mechanism ensures that different regions meet the differentiated accuracy requirements.

[0066] The specific implementation of step S60 involves installing torque sensors at the connection point between the wind tunnel's corner section and the supporting structure. Hydraulic jacks are diagonally positioned at the four quadrants of the corner section. The torque sensors collect the torsion angle signal of the corner section in real time and transmit it to the controller. The controller calculates the deviation between the measured torsion angle and the allowable value based on the torsion angle difference equation. The input parameters of the torsion angle difference equation are the measured torsion angle and the allowable value, and the output parameter is the torsion angle deviation. The torsion angle deviation is then input into the reverse torque equation to calculate the magnitude of the required reverse torque. The input parameters of the reverse torque equation are the torsion angle deviation and the torsional stiffness of the corner section, and the output parameter is the reverse torque. The torsional stiffness of the corner segment is calculated by multiplying the material's shear modulus and the polar moment of inertia of the cross section by the length of the corner segment. The controller generates a control signal based on the calculated reverse torque to drive the hydraulic jack. The hydraulic jack applies reverse torque to compensate for the torsion of the corner segment, forming a closed-loop control system that includes torque sensor measurement, controller calculation, and hydraulic jack execution. When the measured torsion angle is less than or equal to the allowable value, the application of reverse torque is stopped. The allowable value has a reference range of 0.02 to 0.05 degrees. The closed-loop control method is based on the feedback control principle to achieve dynamic balance of the torsion of the corner segment. The diagonal arrangement can provide symmetrical torsional torque to maintain axial stability.

[0067] It should be noted that the key technical ideas of this invention include a multi-scale hierarchical iterative leveling strategy and a temperature field-deformation field coupled prediction mechanism. The multi-scale hierarchical iterative leveling strategy divides the entire wind tunnel loop into three scale levels—coarse, medium, and fine—to achieve progressively refined adjustments from the overall to the local. Compared with the traditional single-scale global leveling method, this strategy quickly determines the overall trend at the coarse scale level, significantly reducing the number of initial measurement points. At the medium scale level, it performs segmented adjustments for deviations in each segment to avoid the dimensionality explosion problem of global calculations. At the fine scale level, it performs high-density measurements only on key areas to reduce measurement costs. The hierarchical strategy uses weighted least squares fitting to assign differentiated weight coefficients at different scale levels to ensure accuracy in key areas. Prioritizing temperature, this method gradually converges from coarse to fine through an iterative process, avoiding the oscillations caused by measurement noise and adjustment coupling in traditional methods. This significantly reduces the number of adjustments and computational complexity. The temperature field-deformation field coupling prediction mechanism collects cave temperature distribution data in real time through a distributed temperature sensing network. The temperature data is then used as a thermal load input to the finite element model for thermal-structural sequential coupling analysis. Compared to traditional reactive adjustment methods, this mechanism can predict the thermal deformation trend caused by temperature gradients in advance and calculate the compensation amount, achieving proactive control to avoid axial drift caused by accumulated thermal deformation. The coupling analysis method considers the combined effects of the material's thermal expansion coefficient and structural constraints to improve the accuracy of deformation prediction.

[0068] The synergistic effect of the multi-scale hierarchical iterative leveling strategy and the temperature field-deformation field coupled prediction mechanism lies in the fact that the former solves the problem of error propagation and accumulation in the spatial domain, while the latter solves the problem of thermal deformation prediction and compensation in the time domain. The combination of the two forms a spatial-temporal dual-domain collaborative control system. In the multi-scale leveling process, the deformation prediction results of the temperature field coupled analysis are introduced as additional constraints for the calculation of the adjustment amount. This ensures that the adjustment amount not only considers the current measured deviation but also reserves the future deformation amount caused by temperature changes, avoiding the degradation of accuracy caused by temperature changes after leveling. In the temperature compensation process, the multi-scale hierarchical measurement point data is used to verify the prediction accuracy of the finite element model and correct the model parameters, forming a two-way feedback closed loop between the measurement data and the simulation model. Compared with the traditional single control method, the collaborative mechanism achieves the comprehensive control of the coaxiality of the wind tunnel loop through the hierarchical optimization in the spatial domain and the prediction compensation in the time domain. This fundamentally solves the problem of coaxiality control under the coupled effects of multiple factors such as error propagation, thermal deformation, and gravity deformation in the multi-segment splicing of long-distance wind tunnels.

[0069] It should be noted that this invention also solves the following technical problem: the problem of cumulative error propagation leading to distortion of the adjustment scheme during wind tunnel loop measurement. Traditional methods, when handling long loop measurements, see errors at each measuring point accumulate and propagate along the measurement path. The deviation data of subsequent measuring points includes the measurement errors of previous measuring points, resulting in a systematic deviation in the adjustment amount of the support points calculated based on this data. This leads to a significant difference between the actual coaxiality after adjustment and the expected result. This invention employs a global optimization adjustment calculation method, establishing an indirect adjustment model that includes position observation equations, elevation observation equations, and azimuth observation equations. It uses the measurement errors and installation errors of each segment as observation values, and solves for the optimal estimates of unknown parameters using the least squares criterion. This achieves a reasonable distribution of errors across different tunnel segments, avoiding the accumulation of errors in a single direction. Simultaneously, a weighted least squares fitting method is introduced to assign higher weight coefficients to the measuring points in the test section area, prioritizing the adjustment accuracy of key areas, effectively solving the problem of cumulative error propagation causing distortion of the adjustment scheme.

[0070] Furthermore, this invention solves the technical problem of the difficulty in independently controlling the coupled torsional deformation and axial deviation in corner sections. In the wind tunnel circuit, corner sections experience uneven loads due to airflow deflection, resulting in both torsional deformation and axial deviation. Traditional methods only adjust the axial direction, failing to effectively suppress torsional deformation. This leads to situations where, although the adjusted axial direction meets the standard, the torsional angle exceeds the limit, or vice versa, but the axial deviation rebounds. This invention, by installing torque sensors in the corner section to monitor torsional angle data in real time, automatically activates diagonally arranged hydraulic jacks to apply reverse torque for dynamic balancing adjustment when the torsional angle exceeds the allowable value. This decouples torsional control and axial adjustment into two independent control loops. The torsional control loop compensates for torsional deformation in real time through closed-loop feedback, while the axial adjustment loop focuses on posture correction. The two loops work together to achieve precise coaxiality control in the corner section, solving the technical problem of the difficulty in independently controlling the coupled torsional deformation and axial deviation.

[0071] Specifically, the principle of this invention is as follows: The core technical problem solved by this invention lies in combining multi-physics field coupling analysis with multi-scale control strategies to construct a quantitative transmission path from the temperature field to the deformation field. Real-time data collected by the temperature sensing network is input into the finite element model, and thermal stress and deformation at each node are calculated through thermo-structure coupling. These thermal deformations are superimposed on the measurement deviation and participate in subsequent adjustment calculations, enabling the adjustment scheme to compensate for temperature effects and ensuring the effectiveness of coaxiality control under fluctuating temperature conditions. The principle of the multi-scale layered iterative leveling method is to dynamically adjust the measurement and control accuracy based on the magnitude of the deviation. The coarse-scale layer covers the entire loop with a large spacing to quickly locate the concentrated deviation area. The medium-scale layer establishes a linear mapping relationship between the support point adjustment amount and the hole pose change through the Jacobian matrix, and uses the matrix pseudo-inverse to solve for the optimal adjustment amount to achieve decoupled control. The fine-scale layer with high-density measurement points ensures that key areas meet strict accuracy requirements. The three layers work together to control the cumulative error within the allowable range. The torsional closed-loop control principle of the corner segment is to sense the torsional angle deviation in real time through a torque sensor. The controller calculates the required reverse torque based on the torsional stiffness of the corner segment and drives the diagonally arranged hydraulic jacks to apply a balancing torque, forming a dynamic balance system of measurement feedback adjustment. This effectively suppresses the coupling effect of torsional deformation on the coaxiality of the axis and ensures the positional stability of the corner segment area.

[0072] 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.

[0073] The specific implementation of step S10 involves establishing a three-dimensional solid finite element model based on the wind tunnel loop design drawings. Geometric parameters and material property parameters for each section of the tunnel are input into the model. An initial measurement point layout scheme for the coarse-scale layer is set. The geometric parameters include the tunnel diameter, tunnel wall thickness, tunnel length, and cross-sectional shape. The material property parameters include the elastic modulus, Poisson's ratio, coefficient of thermal expansion, and density. The spacing between the coarse-scale layer measurement points is calculated using a coarse-scale measurement point spacing function, as shown below:

[0074] ;

[0075] In the formula, The distance between measurement points in the coarse-scale layer is in meters. The diameter of the cave is in meters. The reference diameter is set at 5 meters. The thickness of the cave wall is measured in meters. The baseline wall thickness is 0.02 meters. The baseline measurement point spacing is set at 18 meters. The measurement point spacing for the coarse-scale layer ranges from 15 to 20 meters. The formula for calculating the number of measurement points in the coarse-scale layer is as follows:

[0076] ;

[0077] In the formula, The number of measurement points in the coarse-scale layer is dimensionless. The total length of the loop is in meters. This indicates rounding up. The reference diameter, reference wall thickness, and reference measuring point spacing are determined through a standard tunnel test. The standard tunnel test involves constructing a standard tunnel test section with a diameter of 5 meters, a wall thickness of 0.02 meters, and a length of 100 meters. Measuring point groups are arranged on the standard tunnel test section at intervals of 10 meters, 15 meters, 18 meters, 20 meters, and 25 meters, respectively. The axial deviation data of each measuring point group is measured using a 3D laser scanner, and the measurement coverage rate and error propagation coefficient of each measuring point group are calculated. The formula for calculating the measurement coverage rate is:

[0078] ;

[0079] In the formula, For measuring coverage, dimensionless; The deviation at the measuring point is expressed in millimeters. This represents the maximum deviation between measuring points, expressed in millimeters. The formula for calculating the error propagation coefficient is:

[0080] ;

[0081] In the formula, The error propagation coefficient is dimensionless. The number of adjacent measurement point pairs, dimensionless; For the first The cumulative error of each measuring point is expressed in millimeters. This represents the single-point measurement error, expressed in millimeters, with an empirical value of 0.1 millimeters. This refers to the measurement point number. The maximum spacing with a measurement coverage rate greater than 90% and an error propagation coefficient less than 1.2 is selected as the benchmark measurement point spacing.

[0082] The specific implementation of step S20 involves arranging a distributed temperature sensing network on the surface and inside the wind tunnel to collect temperature data at each measuring point in real time. This temperature data is then input into the finite element model as a thermal load boundary condition for coupled temperature-deformation field analysis. This includes first solving for the steady-state temperature field distribution to obtain the temperature values ​​at each node, then applying these node temperature values ​​as body loads to the structural field to calculate the thermal stress and thermal deformation distribution, and extracting the displacement of key nodes along the tunnel axis. Key nodes are those on the tunnel axis that have the greatest impact on coaxiality, including the center points of connecting flanges in each tunnel section, the center points of corner sections, and the inlet and outlet center points of the test section. The displacement of key nodes includes displacement components along the axial direction. and displacement components perpendicular to the axis All units are in millimeters. The spacing between the temperature sensors on the outer surface of the cave is calculated using an external surface sensor spacing function, as shown below:

[0083] ;

[0084] In the formula, The spacing between temperature sensors on the outer surface of the cave is in meters. The reference outer surface spacing is set at 2.5 meters. This refers to the outer surface area of ​​the cave, in square meters. The base external surface area is set at 500 square meters; This represents the rate of change of ambient temperature, expressed in °C per hour. The baseline temperature change rate is set at 5°C per hour. The spacing between the temperature sensors on the outer surface of the cave is between 2 and 3 meters. The spacing between the temperature sensors on the inner surface of the cave is calculated using an inner surface sensor spacing function, as shown below:

[0085] ;

[0086] In the formula, The spacing between temperature sensors on the inner surface of the cave is in meters. The reference inner surface spacing is set to 4 meters; The area is the internal surface area of ​​the cave, in square meters. The baseline internal surface area is set at 480 square meters. This represents the temperature gradient inside the cave, expressed in °C per meter. The baseline internal temperature gradient is set at 2°C per meter. The spacing between temperature sensors on the inner surface of the cave is between 3 and 5 meters. The baseline outer surface spacing, baseline outer surface area, baseline temperature change rate, baseline inner surface spacing, baseline inner surface area, and baseline internal temperature gradient are determined through temperature field monitoring experiments. These experiments involve deploying temperature sensor networks of different densities on a standard cave test section: networks with spacings of 1 meter, 2.5 meters, 4 meters, and 6 meters. Temperature field distribution data is continuously monitored for 24 hours, and the temperature field reconstruction error is calculated. And thermal deformation prediction error Among them, temperature field reconstruction error The root mean square error between the reconstructed temperature field and the measured temperature field, in °C, represents the thermal deformation prediction error. To determine the root mean square error between the predicted and measured deformation values, in millimeters, the maximum spacing with a temperature field reconstruction error of less than 0.5℃ and a thermal deformation prediction error of less than 0.2 mm was selected as the reference spacing.

[0087] The specific implementation of step S30 is to use a multi-scale layered iterative leveling method to perform layered measurement and adjustment of the tunnel loop. First, the overall axial deviation is measured and the coarse adjustment amount of each support point is calculated in the coarse-scale layer. After coarse adjustment, the measurement point density is increased in the meso-scale layer for segmented measurement. The spacing between the measurement points in the meso-scale layer is calculated by the meso-scale measurement point spacing function, as shown below:

[0088] ;

[0089] In the formula, The distance between mesoscale measurement points is in meters. The baseline mesoscale spacing is set at 6.5 meters. This represents the maximum measured deviation value for the coarse-scale layer, in millimeters. The reference deviation value is 5 mm. The reference stiffness parameter has a value of [value missing]. Megapascal; This represents the stiffness parameter of the tunnel body, measured in megapascals (MPA). The spacing between measuring points in the mesoscale layer ranges from 5 to 8 meters. The spacing between measuring points in the fine-scale layer is calculated using a fine-scale measuring point spacing function, as shown below:

[0090] ;

[0091] In the formula, The distance between measurement points in the fine-scale layer is in meters. The baseline fine-scale spacing is set at 1.5 meters. This represents the maximum measured deviation for the mesoscale layer, in millimeters. The reference fine-scale deviation value is set at 2 mm. The length of the critical area is in meters. The baseline area length is set at 30 meters. The spacing between measuring points in the fine-scale layer ranges from 1 to 2 meters. The formula for calculating the tunnel stiffness parameter is:

[0092] ;

[0093] In the formula, The elastic modulus of the cavity material is expressed in megapascals (MPa). The moment of inertia of the tunnel section is expressed in meters. ; This refers to the length of the tunnel section, in meters. The length of the critical area is the length of the wind tunnel test section. The sum of the lengths of the connecting sections before and after the test section, where the length of the connecting sections before and after the test section refers to the length of the connecting section in front of the test section entrance. Length of the connecting section after the test section exit The sum, in meters, is... The benchmark mesoscale spacing, benchmark deviation value, benchmark stiffness parameter, benchmark fine-scale spacing, benchmark fine-scale deviation value, and benchmark region length were determined through multi-scale leveling tests. These tests involved artificially applying different deviations (2 mm, 5 mm, 8 mm, and 10 mm) to a standard tunnel test section. For each deviation, measurements and leveling were performed using different measuring point spacings, and the number of leveling iterations was recorded. and final accuracy The number of leveling iterations The number of adjustment cycles required to achieve the target accuracy, dimensionless, final accuracy. To determine the maximum deviation of the measuring points after leveling, in millimeters, the spacing between measuring points with less than 5 leveling iterations and a final accuracy of less than 0.5 millimeters was selected as the baseline spacing. A curve relating the deviation value to the spacing was established through data fitting. The coarse adjustment was calculated by subtracting the measured axial coordinates and the designed axial coordinates of each measuring point in the coarse-scale layer to obtain the deviation vector. A global optimization adjustment calculation was then performed on the deviation vector, distributing the accumulated error evenly across each section of the tunnel to obtain the vertical adjustment amount for each support point. and horizontal adjustment amount All units are millimeters, with subscripts indicating the units. This indicates the support point number. The global optimization adjustment calculation uses an indirect adjustment model, establishing a set of observation equations and a set of condition equations. The observation equations include location observation equations, elevation observation equations, and azimuth observation equations. The location observation equations are expressed as follows:

[0094] ;

[0095] ;

[0096] In the formula, and For the first The plane position correction for each measuring point, in millimeters; and The measured plane coordinates of the measuring point are in millimeters. and Design plane coordinates, unit: millimeters; and The residuals of the observed values ​​are expressed in millimeters. The measurement point number is specified, with a value range from 1 to... The elevation observation equation is expressed as follows:

[0097] ;

[0098] In the formula, For the first Elevation correction for each measuring point, in millimeters; The measured elevation of the measuring point is in millimeters. Design elevation, in millimeters; The residuals are expressed in millimeters. The azimuth observation equation is as follows:

[0099] ;

[0100] In the formula, For the first The azimuth correction of the line connecting adjacent measuring points, in degrees; The measured azimuth angle is in degrees. The design azimuth angle is expressed in degrees. The residuals of the observed values ​​are expressed in degrees. The segment number has a value range of 1 to 1. The optimal estimates of the unknown parameters are obtained using the least squares criterion. The objective function is:

[0101] ;

[0102] In the formula, The objective function value is dimensionless. The standard deviation of the position observation is 0.5 mm. The standard deviation of the angle observation is set to 0.01 degrees. This ensures a reasonable distribution of error across different segments. The adjustment amount for each layer is calculated using a weighted least squares fitting method, assigning different weighting coefficients to the deviation values ​​of each measuring point. The weighting coefficients for measuring points in the test section area are as follows. The value ranges from 1.5 to 2.0, and the weighting coefficient for measuring points in non-test section areas is... The value of is between 0.8 and 1.0. The optimal adjustment is solved by minimizing the weighted sum of squared residuals. The objective function is:

[0103] ;

[0104] In the formula, These are the weighted objective function values, in millimeters. ; For the first The weighting coefficients for each measurement point are dimensionless. For the first The measured deviation values ​​at each measuring point are in millimeters. For the first The fitting deviation value for each measuring point, in millimeters; The total number of measuring points is dimensionless. The weighting coefficients were determined through weighting coefficient experiments, which included balancing calculations on the test section and non-test section areas using different weight ratios: 1.2:1, 1.5:1, 1.8:1, 2.0:1, and 2.5:1. The accuracy of the test section corresponding to each weight ratio was calculated. And overall leveling efficiency Among them, the accuracy of the test section This represents the maximum deviation of the measuring points in the test section, in millimeters, and the overall leveling efficiency. The ratio of the number of measuring points whose deviation meets the requirements after leveling to the total number of measuring points is expressed as a percentage. The weighting ratio is selected based on the accuracy of the test section being less than 0.5 mm and the overall leveling efficiency being greater than 85%.

[0105] The specific implementation of step S40 involves identifying the pose deviation of each segment of the cavity at the mesoscale layer, establishing a Jacobian mapping matrix between the rigid body pose parameters of the cavity and the adjustment amounts of each support point, solving for the adjustment amounts of the support points through matrix pseudo-inverse operation, and performing the adjustment. The rigid body pose parameters include three translational degrees of freedom and three rotational degrees of freedom of the cavity. The three translational degrees of freedom are displacements along the three coordinate axes. , , The unit is millimeters, and the three rotational degrees of freedom are rotational angles about the three coordinate axes. , , The unit is degrees. The rigid body pose parameters are obtained by measuring with a 3D laser scanner or total station. The Jacobian mapping matrix is ​​a matrix describing the sensitivity coefficients of the adjustment amount of each support point to the pose change of the cavity. The matrix is ​​represented as:

[0106] ;

[0107] In the formula, The Jacobian mapping matrix; The number of support points is dimensionless. Let be the sensitivity coefficient, representing the th . The unit adjustment amount of the support point affects the first The degree of influence of each degree of pose freedom, when The corresponding translational degrees of freedom, the sensitivity coefficient is dimensionless, when The time corresponds to the rotational degrees of freedom, and the sensitivity coefficient is expressed in degrees per millimeter. The process of establishing the Jacobian mapping matrix involves applying a unit adjustment to each support point. The value is 1 mm, and the change in the hole's pose parameters is recorded. The sensitivity coefficient is calculated using the following formula:

[0108] ;

[0109] In the formula, For the first The change in each pose parameter, when The unit of time is millimeters. The unit of time is degrees; Number the degrees of freedom of pose, with values ​​ranging from 1 to 6; The support point is numbered, with a value ranging from 1 to... After traversing all support points, the complete Jacobian mapping matrix is ​​obtained. The matrix pseudo-inverse operation uses singular value decomposition to handle the ill-conditioned problem of the matrix. A damping factor is set during the singular value decomposition process, and the solution formula is:

[0110] ;

[0111] In the formula, Adjustment vectors for each support point, in millimeters; The normalized cave pose deviation vector is dimensionless, calculated by dividing the translational component by the reference displacement. Multiply the rotation component by the reference length. Divide by the reference displacement To achieve normalization, the reference displacement is... The value is 1 mm, and the reference length is... The value is 5 meters; The damping factor is dimensionless and ranges from 0.01 to 0.05. It is an identity matrix, dimensionless; This is the transpose of the Jacobian matrix. The damping factor was determined through a damping factor selection test, which included leveling calculations on a standard tunnel test section using different damping factors: 0.01, 0.02, 0.03, 0.04, and 0.05. The adjustment magnitude corresponding to each damping factor was recorded. and leveling accuracy Among them, the adjustment range This represents the maximum adjustment amount for all support points, in millimeters, indicating the leveling accuracy. The maximum value of the measurement point deviation after leveling is in millimeters. The recommended value is a damping factor with an adjustment range of less than 5 millimeters and a leveling accuracy of less than 1 millimeter, to avoid over-adjustment caused by excessive adjustment.

[0112] The specific implementation of step S50 involves performing high-density measurements on key areas at the fine-scale layer. When the deviation of a measurement point exceeds a preset threshold, the system returns to the mesoscale layer for readjustment. When the deviation of all measurement points is less than the preset threshold, the system enters the torsional control phase of the corner section. The preset threshold is determined based on the wind tunnel coaxiality design requirements, and includes preset thresholds for the test section area. Preset threshold for non-test section area The preset threshold for the test section area was determined through a test section accuracy test. This test included airflow quality testing on a test section simulator, with comparative tests conducted using different coaxiality deviations of 0.3 mm, 0.5 mm, 0.8 mm, 1.0 mm, and 1.5 mm, respectively, to measure airflow velocity uniformity. and turbulence Among them, airflow velocity uniformity The ratio of the standard deviation of the velocity at the test section cross-section to the average velocity, expressed as a percentage, is the turbulence intensity. The ratio of the root mean square value of the fluctuating velocity to the average velocity, expressed as a percentage, is used. The maximum deviation where the airflow velocity uniformity is greater than 98% and the turbulence intensity is less than 0.1% is selected as the preset threshold for the test section area, with a value range of 0.5 to 1.0 mm. The preset threshold for the non-test section area is twice the preset threshold for the test section area, with a value range of 1.0 to 2.0 mm. When returning to the mesoscale layer, the measurement data of the fine-scale layer is retained for the next iteration.

[0113] The specific implementation of step S60 involves installing a torque sensor and diagonally arranged hydraulic jacks in the corner section of the wind tunnel to monitor the torsion angle data in real time. When the torsion angle exceeds the allowable value, the hydraulic jacks apply reverse torque for dynamic balance adjustment. The corner section is the section of the wind tunnel where the airflow direction changes. The torque sensor is installed at the connection between the corner section and the supporting structure, and the hydraulic jacks are diagonally arranged in the four quadrants of the corner section. Dynamic torque balance is achieved through a closed-loop control system. The closed-loop control system includes a torque sensor, a controller, and hydraulic jacks. The torque sensor transmits the measured torsion angle signal to the controller. The controller calculates the magnitude of the reverse torque based on the difference between the measured torsion angle and the allowable value, and outputs a control signal to drive the hydraulic jacks, forming a measurement-calculation-execution closed loop. The allowable value is the maximum allowable torsion angle of the corner section. The torsion test of the corner segment was used to determine the torsion angle. The torsion test involved applying different torque loads (100 Nm, 200 Nm, 300 Nm, 400 Nm, and 500 Nm) to the corner segment simulator and measuring the corresponding torsion angles. and stress distribution The torsion angle The unit is degrees, and the stress distribution is... The maximum stress value on the surface of the corner segment is expressed in megapascals (MPa). The torsional angle corresponding to a stress value less than 80% of the material's allowable stress is selected as the allowable value, ranging from 0.02 to 0.05 degrees. The magnitude of the reverse torque is determined by the torque calculation equations, and the torsional angle difference equation is expressed as follows:

[0114] ;

[0115] In the formula, This refers to the deviation of the torsion angle, in degrees. The measured torsional angle is given in degrees. The equation for the reverse torque is as follows:

[0116] ;

[0117] In the formula, This refers to the magnitude of the reverse torque, measured in Newton-meters (Nm). This represents the torsional stiffness of the corner segment, expressed in Newton-meters per degree (Nm / degree). The formula for calculating the torsional stiffness of the corner segment is:

[0118] ;

[0119] In the formula, This refers to the shear modulus of the material at the corner, expressed in megapascals (MPA). The polar moment of inertia of the corner section is given in meters. ; The measured torsion angle is the length of the turning segment, in meters. When the measured torsion angle exceeds the allowable value, the control system automatically activates the hydraulic jack to apply reverse torque, forming a closed-loop control of measurement-calculation-execution.

[0120] It should be noted that the variables involved in this embodiment are explained in detail in Tables 1 and 2.

[0121] Table 1. Variable Explanation Table (Part 1)

[0122]

[0123] Table 2. Variable Explanation Table (Part Two)

[0124]

[0125] To better understand and implement this invention, a specific application scenario of this invention is provided below as an example 2: Figure 2 As shown, a wind tunnel loop adopts a closed layout, including four straight sections, four corner sections and one test section. The test section is located on the left vertical section and is 30 meters long.

[0126] The technical team first established a three-dimensional solid finite element model based on the wind tunnel loop design drawings, inputting the geometric parameters and material properties of each section of the tunnel into the model. The baseline diameter was 5 meters, the baseline wall thickness was 0.02 meters, and the baseline measuring point spacing was 18 meters. The calculated spacing of the coarse-scale layer measuring points was also 18 meters. A total of 20 coarse-scale layer measuring points were arranged throughout the loop at 18-meter intervals. Figure 4 As shown, a distributed temperature sensor network was deployed on the outer and inner surfaces of the cave. The outer surface area of ​​the cave is 500 square meters, the ambient temperature change rate is 5℃ per hour, and the sensor spacing on the outer surface is 2.5 meters, with a total of 144 outer surface temperature sensors. The inner surface area of ​​the cave is 480 square meters, the internal temperature gradient is 2℃ per meter, and the sensor spacing on the inner surface is 4 meters, with a total of 90 inner surface temperature sensors. All temperature sensors are connected to a control computer via a data acquisition system, with a sampling frequency set to 1 Hz and a temperature measurement accuracy of ±0.1℃.

[0127] The technical team used a 3D laser scanner to conduct a coarse-scale overall measurement of the tunnel's loop. The measurement revealed a maximum axial deviation of 8 mm, occurring in the middle of the bottom straight section. Real-time temperature data was input into the finite element model as thermal load boundary conditions for coupled temperature and deformation field analysis. First, the steady-state temperature field distribution was solved to obtain the temperature values ​​at each node. Then, these temperature values ​​were applied as volume loads to the structural field to calculate the thermal stress and thermal deformation distribution, extracting the displacement of key nodes along the tunnel's axis. The analysis results showed that due to the diurnal temperature difference, a 2°C temperature gradient existed between the outer and inner surfaces of the tunnel, resulting in a 0.3 mm thermal deformation.

[0128] like Figure 3 As shown, the technical team employed a multi-scale, layered, iterative leveling method to perform layered measurements and adjustments on the tunnel loop. After coarse-scale layer measurements, the deviation vector was obtained by subtracting the measured axis coordinates from the design axis coordinates at 20 measuring points. An indirect adjustment model was used to establish a set of observation equations and condition equations, including position observation equations, elevation observation equations, and azimuth observation equations. Through global optimization adjustment calculations, the cumulative error was evenly distributed to each section of the tunnel, resulting in the vertical and horizontal adjustment amounts for each support point. A total of 12 support points were set at the bottom straight section. The calculated maximum adjustment amount was 4.2 mm, and the minimum adjustment amount was 0.8 mm. After coarse adjustment, the overall axis deviation was reduced to 3 mm.

[0129] After transitioning to the mesoscale layer, the maximum measured deviation in the coarse-scale layer was 3 mm, and the cavity stiffness parameter was... The mesoscale measurement point spacing was calculated to be 6.5 meters (MPa). The technical team increased the measurement point density on each section of the tunnel at a 6.5-meter spacing, adding a total of 35 mesoscale measurement points. Segmented measurements were performed on each tunnel section to identify the pose deviation of each section, and a Jacobian mapping matrix was established between the rigid body pose parameters of the tunnel and the adjustment amounts at each support point. The matrix has 6 rows, representing three translational degrees of freedom and three rotational degrees of freedom, and 12 columns, representing the 12 support points on the bottom straight section. A unit adjustment of 1 mm was applied to each support point using the finite difference method, and the changes in the tunnel pose parameters were recorded. The sensitivity coefficient was calculated and filled into the Jacobian mapping matrix. Singular value decomposition was used to address the ill-conditioned nature of the matrix, with a damping factor set to 0.03. The mid-adjustment amount at each support point was solved using matrix pseudo-inverse operations. After mid-adjustment, the axial deviation of each tunnel section was reduced to 1.5 mm.

[0130] The technical team conducted high-density fine-scale layer measurements on key areas of the test section and its connecting sections. The key area was 30 meters long. The maximum deviation measured at the mesoscale layer was 1.5 mm, and the calculated spacing between the fine-scale measurement points was 1.5 meters. A total of 20 fine-scale layer measurement points were deployed in the key area. Measurements revealed that the deviations of 3 measurement points at the entrance of the test section exceeded the preset threshold. The preset threshold for the test section area was 0.8 mm, and for the non-test section area, it was 1.6 mm. The technical team returned to the mesoscale layer for readjustment, retaining the measurement data from the fine-scale layer for the next iteration. A weighted least squares fitting method was used to calculate the adjustment amount. The weight coefficient for the measurement points in the test section area was set to 1.8, and the weight coefficient for the measurement points in the non-test section area was set to 0.9. The optimal adjustment amount was solved by minimizing the weighted sum of squared residuals. After two iterations of adjustment, the deviations of all measurement points were less than the preset thresholds. The coaxiality accuracy of the test section area reached 0.6 mm, and the coaxiality accuracy of the non-test section area reached 1.2 mm.

[0131] like Figure 5 As shown, the technical team installed torque sensors and diagonally arranged hydraulic jacks at the corner section of the wind tunnel. The torque sensors are installed at the connection between the wind tunnel body and the supporting structure at the corner section, monitoring the torsional angle data in real time with a sampling frequency of 10 Hz and an angle measurement accuracy of 0.001 degrees. Four hydraulic jacks are diagonally arranged in the four quadrants of the wind tunnel body at the corner section, each with a maximum output force of 50 kN and a stroke of 10 mm. The shear modulus of the material at the corner section is... The value is MPa, the polar moment of inertia of the cross section is 0.98 m³, the length of the corner section is 8 m, and the torsional stiffness of the corner section is... The torque is measured in Newton-meters per radian. The allowable torsional angle is set to 0.03 degrees. When the measured torsional angle exceeds 0.03 degrees, the control system automatically activates the hydraulic jacks to apply reverse torque. During wind tunnel operation, the torsional angle of the corner segment was monitored to reach 0.035 degrees. Based on the torsional angle deviation of 0.005 degrees and the torsional stiffness of the corner segment, the controller calculated the magnitude of the reverse torque to be 0.84 Newton-meters. The controller outputs a control signal to drive the two diagonally arranged hydraulic jacks to apply reverse torque for dynamic balance adjustment. The torsional angle decreased to 0.025 degrees within 3 seconds, forming a stable closed-loop control.

[0132] Through a week of continuous temperature field monitoring, the technical team discovered that the cavern underwent periodic thermal deformation under the influence of diurnal temperature variations, with a maximum thermal deformation of 0.5 mm. The error between the thermal deformation predicted by the finite element model and the measured value was less than 0.15 mm, and the temperature field reconstruction error was less than 0.4℃, verifying the accuracy of the temperature-deformation field coupling analysis. The technical team updated the finite element model in real time based on the temperature data, calculated the thermal deformation compensation, and compensated for the impact of thermal deformation by adjusting the height of the support points, ensuring that the cavern maintained its designed coaxiality under different temperature conditions.

[0133] 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 comprehensive control of the coaxiality of wind tunnel circuits, characterized in that, include: A three-dimensional solid finite element model was established based on the wind tunnel's loop design drawings. Geometric and material properties of each section of the tunnel were input into the model, and an initial measurement point layout scheme for the coarse-scale layer was set. A distributed temperature sensor network was deployed on the surface and inside the wind tunnel to collect temperature data from each measurement point in real time. This temperature data was used as the thermal load boundary condition and input into the finite element model for coupled temperature-deformation field analysis. A multi-scale layered iterative leveling method was used to perform layered measurements and adjustments to the tunnel loop. First, the overall axial deviation was measured at the coarse-scale layer, and the coarse adjustment amount for each support point was calculated. After coarse adjustment, the process moved to the meso-scale layer to increase the measurement point density and segment the tunnel. Measurement; Identify the pose deviation of each section of the tunnel at the mesoscale level, establish the Jacobian mapping matrix between the rigid body pose parameters of the tunnel and the adjustment of each support point, solve the adjustment of the support point through matrix pseudo-inverse operation and execute the adjustment; Perform high-density measurement on key areas at the fine-scale level, return to the mesoscale level for readjustment when the measurement point deviation exceeds the preset threshold, and enter the torsion control of the corner section when all measurement point deviations are less than the preset threshold; Install torque sensors and diagonally arranged hydraulic jacks in the corner section of the wind tunnel to monitor the torsion angle data of the corner section in real time, and apply reverse torque through hydraulic jacks to perform dynamic balance adjustment when the torsion angle exceeds the allowable value.

2. The method according to claim 1, characterized in that, The initial measurement point layout scheme for the coarse-scale layer specifically involves setting measurement points on the tunnel loop according to the spacing between the coarse-scale layer measurement points. The spacing between the coarse-scale layer measurement points is calculated using a coarse-scale measurement point spacing function.

3. The method according to claim 2, characterized in that, The calculation of the coarse-scale measuring point spacing function is specifically to multiply the ratio of the tunnel diameter to the reference diameter by the ratio of the tunnel wall thickness to the reference wall thickness, and then multiply by the reference measuring point spacing.

4. The method according to claim 3, characterized in that, The reference diameter, reference wall thickness, and reference measuring point spacing are determined through a standard tunnel test. The standard tunnel test includes constructing a standard tunnel test section, arranging measuring point groups at different spacings on the standard tunnel test section, measuring the axial deviation data of each measuring point group using a three-dimensional laser scanner, calculating the measurement coverage and error propagation coefficient of each measuring point group, and selecting the maximum spacing with a measurement coverage greater than 90% and an error propagation coefficient less than 1.2 as the reference measuring point spacing.

5. The method according to claim 1, characterized in that, The temperature field-deformation field coupled analysis and calculation specifically involves first solving the steady-state temperature field distribution to obtain the temperature values ​​of each node, then applying the temperature values ​​of each node as body loads to the structural field to calculate the thermal stress distribution and thermal deformation distribution, and extracting the displacement of key nodes on the tunnel axis.

6. The method according to claim 1, characterized in that, The multi-scale hierarchical iterative leveling method includes three levels: coarse-scale, meso-scale, and fine-scale. The distance between measurement points in the meso-scale layer is calculated using the meso-scale measurement point distance function, and the distance between measurement points in the fine-scale layer is calculated using the fine-scale measurement point distance function. The adjustment amount for each layer is calculated using the weighted least squares fitting method.

7. The method according to claim 6, characterized in that, The calculation of the mesoscale measuring point spacing function is specifically to multiply the reference mesoscale spacing by the ratio of the measured maximum deviation value of the coarse-scale layer to the reference deviation value, and then multiply by the ratio of the reference stiffness parameter to the tunnel stiffness parameter.

8. The method according to claim 6, characterized in that, The calculation of the fine-scale measurement point spacing function is specifically to multiply the reference fine-scale spacing by the ratio of the measured maximum deviation value of the mesoscale layer to the reference fine-scale deviation value, and then multiply it by the ratio of the reference region length to the key region length.

9. The method according to claim 1, characterized in that, The calculation of the coarse adjustment amount specifically involves subtracting the measured axis coordinates of each measuring point in the coarse-scale layer from the design axis coordinates to obtain the deviation vector. The deviation vector is then subjected to global optimization adjustment calculation, and the accumulated error is evenly distributed to each section of the tunnel to obtain the vertical and horizontal adjustment amounts of each support point.

10. The method according to claim 9, characterized in that, The global optimization adjustment calculation adopts an indirect adjustment model, establishing a set of observation equations and a set of condition equations. The set of observation equations includes position observation equations, elevation observation equations, and azimuth observation equations. The measurement errors and installation errors of each segment are used as observation values, and the optimal estimates of the unknown parameters are solved by the least squares criterion.

Images