A wind tunnel large-span steel grid hybrid vibration reduction method based on TMD and viscous damper

Abstract

This invention provides a hybrid vibration reduction method for large-span steel space frames in wind tunnels based on TMD and viscous dampers, belonging to the field of wind tunnel technology. This invention involves installing magnetorheological elastomer variable stiffness devices and temperature-compensated viscous dampers within each tuned mass damper unit. It collects signals using triaxial accelerometers and performs frequency domain decomposition algorithm mode identification, establishing a two-layer game optimization model to output the optimal frequency distribution vector and optimal damping force distribution vector. Based on the optimization results, the stiffness coefficient and damping coefficient are adjusted using a current controller. The vibration reduction effect is evaluated using a sliding time-window power spectral density analysis method, and the optimization model is re-executed to update the control parameters based on the evaluation results. This solves the technical problem of vibration reduction failure in large-span steel space frame structures due to frequency detuning of tuned mass dampers caused by multimodal coupled vibration under wind load excitation.

Description

Technical Field

[0001] This invention belongs to the field of wind tunnel technology, and more specifically, relates to a hybrid vibration reduction method for large-span steel space frames in wind tunnels based on TMD and viscous dampers. Background Technology

[0002] Large-span steel space frame structures are widely used in wind tunnel laboratory construction. Traditional vibration reduction technology uses tuned mass dampers (TMDs) with fixed parameters in conjunction with viscous dampers for vibration control. Energy dissipation is achieved by tuning the natural frequency of the TMD to the dominant mode frequency of the structure, and the viscous damper provides velocity-dependent damping force to suppress the structural vibration response. This can achieve good vibration reduction effects under single-frequency excitation and steady-state conditions. However, during wind tunnel operation, the natural frequency of the structure drifts with temperature changes, and changes in load conditions cause fluctuations in modal parameters. The viscosity of the viscous damper medium changes significantly with ambient temperature, and the fixed-parameter tuned mass damper cannot adaptively track the frequency drift. When the tuning frequency deviates from the optimal matching state with the structural frequency, the vibration reduction efficiency drops sharply. At the same time, large-span space frames have multiple dense modes, and wind loads may simultaneously excite multiple vibration modes. A single-frequency tuned damper can only control one dominant mode, and its control effect on other modes is limited. In other words, in the existing technology, due to structural frequency drift and changes in environmental parameters, the tuned mass damper becomes detuned with the structural frequency. In addition, multimodal coupled vibration makes it impossible for the single-frequency tuning strategy to fully cover all modes. The existing technology has a technical problem that large-span steel space frame structures fail to reduce vibration due to frequency detuning of the tuned mass damper caused by multimodal coupled vibration under wind load excitation. Summary of the Invention

[0003] In view of this, the present invention provides a hybrid vibration reduction method for large-span steel space frames in wind tunnels based on TMD and viscous dampers, which can solve the technical problem in the prior art where large-span steel space frame structures fail to reduce vibration due to frequency detuning of tuned mass dampers caused by multimodal coupled vibration under wind load excitation.

[0004] This invention is implemented as follows: It provides a hybrid vibration reduction method for large-span steel space frames in wind tunnels based on TMD (tunable mass damper) and viscous dampers. A proportional steel frame jig is constructed on the ground. A 3D laser scanner collects the coordinate point cloud data of the space frame nodes and compares the deviation with the BIM model to generate compensation and correction data. High-precision reference points are set around the space frame, and a spatial 3D reference line network is formed using piano wire tensioning. The support frame of the multi-frequency tuned mass damper group is pre-assembled on the ground, with adjustable connection nodes providing three-dimensional adjustment margins. After hoisting into place, a laser tracker is used to monitor the six-degree-of-freedom pose of the support frame of the multi-frequency tuned mass damper group in real time. Magnetorheological elastomer variable stiffness devices and temperature-compensated viscous dampers are installed in each tuned mass damper unit of the multi-frequency tuned mass damper group. Three-dimensional acceleration devices are deployed at key nodes of the space frame. The system uses a frequency domain decomposition algorithm to perform modal identification on the acceleration signals collected by the triaxial accelerometer and a strain sensor to extract the first five modal frequencies and modal damping ratios. A two-layer game optimization model is established. The upper-layer optimization model outputs the optimal frequency distribution vector and mass distribution ratio vector of the multi-frequency tuned mass damper group, while the lower-layer optimization model outputs the optimal damping force distribution vector of each tuned mass damper unit. Cooperative optimization is achieved through the coupling terms of the upper and lower-layer optimization models. Based on the optimal frequency distribution vector output by the two-layer game optimization model, the stiffness coefficient of each tuned mass damper unit is adjusted by the current controller of the magnetorheological elastomer variable stiffness device, and the damping coefficient of each tuned mass damper unit is adjusted according to the optimal damping force distribution vector. The vibration reduction effect is evaluated using a sliding time window power spectral density analysis method.

[0005] Among them, high-precision reference points are set at the intersection of the diagonals around the grid structure, and ultrasonic base stations are set up for precise coordinate calibration. The three-dimensional coordinates of the ultrasonic base stations are precisely calibrated by a total station.

[0006] Among them, based on the six-degree-of-freedom pose deviation data output by the laser tracker, the spatial position of the multi-frequency tuned mass damper group support frame is adjusted, and a spatial control network is established by joint measurement using the Beidou high-precision positioning system and a total station.

[0007] In this process, semiconductor cooling chips are evenly distributed on the outer wall of the cylinder of the temperature-compensated viscous damper. The temperature of the damping medium is monitored in real time by a temperature sensor, and the temperature compensation control program is activated to adjust the working current of the semiconductor cooling chips.

[0008] Specifically, when the deviation between the first-order modal frequency identified by the frequency domain decomposition algorithm and the initial design frequency exceeds a set value, the frequency adaptive tuning program is triggered. The modal parameters are updated and identified through the frequency domain decomposition algorithm to obtain the updated first 5 modal frequencies and modal damping ratios.

[0009] The upper-level optimization model aims to minimize the root mean square of the multi-point displacement response of the grid structure, while the lower-level optimization model aims to maximize the total energy dissipation of the multi-frequency tuned mass damper group. The coupling term is the inner product of the optimal frequency distribution vector and the optimal damping force distribution vector divided by the product of the magnitudes of the two vectors.

[0010] The stiffness coefficient adjustment time interval is set to a fixed value. When the ambient temperature fluctuation in the wind tunnel exceeds the set value, the required cooling power is calculated based on the damping medium viscosity-temperature characteristic compensation model, and the working current of the semiconductor cooling chip is adjusted.

[0011] In the sliding window power spectral density analysis method, the window length is set to a fixed value. After updating the optimal frequency distribution vector and the optimal damping force distribution vector, the stiffness coefficient and damping coefficient of each tuned mass damper unit are adjusted according to the updated vector.

[0012] Among them, the multi-frequency tuned mass damper group is a device that combines tuned mass damper units with different tuning frequencies in a series and parallel manner. The mass distribution of each tuned mass damper unit follows the Gaussian distribution principle of maximizing the mass of the unit at the center frequency.

[0013] The magnetorheological elastomer variable stiffness device consists of a silicone rubber matrix embedded with carbonyl iron powder, an excitation coil, and a magnetic core. Under the action of an external magnetic field, the shear modulus of the silicone rubber matrix changes, and the natural frequency of each tuned mass damper unit can be continuously adjusted by adjusting the excitation current.

[0014] Among them, the temperature-compensated viscous damper is a velocity-dependent damper that generates damping force through the relative motion between the piston rod and the cylinder. The damping medium is a modified polyalphaolefin base oil compounded with nano-polymers. Thickener, used in semiconductor refrigeration chips Material.

[0015] Among them, the damping medium viscosity-temperature characteristic compensation model is based on the Arrhenius equation, which describes the nonlinear relationship between the viscosity and temperature of the damping medium. The model parameters are obtained by viscosity calibration tests at fixed intervals within the temperature range.

[0016] Among them, the frequency domain decomposition algorithm uses fast Fourier transform to convert the time domain acceleration signal into a frequency domain power spectrum, identifies the frequency of each mode by peak picking method, and calculates the mode damping ratio by half power bandwidth method.

[0017] Among them, the frequency adaptive tuning program uses a proportional-integral control algorithm to adjust the deviation between the real-time modal frequency identified by the frequency domain decomposition algorithm and the current tuning frequency of each tuning mass damper unit in the multi-frequency tuned mass damper group.

[0018] The objective function of the upper-level optimization model is the minimum value of the sum of the weighted sum of the root mean square of the displacement responses of all monitoring points, the product of the modal participation coefficient and the quotient of the frequency deviation coefficient plus 1. The objective function of the lower-level optimization model is the maximum value of the sum of the product of the relative displacement amplitude of each tuned mass damper unit and the damping coefficient and the square root of the stiffness coefficient.

[0019] In this process, the upper-level optimization model and the lower-level optimization model are solved by alternating iterations. When the change of the coupling term in two consecutive iterations is less than the set value, the Nash equilibrium solution is considered to be reached. At this time, the optimal frequency distribution vector and the optimal damping force distribution vector are output.

[0020] This invention employs a multi-frequency tuned mass damper assembly in conjunction with a magnetorheological elastomer variable stiffness device and a temperature-compensated viscous damper. By dynamically adjusting the stiffness coefficient and damping coefficient of each tuned mass damper unit through real-time modal identification and a two-layer game optimization model, adaptive control of multimodal coupled vibration is achieved. This invention utilizes a magnetorheological elastomer variable stiffness device to change the shear modulus under the action of an external magnetic field, enabling the natural frequency of each tuned mass damper unit to be continuously adjustable within ±12% of the design frequency, with a response time of less than 100ms, quickly tracking structural frequency drift. A temperature-compensated viscous damper employs a semiconductor cooling chip to actively control the temperature of the damping medium, eliminating the influence of temperature changes on the damping coefficient. A multi-frequency tuned mass damper group arranges 5 to 9 tuned mass damper units of different frequencies according to a frequency dispersion strategy, covering multiple target modes. A two-layer game optimization model simultaneously optimizes the frequency distribution and damping force allocation; the upper-layer model minimizes the structural response, while the lower-layer model maximizes energy consumption. The two models achieve collaborative optimization through coupling terms, reaching a Nash equilibrium solution where the parameters of each tuned mass damper unit are globally optimal. In summary, this invention achieves frequency adaptive tracking through magnetorheological elastomer variable stiffness technology, ensures stable damping performance through temperature compensation technology, covers multi-mode vibration through multi-frequency tuning, and achieves coordinated parameter configuration through two-layer game optimization. This solves the technical problem mentioned in the background art of vibration reduction failure of large-span steel space frame structures due to frequency detuning of tuned mass dampers caused by multi-mode coupled vibration under wind load excitation. 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 adjustment process of the hydraulic synchronous lifting system in the embodiment.

[0023] Figure 3 The image shows the frequency identification results of the first 5 modes of the space frame in the embodiment.

[0024] Figure 4This is a comparison chart showing the reduction effect of the dominant mode power spectral density in the embodiments. Detailed Implementation

[0025] 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.

[0026] like Figure 1 As shown, this invention provides a hybrid vibration reduction method for large-span steel space frames in wind tunnels based on TMD and viscous dampers, comprising:

[0027] S10. Construct a 1:1 scale steel frame on the ground, collect the coordinate point cloud data of the grid node using a 3D laser scanner, compare and analyze the deviation with the BIM model, generate compensation and correction data, set high-precision reference points around the grid, and use φ1mm piano steel wire to tension and form a spatial 3D reference line network. The tension is controlled to be constant at 500N by a tension sensor.

[0028] S20. The multi-frequency tuned mass damper group support frame is pre-assembled on the ground. An adjustment margin of ±8cm in three directions is reserved through adjustable connection nodes. After hoisting into place, the six-degree-of-freedom position of the multi-frequency tuned mass damper group support frame is monitored in real time using a laser tracker. The hydraulic synchronous jacking system is used to adjust the position synchronously at 8 support points.

[0029] S30. Install a magnetorheological elastomer variable stiffness device and a temperature-compensated viscous damper in each tuned mass damper unit of the multi-frequency tuned mass damper group. The magnetorheological elastomer variable stiffness device adjusts the shear modulus by applying an excitation current of 0 to 2A, so that the stiffness of each tuned mass damper unit can be dynamically adjusted within a range of 15 to 45 kN / m.

[0030] S40. Install triaxial acceleration sensors and strain sensors at key nodes of the grid structure. The total number of sensors shall not be less than 8% of the total number of nodes of the grid structure. The sampling frequency shall be set to 200Hz. Modal identification shall be performed on the acceleration signals collected by the triaxial acceleration sensors through frequency domain decomposition algorithm to extract the first 5 modal frequencies and modal damping ratios.

[0031] S50. Establish a two-layer game optimization model, which includes an upper-layer optimization model and a lower-layer optimization model. The upper-layer optimization model outputs the optimal frequency distribution vector and mass allocation ratio vector of the multi-frequency tuned mass damper group, and the lower-layer optimization model outputs the optimal damping force allocation vector of each tuned mass damper unit.

[0032] S60. Based on the optimal frequency distribution vector output by the two-layer game optimization model, the stiffness coefficient of each tuned mass damper unit is adjusted by the current controller of the magnetorheological elastomer variable stiffness device, and the damping coefficient of each tuned mass damper unit is adjusted according to the optimal damping force distribution vector.

[0033] S70. The vibration reduction effect is evaluated by using the sliding time window power spectral density analysis method to obtain the power spectral density at the peak frequency of the dominant mode. When the reduction rate of the power spectral density at the peak frequency of the dominant mode is less than 40%, the two-layer game optimization model is re-executed to update the optimal frequency distribution vector and the optimal damping force distribution vector.

[0034] In step S10, the high-precision reference points are set at the intersection of the diagonals around the grid structure, and 6 to 8 ultrasonic base stations are set up for precise coordinate calibration. The ultrasonic base stations are precisely calibrated in three dimensions using a total station with a calibration accuracy of ±2mm.

[0035] In step S20, the spatial position of the multi-frequency tuned mass damper group support frame is adjusted according to the six-degree-of-freedom pose deviation data output by the laser tracker to ensure that the installation accuracy is controlled within ±5mm and the installation angle deviation is less than 0.3°. A spatial control network is established by using the Beidou high-precision positioning system and the total station for joint measurement.

[0036] In step S30, 4 to 6 semiconductor cooling chips are evenly installed on the outer wall of the cylinder of the temperature-compensated viscous damper. The temperature of the damping medium is monitored in real time by a temperature sensor. When the temperature of the damping medium output by the temperature sensor deviates from the set value of ±2℃, the temperature compensation control program is started to adjust the working current of the semiconductor cooling chip and stabilize the temperature of the damping medium at 23±1℃.

[0037] In step S40, when the deviation between the first-order modal frequency identified by the frequency domain decomposition algorithm and the initial design frequency exceeds 3%, the frequency adaptive tuning program is triggered. Every 60 seconds, the modal parameter update identification is performed through the frequency domain decomposition algorithm to obtain the updated first 5 modal frequencies and modal damping ratios. When the modal parameter update identification results show that the damping ratio of a certain modality decreases by more than 15% after 3 consecutive updates, the damping coefficient of the corresponding frequency-tuned mass damper unit is increased by 15% to 25% according to the optimal damping force allocation vector output by the two-layer game optimization model.

[0038] The upper-level optimization model in step S50 aims to minimize the root mean square of the multi-point displacement response of the grid structure. Its inputs include the displacement response amplitude, modal participation coefficient, and frequency deviation coefficient of each measuring point. The lower-level optimization model aims to maximize the total energy dissipation of the multi-frequency tuned mass damper group. Its inputs include the relative displacement amplitude, damping coefficient, and stiffness coefficient of each tuned mass damper unit. The upper-level optimization model and the lower-level optimization model are associated through a coupling term, which is the inner product of the optimal frequency distribution vector and the optimal damping force distribution vector divided by the product of the magnitudes of the two vectors.

[0039] In step S60, the adjustment time interval of the stiffness coefficient is set to 10s, and the single adjustment amount does not exceed 5% of the current stiffness coefficient value. When the ambient temperature fluctuation in the wind tunnel exceeds 10℃, the required cooling power is calculated according to the damping medium viscosity-temperature characteristic compensation model, and the working current of the semiconductor cooling chip is adjusted. The working current adjustment range is 0.5 to 3.5A, and the response time does not exceed 30s.

[0040] In step S70, the time window length of the sliding time window power spectral density analysis method is set to 120s, the overlap rate is 50%, and after updating the optimal frequency distribution vector and the optimal damping force distribution vector, the stiffness coefficient and damping coefficient of each tuned mass damper unit are adjusted according to the updated vector.

[0041] The multi-frequency tuned mass damper group is a device that combines 5 to 9 tuned mass damper units with different tuning frequencies in a series and parallel manner. The tuning frequencies of each tuned mass damper unit are distributed in an arithmetic or geometric sequence, with the frequency interval being 2 to 5% of the dominant mode frequency. The total mass ratio is controlled at 1.5 to 3% of the total mass of the grid structure. The mass distribution of each tuned mass damper unit follows the Gaussian distribution principle of maximizing the mass of the unit at the center frequency.

[0042] The magnetorheological elastomer variable stiffness device consists of a silicone rubber matrix embedded with carbonyl iron powder, an excitation coil, and a magnetic core. The carbonyl iron powder has a volume fraction of 30% to 40% and a particle size of 3 to 8 μm. The excitation coil has 400 to 600 turns. Under the action of an external magnetic field, the shear modulus of the silicone rubber matrix increases from the ground state of 0.8 to 1.2 MPa to the saturation state of 2.5 to 4.0 MPa, with a response time of less than 100 ms. The natural frequency of each tuned mass damper unit can be continuously adjusted within ±12% of the design frequency by adjusting the excitation current.

[0043] The temperature-compensated viscous damper is a velocity-dependent damper that generates damping force through the relative motion between the piston rod and the cylinder. The damping medium is a modified polyalphaolefin base oil compounded with nano-polymers. Thickener with a viscosity index not lower than 180 and a kinematic viscosity change rate of less than 35% in a temperature range of 5 to 50°C; the semiconductor refrigeration chip uses... The material has a single-piece cooling capacity of 15 to 25W, and the cooling power is adjusted by pulse width modulation, with a temperature control accuracy of ±0.5℃.

[0044] The damping medium viscosity-temperature characteristic compensation model is based on the Arrhenius equation and describes the nonlinear relationship between the viscosity and temperature of the damping medium. The model parameters are obtained by viscosity calibration tests at 5°C intervals within the temperature range of −10 to 60°C. The input of the damping medium viscosity-temperature characteristic compensation model is the real-time measured temperature of the damping medium and the outer wall temperature of the cylinder. The output is the operating current of the semiconductor cooling chip required to maintain the target damping coefficient. The damping medium viscosity-temperature characteristic compensation model updates the calculation results every 30 seconds.

[0045] The frequency domain decomposition algorithm uses Fast Fourier Transform to convert the time-domain acceleration signal into a frequency-domain power spectrum, identifies the frequency of each mode by peak picking, and calculates the mode damping ratio by half-power bandwidth method. To improve the recognition accuracy, the original acceleration signal is subjected to Hanning window weighting processing, the frequency resolution is set to 0.02Hz, and the error of the first 5 identified mode frequencies does not exceed 0.5%.

[0046] The frequency adaptive tuning program calculates the required stiffness adjustment for each tuned mass damper unit based on the deviation between the real-time modal frequency identified by the frequency domain decomposition algorithm and the current tuning frequency of each tuned mass damper unit in the multi-frequency tuned mass damper group. The adjustment strategy adopts a proportional-integral control algorithm, with the proportional coefficient set to 0.6 to 0.9 and the integral time constant set to 8 to 15s. Through iterative adjustment, the frequency deviation of each tuned mass damper unit is brought to within ±1%, and the entire tuning process lasts no more than 180s.

[0047] The objective function of the upper-level optimization model is the minimum value of the sum of the weighted sum of the root mean square sum of the displacement responses of all monitoring points, the product of the modal participation coefficients, and the sum of the product of the frequency deviation coefficients plus 1. The weight coefficients are assigned according to the importance of each measuring point in the grid structure, with key nodes having a weight of 1.0 and secondary nodes having a weight of 0.6 to 0.8. The constraints include the frequency distribution range of each tuned mass damper unit, the sum of all components in the mass allocation ratio vector being 1, and the mass ratio of a single tuned mass damper unit not exceeding 40% of the total mass ratio.

[0048] The objective function of the lower-level optimization model is the maximum value of the sum of the product of the relative displacement amplitude of each tuned mass damper unit and the damping coefficient, divided by the square root of the stiffness coefficient. The constraints include that the sum of the damping forces of each tuned mass damper unit does not exceed 18% of the maximum inertial force of the grid structure, the damping force of a single tuned mass damper unit does not exceed 3 times the weight of its mass block, and the damping coefficient is adjusted within the range of 0.7 to 1.5 times the initial damping coefficient value.

[0049] The coupling term value reflects the degree of coordination between the optimal frequency distribution vector determined by the upper-level optimization model and the optimal damping force allocation vector optimized by the lower-level optimization model. The closer the coupling term is to 1, the better the coordination. By iteratively solving the upper-level optimization model and the lower-level optimization model, when the change in the coupling term in two consecutive iterations is less than 0.02, it is considered that a Nash equilibrium solution has been reached. At this time, the optimal frequency distribution vector and the optimal damping force allocation vector are output.

[0050] The modal participation coefficient characterizes the contribution of each mode to the overall structural response. It is calculated by dividing the inner product of each mode shape and the load distribution vector by the modal mass. The larger the absolute value of the modal participation coefficient, the more significant the influence of the corresponding mode on the response. The modal participation coefficient is introduced into the objective function of the upper-level optimization model to achieve key control of the dominant mode.

[0051] The frequency deviation coefficient is the absolute value of the difference between the real-time modal frequency identified by the frequency domain decomposition algorithm and the design modal frequency, divided by the design modal frequency. It reflects the degree of structural frequency drift. When the frequency deviation coefficient exceeds 0.05, it is considered that significant detuning has occurred, and the frequency adaptive tuning program needs to be started. In the objective function of the upper-level optimization model, it is used as a denominator term to reflect the weakening effect of frequency deviation on vibration reduction.

[0052] The sliding window power spectral density analysis method divides the long-time-history response signal into multiple time windows, each with a length of 120s and an overlap of 60s between adjacent time windows. Power spectral density is estimated for each time window signal. Random errors are reduced by averaging the power spectra of multiple time windows using the Welch method. The reduction rate of power spectral density at the peak frequency of the dominant mode is the difference between the peak power spectrum after vibration reduction and the peak power spectrum before vibration reduction, divided by the peak power spectrum before vibration reduction. A reduction rate of 40% or more is considered to meet the requirements for vibration reduction.

[0053] The six-degree-of-freedom pose includes three translational degrees of freedom and three rotational degrees of freedom in space. The three translational degrees of freedom are displacement along the X-axis, displacement along the Y-axis, and displacement along the Z-axis. The three rotational degrees of freedom are roll angle around the X-axis, pitch angle around the Y-axis, and yaw angle around the Z-axis. The laser tracker calculates the six-degree-of-freedom pose of the multi-frequency tuned mass damper group support frame in real time by measuring the spatial coordinates of the reflective target ball installed on the support frame of the multi-frequency tuned mass damper group. The measurement accuracy is ±0.05mm for displacement and ±5 arcseconds for angle.

[0054] The displacement response amplitude of each measuring point is obtained by performing a second integration on the acceleration signal collected by the triaxial accelerometer. The integration algorithm adopts the Simpson numerical integration method, and the integration time step is set to 0.005s. In order to eliminate the integration drift error, the displacement response amplitude of each measuring point is subjected to high-pass filtering, and the cutoff frequency is set to 0.1Hz.

[0055] The dominant mode is the mode with the largest absolute value of the mode participation coefficient. When the absolute value of the mode participation coefficient corresponding to the first-order mode frequency is greater than 60% of the sum of the absolute values ​​of the mode participation coefficients corresponding to the other mode frequencies, the first-order mode is considered to be the dominant mode. The peak frequency of the dominant mode is the mode frequency corresponding to the dominant mode.

[0056] The specific implementation methods of the above steps are described in detail below.

[0057] The specific implementation of step S10 is as follows: First, the reference position of the steel frame jig is marked on a flat ground area according to the design drawings. A laser plumb line is used to ensure that the vertical deviation of the reference point is less than 2mm. Then, a 1:1 scale steel frame jig is built according to the actual size of the space frame. The components are connected by bolts, and the pre-tightening torque is controlled between 150 and 200 Nm using a torque wrench. The 3D laser scanner uses the time-of-flight ranging principle to collect point cloud data of the space frame node coordinates at a scanning speed of 500,000 points per second. The scanning angle resolution is set to 0.01°, and the scanning distance is set to 5 to 30m. The collected point cloud data is then exported... The BIM model is registered and compared. The iterative nearest point algorithm is used to calculate the spatial deviation between the point cloud and the model. When the deviation exceeds 5mm, compensation and correction data are generated. Then, high-precision reference points are set at the intersection of the diagonals around the grid structure. Three-dimensional coordinate measurement is performed using a total station with a measurement accuracy of ±1mm. The two ends of the φ1mm piano wire are fixed between the relative reference points. The tension of the wire is monitored in real time by a tension sensor. When the tension reaches 500N, the tensioning device is locked to form a spatial three-dimensional reference line network. The reference line network is used for spatial position transfer during the subsequent installation process to ensure the positional accuracy of each installation node.

[0058] The specific implementation of step S20 is as follows: On the ground, the connection nodes of the multi-frequency tuned mass damper assembly support frame are connected using adjustable bolts, with a three-way adjustment margin of ±8cm reserved for subsequent aerial fine-tuning. An absolute coordinate system for the construction area is established using the BeiDou high-precision positioning system, with a positioning accuracy of ±10mm horizontally and ±20mm vertically. A spatial control network is established through localized densification measurements using a total station, with the control network point spacing set at 15 to 25m. During hoisting, a dual-crane lifting method is adopted, and a synchronous control system ensures that the speed difference between the two cranes is less than 5%. After hoisting into place, reflective target balls are installed at eight key locations on the multi-frequency tuned mass damper assembly support frame, and laser tracking is used. The tracker measures the three-dimensional coordinates of the target ball by emitting a laser beam and receiving the reflected signal. It calculates the six-degree-of-freedom pose of the support frame in real time, with a sampling frequency set to 10Hz. When the pose deviation exceeds the set threshold, the hydraulic synchronous lifting system starts synchronously at the eight support points. The hydraulic pressure at each lifting point is monitored by a pressure sensor, and the lifting speed at each point is kept consistent by a proportional valve. The single adjustment stroke is controlled within 2 to 5 mm. After multiple iterative adjustments, the installation accuracy meets the requirements of ±5 mm and the installation angle deviation is less than 0.3°. The six-degree-of-freedom pose adjustment process adopts the closed-loop feedback control principle, which drives the actuator to make compensation adjustments by measuring the deviation from the target pose in real time.

[0059] The specific implementation of step S30 is as follows: a magnetorheological elastomer stiffness-changing device is installed between the mass block and the guide rail of each tuned mass damper unit. The device consists of an upper and lower silicone rubber matrix sandwiching a magnetic core. Carbonyl iron powder is uniformly dispersed in the silicone rubber matrix, with a volume fraction of 30 to 40% and a particle size of 3 to 8 μm. An excitation coil with 400 to 600 turns is wound around the outside of the iron core. When an excitation current of 0 to 2 A is applied, the magnetic field generated by the iron core passes through the silicone rubber matrix, and the carbonyl iron powder is subjected to the magnetic field. The formation of a chain-like structure increases the shear modulus of the silicone rubber matrix from 0.8 to 1.2 MPa in the ground state to 2.5 to 4.0 MPa in the saturated state, thereby altering the equivalent stiffness of the device. This allows for dynamic adjustment of the stiffness of each tuned mass damper unit within the range of 15 to 45 kN / m. The magnetorheological elastomer has a response time of less than 100 ms, enabling rapid tracking of frequency changes. Temperature-compensated viscous dampers are installed on both sides of the mass block, and the space between the damper piston rod and the cylinder is filled with a modified polyalphaolefin base oil compounded with nanoparticles. The thickener serves as a damping medium, wherein the viscosity index of the damping medium is not less than 180, and 4 to 6 semiconductor cooling chips are evenly distributed and installed on the outer wall of the cylinder body. The semiconductor cooling chips employ... The material has a single-chip cooling capacity of 15 to 25W. A temperature sensor is attached to the cylinder surface to monitor the temperature of the damping medium in real time. The sampling frequency is set to 1Hz. When the detected temperature deviates from the set value of 23℃ by more than ±2℃, the temperature compensation control program is activated. The required cooling power is calculated according to the viscosity-temperature characteristics compensation model of the damping medium. The working current of the semiconductor cooling chip is adjusted by pulse width modulation. The current adjustment range is 0.5 to 3.5A to stabilize the temperature of the damping medium at 23±1℃ and ensure that the damping coefficient fluctuation range is controlled within ±8% of the initial value.

[0060] The specific implementation of step S40 is as follows: Triaxial accelerometers are installed at key locations such as beam-column nodes, mid-span positions, and near supports of the space frame. The sensor measurement range is set to ±5g, and the sensitivity is 500mV / g. The total number of sensors installed is no less than 8% of the total number of nodes in the space frame. Simultaneously, strain sensors are installed in stress concentration areas, with a strain measurement range of ±3000με. The sampling frequency of the data acquisition system is uniformly set to 200Hz to meet the Nyquist sampling theorem requirements. The acquired acceleration signals are first processed to remove baseline drift by detrending term processing. Then, the time-domain signal is converted into a frequency-domain power spectrum using a fast Fourier transform (FFT). The FFT uses a radix-2 algorithm, with 8192 transform points and a frequency resolution of 0.02Hz. The modal frequencies are identified on the power spectrum using a peak-picking method, with a peak-picking threshold set to... The modal damping ratio is calculated using the half-power bandwidth method for each identified modal frequency, with the half-power point defined as the frequency point corresponding to a 3dB drop in peak power. The frequency domain decomposition algorithm can simultaneously identify the first 5 modal frequencies and modal damping ratios. The identification accuracy is affected by the signal length and frequency resolution. When the deviation of the first modal frequency from the initial design frequency exceeds 3%, frequency detuning is determined, triggering the frequency adaptive tuning program. During the operation of the vibration reduction system, modal parameter update identification is automatically performed every 60 seconds. The changes in structural dynamic characteristics are continuously monitored through a sliding data window. When the identification results show that the damping ratio of a certain modality drops by more than 15% after 3 consecutive identifications, it indicates that the vibration reduction efficiency of the tuned mass damper unit at the corresponding frequency has decreased, and its damping coefficient needs to be increased by 15% to 25% for compensation.

[0061] The specific implementation of step S50 is as follows: a two-layer game optimization model consisting of an upper-layer optimization model and a lower-layer optimization model is established. The upper-layer optimization model aims to minimize the root mean square of the multi-point displacement response of the space frame. The input parameters include the displacement response amplitude, modal participation coefficient, and frequency deviation coefficient of each measuring point. The displacement response amplitude of each measuring point is obtained by performing a second integration on the acceleration signal. The integration algorithm adopts the Simpson numerical integration method with an integration time step of 0.005s. The integration result is filtered by a high-pass filter with a cutoff frequency of 0.1Hz to eliminate integration drift error. The modal participation coefficient is calculated by dividing the inner product of each mode shape and the load distribution vector by the modal mass, representing the contribution of each mode to the overall structural response. The frequency deviation coefficient is the absolute value of the difference between the real-time modal frequency and the design modal frequency divided by the design modal frequency. The output of the upper-layer optimization model is the optimal frequency distribution vector and the mass distribution ratio. The lower-level optimization model aims to maximize the total energy dissipation of the multi-frequency tuned mass damper group. Input parameters include the relative displacement amplitude, damping coefficient, and stiffness coefficient of each tuned mass damper unit. The output of the lower-level optimization model is the optimal damping force distribution vector for each tuned mass damper unit. The two models are linked by a coupling term, which is the inner product of the optimal frequency distribution vector and the optimal damping force distribution vector divided by the product of the magnitudes of the two vectors. The value of the coupling term reflects the degree of synergy between the two optimization results; the closer the coupling term is to 1, the better the synergy. An alternating iterative approach is used to solve the two-level game optimization model. First, the lower-level variables are fixed to solve the upper-level model. Then, the upper-level results are substituted into the lower-level model for further solving. This process is repeated iteratively. When the change in the coupling term between two consecutive iterations is less than 0.02, a Nash equilibrium solution is considered to have been reached. At this point, the output optimal frequency distribution vector and optimal damping force distribution vector are used to guide subsequent parameter adjustments.

[0062] The specific implementation of step S60 is as follows: Based on the optimal frequency distribution vector output by the two-layer game optimization model, the required stiffness coefficient of each tuned mass damper unit is calculated. The excitation current of the magnetorheological elastomer variable stiffness device is adjusted by a current controller. The relationship between the current and the stiffness coefficient is established through pre-calibration tests. The adjustment time interval is set to 10 seconds to avoid frequent adjustments that could lead to system instability. The single adjustment amount is limited to within 5% of the current stiffness coefficient value. A gradual adjustment strategy is adopted, and the damping coefficient of each tuned mass damper unit is adjusted according to the optimal damping force distribution vector. The damping coefficient is adjusted by changing the temperature of the damping medium of the temperature-compensated viscous damper. When the ambient temperature inside the tunnel fluctuates by more than 10°C, the viscosity of the damping medium changes significantly. The cooling power required to maintain the target damping coefficient is calculated based on the viscosity-temperature characteristic compensation model of the damping medium. The viscosity-temperature characteristic compensation model is based on the Arrhenius equation, which describes the exponential decay of viscosity with temperature. The model parameters are obtained by viscosity calibration tests at 5°C intervals within the temperature range of −10 to 60°C. The input of the compensation model is the real-time measured temperature of the damping medium and the outer wall temperature of the cylinder. The output is the operating current required by the semiconductor cooling chip. The current adjustment range is 0.5 to 3.5A. Precise control is achieved through pulse width modulation, with a response time of no more than 30s, ensuring the stability of the damping coefficient.

[0063] The specific implementation of step S70 is as follows: A sliding time-window power spectral density analysis method is used to continuously evaluate the vibration reduction effect. The long-time-history response signal is divided into 120-second time windows, with adjacent time windows overlapping by 60 seconds. Power spectral density is estimated for each time window signal. The power spectra of multiple time windows are averaged using the Welch method to reduce the influence of random errors. The power spectral density value at the peak frequency of the dominant mode is extracted. The dominant mode is the mode with the largest absolute value of the modal participation coefficient. The power spectral density reduction rate is calculated as the difference between the peak power spectral density after vibration reduction and the peak power spectral density before vibration reduction, divided by the peak power spectral density before vibration reduction. When the reduction rate is less than 40%, the vibration reduction effect is determined to be unsatisfactory. The two-layer game optimization model is re-executed to update the optimal frequency distribution vector and the optimal damping force distribution vector. The stiffness coefficient and damping coefficient of each tuned mass damper unit are adjusted according to the updated vectors. Closed-loop feedback control is used to achieve adaptive optimization of the vibration reduction system, ensuring that the vibration reduction system always maintains its optimal working state under various disturbances such as structural frequency drift, ambient temperature changes, and load condition changes.

[0064] It should be noted that the key technical concepts of this invention include magnetorheological elastomer variable stiffness adaptive tuning technology, temperature-compensated viscous damper isothermal control technology, and multi-modal collaborative control technology with dual-layer game optimization. The magnetorheological elastomer variable stiffness adaptive tuning technology achieves continuous adjustment of the shear modulus by regulating the chain structure of carbonyl iron powder within the silicone rubber matrix through an external magnetic field. With a response time of less than 100ms, it can quickly track structural frequency drift, solving the problem of detuning and reduced vibration reduction efficiency caused by structural frequency changes in traditional fixed-parameter tuned mass dampers. Compared to mechanical stiffness adjustment, the magnetorheological elastomer variable stiffness technology has no mechanical wear, high adjustment accuracy, and low energy consumption. The temperature-compensated viscous damper isothermal control technology uses a semiconductor refrigeration chip and a viscosity-temperature characteristic compensation model to stabilize the damping medium temperature within a set value ±1℃. This overcomes the defect of traditional viscous dampers where damping coefficient fluctuations occur due to ambient temperature changes. Compared to passive temperature isolation methods, active temperature compensation can adapt to temperature differences exceeding 40℃ in wind tunnels, ensuring damping coefficient stability. The dual-level game-theoretic optimization multimodal collaborative control technology minimizes structural response through an upper-level optimization model and maximizes energy dissipation through a lower-level optimization model. The two models interact through coupling terms, achieving a Nash equilibrium solution that simultaneously satisfies optimal vibration reduction and maximum energy dissipation. Compared to single-objective optimization methods, dual-level game-theoretic optimization can consider multiple performance indicators, avoiding trade-offs. The synergistic effect of three key technologies is reflected in the frequency adaptation capability provided by the magnetorheological elastomer variable stiffness technology, the temperature compensation technology ensuring damping performance stability, and the dual-level game-theoretic optimization technology coordinating global parameter allocation. Together, they constitute a closed-loop adaptive control system. Compared to traditional passive tuned mass dampers and single control strategies, this invention can achieve multimodal collaborative vibration reduction in complex and variable wind tunnel environments, significantly improving the vibration control effect and system robustness of large-span steel space frames.

[0065] It should be noted that this invention also solves the following technical problems: First, it solves the technical problem of unstable vibration reduction performance caused by drastic fluctuations in the damping coefficient of the viscous damper with ambient temperature. This invention installs a semiconductor cooling chip on the outer wall of the cylinder body to monitor the temperature of the damping medium in real time. When the temperature deviates from the set value, a temperature compensation control program is activated. The required cooling power is calculated based on the viscosity-temperature characteristics compensation model of the damping medium, and the working current of the semiconductor cooling chip is precisely adjusted by pulse width modulation to stabilize the temperature of the damping medium within the range of 23±1℃. The damping medium uses modified polyalphaolefin base oil compounded with nano-polymers. The thickener has a viscosity index of not less than 180. Under temperature control conditions, the damping coefficient fluctuation range is controlled within ±8% of the initial value, overcoming the defect of traditional viscous dampers where the damping coefficient changes by more than 50% due to temperature differences exceeding 40℃ in wind tunnel environments. Secondly, it solves the technical problem of difficulty in ensuring the installation accuracy of multi-frequency tuned mass damper assemblies on large-span flexible space frame structures. This invention uses a 3D laser scanner to collect the coordinate point cloud data of the space frame nodes and compares the deviation with the BIM model to generate compensation and correction data. After hoisting into place, a laser tracker is used to monitor the six-degree-of-freedom pose of the support frame in real time. A hydraulic synchronous jacking system is used to make millimeter-level fine adjustments at eight support points. Using a closed-loop feedback control principle, the actuator is driven to make compensation adjustments by measuring the deviation from the target pose in real time, ensuring that the installation accuracy is controlled within ±5mm and the installation angle deviation is less than 0.3°. This provides a precise installation benchmark for the normal operation of magnetorheological elastomer variable stiffness devices and temperature-compensated viscous dampers.

[0066] Specifically, the principle of this invention is as follows: The invention solves the aforementioned technical problems by using a magnetorheological elastomer variable stiffness device to adjust the excitation current and change the chain-like arrangement of carbonyl iron powder within the silicone rubber matrix. This allows the shear modulus to be continuously adjustable within the range of 0.8 to 4.0 MPa, thereby altering the equivalent stiffness of the tuned mass damper unit. When the frequency domain decomposition algorithm detects structural frequency drift, the frequency adaptive tuning program calculates the required stiffness adjustment based on the deviation between the real-time modal frequency and the current tuning frequency. The current controller adjusts the excitation current to achieve dynamic stiffness tracking, ensuring the tuning frequency remains within ±1% of the structural frequency. The temperature-compensated viscous damper calculates the cooling power required to maintain the target damping coefficient based on the viscosity-temperature characteristic compensation model established by the Arrhenius equation. The damping medium temperature is stabilized at a set value of ±1℃ by a semiconductor cooling chip, eliminating the influence of temperature on viscosity. The multi-frequency tuned mass damper group configures the mass damper units of different frequencies according to the Gaussian distribution principle, with a frequency interval of 2 to 5% of the dominant mode frequency, which can simultaneously control multiple dense modes. The two-layer game optimization model solves the two objectives of minimizing the upper structure response and maximizing the lower energy consumption through alternating iterations. When the coupling term converges to the Nash equilibrium solution, the output optimal frequency distribution vector and optimal damping force distribution vector guide the parameter adjustment of each tuned mass damper unit, realizing the organic combination of frequency adaptation, temperature compensation and multi-mode collaborative control. Therefore, the technical solution of this invention is logical and can effectively solve the problems of frequency detuning and multi-mode coupled vibration.

[0067] 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.

[0068] The specific implementation of step S10 involves constructing a 1:1 scale steel frame on the ground, collecting point cloud data of the grid node coordinates using a 3D laser scanner, comparing and analyzing the deviations with the building information model, generating compensation and correction data, setting high-precision reference points at the intersection of the diagonals around the grid, and deploying 6 to 8 ultrasonic base stations for precise coordinate calibration. The ultrasonic base stations are precisely calibrated with their 3D coordinates using a total station with a calibration accuracy of ±2mm. High-precision reference points are set around the grid, and a spatial 3D reference line network is formed by tensioning φ1mm piano steel wire, with the tension controlled to be constant at 500N by a tension sensor.

[0069] The specific implementation of step S20 involves pre-assembling the multi-frequency tuned mass damper assembly support frame on the ground, reserving a three-way adjustment margin of ±8cm through adjustable connection nodes, and using a laser tracker to monitor the six-degree-of-freedom pose of the multi-frequency tuned mass damper assembly support frame in real time after hoisting into place. The six-degree-of-freedom pose includes three translational degrees of freedom and three rotational degrees of freedom in space. The three translational degrees of freedom are displacement along the X-axis, displacement along the Y-axis, and displacement along the Z-axis; the three rotational degrees of freedom are roll angle around the X-axis, pitch angle around the Y-axis, and yaw angle around the Z-axis. The laser tracker measures the position of the frame by measuring the position of the frame. The spatial coordinates of the reflective target sphere on the support frame of the multi-frequency tuned mass damper assembly are calculated in real time, and the six-degree-of-freedom pose of the support frame is calculated with a measurement accuracy of ±0.05mm displacement and ±5 arcseconds angle. The hydraulic synchronous jacking system is used to adjust the support frame synchronously at 8 support points. Based on the six-degree-of-freedom pose deviation data output by the laser tracker, the spatial position of the support frame of the multi-frequency tuned mass damper assembly is adjusted to ensure that the installation accuracy is controlled within ±5mm and the installation angle deviation is less than 0.3°. A spatial control network is established by combining the Beidou high-precision positioning system and the total station for measurement.

[0070] The specific implementation of step S30 involves installing a magnetorheological elastomer variable stiffness device and a temperature-compensated viscous damper in each tuned mass damper unit of the multi-frequency tuned mass damper group. The magnetorheological elastomer variable stiffness device adjusts the shear modulus by applying an excitation current of 0 to 2A, thereby achieving a dynamic stiffness adjustment range of 15 to 45 kN / m for each tuned mass damper unit. The stiffness coefficient of each tuned mass damper unit is described as follows:

[0071] ;

[0072] In the formula, For the first The real-time stiffness coefficient of each tuned mass damper element, in kN / m; For the first The ground state stiffness coefficient of each tuned mass damper element, in units of kN / m, is typically taken as 20 to 35 kN / m. To the excitation current The shear modulus of a magnetorheological elastic body under action, in MPa; This is the ground-state shear modulus of the magnetorheological elastomer, expressed in MPa, and typically ranges from 0.8 to 1.2 MPa. To apply in the The excitation current on each tuned mass damper unit is expressed in amperes (A). The serial number of the tuned mass damper unit, with a value ranging from 1 to... ; The total number of tuning mass damper units is 5 to 9.

[0073] The relationship between the shear modulus of a magnetorheological elastic body and the excitation current is expressed as follows:

[0074] ;

[0075] In the formula, This represents the maximum increment of the shear modulus, expressed in MPa, with an empirical value of 1.7 to 2.8 MPa. It is the saturation current of the magnetic field, measured in amperes (A), and typically ranges from 1.5 to 2.0 A.

[0076] The magnetorheological elastomer variable stiffness device consists of a silicone rubber matrix embedded with carbonyl iron powder, an excitation coil, and a magnetic core. The carbonyl iron powder has a volume fraction of 30% to 40% and a particle size of 3 to 8 μm. The excitation coil has 400 to 600 turns. Under the action of an external magnetic field, the shear modulus of the silicone rubber matrix increases from 0.8 to 1.2 MPa in the ground state to 2.5 to 4.0 MPa in the saturation state, with a response time of less than 100 ms. The natural frequency of each tuned mass damper unit can be continuously adjusted within ±12% of the design frequency by adjusting the excitation current.

[0077] Four to six thermoelectric cooling chips are evenly distributed on the outer wall of the cylinder of the temperature-compensated viscous damper. The temperature of the damping medium is monitored in real time by a temperature sensor. When the temperature of the damping medium output by the temperature sensor deviates from the set value by ±2℃, the temperature compensation control program is activated to adjust the working current of the thermoelectric cooling chips and stabilize the temperature of the damping medium at 23±1℃. The viscosity-temperature characteristic compensation model of the damping medium is based on the Arrhenius equation and describes the nonlinear relationship between the viscosity of the damping medium and the temperature, as detailed below:

[0078] ;

[0079] In the formula, For temperature The dynamic viscosity of the damped medium, expressed in Pa·s; Reference temperature The dynamic viscosity of the damped medium, expressed in Pa·s, is typically between 0.05 and 0.15 Pa·s. The viscosity activation energy is expressed in J / mol and is obtained through viscosity calibration tests conducted at 5°C intervals within a temperature range of −10 to 60°C. The empirical value is 2500 to 4500 J / mol. The universal gas constant is 8.314 J / (mol·K); The temperature of the damping medium is measured in real time, and the unit is K. This is a reference temperature in Kelvin (K), with a default value of 296K.

[0080] The formula for calculating the required operating current of a thermoelectric cooler is as follows:

[0081] ;

[0082] In the formula, This represents the operating current of the thermoelectric cooler, measured in amperes (A). The total cooling power required to maintain the target damping coefficient, in watts (W). The cooling capacity of a single semiconductor refrigeration chip is expressed in W, with an empirical value of 15 to 25 W. This is the maximum operating current of the thermoelectric cooler, measured in amperes (A), typically 3.5A. The number of installed thermoelectric coolers, ranging from 4 to 6.

[0083] Temperature-compensated viscous dampers are velocity-dependent dampers that generate damping force through the relative motion between the piston rod and the cylinder. The damping medium is a modified polyalphaolefin base oil compounded with nano-polymers. Thickener with a viscosity index of not less than 180 and a kinematic viscosity change rate of less than 35% in a temperature range of 5 to 50°C; used in semiconductor refrigeration chips. The material has a single-chip cooling capacity of 15 to 25W. The cooling power is adjusted by pulse width modulation. The temperature control accuracy is ±0.5℃. The damping medium viscosity-temperature characteristic compensation model takes the real-time measured temperature of the damping medium and the outer wall temperature of the cylinder as input and outputs the semiconductor cooling chip operating current required to maintain the target damping coefficient. The damping medium viscosity-temperature characteristic compensation model updates the calculation results every 30 seconds.

[0084] The specific implementation of step S40 is to deploy triaxial acceleration sensors and strain sensors at key nodes of the grid structure. The total number of sensors is not less than 8% of the total number of nodes in the grid structure. The sampling frequency is set to 200Hz. The acceleration signals collected by the triaxial acceleration sensors are modally identified using a frequency domain decomposition algorithm. The first 5 modal frequencies and modal damping ratios are extracted. The frequency domain decomposition algorithm uses a fast Fourier transform to convert the time-domain acceleration signal into a frequency-domain power spectrum. To improve the identification accuracy, the original acceleration signal is weighted using a Hanning window. The frequency resolution is set to 0.02Hz. The error of the first 5 identified modal frequencies does not exceed 0.5%.

[0085] The formula for calculating power spectral density is as follows:

[0086] ;

[0087] In the formula, For frequency Power spectral density at , in units of / Hz; This represents the number of sampling points; This is the sampling time interval, in seconds (s), with a default value of 0.005s. Acceleration signal Fourier transform, unit is m / ; Frequency, unit: Hz, subscript Frequency number; This is a time-domain acceleration signal, with units of 1. ; For time variables, the unit is seconds (s), and the subscript is... This is the time sequence number.

[0088] The modal frequencies are identified using the peak-picking method, and the modal damping ratio is calculated using the half-power bandwidth method. The formula for calculating the modal damping ratio is as follows:

[0089] ;

[0090] In the formula, For the first First-order modal damping ratio; For the first First-order modal frequency, in Hz; and For the first The power spectral density decreases to the peak value on both sides of the first modal frequency peak. The frequency corresponding to the multiple, in Hz; The modal order is denoted by 1 to 5.

[0091] The displacement response amplitude at each measuring point is obtained by performing a double integration on the acceleration signal acquired by the triaxial accelerometer. The integration algorithm adopts the Simpson numerical integration method, and the integration time step is set to 0.005s. To eliminate integration drift error, the displacement response amplitude at each measuring point is subjected to high-pass filtering, and the cutoff frequency is set to 0.1Hz. The displacement response calculation formula is expressed as follows:

[0092] ;

[0093] In the formula, For the first Each measuring point at time [time] The displacement response amplitude, in meters; For the first Acceleration signals at each measuring point, in units of ; and The variable is the integral variable, and the unit is seconds (s). The measurement point number is 1 to 1. ; This represents the total number of monitoring points. The time for displacement response calculation is given in seconds (s), and the subscript is used. This is the time sequence number.

[0094] When the deviation between the first-order modal frequency identified by the frequency domain decomposition algorithm and the initial design frequency exceeds 3%, the frequency adaptive tuning program is triggered. Every 60 seconds, the modal parameters are updated and identified through the frequency domain decomposition algorithm to obtain the updated first 5 modal frequencies and modal damping ratios. When the results of three consecutive modal parameter update identifications show that the damping ratio of a certain modality has decreased by more than 15%, the damping coefficient of the corresponding frequency-tuned mass damper unit is increased by 15% to 25% according to the optimal damping force allocation vector output by the two-layer game optimization model.

[0095] The frequency adaptive tuning program calculates the required stiffness adjustment for each tuned mass damper unit based on the deviation between the real-time modal frequency identified by the frequency domain decomposition algorithm and the current tuning frequency of each tuned mass damper unit in the multi-frequency tuned mass damper group. The adjustment strategy adopts a proportional-integral control algorithm, and the formula for calculating the stiffness adjustment is as follows:

[0096] ;

[0097] In the formula, For the first A tuned mass damper unit at time... The stiffness adjustment amount, in kN / m; This is a proportionality coefficient, typically ranging from 0.6 to 0.9; The integral coefficient is expressed in units of 1000 ppm. The integral time constant Set to 8 to 15 seconds, integral coefficient ; For the first A tuned mass damper unit at time... The frequency deviation, in Hz, is calculated using the following formula: ; For the first The target tuning frequency of each tuned mass damper unit is in Hz. For the first The current tuning frequency of each tuned mass damper unit, in Hz; The time for stiffness adjustment is expressed in seconds (s), and the subscript is used. To adjust the time sequence number; The variable is an integral variable, and the unit is seconds. Through iterative adjustment, the frequency deviation of each tuned mass damper unit is brought to within ±1%, and the entire tuning process lasts no more than 180 seconds.

[0098] The specific implementation of step S50 is to establish a two-layer game optimization model, which includes an upper-layer optimization model and a lower-layer optimization model. The upper-layer optimization model outputs the optimal frequency distribution vector and mass allocation ratio vector of the multi-frequency tuned mass damper group, and the lower-layer optimization model outputs the optimal damping force allocation vector of each tuned mass damper unit.

[0099] The upper-level optimization model aims to minimize the root mean square of the displacement response at multiple points on the grid structure. The inputs include the displacement response amplitude, modal participation coefficient, and frequency deviation coefficient at each measuring point. The objective function of the upper-level optimization model is expressed as follows:

[0100] ;

[0101] In the formula, The objective function value of the upper-level optimization model is expressed in m. The optimal frequency distribution vector has a dimension of , of which element Indicates the first The tuning frequency of each tuned mass damper unit, in Hz; Assign a quality proportion vector with dimension . , of which element Indicates the first The mass ratio of each tuned mass damper unit; This represents the total number of monitoring points. For the first The weighting coefficients for each measurement point are as follows: the weight of a key node is 1.0, and the weight of a secondary node is 0.6 to 0.8. For the first The root mean square of the displacement response at each measuring point is expressed in meters. For the first First-order modal participation coefficients; For the first Frequency deviation coefficient.

[0102] The formula for calculating the root mean square of the displacement response is as follows:

[0103] ;

[0104] In the formula, The sampling time is expressed in seconds. The observation time variable is expressed in seconds (s), and the subscript is... Indicates observation.

[0105] The modal participation factor (MFI) characterizes the contribution of each mode to the overall structural response. It is calculated by dividing the inner product of the mode shape and the load distribution vector by the modal mass. A larger absolute value of the MFI indicates a more significant influence of the corresponding mode on the response. The MFI is introduced into the objective function of the upper-level optimization model to achieve focused control of the dominant modes. The formula for calculating the MFI is as follows:

[0106] ;

[0107] In the formula, For the first The first-order mode shape vector has a dimension of ; Let be the load distribution vector, with dimension . The unit is N; For the first Modal mass, in kg; superscript This represents the transpose of a vector.

[0108] The frequency deviation coefficient is calculated by dividing the absolute value of the difference between the real-time modal frequency identified by the frequency domain decomposition algorithm and the design modal frequency by the design modal frequency. It reflects the degree of structural frequency drift. When the frequency deviation coefficient exceeds 0.05, significant detuning is considered to have occurred, requiring the initiation of a frequency adaptive tuning program. In the objective function of the upper-level optimization model, it serves as a denominator term, reflecting the weakening effect of frequency deviation on vibration reduction. The formula for calculating the frequency deviation coefficient is as follows:

[0109] ;

[0110] In the formula, The first one identified by the frequency domain decomposition algorithm Real-time modal frequencies, in Hz; For the first Design mode frequency, in Hz.

[0111] The constraints of the upper-level optimization model include the frequency distribution range of each tuned mass damper element, the sum of all components in the mass distribution ratio vector being 1, and the mass ratio of a single tuned mass damper element not exceeding 40% of the total mass ratio. The constraints are stated as follows:

[0112] , , ;

[0113] In the formula, This is the lower limit of the tuning frequency, expressed in Hz. This represents the upper limit of the tuning frequency, measured in Hz.

[0114] The lower-level optimization model aims to maximize the total energy dissipation of the multi-frequency tuned mass damper group. The inputs include the relative displacement amplitude, damping coefficient, and stiffness coefficient of each tuned mass damper element. The objective function of the lower-level optimization model is expressed as follows:

[0115] ;

[0116] In the formula, This represents the objective function value of the lower-level optimization model, in N·s. Assign a vector to the optimal damping force, with dimension . , of which element Indicates the first The damping force of a tuned mass damper unit, in N; For the first The relative displacement amplitude of each tuned mass damper unit, in meters; For the first The damping coefficient of each tuned mass damper unit, in N·s / m; For the first The stiffness coefficient of each tuned mass damper element, in N / m; This is a reference stiffness coefficient, in N / m, with an empirical value of 25000 N / m.

[0117] The constraints of the lower-level optimization model include: the sum of the damping forces of each tuned mass damper unit does not exceed 18% of the maximum inertial force of the grid structure; the damping force of a single tuned mass damper unit does not exceed 3 times the weight of its mass block; and the damping coefficient can be adjusted from 0.7 to 1.5 times the initial damping coefficient value. The constraints are stated as follows:

[0118] , , ;

[0119] In the formula, The maximum inertial force of the space frame is expressed in Newtons (N). For the first The mass of a tuned mass damper unit is expressed in kg. The acceleration due to gravity is taken as 9.8. ; For the first The initial damping coefficient of each tuned mass damper unit, in N·s / m.

[0120] The upper-level optimization model and the lower-level optimization model are linked through a coupling term. The coupling term is calculated by dividing the inner product of the optimal frequency distribution vector and the optimal damping force allocation vector by the product of the magnitudes of the two vectors. The value of the coupling term reflects the degree of coordination between the optimal frequency distribution vector determined by the upper-level optimization model and the optimal damping force allocation vector optimized by the lower-level optimization model. The closer the coupling term is to 1, the better the coordination. The upper-level and lower-level optimization models are solved iteratively. When the change in the coupling term between two consecutive iterations is less than 0.02, the Nash equilibrium solution is considered to be reached. At this point, the optimal frequency distribution vector and the optimal damping force allocation vector are output. The formula for calculating the coupling term is as follows:

[0121] ;

[0122] In the formula, The numerical value of the coupling term; The magnitude of the optimal frequency distribution vector is calculated using the following formula: The unit is Hz; The modulus of the optimal damping force allocation vector is calculated using the following formula: The unit is N; Normalized damping force, unit: N, empirical value: 1000 N; symbol: This represents the vector dot product operation. The unit is Hz·N.

[0123] The specific implementation of step S60 is as follows: based on the optimal frequency distribution vector output by the two-layer game optimization model, the stiffness coefficient of each tuned mass damper unit is adjusted by the current controller of the magnetorheological elastomer variable stiffness device. The damping coefficient of each tuned mass damper unit is adjusted according to the optimal damping force distribution vector. The adjustment time interval of the stiffness coefficient is set to 10s, and the single adjustment amount does not exceed 5% of the current stiffness coefficient value. When the ambient temperature fluctuation in the wind tunnel exceeds 10℃, the required cooling power is calculated according to the damping medium viscosity-temperature characteristic compensation model, and the working current of the semiconductor cooling chip is adjusted. The working current adjustment range is 0.5 to 3.5A, and the response time does not exceed 30s.

[0124] The specific implementation of step S70 is to use the sliding window power spectral density analysis method to evaluate the vibration reduction effect and obtain the power spectral density at the peak frequency of the dominant mode. The time window length of the sliding window power spectral density analysis method is set to 120s, and the overlap rate is 50%. The sliding window power spectral density analysis method divides the long time history response signal into multiple time windows, each with a length of 120s and an overlap of 60s between adjacent time windows. The power spectral density of each time window signal is estimated separately, and the random error is reduced by averaging the power spectra of multiple time windows using the Welch method.

[0125] The dominant mode is the mode with the largest absolute value of the modal participation coefficient. When the absolute value of the modal participation coefficient corresponding to the first modal frequency is greater than 60% of the sum of the absolute values ​​of the modal participation coefficients corresponding to the other modal frequencies, the first mode is considered to be the dominant mode. The peak frequency of the dominant mode is the modal frequency corresponding to the dominant mode.

[0126] The reduction rate of power spectral density at the peak frequency of the dominant mode is defined as the difference between the peak power spectral density after vibration reduction and the peak power spectral density before vibration reduction, divided by the peak power spectral density before vibration reduction. A reduction rate of 40% or higher is considered to meet the vibration reduction requirements. The formula for calculating the reduction rate of power spectral density at the peak frequency of the dominant mode is as follows:

[0127] ;

[0128] In the formula, The rate of decrease in power spectral density at the peak frequency of the dominant mode; Peak frequency of the dominant mode before vibration reduction Power spectral density at , in units of / Hz; Peak frequency of the dominant mode after vibration reduction Power spectral density at , in units of / Hz; The dominant mode peak frequency is expressed in Hz.

[0129] When the power spectral density reduction rate at the peak frequency of the dominant mode is less than 40%, the two-layer game optimization model calculation is re-executed to update the optimal frequency distribution vector and the optimal damping force distribution vector. After updating the optimal frequency distribution vector and the optimal damping force distribution vector, the stiffness coefficient and damping coefficient of each tuned mass damper unit are adjusted according to the updated vector.

[0130] The multi-frequency tuned mass damper group is a device that combines 5 to 9 tuned mass damper units with different tuning frequencies in a series and parallel manner. The tuning frequencies of each tuned mass damper unit are distributed in an arithmetic or geometric sequence, with the frequency interval being 2 to 5% of the dominant mode frequency. The total mass ratio is controlled at 1.5 to 3% of the total mass of the grid structure. The mass distribution of each tuned mass damper unit follows the Gaussian distribution principle of maximizing the mass of the unit at the center frequency.

[0131] To better understand and implement this invention, the following is a specific application scenario of this invention, Example 2:

[0132] The technical team first constructed a 1:1 scale steel frame on the ground of the wind tunnel test section. Using a FARO FocusS350 3D laser scanner, they scanned the nodes of the grid structure, collecting 2847 node coordinate point cloud data. A deviation analysis was performed between the point cloud data and the BIM model, revealing a maximum deviation of 23mm. After generating compensation and correction data, eight high-precision benchmark points were established around the grid structure. The team used 1mm diameter high-strength piano steel wire to tension and form a spatial 3D benchmark network. Tension was monitored in real-time using an HBM U9C tension sensor and adjusted to a constant 500N. Eight Sonitor ultrasonic base stations were deployed at the diagonal intersections of the grid structure. A Leica TS60 total station was used to calibrate the 3D coordinates of the base stations, achieving a calibration accuracy of ±1.8mm.

[0133] The support frame for the multi-frequency tuned mass damper assembly was pre-assembled on the ground. This support frame uses a box-section steel structure, with adjustable connection nodes providing a three-dimensional adjustment margin of ±8cm. After hoisting into place, the technical team used an API Radian laser tracker to monitor the six-degree-of-freedom pose of the support frame in real time, with measurement accuracy of displacement ±0.04mm and angle ±4 arcseconds. Figure 2 As shown, the hydraulic synchronous jacking system adjusts synchronously at eight support points, with a maximum adjustment stroke of 120mm at a single support point and an adjustment speed of 2mm / s. After precise adjustments, the installation accuracy of the support frame is controlled within ±4.2mm, and the installation angle deviation is only 0.26°. The technical team used a BeiDou high-precision positioning system and a total station to jointly measure and establish a spatial control network, achieving real-time monitoring of the support frame's position and attitude.

[0134] In the design of the multi-frequency tuned mass damper assembly, the technical team combined and arranged seven tuned mass damper units with different tuning frequencies in a series-parallel configuration. The tuning frequencies of each unit are distributed in a geometric progression, with a center frequency of 2.38 Hz, a frequency interval of 3.2% of the dominant mode frequency, and a frequency distribution range of 2.16 to 2.62 Hz. The total mass ratio is controlled at 2.3% of the total mass of the grid structure, i.e., 12.88 t. The mass distribution of each tuned mass damper unit follows the Gaussian distribution principle, with the center frequency unit having a mass of 2.85 t, and the mass of the units on both sides decreasing sequentially.

[0135] Magnetorheological elastomer (MEE) variable stiffness devices and temperature-compensated viscous dampers were installed within each tuned mass damper unit. The MEE variable stiffness device consisted of a silicone rubber matrix embedded with carbonyl iron powder, an excitation coil, and a magnetic core. The carbonyl iron powder had a volume fraction of 35% and a particle size of 5 μm, and the excitation coil had 520 turns. By applying an excitation current from 0 to 2 A, the shear modulus of the silicone rubber matrix increased from the ground state of 1.05 MPa to the saturation state of 3.2 MPa, with a response time of less than 85 ms. The technical team tested the dynamic stiffness adjustment range of each tuned mass damper unit, and the results showed that the stiffness could be continuously adjusted within the range of 18 to 42 kN / m, meeting the requirement of continuous frequency adjustment within ±12% of the design frequency.

[0136] The temperature-compensated viscous damper employs a velocity-dependent design where the damping force is generated by the relative motion between the piston rod and the cylinder. The damping medium is a modified polyalphaolefin base oil compounded with nano-polymers. The thickener has a viscosity index of 195 and a kinematic viscosity change rate of only 28% within a temperature range of 5 to 50°C. Five sheets are evenly distributed on the outer wall of the damper cylinder. The material is a semiconductor refrigeration chip with a single chip cooling capacity of 20W. The technical team deployed a PT100 temperature sensor to monitor the temperature of the damping medium in real time. When the temperature deviates from the set value of 23℃ by more than ±2℃, the temperature compensation control program is activated. The operating current of the semiconductor refrigeration chip is adjusted through pulse width modulation. The current adjustment range is 0.5 to 3.5A, the response time is only 26s, and the temperature control accuracy reaches ±0.4℃.

[0137] Triaxial accelerometers and strain sensors were deployed at key nodes of the space frame, totaling 236 sensors, accounting for 9.2% of the total number of nodes. The sampling frequency was set to 200Hz. The technical team used PCB 356A15 triaxial accelerometers with a sensitivity of 100mV / g and a frequency response range of 0.5 to 3000Hz. The time-domain acceleration signal was converted into a frequency-domain power spectrum using Fast Fourier Transform (FFT). Peak picking was used to identify the modal frequencies, and the modal damping ratio was calculated using the half-power bandwidth method. To improve identification accuracy, the original acceleration signal was weighted using a Hanning window, and the frequency resolution was set to 0.02Hz. Figure 3 As shown, the first five modal frequencies were identified as 2.38Hz, 3.76Hz, 4.95Hz, 6.23Hz, and 7.58Hz, with corresponding modal damping ratios of 0.018, 0.022, 0.025, 0.019, and 0.021. The deviation of the first modal frequency from the initial design frequency of 2.45Hz was 2.86%, which did not exceed the 3% threshold, therefore the frequency adaptive tuning procedure was not triggered.

[0138] The technical team established a two-layer game-theoretic optimization model. The upper-layer optimization model aims to minimize the root mean square (RMS) of the displacement response at multiple points on the grid structure, while the lower-layer optimization model aims to maximize the total energy dissipation of the multi-frequency tuned mass damper array. The objective function of the upper-layer optimization model is the minimum value of the sum of the weighted sum of the RMS of the displacement responses at all monitoring points, multiplied by the modal participation coefficient, and then divided by the frequency deviation coefficient plus 1. Weight coefficients are assigned according to the importance of each measuring point in the grid structure, with critical nodes having a weight of 1.0 and secondary nodes having a weight of 0.7. Constraints include a frequency distribution range of 2.16 to 2.62 Hz for each tuned mass damper element, a sum of 1 for all components in the mass distribution ratio vector, and a mass ratio of no more than 40% of the total mass ratio for a single tuned mass damper element. The objective function of the lower-layer optimization model is the maximum value of the sum of the square roots of the products of the relative displacement amplitude and damping coefficient of each tuned mass damper element, divided by the stiffness coefficient. The constraints include that the sum of the damping forces of each tuned mass damper unit does not exceed 18% of the maximum inertial force of the space frame, the damping force of a single tuned mass damper unit does not exceed 3 times the weight of its mass block, and the damping coefficient can be adjusted from 0.7 to 1.5 times the initial damping coefficient value.

[0139] The upper-level optimization model and the lower-level optimization model are linked through a coupling term, which is the inner product of the optimal frequency distribution vector and the optimal damping force allocation vector divided by the product of the magnitudes of the two vectors. Through alternating iterative solutions, in the 8th iteration, the change in the coupling term between two consecutive iterations is 0.016, which is less than the threshold of 0.02, indicating that a Nash equilibrium solution has been reached. The output optimal frequency distribution vector and optimal damping force allocation vector at this point are shown in Table 1.

[0140] Table 1 Optimal Frequency Distribution and Damping Force Allocation

[0141]

[0142] Based on the optimal frequency distribution vector, the technical team adjusted the stiffness coefficient of each tuned mass damper unit using the current controller of the magnetorheological elastomer variable stiffness device. The adjustment time interval for the stiffness coefficient was set to 10 seconds, and the single adjustment amount did not exceed 5% of the current stiffness coefficient value. The damping coefficient of each tuned mass damper unit was adjusted according to the optimal damping force distribution vector. When the ambient temperature fluctuation in the wind tunnel exceeded 10°C, the required cooling power was calculated based on the damping medium viscosity-temperature characteristic compensation model established according to the Arrhenius equation. The model parameters were obtained through viscosity calibration tests conducted at 5°C intervals within the temperature range of −10 to 60°C. The model input consisted of real-time measured damping medium temperature and cylinder outer wall temperature, and the output was the operating current of the semiconductor cooling chip required to maintain the target damping coefficient. The model updated the calculation results every 30 seconds.

[0143] The technical team employed a sliding window power spectral density analysis method to evaluate the vibration reduction effect, with a time window length of 120 seconds and an overlap rate of 50%. Power spectral density was estimated for each time window signal, and the Welch method was used to average the power spectra across multiple time windows to reduce random errors. When the wind tunnel was running at a wind speed of 80 m / s, the power spectral density at the peak frequency of the dominant mode decreased from 0.0285 before vibration reduction. / Hz decreased to 0.0162 after vibration reduction. / Hz, the reduction rate reaches 43.2%, meeting the requirement of over 40%.

[0144] During wind tunnel operation, the technical team updated and identified modal parameters every 60 seconds using a frequency domain decomposition algorithm. At the 1800th second, three consecutive modal parameter updates showed that the first-order modal damping ratio had decreased from an initial 0.018 to 0.015, a decrease of 16.7%, exceeding the 15% threshold. Based on the optimal damping force distribution vector output by the two-layer game optimization model, the team increased the damping coefficient of the corresponding frequency-tuned mass damper unit by 20%, adjusting it from 5.45 kN·s / m to 6.54 kN·s / m. After the adjustment, the first-order modal damping ratio recovered to 0.019 within 300 seconds, effectively suppressing the increase in vibration response.

[0145] At 3600 seconds, the frequency domain decomposition algorithm identified the first-order mode frequency as 2.46 Hz, which deviated from the initial design frequency of 2.45 Hz by 0.41%, but deviated from the current tuning frequency of 2.38 Hz by 3.36%, exceeding the 3% threshold and triggering the frequency adaptive tuning program. The technical team used a proportional-integral control algorithm for tuning, setting the proportional coefficient to 0.75 and the integral time constant to 12 s. Through iterative adjustment, the frequency deviation of each tuned mass damper unit converged to within ±0.8% within 150 seconds, and the entire tuning process lasted 150 seconds. The frequency distribution and damping force distribution after tuning are shown in Table 2.

[0146] Table 2 Frequency distribution and damping force distribution after tuning

[0147]

[0148] After frequency adaptive tuning, the power spectral density at the peak frequency of the dominant mode was further reduced to 0.0138. / Hz, a reduction rate of 51.6%. The technical team conducted statistical analysis on the displacement response of key nodes of the space frame. The displacement response amplitude of each measuring point was obtained by double integration of the acceleration signals collected by the triaxial accelerometer. The integration algorithm adopted the Simpson numerical integration method, and the integration time step was set to 0.005s. To eliminate integration drift error, the displacement response amplitude of each measuring point was high-pass filtered, and the cutoff frequency was set to 0.1Hz. Statistical results show that after vibration reduction, the average displacement response amplitude of each measuring point decreased by 48.5%, and the maximum displacement response decreased from 12.6mm before vibration reduction to 6.3mm after vibration reduction.

[0149] It should be noted that the variables involved in this invention are explained in detail in Tables 3 and 4.

[0150] Table 3. Variable Explanation Table (Part 1)

[0151]

[0152] Table 4. Variable Explanation Table (Part Two)

[0153]

[0154] 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 hybrid vibration reduction method for large-span steel space frames in wind tunnels based on TMD and viscous dampers, characterized in that, A proportional steel frame was erected on the ground. A 3D laser scanner was used to collect the coordinate point cloud data of the space frame nodes, and the data was compared with the BIM model to generate compensation and correction data. High-precision reference points were set around the space frame, and a spatial 3D reference line network was formed using piano wire tensioning. The multi-frequency tuned mass damper assembly support frame was pre-assembled on the ground, with adjustable connection nodes allowing for three-dimensional adjustment. After hoisting into place, a laser tracker was used to monitor the six-degree-of-freedom pose of the multi-frequency tuned mass damper assembly support frame in real time. Magnetorheological elastomer variable stiffness devices and temperature-compensated viscous dampers were installed in each tuned mass damper unit of the multi-frequency tuned mass damper assembly. Three-dimensional accelerometers and strain sensors were deployed at key nodes of the space frame, and frequency domain decomposition algorithms were used to analyze the three-dimensional accelerometer data. Modal identification is performed on the acceleration signals collected by the accelerometer to extract the first five modal frequencies and modal damping ratios. A two-layer game optimization model is established. The upper-layer optimization model outputs the optimal frequency distribution vector and mass allocation ratio vector of the multi-frequency tuned mass damper group, while the lower-layer optimization model outputs the optimal damping force allocation vector of each tuned mass damper unit. Cooperative optimization is achieved through the coupling terms of the upper and lower-layer optimization models. Based on the optimal frequency distribution vector output by the two-layer game optimization model, the stiffness coefficient of each tuned mass damper unit is adjusted by the current controller of the magnetorheological elastomer variable stiffness device, and the damping coefficient of each tuned mass damper unit is adjusted according to the optimal damping force allocation vector. The vibration reduction effect is evaluated using the sliding time window power spectral density analysis method.

2. The method according to claim 1, characterized in that, High-precision reference points are set up at the intersection of the diagonals around the grid structure. Ultrasonic base stations are set up for precise coordinate calibration. The three-dimensional coordinates of the ultrasonic base stations are precisely calibrated using a total station.

3. The method according to claim 2, characterized in that, Based on the six-degree-of-freedom pose deviation data output by the laser tracker, the spatial position of the multi-frequency tuned mass damper assembly support frame is adjusted, and a spatial control network is established by combining the BeiDou high-precision positioning system with a total station for measurement.

4. The method according to claim 3, characterized in that, Semiconductor cooling chips are evenly installed on the outer wall of the cylinder of the temperature-compensated viscous damper. The temperature of the damping medium is monitored in real time by a temperature sensor, and the temperature compensation control program is activated to adjust the working current of the semiconductor cooling chips.

5. The method according to claim 4, characterized in that, When the deviation between the first-order modal frequency identified by the frequency domain decomposition algorithm and the initial design frequency exceeds the set value, the frequency adaptive tuning program is triggered. The modal parameters are updated and identified through the frequency domain decomposition algorithm to obtain the updated first 5 modal frequencies and modal damping ratios.

6. The method according to claim 5, characterized in that, The upper-level optimization model aims to minimize the root mean square of the multi-point displacement response of the grid structure, while the lower-level optimization model aims to maximize the total energy dissipation of the multi-frequency tuned mass damper group. The coupling term is the inner product of the optimal frequency distribution vector and the optimal damping force distribution vector divided by the product of the magnitudes of the two vectors.

7. The method according to claim 6, characterized in that, The stiffness coefficient adjustment time interval is set to a fixed value. When the ambient temperature fluctuation in the wind tunnel exceeds the set value, the required cooling power is calculated based on the damping medium viscosity-temperature characteristic compensation model, and the working current of the semiconductor cooling chip is adjusted.

8. The method according to claim 7, characterized in that, The time window length of the sliding time window power spectral density analysis method is set to a fixed value. After updating the optimal frequency distribution vector and the optimal damping force distribution vector, the stiffness coefficient and damping coefficient of each tuned mass damper unit are adjusted according to the updated vector.

9. The method according to claim 8, characterized in that, A multi-frequency tuned mass damper group is a device that combines tuned mass damper units with different tuning frequencies in a series and parallel manner. The mass distribution of each tuned mass damper unit follows the Gaussian distribution principle that maximizes the mass of the unit at the center frequency.

10. The method according to claim 9, characterized in that, The magnetorheological elastomer variable stiffness device consists of a silicone rubber matrix embedded with carbonyl iron powder, an excitation coil, and a magnetic core. Under the action of an external magnetic field, the shear modulus of the silicone rubber matrix changes, and the natural frequency of each tuned mass damper unit can be continuously adjusted by adjusting the excitation current.

Images