A method for testing the vibration response of a wind tunnel facility installation

Abstract

The application provides a method for installing a wind tunnel equipment vibration response test, and belongs to the technical field of wind tunnels.The method installs multiple three-axis acceleration sensors on the main structure nodes of the wind tunnel equipment and establishes a vibration response game optimization model, processes original vibration signals by using a vibration signal enhancement model, constructs a micro-amplitude vibration feature matrix, a signal superposition weight matrix and a false vibration index matrix for signal analysis, calculates a comprehensive vibration evaluation index by using a vibration response evaluation function, adjusts input size parameters according to different numerical ranges by using corresponding optimization strategies, realizes adaptive signal processing and temperature compensation, and finally outputs a test report containing vibration response data, false signal identification results and structure safety evaluation conclusions, thereby solving the technical problem that the prior art cannot accurately distinguish between real structure vibration of the wind tunnel equipment and false vibration signals caused by temperature changes.

Description

Technical Field

[0001] This invention belongs to the field of wind tunnel technology, and more specifically, relates to a method for testing the vibration response of wind tunnel equipment installation. Background Technology

[0002] As core equipment in aerospace and engineering fluid mechanics research, wind tunnel equipment's vibration response characteristics directly affect testing accuracy and equipment safety. Traditional vibration testing methods primarily employ simple accelerometer placement and time-frequency domain signal analysis to monitor the equipment's vibration state. However, these traditional techniques face significant limitations in practical applications. Due to the drastic temperature changes that occur during wind tunnel operation, leading to structural thermal expansion and contraction, accelerometers simultaneously receive both genuine structural vibration signals and spurious vibration signals caused by temperature variations. Existing signal processing methods cannot effectively distinguish between these two different types of signals. Current wind tunnel vibration monitoring technology lacks temperature compensation mechanisms and intelligent signal recognition capabilities. Testers often misinterpret sensor displacement caused by thermal expansion as genuine structural vibration. This misinterpretation not only affects the accuracy of vibration test results but may also lead to incorrect assessments of the equipment's safety status, thus failing to provide reliable technical basis for equipment maintenance and optimization. Summary of the Invention

[0003] This invention provides a method for testing the vibration response of wind tunnel equipment installation, which can solve the technical problem in the prior art that it is impossible to accurately distinguish between the actual structural vibration of wind tunnel equipment and the false vibration signals caused by temperature changes.

[0004] To achieve the above objectives, the present invention adopts the following technical solution: A method for testing the vibration response of wind tunnel equipment includes installing multiple triaxial accelerometers at key structural nodes of the wind tunnel equipment. The installation locations of the triaxial accelerometers cover the support structure, air supply system, and test section of the wind tunnel equipment. A game-theoretic optimization model for the wind tunnel equipment's vibration response is established, comprising an upper-level model aimed at suppressing small-amplitude vibrations and a lower-level model aimed at eliminating spurious signals. The upper-level and lower-level models are linked through a structural temperature-vibration coupling term to achieve coordinated optimization of temperature compensation and vibration control. The wind tunnel equipment is started and multiple operating conditions are set, including different wind speeds and different temperatures. Temperature sensors are used to measure the vibration response. Sensors monitor temperature changes within the wind tunnel, establishing a temperature-vibration response correlation database. A vibration signal enhancement model processes the raw vibration signals collected by the triaxial accelerometer, optimizing signal recognition accuracy by adjusting input size parameters. Three key matrices are constructed for vibration signal analysis: a micro-amplitude vibration characteristic matrix, a signal superposition weight matrix, and a spurious vibration index matrix. A comprehensive vibration evaluation index is calculated using a vibration response evaluation function. Different optimization strategies are employed to adjust the input size parameters of the vibration signal enhancement model when the comprehensive vibration evaluation index falls within different numerical ranges, achieving adaptive signal processing. A wind tunnel equipment installation vibration response test report is output.

[0005] The number of triaxial accelerometers shall not be less than 8, and the sampling frequency of the triaxial accelerometers shall be set to 2000Hz.

[0006] The objective function of the upper-level model of the wind tunnel equipment vibration response game optimization model is the function for minimizing the superimposed vibration response of the wind tunnel structure, which is used to minimize the overall vibration response of the wind tunnel structure. The inputs of the upper-level model objective function include the structural damping coefficient, natural frequency, excitation force amplitude, temperature gradient and material elastic modulus. The output of the upper-level model objective function is the optimal control strategy for the structural vibration response.

[0007] The objective function of the lower-level model of the wind tunnel equipment vibration response game optimization model is a thermal expansion spurious signal filtering function, which is used to eliminate spurious vibration signals caused by temperature changes. The inputs of the lower-level model objective function include the temperature change rate, thermal expansion coefficient, structural geometric parameters and signal frequency characteristics. The outputs of the lower-level model objective function are the spurious signal identification parameters and filtering parameters.

[0008] The specific structure of the vibration signal enhancement model is a signal processing network based on a vision transformer architecture. The vibration signal enhancement model includes multiple coding layers and an adaptive input size adjustment mechanism. The adaptive input size adjustment mechanism balances computational resources and the degree of signal detail preservation by dynamically adjusting the resolution of the original vibration signal.

[0009] The core of the vibration signal enhancement model includes a position encoding module, a multi-head self-attention mechanism, and a feedforward neural network layer. The vibration signal enhancement model uses a supervised learning method to train parameters, and the loss function combines signal reconstruction error and classification accuracy. The Adam optimizer is used to update the parameters.

[0010] The micro-amplitude vibration characteristic matrix is ​​used to describe the vibration modal characteristics of the wind tunnel equipment structure under micro-excitation. The matrix elements of the micro-amplitude vibration characteristic matrix include the frequency, damping ratio and mode shape coefficient of each mode. The dynamic characteristic parameters of the structure are obtained by modal analysis.

[0011] The signal superposition weight matrix is ​​used to quantify the contribution of multiple vibration signal sources to the overall response. The matrix elements of the signal superposition weight matrix represent the weight coefficients of different excitation sources in each frequency band. The signal superposition weight matrix determines the superposition relationship of each signal component through frequency domain analysis and signal decomposition techniques.

[0012] The spurious vibration index matrix is ​​used to identify and quantify non-real vibration signals caused by temperature changes. The matrix elements of the spurious vibration index matrix include the correspondence between temperature gradient, thermal expansion displacement and spurious acceleration. The spurious vibration index matrix establishes a mathematical model of spurious signals through temperature vibration correlation analysis.

[0013] The vibration response evaluation function is used to comprehensively evaluate the vibration state and signal processing effect of the wind tunnel equipment. The inputs of the vibration response evaluation function include the amplitude of micro-vibration, signal superposition coefficient, spurious vibration index, temperature change rate and structural natural frequency. The output of the vibration response evaluation function is a comprehensive vibration evaluation index between 0 and 1.

[0014] Specifically, when the comprehensive vibration assessment index ∈ [0, 0.249], high-resolution input size parameters are used to obtain fine signal characteristics; when the comprehensive vibration assessment index ∈ (0.249, 0.683), medium-resolution input size parameters are used to balance processing accuracy and computational efficiency; when the comprehensive vibration assessment index ∈ [0.683, 1], low-resolution input size parameters are used to quickly identify obvious vibration anomalies.

[0015] Specifically, the high-resolution input size parameter corresponds to an input data resolution of 2048 sampling points for the vibration signal enhancement model, the medium-resolution input size parameter corresponds to an input data resolution of 1024 sampling points for the vibration signal enhancement model, and the low-resolution input size parameter corresponds to an input data resolution of 512 sampling points for the vibration signal enhancement model.

[0016] The false acceleration refers to the non-structural vibration acceleration signal generated by the sensor position change caused by the thermal expansion and contraction of the wind tunnel equipment. False acceleration causes structural deformation due to temperature changes, resulting in misjudgment signals from the triaxial accelerometer.

[0017] The temperature vibration response association database stores the vibration response characteristics of wind tunnel equipment under different temperature conditions. The temperature vibration response association database is used to establish the mapping relationship between temperature change and vibration response. The structural temperature vibration coupling term represents the interaction mechanism between the temperature field and the vibration field. The structural temperature vibration coupling term is mathematically expressed through thermoelastic theory.

[0018] The adaptive signal processing dynamically adjusts the signal processing strategy based on the real-time vibration state, and determines the optimal processing parameters through the output of the vibration response evaluation function.

[0019] The wind tunnel equipment installation vibration response test report includes vibration response data under various operating conditions, false signal identification results, and structural safety assessment conclusions.

[0020] Specifically, the principle of this invention is as follows: The core of this invention's ability to distinguish between real and spurious vibration signals lies in establishing a coupling relationship model between temperature and vibration response. Through game theory optimization, vibration control and temperature compensation are unified within a mathematical framework for collaborative processing. The upper-level model aims to minimize the overall vibration response of the wind tunnel structure, taking into input parameters such as structural damping coefficient, natural frequency, excitation force amplitude, temperature gradient, and material elastic modulus, and outputting the optimal control strategy for the structural vibration response. The lower-level model aims to eliminate spurious vibration signals caused by temperature changes, taking into input parameters such as temperature change rate, thermal expansion coefficient, structural geometric parameters, and signal frequency characteristics, and outputting parameters for identifying and filtering spurious signals. The two models are linked through a structural temperature-vibration coupling term, achieving active compensation for temperature effects. The vibration signal enhancement model adopts a visual transformer architecture, using a position encoding module, a multi-head self-attention mechanism, and a feedforward neural network layer to automatically learn the time-frequency domain characteristics of the vibration signal. Furthermore, it dynamically adjusts the processing accuracy based on signal complexity through an adaptive input size adjustment mechanism. The establishment of three key matrices provides a quantitative tool for signal analysis: the micro-amplitude vibration feature matrix describes the structural dynamics, the signal superposition weight matrix quantifies the contribution of multi-source excitation, and the spurious vibration index matrix identifies temperature interference components. The vibration response evaluation function comprehensively considers multiple parameters such as micro-amplitude vibration amplitude, signal superposition coefficient, and spurious vibration index. Through weighted summation and normalization, it outputs a comprehensive evaluation index between 0 and 1. The signal processing strategy is automatically adjusted according to different value ranges of this index to ensure optimal signal recognition results under different operating conditions.

[0021] First, the two-layer game theory optimization model transforms vibration control and signal processing problems into a game theory framework, enabling collaborative solutions for multi-objective optimization. Traditional methods typically treat vibration control and signal processing as independent problems, lacking a systematic coordination mechanism. The game theory optimization model, through the interaction between the upper and lower layers, can find the optimal balance between vibration suppression and spurious signal elimination, significantly improving overall control performance. Compared to traditional single-objective optimization, this method exhibits greater robustness and adaptability, enabling it to cope with the complex and ever-changing wind tunnel operating environment.

[0022] Secondly, the signal enhancement technique based on a visual transformer introduces advanced deep learning architectures from the field of computer vision into vibration signal processing, overcoming the limitations of traditional signal processing methods. Traditional methods such as Fourier transform and wavelet analysis are insufficient when processing non-stationary and nonlinear signals, while the visual transformer, through its self-attention mechanism, can capture long-range dependencies and complex patterns in the signal, achieving accurate differentiation between real and spurious vibration signals. Compared to traditional methods, this technique has stronger feature learning capabilities and higher processing accuracy, especially demonstrating significant advantages in processing complex spurious signals caused by temperature changes.

[0023] Third, the adaptive input size adjustment technology achieves an intelligent balance between computational efficiency and processing accuracy by dynamically adjusting the signal processing resolution. Traditional fixed-resolution processing methods cannot adapt to vibration signals of different intensities and types, resulting in either insufficient processing accuracy or wasted computational resources. The adaptive adjustment mechanism selects the optimal processing parameters in real time based on the vibration state, maximizing computational efficiency while ensuring processing quality, enabling the system to adapt to the full range of monitoring needs from weak vibrations to strong anomalies.

[0024] The multidimensional correlation matrix analysis technique achieves multidimensional quantitative analysis of vibration signals by constructing a micro-amplitude vibration characteristic matrix, a signal superposition weight matrix, and a spurious vibration index matrix. Traditional analysis methods are often based on single-dimensional or simple time-frequency analysis, which is difficult to fully reflect the vibration characteristics of complex structures. The multidimensional matrix analysis method can comprehensively evaluate the vibration state from multiple perspectives such as structural dynamics, signal superposition, and temperature coupling, providing more accurate and comprehensive analysis results.

[0025] The synergistic effect of these key technological approaches forms a complete intelligent vibration monitoring and analysis system. The game theory optimization model provides the theoretical framework and optimization objectives for the entire system; the visual transformer model offers powerful signal processing capabilities; the adaptive adjustment mechanism ensures the system's real-time performance and efficiency; and multi-dimensional matrix analysis provides comprehensive state assessment. These four technological approaches complement each other, jointly solving key problems in traditional wind tunnel vibration monitoring such as low accuracy, poor adaptability, and low processing efficiency. This achieves a technological leap from passive monitoring to active analysis, from qualitative judgment to quantitative assessment, and from single indicators to comprehensive evaluation, providing advanced technical means for the safe operation and performance optimization of wind tunnel equipment.

[0026] Furthermore, this invention addresses the technical problems of insufficient vibration signal feature extraction and difficulty in quantifying the superposition relationship of multi-source vibration signals in existing technologies. Traditional methods mainly rely on basic statistical features in the time and frequency domains, lacking the ability to mine deep modal features of vibration signals, and also lack effective means to separate and quantify the superposition effect of vibrations from multiple excitation sources. This invention constructs three key analysis matrices: a micro-amplitude vibration feature matrix to obtain structural dynamic characteristic parameters through modal analysis; a signal superposition weight matrix to determine the superposition relationship of each signal component through frequency domain analysis and signal decomposition techniques; and a spurious vibration index matrix to establish a mathematical model of spurious signals through temperature-vibration correlation analysis. This achieves a comprehensive quantitative description and multi-dimensional accurate analysis of the vibration state of wind tunnel equipment. Attached Figure Description

[0027] Figure 1 This is a flowchart of the method of the present invention.

[0028] Figure 2 This is a schematic diagram of the neural network structure of the vibration signal enhancement model involved in the present invention. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0030] Example 1 like Figure 1 The diagram shown illustrates the process of this embodiment, including the following steps: S01. Install multiple triaxial accelerometers at the main structural nodes of the wind tunnel equipment. The installation positions of the triaxial accelerometers cover the support structure, air supply system and test section of the wind tunnel equipment. The number of triaxial accelerometers shall not be less than 8, and the sampling frequency of the triaxial accelerometers shall be set to 2000Hz. S02. Establish a game-theoretic optimization model for the vibration response of wind tunnel equipment. The game-theoretic optimization model for the vibration response of wind tunnel equipment includes an upper-level model with the goal of suppressing small-amplitude vibration and a lower-level model with the goal of eliminating false signals. The objective function of the upper-level model is the function of minimizing the superimposed vibration response of the wind tunnel structure, and the objective function of the lower-level model is the function of filtering false signals of thermal expansion. S03. Start the wind tunnel equipment and set multiple operating conditions, including different wind speed conditions and different temperature conditions. At the same time, use a temperature sensor to monitor the temperature change inside the wind tunnel and establish a temperature vibration response correlation database. S04. The original vibration signal collected by the triaxial accelerometer is processed using a vibration signal enhancement model. The vibration signal enhancement model optimizes the signal recognition accuracy by adjusting the input size parameters. The input size parameters are determined based on the wind tunnel operating temperature variation range, the thermal expansion coefficient of the structural material, and the vibration frequency range. S05. Construct three key matrices for vibration signal analysis: the micro-amplitude vibration feature matrix, the signal superposition weight matrix, and the spurious vibration index matrix. The micro-amplitude vibration feature matrix, the signal superposition weight matrix, and the spurious vibration index matrix are used to quantify the real vibration response of the wind tunnel equipment under different operating conditions. S06. Calculate the comprehensive vibration evaluation index through the vibration response evaluation function. When the comprehensive vibration evaluation index falls within different numerical ranges, use different optimization strategies to adjust the input size parameters of the vibration signal enhancement model to achieve adaptive signal processing. S07. Output wind tunnel equipment installation vibration response test report, which includes vibration response data under various operating conditions, false signal identification results, and structural safety assessment conclusions.

[0031] The upper-level objective function of the wind tunnel equipment vibration response game-theoretic optimization model minimizes the overall vibration response of the wind tunnel structure. The inputs to the upper-level objective function include the structural damping coefficient, natural frequency, excitation force amplitude, temperature gradient, and material elastic modulus. The output of the upper-level objective function is the optimal control strategy for the structural vibration response. The lower-level objective function eliminates spurious vibration signals caused by temperature changes. The inputs to the lower-level objective function include the rate of temperature change, coefficient of thermal expansion, structural geometric parameters, and signal frequency characteristics. The output of the lower-level objective function consists of spurious signal identification parameters and filtering parameters. The upper-level and lower-level models are linked through a structural temperature-vibration coupling term, achieving coordinated optimization of temperature compensation and vibration control.

[0032] The specific structure of the vibration signal enhancement model is a signal processing network based on a vision converter architecture. The model includes multiple encoding layers and an adaptive input size adjustment mechanism. This mechanism balances computational resources with signal detail preservation by dynamically adjusting the resolution of the original vibration signal. The core of the vibration signal enhancement model includes a position encoding module, a multi-head self-attention mechanism, and a feedforward neural network layer. The steps for establishing the training dataset for the vibration signal enhancement model specifically include collecting vibration signal data from different types of wind tunnel equipment under various operating conditions; standardizing and preprocessing the vibration signal data and labeling real and spurious vibration signals; and establishing a multidimensional dataset containing temperature changes, structural parameters, and vibration responses. The multidimensional dataset has no fewer than 10,000 sample points and covers various states such as normal operation, abnormal vibration, and thermal expansion interference. The specific steps for training the vibration signal enhancement model include training the model parameters using a supervised learning method, combining the loss function with the signal reconstruction error and classification accuracy, updating the parameters using the Adam optimizer, employing an early stopping mechanism to prevent overfitting during training, and performing cross-validation after the vibration signal enhancement model converges to ensure generalization performance. Ultimately, the vibration signal enhancement model achieves a signal recognition accuracy of over 95% on the test set.

[0033] The micro-amplitude vibration characteristic matrix is ​​used to describe the vibration modal characteristics of the wind tunnel equipment structure under micro-excitation. The matrix elements of the micro-amplitude vibration characteristic matrix include the frequency, damping ratio, and mode shape coefficient of each mode. The dynamic characteristic parameters of the structure are obtained through modal analysis. The signal superposition weight matrix is ​​used to quantify the contribution of multiple vibration signal sources to the overall response. The matrix elements of the signal superposition weight matrix represent the weight coefficients of different excitation sources in each frequency band. The superposition relationship of each signal component is determined through frequency domain analysis and signal decomposition techniques. The spurious vibration index matrix is ​​used to identify and quantify non-real vibration signals caused by temperature changes. The matrix elements of the spurious vibration index matrix include the correspondence between temperature gradient, thermal expansion displacement, and spurious acceleration. The spurious vibration index matrix establishes a mathematical model of spurious signals through temperature-vibration correlation analysis.

[0034] The vibration response evaluation function is used to comprehensively evaluate the vibration state and signal processing effect of the wind tunnel equipment. The inputs to the vibration response evaluation function include the amplitude of the micro-vibration, the signal superposition coefficient, the spurious vibration index, the temperature change rate, and the natural frequency of the structure. The output of the vibration response evaluation function is a comprehensive vibration evaluation index between 0 and 1. The vibration response evaluation function achieves comprehensive evaluation of multiple parameters through weighted summation, exponential operation, and normalization. When the comprehensive vibration evaluation index a∈[0, 0.249], high-resolution input size parameters are used to obtain fine signal characteristics; when the comprehensive vibration evaluation index a∈(0.249, 0.683), medium-resolution input size parameters are used to balance processing accuracy and computational efficiency; when the comprehensive vibration evaluation index a∈[0.683, 1], low-resolution input size parameters are used to quickly identify obvious vibration anomalies. The adjustment of the input size parameters directly affects the signal processing accuracy and computational resource consumption of the vibration signal enhancement model.

[0035] The structural damping coefficient was obtained through free vibration decay testing. During the test, an initial displacement excitation was applied to the wind tunnel equipment, and the vibration decay process was measured. The structural damping coefficient ranged from 0.01 to 0.05. The natural frequencies were determined through modal testing. Force hammer excitation and acceleration response measurements were used to determine the natural frequencies of the wind tunnel equipment, with a range of 5 Hz to 200 Hz. The excitation force amplitude was calculated from the aerodynamic forces during wind tunnel operation. The magnitude of the excitation force was determined based on wind speed and blade parameters, with a range of 100 N to 5000 N. The temperature gradient was measured by a temperature sensor array. The temperature gradient represents the temperature difference at different locations within the wind tunnel, with a range of 1℃ / m to 10℃ / m. The material elastic modulus is an inherent property of the wind tunnel structural material, determined through material mechanics testing. The material elastic modulus ranged from [missing value]. Pa to Pa.

[0036] The temperature change rate is calculated through continuous monitoring by a temperature sensor. It represents the degree of temperature change per unit time, and its value ranges from 0.1℃ / min to 5℃ / min. The coefficient of thermal expansion is a thermophysical property parameter of the structural material. It is determined by consulting material handbooks or through thermal expansion tests, and its value ranges from [missing value]. / ℃ to / ℃. The structural geometric parameters include dimensional parameters such as the length, width, height, and wall thickness of the wind tunnel equipment, which are obtained from design drawings or determined through on-site measurements. The signal frequency characteristics are obtained through Fast Fourier Transform analysis, and include parameters such as the dominant frequency, frequency distribution, and bandwidth energy.

[0037] The amplitude of the micro-vibration was obtained by measuring with a triaxial accelerometer. The amplitude of the micro-vibration represents the vibration intensity of the structure under small-amplitude excitation, and the range of the micro-vibration amplitude is [value missing]. arrive The signal superposition coefficient is calculated through multi-signal source analysis, reflecting the contribution ratio of different vibration sources to the total response, and its value ranges from 0.1 to 1.0. The spurious vibration index is calculated using a temperature compensation algorithm, quantifying the non-structural vibration components caused by temperature changes, and its value ranges from 0 to 0.5. The structural natural frequency is the inherent dynamic characteristic of the wind tunnel equipment, determined through modal testing.

[0038] The high-resolution input size parameter corresponds to an input data resolution of 2048 sampling points for the vibration signal enhancement model. This high-resolution parameter is used to process low-amplitude, fine vibration signals, ensuring accurate identification of minute vibration features. The medium-resolution input size parameter corresponds to an input data resolution of 1024 sampling points for the vibration signal enhancement model. This medium-resolution parameter reduces computational load while maintaining reasonable processing accuracy. The low-resolution input size parameter corresponds to an input data resolution of 512 sampling points for the vibration signal enhancement model. This low-resolution parameter is used to quickly identify high-amplitude, abnormal vibration signals, prioritizing processing speed.

[0039] The spurious acceleration refers to the non-structural vibration acceleration signal generated by sensor position changes due to thermal expansion and contraction of the wind tunnel equipment. False acceleration is caused by structural deformation due to temperature changes, leading to misjudgments by the triaxial accelerometer. The range of spurious acceleration values ​​is... arrive The temperature-vibration response association database stores the vibration response characteristics of wind tunnel equipment under different temperature conditions, and is used to establish the mapping relationship between temperature changes and vibration response. The structural temperature-vibration coupling term represents the interaction mechanism between the temperature field and the vibration field, and its mathematical expression is established using thermoelastic theory. The adaptive signal processing dynamically adjusts the signal processing strategy according to the real-time vibration state, and determines the optimal processing parameters based on the output of the vibration response evaluation function.

[0040] The specific implementation methods of the above steps are described in detail below.

[0041] The specific implementation of step S01 involves achieving comprehensive monitoring of key structural nodes of the wind tunnel equipment through a systematic sensor deployment strategy. First, the sensor installation locations are determined based on the structural characteristics of the wind tunnel equipment and vibration propagation path analysis. Modal analysis theory is used to identify the main vibration nodes and inverse nodes of the structure, ensuring that the sensors are placed in vibration-sensitive areas. At least three triaxial accelerometers are installed in the support structure, located at the main support column, auxiliary support beam, and foundation connection points, respectively, to capture the vibration characteristics of the main structure. At least three sensors are installed in the air supply system, distributed near the fan unit, duct connection points, and rectifier, to monitor the vibration response caused by aerodynamic excitation. At least two sensors are installed in the test section, located at the inlet and outlet of the test section, to record the vibration state during the test. All sensors are installed using bolt fixing to ensure a rigid connection between the sensors and the structure, avoiding measurement errors caused by loose installation. The sensor sampling frequency is set to 2000Hz based on the Nyquist sampling theorem, ensuring accurate capture of all vibration signals within the operating frequency range of the wind tunnel equipment. This sampling frequency can effectively monitor the vibration response within the range of 0 to 1000Hz.

[0042] The specific implementation of step S02 involves constructing a two-layer game-theoretic optimization mathematical model to achieve synergistic optimization of vibration control and signal processing. The upper-layer model establishes a vibration response minimization objective function based on structural dynamics theory. This function uses structural damping coefficient, natural frequency, excitation force amplitude, temperature gradient, and material elastic modulus as input parameters, and solves for the optimal control strategy using the Lagrange multiplier method and gradient descent algorithm. The core of the upper-layer model is to transform the motion equations of the multi-degree-of-freedom vibration system into an optimization problem, with the goal of minimizing the overall vibration response of the structure under given constraints. The lower-layer model constructs a spurious signal filtering objective function based on signal processing theory. Input parameters include the rate of temperature change, thermal expansion coefficient, structural geometric parameters, and signal frequency characteristics. Output parameters for identifying and filtering spurious signals are provided. The lower-layer model employs an adaptive filtering algorithm and frequency domain analysis method to identify and eliminate spurious vibration signals caused by thermal expansion by establishing a mathematical mapping relationship between temperature and vibration signals. The two-layer model establishes a mathematical connection through a structural temperature-vibration coupling term. This coupling term, based on thermoelasticity theory, describes the interaction between the temperature field and the stress field, realizing a unified optimization framework for temperature compensation and vibration control.

[0043] The specific implementation of step S03 involves establishing a multi-condition operation test system and a temperature-vibration correlation database. The operating conditions are designed to cover multiple wind speed levels from 5 m / s to 50 m / s, with different temperature conditions set for each wind speed level, ranging from -10℃ to 60℃, forming a two-dimensional wind speed-temperature matrix. The wind tunnel equipment is continuously run for at least 30 minutes under each condition to ensure the system reaches thermodynamic equilibrium. Temperature sensors are distributed, with high-precision sensors installed inside and outside the wind tunnel and at key structural nodes, achieving a measurement accuracy of 0.1℃ and a sampling frequency of 10Hz. The temperature-vibration response correlation database uses a relational database structure, storing fields such as timestamps, operating parameters, temperature data, vibration data, and environmental parameters. Data preprocessing techniques are employed during database establishment, including outlier detection, data interpolation, and standardization, to ensure data quality and consistency. Correlation analysis uses correlation and regression analysis methods to establish a quantitative relationship model between temperature changes and vibration response, providing fundamental data support for subsequent false signal identification.

[0044] The specific implementation of step S04 involves using a deep learning model based on a vision transformer architecture to intelligently process the original vibration signal. The vibration signal enhancement model receives the time-domain signal collected by a triaxial accelerometer as input and first performs signal preprocessing, including denoising, filtering, and normalization. An input size parameter adjustment mechanism dynamically determines the resolution level of the signal processing based on the wind tunnel's operating status. This mechanism is based on the signal fidelity principle in information theory, seeking an optimal balance between computational efficiency and signal detail preservation. The determination of the input size parameters is based on factors including the wind tunnel's operating temperature variation range; when the temperature variation rate exceeds 2℃ / min, a high-resolution processing mode is used. The thermal expansion coefficient of the structural material affects the intensity of false signals; materials with a larger thermal expansion coefficient require more refined signal analysis. The vibration frequency range determines the required sampling resolution. The model uses an attention mechanism to identify key features in the signal, calculating the correlation between different time points and frequency components through multi-head self-attention, thereby effectively distinguishing between real and false vibration signals.

[0045] The specific implementation of step S05 involves constructing three key analysis matrices to achieve multi-dimensional quantitative evaluation of vibration signals. The micro-amplitude vibration characteristic matrix is ​​established based on modal analysis theory. By conducting modal tests on the structure, the frequencies, damping ratios, and mode shapes of each mode are obtained. The matrix dimension is the modal order multiplied by the number of characteristic parameters, typically a 20×3 matrix structure. This matrix uses the least squares method and frequency domain analysis techniques to extract the inherent dynamic characteristics of the structure from the measured vibration data. The signal superposition weight matrix is ​​constructed using independent component analysis and blind source separation techniques. Matrix elements represent the contribution weights of different excitation sources in each frequency band. The matrix dimension is the number of excitation sources multiplied by the number of frequency bands. This matrix uses a signal decomposition algorithm to separate the composite vibration signal into multiple independent signal components, quantifying the influence of each component on the overall response. The spurious vibration index matrix is ​​established based on a temperature compensation algorithm. Matrix elements include temperature gradient, thermal expansion displacement, and corresponding spurious acceleration values. Thermodynamic analysis and the finite element method are used to calculate the structural deformation caused by temperature changes, thereby determining the intensity of spurious vibration signals caused by sensor position changes.

[0046] The specific implementation of step S06 involves adaptive optimization of vibration signal processing parameters through a comprehensive evaluation algorithm. The vibration response evaluation function employs a weighted fusion method, normalizing multiple parameters such as the amplitude of micro-vibration, signal superposition coefficient, spurious vibration index, temperature change rate, and structural natural frequency, and then linearly combining them according to weighted coefficients. The weighted coefficients are determined based on the importance and sensitivity of the parameters: the amplitude of micro-vibration has a weight of 0.3, the signal superposition coefficient has a weight of 0.25, the spurious vibration index has a weight of 0.2, the temperature change rate has a weight of 0.15, and the structural natural frequency has a weight of 0.1. The evaluation function uses exponential transformation and the sigmoid function to map the output value between 0 and 1, ensuring the numerical stability and comparability of the evaluation index. The adaptive parameter adjustment strategy is designed based on fuzzy control theory. When the comprehensive vibration evaluation index a∈[0, 0.249], it indicates that the vibration signal is weak and requires fine analysis. A high-resolution mode with 2048 sampling points is adopted. When the comprehensive vibration evaluation index a∈(0.249, 0.683), it indicates that the vibration state is moderate. A medium-resolution mode with 1024 sampling points is adopted. When the comprehensive vibration evaluation index∈[0.683, 1], it indicates that there is obvious vibration anomaly. A low-resolution fast processing mode with 512 sampling points is adopted.

[0047] The specific implementation of step S07 involves generating a standardized vibration response test report to provide technical support for the safety assessment and maintenance decisions of wind tunnel equipment. The report generation system employs a template-based design, including a data aggregation module, a chart generation module, and a conclusion analysis module. The data aggregation module extracts vibration response data under various operating conditions from the database, including time-domain and frequency-domain analysis results, statistical parameters, and peak information. The chart generation module uses data visualization technology to generate vibration time-history curves, spectrum analysis diagrams, modal parameter comparison diagrams, and temperature-vibration correlation diagrams, visually displaying the vibration response characteristics and trends. The false signal identification results section details the processing effect of the temperature compensation algorithm, including the accuracy of false signal identification, signal comparison before and after filtering, and the optimization process of temperature compensation parameters. The structural safety assessment conclusion is based on a comparative analysis of vibration response data and design limits. The safety factor method and fatigue life assessment method are used to determine the structural safety status. When the vibration amplitude exceeds 80% of the design limit, a warning signal is issued; when it exceeds 90%, a shutdown inspection is recommended to ensure the safe and reliable operation of the wind tunnel equipment.

[0048] like Figure 2 As shown, the detailed structure of the vibration signal enhancement model is based on the Transformer architecture, specifically comprising four main components: an input embedding layer, a position encoding module, a multi-layer encoder, and an output decoding layer. The input embedding layer converts the one-dimensional vibration signal into a high-dimensional feature vector, using a one-dimensional convolutional layer to extract local features of the signal. The convolutional kernel size is 16, the stride is 8, and the number of output channels is 256. The position encoding module uses sine and cosine functions to generate position encoding vectors, assigning a unique code to each position in the sequence, enabling the model to understand the temporal relationship of the signal. The multi-layer encoder contains eight identical encoder layers, each containing two sub-layers: a multi-head self-attention mechanism and a feedforward neural network. The multi-head attention mechanism uses eight attention heads, each with a dimension of 64. The hidden layer dimension of the feedforward network is 1024, and the activation function is ReLU. The adaptive input size adjustment mechanism is implemented through dynamic pooling and upsampling techniques, automatically adjusting the length and resolution of the input signal according to the evaluation exponent to ensure the model's processing performance under different operating conditions. The output decoding layer uses linear transformation and the Softmax function to classify signals and outputs the probability distributions of real and spurious vibration signals.

[0049] The detailed steps for building the training dataset begin with data collection. Vibration signal data are acquired from various types of wind tunnel equipment, including open-type, closed-type, and supersonic wind tunnels, covering a variety of test conditions such as low-speed, subsonic, and supersonic. Detailed operating parameters are recorded during data collection, including environmental conditions such as wind speed, temperature, pressure, and humidity, as well as the wind tunnel's operating status and structural parameters. Data preprocessing steps include signal denoising, standardization, and segmentation. Butterworth filters are used to remove high-frequency noise, Z-score standardization eliminates magnitude differences between different sensors, and continuous signals are segmented into fixed-length sample segments. Data annotation is performed by domain experts, who manually annotate real and spurious vibration signals based on temperature records and vibration characteristics, aiming for an accuracy rate of over 98%. During the construction of the multidimensional dataset, the correlation between temperature, vibration, and structural parameters is established. Principal component analysis is used to reduce data dimensionality while retaining over 95% of the information. The dataset size is controlled to be above 12,000 sample points, with the training set accounting for 70%, the validation set for 15%, and the test set for 15%, ensuring the balanced and representative distribution of the data. Data augmentation techniques include time stretching, amplitude scaling, and noise addition to increase the diversity of the dataset and the generalization ability of the model.

[0050] The model training process employs supervised learning. The loss function combines two objectives: signal reconstruction error and classification accuracy. Reconstruction error is calculated using mean squared error, while classification accuracy is achieved using cross-entropy loss, with weights of 0.6 and 0.4 respectively. The optimizer uses the Adam algorithm with a learning rate of 0.001, employing a learning rate decay strategy, decreasing by 10% every 50 epochs. During training, a batch size of 32 is used, with a total of 200 training epochs. An early stopping mechanism is employed to prevent overfitting; training stops when the validation set loss shows no improvement for 10 consecutive epochs. Five-fold cross-validation is used to ensure model stability across different data subsets. Model evaluation metrics include signal reconstruction accuracy, classification accuracy, recall, and F1 score, requiring a combined performance metric of over 95% on the test set to ensure the model's practicality and reliability.

[0051] Example 2 Based on Embodiment 1, this embodiment is carried out as follows: A game-theoretic optimization model for the vibration response of wind tunnel equipment is established, comprising two levels of objective functions: an upper-level model and a lower-level model. The objective function of the upper-level model minimizes the overall vibration response of the wind tunnel structure, as specifically expressed below: In the formula, This represents the objective function value of the upper-level model. This represents the number of structural vibration modes. For the first Weighting coefficients for the first mode; For the first The vibration amplitude of the first mode; For structural displacement response; It is a time variable; For temperature gradient; The elastic modulus of the material; External incentives; For weight parameters, It is a natural number. The partial derivatives are consistent.

[0052] The parameter acquisition method is as follows: The modal parameters are obtained through modal analysis experiments, including step 1: conducting pulse excitation tests on the wind tunnel structure; step 2: extracting modal parameters of each order using the frequency domain decomposition method; and step 3: calculating weighting coefficients based on modal mass and stiffness. The accelerometer is used to measure and acquire data in real time. The time-domain signal is converted into a frequency-domain signal by a fast Fourier transform, and the amplitude of each mode is extracted. The temperature is measured using a distributed temperature sensor array, and the calculation formula is as follows: In the formula, They are respectively Temperature gradient components in three directions.

[0053] The objective function of the lower-level model is used to eliminate spurious vibration signals caused by temperature changes, and is specifically expressed as follows: In the formula, The objective function value of the lower-level model; This is a spurious acceleration signal vector; For spurious signal prediction function; The rate of temperature change; The coefficient of thermal expansion; This is a vector of structural geometric parameters; This is the signal frequency feature vector; For regularization parameters; For regularization terms; Here is the filter parameter matrix. Square brackets.

[0054] in, The calculation formula is: In the formula, for Temperature value at time, For time increments, This represents the temperature value corresponding to the time increment.

[0055] In actual calculations, the finite difference method is used: in For the first Temperature values ​​at each sampling time, The sampling interval is denoted as .

[0056] The original vibration signal is processed using a vibration signal enhancement model. The adjustment formula for the input dimensional parameters is as follows: In the formula, The length of the input signal; This is the base signal length, with a default value of 1024. This is a size adjustment function; The range of temperature change; The coefficient of thermal expansion of the structural material; This refers to the range of vibration frequencies.

[0057] The specific expression for the size adjustment function is: In the formula, The adjustment coefficients are set to values ​​of 0.5, 0.3, and 0.2, respectively. For reference values, 10℃ and 10℃ are respectively. / ℃, 100Hz.

[0058] Three key matrices are constructed for vibration signal analysis. The first is the characteristic matrix of small-amplitude vibration. The expression is: In the formula, For the first First modal frequency; For the first First-order modal damping ratio; For the first The first mode in the 1st order The mode shape coefficients of each node; The modal order; Let j be the number of measurement points, and j be a natural number.

[0059] Signal superposition weight matrix The expression is:

[0060] In the formula, For the first The incentive source is in the first... Weighting coefficients for each frequency band; For the number of incentive sources; This refers to the number of frequency bands. The formula for calculating the weighting coefficient is: In the formula, For the first The incentive source is in the first... Power spectral density of each frequency band.

[0061] False vibration index matrix The expression is: In the formula, For the first Temperature gradient at each measuring point; For the first Thermal expansion displacement at each measuring point; For the first False acceleration at each measuring point; This represents the total number of measuring points. The formula for calculating thermal expansion displacement is: In the formula, For the first The distance from each measuring point to the reference point; For the first Temperature change at each measuring point.

[0062] The comprehensive vibration assessment index is calculated using the vibration response assessment function. The expression for the assessment function is: In the formula, A comprehensive vibration assessment index; For the Sigmoid function; For the first The weighting coefficients of each parameter; For the first One normalized parameter.

[0063] The expression for the Sigmoid function is: The formula for calculating the normalization parameter is: In the formula, For the first One original parameter value; The first The minimum and maximum values ​​of each parameter.

[0064] The weighting coefficients are as follows: (Amplitude of micro-vibration) (Signal superposition coefficient) (False vibration index) (Rate of temperature change) (Structure natural frequency).

[0065] The piecewise function of the adaptive input size parameter adjustment strategy is: In the formula, The length of the input signal after adaptive adjustment.

[0066] The principles and effects of each formula are explained below. The upper-level model objective function, based on structural dynamics theory and multi-objective optimization principles, achieves a comprehensive assessment of the overall vibration state of the wind tunnel equipment by weightedly combining structural vibration response, temperature effects, and external excitations. Compared to traditional single-index evaluation methods, this function can simultaneously consider the coupling effects of multiple physical factors, significantly improving the comprehensiveness and effectiveness of the vibration control strategy and avoiding the overall performance degradation problem caused by local optimization. The lower-level model objective function employs the least squares optimization principle combined with regularization technology. By establishing a mathematical mapping relationship between spurious signals and temperature parameters, it achieves accurate identification and effective filtering of spurious vibration signals caused by thermal expansion. Compared to traditional fixed-threshold filtering methods, this function has adaptive learning capabilities, dynamically adjusting filtering parameters according to actual operating conditions, significantly improving the accuracy and robustness of signal processing.

[0067] The temperature gradient calculation formula, based on vector differential theory, accurately quantifies the non-uniformity of temperature distribution inside the wind tunnel by calculating the magnitude of the gradient vector of the temperature field in three-dimensional space. Compared to traditional single-point temperature measurement methods, this formula can comprehensively reflect the spatial variation characteristics of the temperature field, providing a more reliable physical basis for false signal identification and effectively avoiding signal misjudgment caused by inaccurate temperature measurements. The input size parameter adjustment formula, based on the signal fidelity principle in information theory, establishes a quantitative relationship between signal length and temperature changes, material properties, and frequency characteristics, enabling intelligent selection of signal processing resolution. Compared to traditional fixed-resolution processing methods, this formula can dynamically optimize the allocation of computational resources according to actual signal characteristics, significantly improving the system's real-time performance while ensuring processing accuracy.

[0068] The micro-amplitude vibration characteristic matrix, based on modal analysis theory, systematically organizes and stores the dynamic characteristic parameters of a structure in matrix form, achieving a complete description of the vibration characteristics of complex structures. Compared to traditional discrete parameter storage methods, this matrix structure facilitates matrix operations and data processing, significantly improving the computational efficiency and accuracy of vibration analysis results. The signal superposition weight matrix, based on signal decomposition theory and power spectrum analysis principles, quantifies the contribution of different excitation sources in each frequency band, achieving accurate decomposition of composite vibration signals. Compared to traditional subjective judgment methods, this matrix provides objective mathematical evidence, significantly improving the accuracy and reliability of signal source identification.

[0069] The spurious vibration index matrix, based on thermoelasticity theory, establishes a quantitative relationship between temperature gradient, thermal expansion displacement, and spurious acceleration, enabling accurate modeling of temperature-induced non-structural vibration signals. Compared to traditional empirical judgment methods, this matrix provides a rigorous mathematical expression, accurately predicting and compensating for temperature effects, effectively improving the identification accuracy of real vibration signals. The comprehensive vibration assessment index, based on fuzzy mathematics theory and a weighted fusion method, uses the sigmoid function to achieve nonlinear mapping of multiple parameters, quantifying complex vibration states into standardized exponents between 0 and 1. Compared to traditional linear assessment methods, this function better handles nonlinear relationships between parameters, providing more accurate and intuitive vibration state assessment results, and offering a reliable mathematical foundation for subsequent decision support.

[0070] Example 3 Based on Embodiments 1 and 2, a specific application scenario is provided to better understand and implement the present invention: This embodiment first installs 12 triaxial accelerometers (PCB356A15 model) at key structural nodes of the wind tunnel equipment, with a measurement range of ±50g and a sampling frequency set to 2000Hz. The sensor placement strategy is based on structural dynamics analysis results. Three sensors are installed on the main support column, located at the foundation connection, middle, and top, respectively, to monitor the overall vibration characteristics of the support structure. Four sensors are installed in the air supply system, distributed at the axial flow fan unit, the first-stage rectifier, the second-stage rectifier, and the inlet of the contraction section, focusing on monitoring the vibration response caused by aerodynamic excitation. Three sensors are installed in the test section, located at the inlet, the middle observation window, and the outlet diffuser section, respectively, to record the local vibration state during the test. Two sensors are installed in the auxiliary system, located in the cooling system and the electrical control room, to ensure the integrity of the overall monitoring.

[0071] This embodiment establishes a game-theoretic optimization model for the vibration response of wind tunnel equipment. The upper-level model aims to suppress small-amplitude vibrations. The input parameters include a structural damping coefficient of 0.025, a first-order natural frequency of 12.5 Hz, a second-order natural frequency of 38.7 Hz, an excitation force amplitude of 2800 N, a temperature gradient of 4.2 ℃ / m, and the material elastic modulus. Pa. The lower-level model aims to eliminate spurious signals, and its input parameters include the temperature change rate of 1.8℃ / min and the thermal expansion coefficient of the steel structure. The temperature range is ℃, the main structural geometric parameters are 5–85 Hz, and the main signal frequency range is 5–85 Hz. The two-layer model is linked by a thermoelastic coupling coefficient of 0.043 to achieve synergistic optimization of temperature compensation and vibration control.

[0072] Regarding the operational condition settings, this embodiment established a test matrix covering 7 wind speed levels and 5 temperature conditions. The wind speed levels were 10 m / s, 25 m / s, 40 m / s, 55 m / s, 70 m / s, 85 m / s, and 100 m / s, and the temperature conditions were 5℃, 15℃, 25℃, 35℃, and 45℃, forming 35 basic operating condition combinations. Each condition was run continuously for 45 minutes, with the first 15 minutes for system stabilization and the last 30 minutes for data acquisition. This embodiment installed 16 PT100 platinum resistance temperature sensors inside and outside the wind tunnel, with a measurement accuracy of 0.1℃ and a sampling frequency of 10 Hz, establishing a temperature vibration response correlation database containing 8750 valid data points.

[0073] This embodiment employs a vibration signal enhancement model based on a visual transformer architecture to process the raw vibration data. The model comprises eight coding layers, each containing 512 neurons, and a multi-head self-attention mechanism with 16 attention heads. The model training dataset contains 12,000 sample points from different types of wind tunnel equipment, covering 65% normal operation, 25% minor abnormal vibration, and 10% thermal expansion interference. The training process uses the Adam optimizer with a learning rate of 0.0001 and a batch size of 64. After 150 training epochs, the model converged, achieving a signal recognition accuracy of 96.8% on the independent test set.

[0074] In the vibration signal analysis process, this embodiment constructs three key analysis matrices. The micro-amplitude vibration characteristic matrix is ​​15×3 in dimension, containing the frequencies, damping ratios, and mode shape coefficients of the first 15 modes, as shown in Table 1.

[0075] Table 1. First 8 modal parameters of the wind tunnel equipment 1 12.47 0.023 Overall bending 2 18.92 0.019 Torsional vibration 3 26.35 0.027 Local bending 4 38.74 0.031 Vibration of supporting structure 5 45.18 0.025 Vibration of test section 6 52.63 0.029 Vibration of air supply system 7 67.91 0.034 Composite vibration 8 78.45 0.038 Higher-order local vibrations The signal superposition weight matrix is ​​4×8 in dimension, corresponding to the weight distribution of aerodynamic excitation, mechanical excitation, thermal excitation and environmental excitation in 8 main frequency bands, as shown in Table 2.

[0076] Table 2 Weighting coefficients of different excitation sources in each frequency band pneumatic excitation 0.45 0.52 0.38 0.29 0.21 0.16 0.12 0.08 Mechanical excitation 0.32 0.28 0.41 0.47 0.53 0.48 0.35 0.27 thermal excitation 0.18 0.15 0.16 0.19 0.21 0.28 0.35 0.42 Environmental incentives 0.05 0.05 0.05 0.05 0.05 0.08 0.18 0.23 The spurious vibration index matrix is ​​5×4 in dimension, corresponding to the spurious vibration intensity of the 5 main temperature monitoring points under different temperature gradient conditions, as shown in Table 3.

[0077] Table 3 Distribution of False Vibration Index at Each Monitoring Point Main support structure 0.015 0.028 0.041 0.063 Test segment entry 0.023 0.035 0.052 0.078 Middle of the test section 0.018 0.031 0.046 0.069 Air supply system 0.012 0.025 0.038 0.057 auxiliary structure 0.008 0.016 0.025 0.038

[0078] This embodiment calculates the comprehensive vibration assessment index under 35 operating conditions using a vibration response evaluation function. The assessment index ranges from 0.127 to 0.856. When the wind speed is 10 m / s and the temperature is 25°C, the comprehensive assessment index is 0.127. Using a high-resolution processing mode with 2048 sampling points, the detected micro-vibration amplitude is... When the wind speed is 55 m / s and the temperature is 35℃, the comprehensive evaluation index is 0.485. Using a medium-resolution processing mode with 1024 sampling points, the amplitude of the micro-vibration is... When the wind speed is 100 m / s and the temperature is 45℃, the comprehensive evaluation index is 0.782. Using a low-resolution processing mode with 512 sampling points, significant vibration anomalies were detected, with the amplitude of minor vibrations reaching [value missing]. .

[0079] During the specific data processing, this embodiment found that under the conditions of wind speed of 85 m / s and temperature of 40°C, the original vibration signal contained significant temperature interference components. Through calculation using the spurious vibration index matrix, the spurious vibration index under this condition was determined to be 0.074, and the spurious acceleration amplitude was [value missing]. The vibration signal enhancement model successfully identified and filtered out these spurious signals, and the processed true structural vibration amplitude was [value missing]. Compared to the original vibration signal amplitude The accuracy was reduced by 20.9%. The application of the temperature compensation algorithm improved the signal processing accuracy from 82.1% to 96.8%, an increase of 17.9%.

[0080] The final test results showed that the structural vibration response of the wind tunnel equipment was within the design limits under all operating conditions. The maximum vibration amplitude occurred at a wind speed of 100 m / s, with a value of [value missing]. Design limits only The accuracy was 49.5%, with a safety factor of 2.02. The deviation of the structure's first-order natural frequency (12.47Hz) from the design value of 12.50Hz was only 0.24%, indicating that the structural stiffness met the design requirements. The false signal identification accuracy reached 94.6%, and the temperature compensation effect was significant, effectively eliminating measurement errors caused by thermal expansion.

[0081] The vibration response test report generated in this embodiment records detailed vibration characteristic data under 35 operating conditions, including time-domain analysis results, frequency-domain characteristic parameters, modal parameter variation trends, and temperature-vibration correlation analysis. The report shows that the wind tunnel equipment has good dynamic stability, and all vibration indicators meet design requirements, allowing for safe operation. The structural safety assessment concludes that, within the current operating parameter range, the structural fatigue life is expected to exceed 25 years, meeting long-term operation requirements.

[0082] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for testing the vibration response of wind tunnel equipment installation, characterized in that, Includes the following steps: Multiple triaxial accelerometers are installed at the main structural nodes of the wind tunnel equipment. The installation locations of the triaxial accelerometers cover the support structure, air supply system and test section of the wind tunnel equipment. A game-theoretic optimization model for the vibration response of wind tunnel equipment is established, comprising an upper-level model aimed at suppressing small-amplitude vibrations and a lower-level model aimed at eliminating spurious signals. When the objective function is the upper-level model of the function that minimizes the superimposed vibration response of a wind tunnel structure, it is used to minimize the overall vibration response of the wind tunnel structure. The inputs include the structural damping coefficient, natural frequency, excitation force amplitude, temperature gradient, and material elastic modulus. The output is the optimal control strategy for the structural vibration response. When the objective function is a lower-level model of a thermal expansion spurious signal filtering function, used to eliminate spurious vibration signals caused by temperature changes, the inputs include the temperature change rate, thermal expansion coefficient, structural geometric parameters, and signal frequency characteristics. The outputs are spurious signal identification parameters and filtering parameters. The upper-level model and the lower-level model are linked through a structural temperature-vibration coupling term to achieve coordinated optimization of temperature compensation and vibration control; Start the wind tunnel equipment and set multiple operating conditions, including different wind speed conditions and different temperature conditions. At the same time, use temperature sensors to monitor temperature changes inside the wind tunnel and establish a temperature vibration response correlation database. The original vibration signals collected by the triaxial accelerometer are processed using a vibration signal enhancement model. The vibration signal enhancement model optimizes the signal recognition accuracy by adjusting the input size parameters. Three key matrices—a micro-amplitude vibration characteristic matrix, a signal superposition weight matrix, and a spurious vibration index matrix—are constructed for vibration signal analysis. The comprehensive vibration evaluation index is calculated by a vibration response evaluation function. The vibration response evaluation function takes the amplitude of the small vibration, the signal superposition coefficient, the spurious vibration index, the temperature change rate and the natural frequency of the structure as inputs, and outputs a comprehensive vibration evaluation index between 0 and 1. When the comprehensive vibration evaluation index belongs to different numerical ranges, different optimization strategies are used to adjust the input size parameters of the vibration signal enhancement model to achieve adaptive signal processing. Output wind tunnel equipment installation vibration response test report.

2. The method for testing the vibration response of wind tunnel equipment installation according to claim 1, characterized in that, The number of triaxial accelerometers shall not be less than 8, and the sampling frequency of the triaxial accelerometers shall be set to 2000Hz.

3. The method for testing the vibration response of wind tunnel equipment installation according to claim 1, characterized in that, The vibration signal enhancement model is a signal processing network based on a vision converter architecture, including: Multiple coding layers and an adaptive input size adjustment mechanism, The adaptive input size adjustment mechanism balances computational resources with the degree of signal detail preservation by dynamically adjusting the resolution of the original vibration signal.

4. The method for testing the vibration response of wind tunnel equipment installation according to claim 3, characterized in that, The vibration signal enhancement model also includes: The core location encoding module, multi-head self-attention mechanism, and feedforward neural network layer are trained using supervised learning methods. The loss function combines signal reconstruction error and classification accuracy, and the Adam optimizer is used for parameter updates.

5. The method for testing the vibration response of wind tunnel equipment installation according to claim 1, characterized in that, The micro-amplitude vibration characteristic matrix is ​​used to describe the vibration modal characteristics of the wind tunnel equipment structure under micro-excitation, including: Frequency, damping ratio, and mode shape coefficient of each mode. The micro-amplitude vibration characteristic matrix is ​​used to obtain the dynamic characteristic parameters of the structure through modal analysis.

6. The method for testing the vibration response of wind tunnel equipment installation according to claim 1, characterized in that, The signal superposition weight matrix is ​​used to quantify the contribution of multiple vibration signal sources to the overall response. The matrix elements of the signal superposition weight matrix represent the weighting coefficients of different excitation sources in each frequency band. The signal superposition weight matrix determines the superposition relationship of each signal component through frequency domain analysis and signal decomposition techniques.

7. The method for testing the vibration response of wind tunnel equipment installation according to claim 1, characterized in that, The spurious vibration index matrix is ​​used to identify and quantify non-real vibration signals caused by temperature changes. The matrix elements of the spurious vibration index matrix include the correspondence between temperature gradient, thermal expansion displacement, and spurious acceleration. The spurious vibration index matrix establishes a mathematical model of the spurious signal through temperature-vibration correlation analysis.

Images