Structural member modal test method and device, electronic equipment and storage medium

Abstract

The invention discloses a structural member modal test method and device, electronic equipment and a storage medium, and relates to the technical field of structural member testing, and the method comprises the steps: carrying out the boundary pre-verification of a structural member; obtaining a three-dimensional point cloud of the surface of the structural member, and determining an area containing nodes as an avoidance area according to the three-dimensional point cloud; distribution is carried out on the test area to obtain test points; wherein the test area is an area which does not contain nodes on the surface of the structural member; constructing a model of the structural member according to a boundary pre-verification result; selecting excitation points on the test area based on reinforcement learning; arranging a sensor on an excitation point of the model by using a mechanical arm; utilizing a sensor to generate excitation on the corresponding excitation point so as to obtain measurement data of each measurement point of the model; and determining a modal test result of the structural member according to the measurement data. According to the modal test method, the step of generating the fault can be determined according to the modal test result, and then the corresponding step is returned to test again, so that closed-loop verification is realized, and the test efficiency and precision can be improved.

Description

Technical Field

[0001] This application relates to the field of structural component testing technology, and in particular to a structural component modal testing method, apparatus, electronic device, and storage medium. Background Technology

[0002] The following technical problems exist in the current modal testing of structural components:

[0003] 1. Linearized procedures: Traditional experimental processes (excitation-acquisition-analysis) lack closed-loop verification, leading to easy accumulation of errors. 2. High reliance on manual intervention: Key steps such as measurement point setup and excitation point selection rely on operator experience, resulting in poor consistency (large repeatability errors in actual measurements). 3. Delayed quality verification: Data quality issues (such as insufficient coherence) are only discovered in the final step, leading to a high retest rate throughout the entire process. Summary of the Invention

[0004] The main objective of this application is to provide a method, apparatus, electronic device, and storage medium for modal testing of structural components, so as to improve the efficiency and accuracy of modal testing of structural components.

[0005] To achieve the above objectives, one aspect of this application proposes a modal testing method for structural components, the method comprising the following steps:

[0006] Perform boundary pre-verification on structural components;

[0007] Obtain a three-dimensional point cloud of the surface of the structural component, and determine the area containing nodes as the avoidance area based on the three-dimensional point cloud;

[0008] Each test point is assigned to a test area; wherein, the test area is the area on the surface of the structural component that does not contain the node;

[0009] The model of the structural component is constructed based on the results of the boundary pre-verification;

[0010] The incentive points are selected in the test area based on reinforcement learning;

[0011] Sensors are placed at the excitation points of the model using a robotic arm;

[0012] The sensor is used to generate excitation at the corresponding excitation point to obtain measurement data of each of the measurement points of the model;

[0013] The modal test results of the structural component are determined based on the measurement data.

[0014] In some embodiments, the boundary pre-verification of the structural components includes the following steps:

[0015] The difference between the experimental frequency response function and the finite element model is compared by using the frequency response function inversion verification method to quantify whether the boundary conditions of the structure meet expectations.

[0016] The expression for the frequency response function inversion verification method is as follows:

[0017]

[0018] Among them, R valid H represents the difference. exp H represents the experimental frequency response function. FEM This represents the finite element model;

[0019] If the difference reaches a preset threshold, the state parameters of the structural component are adjusted to make the difference lower than the preset threshold.

[0020] The expression for adjusting the state parameters of the structural component is:

[0021]

[0022] Where, ξ new ξ represents the corrected damping ratio. init This represents the initial damping ratio.

[0023] In some embodiments, the boundary pre-verification of the structural components includes the following steps:

[0024] Dynamically acquire the rigid body modal frequencies of the structural component;

[0025] When the proportion of the rigid body modal frequency to the overall test frequency band exceeds a preset proportion threshold, the constraint force of the structural component is determined to be insufficient.

[0026] The preload of the electromagnetic locking structure is adjusted using a PID control algorithm to improve the constraint force of the structure.

[0027] In some embodiments, the allocation of various test points on the test area includes the following steps:

[0028] Each measurement point is assigned to the test area based on the mode curvature.

[0029] In some embodiments, the selection of activation points on the test region based on reinforcement learning includes the following steps:

[0030] The excitation point optimization index is calculated using the self-optimization algorithm of Q-learning;

[0031] The excitation point is selected in the test area according to the excitation point preference index;

[0032] The expression for the incentive point optimization index is:

[0033] I e =α1·Stiffness+β·FRF_Coherence+γ·Mode_Observability;

[0034] Among them, I e The excitation point preference index is represented by α1, β, and γ, where α1, β, and γ are the coefficients of the corresponding terms.

[0035] The steps for constructing the self-optimization algorithm of Q-learning include the following:

[0036] Define the state space: use the candidate excitation point positions, current modal parameters, and environmental noise levels as state variables to form a multidimensional state vector;

[0037] Define the action space: Discretize the positions of each candidate stimulus point, and expand the continuous space using the policy gradient method; wherein each candidate stimulus point position corresponds to an action option;

[0038] The reward function is defined as follows:

[0039]

[0040] Where R represents the reward function;

[0041] The initial Q values ​​in the Q table were pre-filled using Monte Carlo sampling based on historical experimental data.

[0042] The strategy expressions for Monte Carlo sampling include:

[0043] ∈=0.9·e -0.01-episode ;

[0044] Where ∈ represents the exploration probability; episode represents the training round;

[0045] The learning rate is dynamically adjusted as follows:

[0046] α2=0.1·(1+cos(episode / 100));

[0047] Wherein, α2 represents the learning rate;

[0048] When the excitation point optimization index is greater than the preset strategy threshold, the Q value is frozen, and a Gaussian perturbation is performed centered on the optimal point to verify the local optimum.

[0049] In some embodiments, the method further includes the following steps:

[0050] The bandwidth for obtaining the test data is adjusted according to the piecewise function;

[0051] The piecewise function includes:

[0052]

[0053] Among them, f new This represents the dynamically adjusted sampling bandwidth, and Coherence represents the coherence function.

[0054] The attenuation slope, noise spectrum ratio, and main frequency energy concentration of the time-domain signal corresponding to the test data are calculated as signal characteristics, and then a window function for observing the time-domain signal is selected based on the signal characteristics.

[0055] In some embodiments, determining the modal test results of the structural component based on the measurement data includes the following steps:

[0056] A comprehensive score is calculated based on the attenuation slope, coherence function, window function matching degree, and signal-to-noise ratio in the measured data.

[0057] The expression for the comprehensive score is:

[0058]

[0059] Wherein, Score represents the overall score; w i Indicates the indicator weight, x i X represents the measured index value. thresh The threshold values ​​represent the metrics; these metrics include attenuation slope, coherence function, window function matching degree, and signal-to-noise ratio.

[0060] Whether the modal test of the structural component meets the conditions is determined based on whether the comprehensive score is lower than the preset score threshold;

[0061] If the conditions are not met, the test will be reverted to the corresponding fault step and repeated.

[0062] To achieve the above objectives, another aspect of this application provides a structural component modal testing apparatus, the apparatus comprising:

[0063] Boundary pre-verification unit, used to perform boundary pre-verification of structural components;

[0064] A node determination unit is used to acquire a three-dimensional point cloud of the surface of the structural component and determine the area containing nodes as an avoidance area based on the three-dimensional point cloud.

[0065] A test point allocation unit is used to allocate various test points on a test area; wherein, the test area is the area on the surface of the structural component that does not contain the node;

[0066] A model building unit is used to build a model of the structural component based on the results of the boundary pre-verification.

[0067] An incentive point selection unit is used to select incentive points on the test area based on reinforcement learning.

[0068] A sensor placement unit is used to place sensors at the excitation points of the model using a robotic arm;

[0069] An excitation unit is used to generate excitation at the corresponding excitation point using the sensor, thereby acquiring measurement data of each of the measurement points of the model;

[0070] The test analysis unit is used to determine the modal test results of the structural component based on the measurement data.

[0071] To achieve the above objectives, another aspect of this application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described method.

[0072] To achieve the above objectives, another aspect of the embodiments of this application proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0073] The embodiments of this application include at least the following beneficial effects:

[0074] This application enables boundary pre-verification of structural components; acquisition of a 3D point cloud of the structural component surface; determination of areas containing nodes as avoidance zones based on the 3D point cloud; allocation of various measurement points on a test area, wherein the test area is a region on the structural component surface that does not contain nodes; construction of a model of the structural component based on the boundary pre-verification results; selection of excitation points on the test area based on reinforcement learning; placement of sensors on the excitation points of the model using a robotic arm; generation of excitations at the corresponding excitation points using the sensors to acquire measurement data of each measurement point of the model; and determination of the modal test results of the structural component based on the measurement data. This application can determine the steps that cause failures based on the modal test results, and then backtrack to the corresponding steps for retesting, achieving closed-loop verification, thus improving test efficiency and accuracy. Attached Figure Description

[0075] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0076] Figure 1 A schematic flowchart illustrating a modal testing method for structural components provided in this application embodiment;

[0077] Figure 2 An example flowchart of a modal testing method for structural components provided in this application embodiment;

[0078] Figure 3 Example flowchart of the boundary pre-verification loop provided in the embodiments of this application;

[0079] Figure 4 Example flowchart of the measurement point self-optimization loop provided in the embodiments of this application;

[0080] Figure 5 Example flowchart of the excitation dynamic adjustment loop provided in the embodiments of this application;

[0081] Figure 6 Example flowchart of the full-process backtracking loop provided in the embodiments of this application;

[0082] Figure 7 Example flowchart of a partial backtracking loop provided in an embodiment of this application;

[0083] Figure 8 This is a schematic diagram of a structural modal testing device for structural components provided in an embodiment of this application;

[0084] Figure 9 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0085] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit it. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this application; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this application as detailed in the appended claims.

[0086] It is understood that the terms “first,” “second,” etc., used in this application may be used herein to describe various concepts, but unless otherwise stated, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the words “if,” “when,” or “in response to a determination” as used herein may be interpreted as “when…” or “when…” or “in response to a determination.”

[0087] As used in this application, the terms "at least one", "multiple", "each", "any", etc., "at least one" includes one, two or more, "multiple" includes two or more, "each" refers to each of the corresponding multiples, and "any" refers to any one of the multiples.

[0088] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0089] This application provides a structural component modal testing method, apparatus, electronic device, and storage medium, relating to the field of structural component testing technology. The structural component modal testing method, apparatus, electronic device, and storage medium provided in this application can be applied to a terminal, a server, or software running on a terminal or server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, or vehicle terminal, but is not limited thereto; the server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network; the software can be an application implementing a structural component modal testing method, but is not limited to the above forms.

[0090] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0091] Reference Figure 1 This application provides a method for modal testing of structural components, which may include, but is not limited to, steps S100 to S170, as detailed below:

[0092] S100: Perform boundary pre-verification on structural components;

[0093] S110: Obtain a three-dimensional point cloud of the surface of the structural component, and determine the area containing nodes as the avoidance area based on the three-dimensional point cloud;

[0094] S120: Various test points are assigned on the test area; wherein, the test area is the area on the surface of the structural component that does not contain the node;

[0095] S130: Construct a model of the structural component based on the results of the boundary pre-verification;

[0096] S140: Selecting activation points in the test area based on reinforcement learning;

[0097] S150: Use a robotic arm to place sensors at the excitation points of the model;

[0098] S160: The sensor is used to generate an excitation at the corresponding excitation point to obtain measurement data of each of the measurement points of the model;

[0099] S170: Determine the modal test results of the structural component based on the measurement data.

[0100] Optionally, the boundary pre-verification of the structural components includes the following steps:

[0101] The difference between the experimental frequency response function and the finite element model is compared by using the frequency response function inversion verification method to quantify whether the boundary conditions of the structure meet expectations.

[0102] The expression for the frequency response function inversion verification method is as follows:

[0103]

[0104] Among them, R valid H represents the difference. exp H represents the experimental frequency response function. FEM This represents the finite element model;

[0105] If the difference reaches a preset threshold, the state parameters of the structural component are adjusted to make the difference lower than the preset threshold.

[0106] The expression for adjusting the state parameters of the structural component is:

[0107]

[0108] Where, ξ new ξ represents the corrected damping ratio. init This represents the initial damping ratio.

[0109] Optionally, the boundary pre-verification of the structural components includes the following steps:

[0110] Dynamically acquire the rigid body modal frequencies of the structural component;

[0111] When the proportion of the rigid body modal frequency to the overall test frequency band exceeds a preset proportion threshold, the constraint force of the structural component is determined to be insufficient.

[0112] The preload of the electromagnetic locking structure is adjusted using a PID control algorithm to improve the constraint force of the structure.

[0113] Optionally, the process of assigning various test points on the test area includes the following steps:

[0114] Each measurement point is assigned to the test area based on the mode curvature.

[0115] Optionally, the reinforcement learning-based selection of activation points in the test area includes the following steps:

[0116] The excitation point optimization index is calculated using the self-optimization algorithm of Q-learning;

[0117] The excitation point is selected in the test area according to the excitation point preference index;

[0118] The expression for the incentive point optimization index is:

[0119] I e=α1·Stiffness+β·FRF_Coherence+γ·Mode_Observability;

[0120] Among them, I e The excitation point preference index is represented by α1, β, and γ, where α1, β, and γ are the coefficients of the corresponding terms.

[0121] The steps for constructing the self-optimization algorithm of Q-learning include the following:

[0122] Define the state space: use the candidate excitation point positions, current modal parameters, and environmental noise levels as state variables to form a multidimensional state vector;

[0123] Define the action space: Discretize the positions of each candidate stimulus point, and expand the continuous space using the policy gradient method; wherein each candidate stimulus point position corresponds to an action option;

[0124] The reward function is defined as follows:

[0125]

[0126] Where R represents the reward function;

[0127] The initial Q values ​​in the Q table were pre-filled using Monte Carlo sampling based on historical experimental data.

[0128] The strategy expressions for Monte Carlo sampling include:

[0129] ∈=0.9·e -0.01-episode ;

[0130] Where ∈ represents the exploration probability; episode represents the training round;

[0131] The learning rate is dynamically adjusted as follows:

[0132] α2=0.1·(1+cos(episode / 100));

[0133] Wherein, α2 represents the learning rate;

[0134] When the excitation point optimization index is greater than the preset strategy threshold, the Q value is frozen, and a Gaussian perturbation is performed centered on the optimal point to verify the local optimum.

[0135] Optionally, the method further includes the following steps:

[0136] The bandwidth for obtaining the test data is adjusted according to the piecewise function;

[0137] The piecewise function includes:

[0138]

[0139] Among them, f new This represents the dynamically adjusted sampling bandwidth, and Coherence represents the coherence function.

[0140] The attenuation slope, noise spectrum ratio, and main frequency energy concentration of the time-domain signal corresponding to the test data are calculated as signal characteristics, and then a window function for observing the time-domain signal is selected based on the signal characteristics.

[0141] Optionally, determining the modal test results of the structural component based on the measurement data includes the following steps:

[0142] A comprehensive score is calculated based on the attenuation slope, coherence function, window function matching degree, and signal-to-noise ratio in the measured data.

[0143] The expression for the comprehensive score is:

[0144]

[0145] Wherein, Score represents the overall score; w i Indicates the indicator weight, x i X represents the measured index value. thresh The threshold values ​​represent the metrics; these metrics include attenuation slope, coherence function, window function matching degree, and signal-to-noise ratio.

[0146] Whether the modal test of the structural component meets the conditions is determined based on whether the comprehensive score is lower than the preset score threshold;

[0147] If the conditions are not met, the test will be reverted to the corresponding fault step and repeated.

[0148] The following section will provide a detailed introduction and explanation of the solutions in the embodiments of this application, using specific application examples.

[0149] First, refer to Figure 2 First, a comparison and explanation of this embodiment and related technologies will be provided, as shown in Table 1 below:

[0150]

[0151]

[0152] Table 1

[0153] Specifically, this embodiment includes the following steps:

[0154] 1. Reference Figure 3 Pre-verification stage (steps 1-3 innovation).

[0155] (1) Frequency response function inversion verification method:

[0156] By comparing the experimental frequency response function (H) exp ) and finite element model (H FEM The difference in ) quantifies the accuracy of boundary conditions.

[0157] Formula Explanation:

[0158]

[0159] If R valid ≥5% indicates that the actual boundary conditions do not match the theoretical model, and the tooling state needs to be readjusted, triggering the tooling reconfiguration command.

[0160] Formula for automatic update of model parameters:

[0161]

[0162] ξ new : Corrected damping ratio;

[0163] ξ init Initial damping ratio (from finite element model);

[0164] R valid : Boundary verification error percentage (calculated by frequency response function inversion).

[0165] Dynamic calibration:

[0166] When the experimental data is found to deviate from the model by ΔR ≥ 5% during the pre-verification phase, the damping ratio is automatically adjusted to match the actual working conditions.

[0167] Example:

[0168] Initial damping ratio ξ init =0.02;

[0169] The measured error ΔR = 8%;

[0170] Corrected ξ new =0.02×(1+8 / 10)=0.036.

[0171] Results: The corrected model better reflects the actual structural damping characteristics, reducing modal frequency errors. After several iterations, the damping ratio converged to a stable value.

[0172] Gradual correction:

[0173] The 1 / 10 coefficient balances the adjustment range and prevents the model from becoming unstable due to excessive single errors.

[0174] (2) Rigid body modal frequency threshold triggering mechanism:

[0175] When the rigid body modal frequency (such as the first bending / torsional mode) accounts for more than 15% of the overall test frequency band, the tooling constraint is deemed insufficient.

[0176] Project Implementation:

[0177] The tooling incorporates a force sensor to monitor constraint forces in real time.

[0178] The preload of the electromagnetic locking fixture is adjusted by combining a PID control algorithm.

[0179] 2. Reference Figure 4 Intelligent measurement point layout (innovation in steps 2-4-6).

[0180] (1) Generation of 3D point cloud heatmap:

[0181] Point cloud data of the structural component surface was obtained by laser scanning (accuracy ±0.1mm).

[0182] The avoidance area (highlighted in red) is generated based on the distribution of modal vibration nodes.

[0183] The density of measuring points is allocated according to the energy distribution weight (the density is increased by 50% in areas with large mode curvature).

[0184] The heatmap generation algorithm process is as follows:

[0185]

[0186] (2) Automatic pasting by a snake-shaped robotic arm:

[0187] Flatness control: The end effector integrates pressure feedback (0-10N adjustable).

[0188] Path planning: Optimize the robotic arm's motion trajectory based on genetic algorithm (reducing idle travel by 60%).

[0189] 3. Reference Figure 5 Adaptive incentive stage (steps 7-8 innovation).

[0190] (1) Incentive Point Optimization Index I e Design principles:

[0191] I e =α1·Stiffness+β·FRF_Coherence+γ·Mode_Observability;

[0192] α1=0.5, β=0.3, γ=0.2;

[0193] Stiffness: Local stiffness (unit N / mm) is quantified through static deformation testing, and points with higher stiffness are selected to improve the efficiency of excitation energy transfer.

[0194] FRF_Coherence: Coherence coefficient of the frequency response function (average value of 0-200Hz), reflecting the linear correlation between the excitation signal and the response.

[0195] Mode_Observability: Modal observability factor (calculated based on mode slope). The observability factor is calculated based on the mode slope, which quantifies the sensitivity of the excitation point to the target mode.

[0196] (2) Development process of self-optimization algorithm based on Q-learning.

[0197] 1. Algorithm framework design.

[0198] State space definition:

[0199] The candidate excitation point location, current modal parameters (such as damping ratio and frequency error), and environmental noise level are used as state variables to form a multidimensional state vector.

[0200] Action space definition:

[0201] Discretize the candidate activation points (e.g., by meshing), with each location corresponding to an action option. For continuous space problems, this can be extended by combining policy gradient methods.

[0202] Reward function design:

[0203]

[0204] When I e When the value is greater than 75, an exponential reward is given to accelerate convergence to the optimal solution.

[0205] 2. Optimization of core parameters for Q-learning.

[0206] Q table initialization:

[0207] Monte Carlo sampling was used to pre-fill the initial Q value, and historical experimental data were combined to accelerate convergence.

[0208] Explore - Utilize Balance:

[0209] Improved 8-greedy strategy:

[0210] ∈=0.9·e -0.01-episode ;

[0211] Early stages focus on exploration (high ε), later stages lean towards exploitation (low ε).

[0212] Dynamic learning rate adjustment:

[0213] α2=0.1·(1+cos(episode / 100));

[0214] The learning rate is periodically adjusted to escape local optima.

[0215] 3. Locking mechanism implementation.

[0216] When I e Triggering the double-insurance strategy at 75:

[0217] Q-value freezing: Stop updating the Q-value of the current optimal stimulus point to prevent overfitting.

[0218] Neighborhood search: Using the optimal point as the center, perform a Gaussian perturbation (σ = 5mm) to verify the local optimum.

[0219] 4. Summary of the development of a reinforcement learning-based incentive point self-optimization algorithm.

[0220] Compared with traditional methods, the reinforcement learning-based incentive point self-optimization algorithm can improve the completeness of modality order recognition, reduce the time spent on incentive point optimization, and lower the retest rate.

[0221] Key innovations:

[0222] Multi-objective fusion reward mechanism:

[0223] Stiffness, coherence, and modal observability are dynamically weighted to avoid a single indicator dominating the decision.

[0224] Adaptive termination condition:

[0225] Introducing a sliding window evaluation (window size = 10 iterations), when I... e Training will be terminated early if the fluctuation is less than 2% for 5 consecutive times.

[0226] Offline-online blended learning:

[0227] The Q-table is optimized using historical data during the pre-training phase and fine-tuned in real time during the online phase, balancing efficiency and adaptability.

[0228] (3) Dynamic bandwidth adjustment:

[0229] Piecewise function:

[0230]

[0231] Design basis:

[0232] Coherence threshold:

[0233] 85%: The signal linearity is basically reliable, allowing for moderate bandwidth expansion.

[0234] 90%: High-confidence data that can capture higher-order modalities.

[0235] Exponential expansion strategy:

[0236] The bandwidth is expanded by 1.5 times each time (1024→1536→2048Hz) to avoid a sudden drop in frequency domain resolution.

[0237] Initial bandwidth 1024Hz → if coherence > 90% → extend to 2048Hz.

[0238] Mathematical basis:

[0239]

[0240] Among them, f new f represents the dynamically adjusted sampling bandwidth. base Indicates the sampling bandwidth before adjustment; Coherence represents the coherence function.

[0241] By increasing the sampling rate, higher-order modes can be captured.

[0242] (4) Feature-driven window function matching.

[0243] The dynamic matching rules are shown in Table 2 below:

[0244]

[0245] Table 2

[0246] The window function dynamic matching module implements window function matching through the following steps:

[0247] (1) Signal feature extraction: Calculate the attenuation slope k of the time-domain signal. d (Unit: dB / oct) and noise spectrum percentage, main frequency energy concentration.

[0248] (2) Example code for window function matching is as follows:

[0249]

[0250]

[0251] (3) Parameter tuning unit: By evaluating the signal-to-noise ratio (SNR) and frequency leakage of the windowed signal through trial acquisition, the window parameters are iteratively optimized until the coherence function γ is obtained. 2 >0.85.

[0252] 4. Multi-criteria verification (innovation in step 9).

[0253] (1) The design of the four-dimensional mass matrix is ​​shown in Table 3 below:

[0254] index threshold Weight Detection methods attenuation slope >4dB / oct 30％ Perform FFT spectrum analysis on time-domain signals coherence function >0.85 25％ Calculate the input-output cross power spectral density Window function matching degree RMS<0.05 20％ Compare the signal energy difference before and after windowing Signal-to-noise ratio >45dB 25％ Logarithmic calculation of signal power / noise power

[0255] Table 3

[0256] The expression for the overall score:

[0257]

[0258] (2) After multiple criteria are verified to obtain a comprehensive score, the system automatically starts the backtracking process mechanism.

[0259] Analysis of the two-layer architecture of the backtracking mechanism:

[0260] Reference Figure 6 First layer: If Score < 85, global backtracking is triggered (step 9 → 1 / 2 / 3 / 7).

[0261] Triggering condition: When the overall score of the four-dimensional quality matrix is ​​less than 85 points, the system determines that a global process reset is required.

[0262] Backtracking target:

[0263] Step 1: Boundary conditions may be mismatched (e.g., tooling is loose).

[0264] Step 2: There is a systematic deviation in the layout of the measuring points (such as the node area not being avoided).

[0265] Step 3: Incorrect model parameter settings (e.g., incorrect material property input).

[0266] Step 7: The incentive strategy is fundamentally flawed (e.g., the incentive point is located at a modal node).

[0267] Reference Figure 7 Second layer: Local backtracking trigger (step 9 → 4 / 7 / 8).

[0268] Triggering condition: When a single indicator in the four-dimensional matrix fails to meet the standard but the total score is ≥85, a precise local correction is executed.

[0269] Backtracking target:

[0270] Step 4: Sensor installation issues (such as poor contact leading to low signal-to-noise ratio).

[0271] Step 7: Insufficient excitation energy (e.g., if the attenuation slope is not up to standard, the impact force needs to be increased).

[0272] Step 8: Incorrect parameter settings (e.g., incorrect window function type needs to be selected again).

[0273] The essential explanation of logic

[0274] Differences in expression levels:

[0275] (1) Steps 9 → 1 / 2 / 3 / 7 belong to the handling path of a major global error.

[0276] (2) Steps 4 / 7 / 8 belong to the fast correction path of local parameter optimization.

[0277] The severity levels of the faults are shown in Table 4 below:

[0278]

[0279] Table 4

[0280] Engineering implementation of backtracking logic:

[0281] Example of a decision tree (pseudocode description):

[0282]

[0283] (3) The correspondence between the integrity design of the quality matrix and the backtracking logic.

[0284] The explanation for why the coherence index is not explicitly reflected in the local backtracking path in the four-dimensional quality matrix:

[0285] 1. The root cause of the coherence function anomaly.

[0286] coherence function γ 2 (f) Reflects the linear correlation between input and output signals; anomalies in this correlation are typically caused by the following factors:

[0287] Insufficient excitation energy (adjustment required in step 7).

[0288] Excessive noise interference (requires optimization in step 4).

[0289] Insufficient frequency resolution (step 8 is required to increase bandwidth).

[0290] 2. Implicit handling of existing backtracking mechanisms.

[0291] Although the local backtracking path is not directly labeled "step → coherent function", indirect repair is achieved through the following coupling correction mechanism, as shown in Table 5 below:

[0292]

[0293] Table 5

[0294] Therefore, when the coherence function index in the four-dimensional matrix does not meet the standard but the total score is ≥85, local closed-loop optimization in steps 4 / 7 / 8 can be performed.

[0295] Engineering implementation of backtracking logic when the coherence function index in a four-dimensional matrix is ​​below the standard but the total score is ≥85:

[0296]

[0297] (4) Design rationality verification

[0298] Efficiency optimization: Avoid minor issues triggering global backtracking.

[0299] Fault isolation: Locating specific problematic links through the quality matrix.

[0300] Resource conservation: Localized corrections require only a smaller proportion of the original test resources.

[0301] For example, the technical parameters are designed based on the following table 6:

[0302]

[0303] Table 6

[0304] This embodiment has the following beneficial effects:

[0305] Model Adaptability: By dynamically adjusting the damping ratio, the "model-physical" mismatch problem in traditional methods is eliminated.

[0306] Intelligent bandwidth expansion: Based on coherence threshold hierarchical expansion, it takes into account both high-frequency mode capture and data reliability.

[0307] Precise window function matching: Feature-driven window selection strategy improves signal processing performance.

[0308] Reference Figure 8 This application also provides a structural component modal testing apparatus, which can implement the above-described structural component modal testing method. The apparatus includes:

[0309] Boundary pre-verification unit, used to perform boundary pre-verification of structural components;

[0310] A node determination unit is used to acquire a three-dimensional point cloud of the surface of the structural component and determine the area containing nodes as an avoidance area based on the three-dimensional point cloud.

[0311] A test point allocation unit is used to allocate various test points on a test area; wherein, the test area is the area on the surface of the structural component that does not contain the node;

[0312] A model building unit is used to build a model of the structural component based on the results of the boundary pre-verification.

[0313] An incentive point selection unit is used to select incentive points on the test area based on reinforcement learning.

[0314] A sensor placement unit is used to place sensors at the excitation points of the model using a robotic arm;

[0315] An excitation unit is used to generate excitation at the corresponding excitation point using the sensor, thereby acquiring measurement data of each of the measurement points of the model;

[0316] The test analysis unit is used to determine the modal test results of the structural component based on the measurement data.

[0317] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0318] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the method of this application. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0319] It is understood that the content of the above method embodiments is applicable to the device embodiments. The specific functions implemented by the device embodiments are the same as those of the methods of this application, and the beneficial effects achieved are the same as those achieved by the methods of this application.

[0320] Please see Figure 9 , Figure 9 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes:

[0321] The processor 901 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application.

[0322] The memory 902 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 902 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 902 and is called and executed by the processor 901.

[0323] The input / output interface 903 is used to implement information input and output;

[0324] The communication interface 904 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0325] Bus 905 transmits information between various components of the device (e.g., processor 901, memory 902, input / output interface 903, and communication interface 904);

[0326] The processor 901, memory 902, input / output interface 903, and communication interface 904 are connected to each other within the device via bus 905.

[0327] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method of this application.

[0328] It is understood that the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0329] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0330] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0331] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0332] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0333] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0334] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0335] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0336] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0337] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0338] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0339] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0340] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A modal testing method for structural components, characterized in that, The experimental method includes the following steps: Perform boundary pre-verification on structural components; Obtain a three-dimensional point cloud of the surface of the structural component, and determine the area containing nodes as the avoidance area based on the three-dimensional point cloud; Each test point is assigned to a test area; wherein, the test area is the area on the surface of the structural component that does not contain the node; The model of the structural component is constructed based on the results of the boundary pre-verification; The incentive points are selected in the test area based on reinforcement learning; Sensors are placed at the excitation points of the model using a robotic arm; The sensor is used to generate excitation at the corresponding excitation point to obtain measurement data of each of the measurement points of the model; The modal test results of the structural component are determined based on the measurement data.

2. The modal testing method for structural components according to claim 1, characterized in that, The boundary pre-verification of structural components includes the following steps: The difference between the experimental frequency response function and the finite element model is compared by using the frequency response function inversion verification method to quantify whether the boundary conditions of the structure meet expectations. The expression for the frequency response function inversion verification method is as follows: Among them, R valid H represents the difference. exp H represents the experimental frequency response function. FEM This represents the finite element model; If the difference reaches a preset threshold, the state parameters of the structural component are adjusted to make the difference lower than the preset threshold. The expression for adjusting the state parameters of the structural component is: Where, ξ new ξ represents the corrected damping ratio. init This represents the initial damping ratio.

3. The modal testing method for structural components according to claim 1, characterized in that, The boundary pre-verification of structural components includes the following steps: Dynamically acquire the rigid body modal frequencies of the structural component; When the proportion of the rigid body modal frequency to the overall test frequency band exceeds a preset proportion threshold, the constraint force of the structural component is determined to be insufficient. The preload of the electromagnetic locking structure is adjusted using a PID control algorithm to improve the constraint force of the structure.

4. The modal testing method for structural components according to claim 1, characterized in that, The process of assigning various test points on the test area includes the following steps: Each measurement point is assigned to the test area based on the mode curvature.

5. The modal testing method for structural components according to claim 1, characterized in that, The step of selecting activation points in the test region based on reinforcement learning includes the following steps: The excitation point optimization index is calculated using the self-optimization algorithm of Q-learning; The excitation point is selected in the test area according to the excitation point preference index; The expression for the incentive point optimization index is: I e =α1·Stiffness+β·FRF_Coherence+γ·Mode_Observability; Among them, I e The excitation point preference index is represented by α1, β, and γ, where α1, β, and γ are the coefficients of the corresponding terms. The steps for constructing the self-optimization algorithm of Q-learning include the following: Define the state space: use the candidate excitation point positions, current modal parameters, and environmental noise levels as state variables to form a multidimensional state vector; Define the action space: Discretize the positions of each candidate stimulus point, and expand the continuous space using the policy gradient method; wherein each candidate stimulus point position corresponds to an action option; The reward function is defined as follows: Where R represents the reward function; The initial Q values ​​in the Q table were pre-filled using Monte Carlo sampling based on historical experimental data. The strategy expressions for Monte Carlo sampling include: ∈＝0.9·e -0.01-episode ； Where ∈ represents the exploration probability; episode represents the training round; The learning rate is dynamically adjusted as follows: α2=0.1·(1+cos(episode / 100)); Wherein, α2 represents the learning rate; When the excitation point optimization index is greater than the preset strategy threshold, the Q value is frozen, and a Gaussian perturbation is performed centered on the optimal point to verify the local optimum.

6. The modal testing method for structural components according to claim 1, characterized in that, The method further includes the following steps: The bandwidth for obtaining the test data is adjusted according to the piecewise function; The piecewise function includes: Among them, f new This represents the dynamically adjusted sampling bandwidth, and Coherence represents the coherence function. The attenuation slope, noise spectrum ratio, and main frequency energy concentration of the time-domain signal corresponding to the test data are calculated as signal characteristics, and then a window function for observing the time-domain signal is selected based on the signal characteristics.

7. A method for modal testing of structural components according to any one of claims 1 to 6, characterized in that, Determining the modal test results of the structural component based on the measurement data includes the following steps: A comprehensive score is calculated based on the attenuation slope, coherence function, window function matching degree, and signal-to-noise ratio in the measured data. The expression for the comprehensive score is: Wherein, Score represents the overall score; w i Indicates the indicator weight, x i X represents the measured index value. thresh The threshold values ​​represent the metrics; these metrics include attenuation slope, coherence function, window function matching degree, and signal-to-noise ratio. Whether the modal test of the structural component meets the conditions is determined based on whether the comprehensive score is lower than the preset score threshold; If the conditions are not met, the test will be reverted to the corresponding fault step and repeated.

8. A modal testing device for structural components, characterized in that, The device includes: Boundary pre-verification unit, used to perform boundary pre-verification of structural components; A node determination unit is used to acquire a three-dimensional point cloud of the surface of the structural component and determine the area containing nodes as an avoidance area based on the three-dimensional point cloud. A test point allocation unit is used to allocate various test points on a test area; wherein, the test area is the area on the surface of the structural component that does not contain the node; A model building unit is used to build a model of the structural component based on the results of the boundary pre-verification. An incentive point selection unit is used to select incentive points on the test area based on reinforcement learning. A sensor placement unit is used to place sensors at the excitation points of the model using a robotic arm; An excitation unit is used to generate excitation at the corresponding excitation point using the sensor, thereby acquiring measurement data of each of the measurement points of the model; The test analysis unit is used to determine the modal test results of the structural component based on the measurement data.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

Images