A gimbal scanning vibration measurement and modal analysis method and system

Abstract

The application discloses a gimbal scanning vibration measurement and modal analysis method and system, and belongs to the technical field of non-contact vibration testing. The method sequentially executes transactions including movement, stabilization and collection stages for multiple measurement points. In the stabilization stage, whether the positioning mechanism meets the preset stabilization condition is judged based on the motion feedback of the positioning mechanism, and the judgment result is used as a gating condition for the synchronous data acquisition device to enter the collection stage. In the collection stage, the vibration response and the reference signal of the measurement point are synchronously acquired. Data version information including a version number is generated and is stored in association, wherein retesting of the same measurement point generates a new data version. Quality check is performed on the data version. The type or signal-to-noise ratio of the reference signal is used to restrict the type of subsequent modal analysis, and only the data version that passes the quality check is processed. Through stabilization gating, synchronous reference, versioned storage and quality check, the application realizes full-process closed-loop quality control, improves the reliability of data acquisition and the accuracy of modal analysis.

Description

Technical Field

[0001] This invention relates to the field of non-contact vibration testing technology, and in particular to a pan-tilt-zoom scanning vibration measurement and modal analysis method. Background Technology

[0002] Existing technologies for scanning vibration measurement of structures mainly include galvanometer laser vibration measurement, robotic arm vibration measurement, and simple gimbal splicing. Among these, the galvanometer approach uses optical galvanometers to rapidly change the beam direction, resulting in fast scanning speeds. However, it is typically limited by dedicated scanning heads, small field of view, fixed working distances, and closed hardware and software ecosystems. While the robotic arm approach offers greater reach and flexibility, it requires large equipment footprints, complex calibration chains, and stringent path planning and safety requirements, resulting in high procurement and maintenance costs. The simple gimbal splicing approach, although lower in cost, usually simply combines a single-point vibration measurement device with a general-purpose gimbal. The lack of deep coupling between subsystems generally results in a lack of precise transmission feedback, unified visual calibration, reliable positioning stability control, and synchronous references tied to the measurement nodes. This leads to problems such as incorrect data acquisition, data phase mismatch, and inconsistent remeasurement results during the measurement process, making it difficult to guarantee the reliability of the final analysis results.

[0003] Therefore, existing technologies generally suffer from problems such as isolated functions of each subsystem, lack of in-depth causal constraints and closed-loop control. This directly leads to systemic defects such as low measurement data quality, unreliable results, and lack of traceability throughout the process. There is an urgent need for a new technology solution that can fundamentally guarantee data quality and analytical reliability. Summary of the Invention

[0004] This invention aims to solve systemic problems arising from the lack of causal constraints between subsystems, such as incorrect data acquisition, phase mismatch, retest coverage, and erroneous data entering analysis. This invention addresses these problems through a closed-loop technology chain: visual calibration simultaneously provides the target's absolute angle and applicable domain; topology freezing simultaneously constrains the path, reference coverage, and HDF5 matrix dimensions; synchronous gimbal feedback and stability criteria directly gate DAQ deployment; only data that has undergone quality verification and transaction commit can enter the review process; retesting generates new revisions without overwriting the original LOS; modal algorithms can only read versions that meet reference conditions and matrix integrity requirements. The outputs of each stage become prerequisites for the next physical action or data processing, rather than being parallel functions.

[0005] This invention provides a pan-tilt-zoom (PTZ) scanning vibration measurement and modal analysis method. The system utilizes a vision device, a synchronous two-dimensional angle or position positioning mechanism, a non-contact vibration measurement device, a synchronous data acquisition device, and a control terminal. The method includes: sequentially executing transactions comprising movement, stabilization, and acquisition phases for multiple measurement points; in the stabilization phase, determining whether a preset stabilization condition is met based on the motion feedback of the positioning mechanism, and using this determination result as a gate condition for the synchronous data acquisition device to enter the acquisition phase; in the acquisition phase, synchronously acquiring the vibration response and reference signal of the measurement points; acquiring geometric information associated with the measurement points using the vision device; generating data version information containing a version number for each measurement, and storing the geometric information associated with the measurement points, measurement data, and the data version information in association, wherein re-measurement of the same measurement point generates a new data version that does not overwrite the original record; performing quality verification on the associated stored data versions; constraining the subsequent modal analysis type based on the type of the reference signal or whether its signal-to-noise ratio meets preset conditions, and processing only the data versions that pass the quality verification.

[0006] Optionally, the yaw transmission component of the two-dimensional angle or position positioning mechanism includes an annular synchronous belt and an arc-shaped synchronous gear ring; and its pitch transmission component includes a pitch synchronous belt and a synchronous pulley pair.

[0007] Optionally, the two-dimensional angle or position positioning mechanism may be a synchronous belt two-dimensional angle positioning mechanism, a harmonic reducer angle positioning mechanism, a lead screw guide rail type two-dimensional position positioning mechanism, etc.

[0008] Optionally, the position error of the two axes of the two-dimensional angle or position positioning mechanism is continuously less than a first threshold within a preset time; or the jitter index of the position feedback signal of the two axes of the two-dimensional angle or position positioning mechanism is continuously less than a second threshold within a preset time.

[0009] Optionally, the data version quality verification includes "hard failure" and "warning." Hard failures include incomplete records, channel / length mismatch, NaN / Inf, all zeros, no true trigger edge, severe clipping, and lack of specified common references, which can prevent release. Warnings include low backlight, pre-trigger offset, multiple threshold exceedances / suspected double-clicks, approaching the range, and insufficient attenuation / repeatability / coherence. These should generally generate warnings and require manual review, rather than being automatically judged as failures. The relationship between "quality verification passed," "approval released," and "can proceed to analysis" is defined accordingly.

[0010] Optionally, the constraint on the subsequent modal analysis type based on the type of the reference signal specifically means: when the reference signal is a fixed response reference, select to run deformation analysis and run modal analysis; when the reference signal is a force reference, select to run deformation analysis and experimental modal analysis; when the reference signal is missing, or its signal-to-noise ratio is lower than a preset threshold, only run deformation analysis is allowed, and only amplitude or relative display can be given.

[0011] Optionally, the associated storage includes: retaining the original line-of-sight response data, and generating a traceable normal compensation derived data version based on a preset calibration matrix that characterizes the transformation relationship between the measurement coordinate system and the global coordinate system and the surface normal of the measured object extracted from the geometric information; in the calibration matrix, the planar calibration matrix is ​​the homography matrix H, and the curved surface is the camera-gimbal extrinsic parameter R / T.

[0012] The present invention also provides a pan-tilt-zoom scanning vibration measurement and modal analysis system, including a vision device, a synchronous two-dimensional angle positioning mechanism, a non-contact vibration measurement device, a synchronous data acquisition device, and a control terminal, characterized in that the control terminal is configured to perform the method described in any of the preceding claims.

[0013] Optionally, the synchronous belt two-dimensional angle positioning mechanism includes: a yaw transmission assembly, the yaw transmission assembly including an arc-shaped synchronous gear ring connected to a slewing frame rotatable about a yaw axis, and an annular synchronous belt for driving the arc-shaped synchronous gear ring; and a pitch transmission assembly, the pitch transmission assembly including a pitch synchronous belt and a synchronous pulley pair for driving an instrument mounting plate for mounting the non-contact vibration measuring device.

[0014] Optionally, the software architecture of the control terminal includes a device adaptation layer, which is used to decouple the hardware driver from the upper-layer application logic so as to call different hardware through a unified interface.

[0015] Optionally, the control terminal may further include a simulation or playback branch that shares the same business logic as the real hardware, for testing and verifying the upper-layer application logic when there is no physical hardware connection.

[0016] The beneficial effects of this invention are as follows: 1. Improved reliability and accuracy of data acquisition: By setting a stabilization phase in the acquisition process and using the actual motion feedback of the positioning mechanism as the gating condition for acquisition, it is ensured that each data acquisition is carried out after the measurement beam has accurately and stably pointed to the target position. This eliminates measurement position errors caused by asynchronous motion and acquisition from a physical perspective, significantly improving the reliability of the data source.

[0017] 2. Ensures the effectiveness and accuracy of modal analysis: By synchronously acquiring reference signals and constraining subsequent analysis types based on the type or quality of the reference signals, the pain point of difficulty in ensuring phase consistency between measurement points in traditional sequential scanning is effectively solved. This makes it possible to perform accurate operational modal analysis or experimental modal analysis, and can intelligently avoid performing incorrect analysis when the data does not meet the conditions, thus ensuring the scientific nature of the final analysis results.

[0018] 3. Achieved end-to-end quality control and data traceability: By generating a new data version for each measurement that does not overwrite the original record, and by storing geometric information, measurement data, and quality verification results in a linked manner, a complete data traceability chain was constructed. Quality verification of the data versions was performed, and only data that passed verification was processed, achieving closed-loop quality control throughout the entire process from data acquisition to analysis, ensuring the reliability and trustworthiness of the final results. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the 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.

[0020] Figure 1 This is a schematic diagram of the two-dimensional pan-tilt synchronous belt drive mechanism and synchronous acquisition system according to an embodiment of the present invention.

[0021] Figure 2 This is a complete operation flowchart of the gimbal-type vibration measurement and modal analysis according to an embodiment of the present invention.

[0022] Figure 3 This is a system hardware and software modular topology block diagram according to an embodiment of the present invention.

[0023] Figure 4 This is a schematic diagram of the scanning measurement and real-time monitoring interface according to an embodiment of the present invention.

[0024] Figure 5 This is a schematic diagram of the data review and retest decision-making interface according to an embodiment of the present invention.

[0025] Figure 6 This is a schematic diagram of the interface for displaying modal analysis results and mode shape contour maps according to an embodiment of the present invention.

[0026] The reference numerals in the attached diagrams are explained as follows: 101 Gimbal base; 102 Yaw arc-shaped synchronous gear ring; 103 Annular synchronous belt; 104 Yaw drive motor; 105 Yaw small synchronous pulley; 106 Yaw swivel frame; 107 Pitch axis; 108 Pitch synchronous belt; 109 Pitch drive motor; 110 Pitch synchronous pulley pair; 111 Instrument mounting plate; 112 Laser vibration measurement front end; 113 RGB-D vision camera; 114 Encoder and limit assembly; 115 Cable guide; 116 Object under test; 117 Synchronous DAQ device; 118 Industrial control terminal; 119 Synchronous reference sensor; 120 Two-axis driver; S101 Create project; S104 Perform vision calibration; S105 Generate measurement point mesh; S107 Perform six-stage transaction; S108 Perform quality verification; S111 Enable analysis module; 301 Visual Acquisition Module; 302 Synchronous 2D Pan-Tilt Unit; 303 Laser Vibration Measurement Front-End; 304 Synchronous DAQ Module; 310 Equipment Adaptation Layer; 311 Mock / Replay Simulation Branch; 320 Calibration Layer; 321 Scanning Transaction Layer; 322 Data Review Layer; 323 Modal Analysis Layer; 330 HDF5 Versioned Engineering Data; 400 Scanning Measurement and Real-Time Monitoring Interface; 401 Status Monitoring Area; 402 Real-Time Data Display Area; 403 Scanning Progress Area; 404 Control Area; 411 Current Measurement Point Display; 412 Transaction Status Display; 421 Real-Time Vibration Response Curve; 431 Scanning Progress Display; 441 Pause / Continue Button; 442 Stop Button; 500 Data Review and Retest Decision Interface; 501 Measurement Point Data Version List; 502 Quality Verification Details Area; 503 Decision Operation Area; 511 Data Version Entries; 512 513 Data Version Entry; 521 Selected Item Display; 522 Verification Result Display; 531 Accept Version Button; 532 Retest This Point Button; 533 Delete Version Button; 600 Modal Analysis Results and Mode Shape Contour Display Interface; 601 Mode Shape Selection Area; 602 Modal Parameter List; 603 Mode Shape Contour Display Area; 611 Mode Shape Order Selector; 621 Modal Parameter Entry; 622 Modal Parameter Entry; 631 Mode Shape Contour. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be considered as limitations on this application. All other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of this application. Unless otherwise defined, all technical and scientific terms used in the embodiments of this application have the same meaning as commonly understood by those skilled in the art. The terminology used in the embodiments of this application is for the purpose of describing the embodiments of this application only and is not intended to limit this application. Before further detailed description of the embodiments of this application, the nouns and terms involved in the embodiments of this application are explained, and the nouns and terms involved in the embodiments of this application are subject to the following interpretations.

[0028] (1) Six-stage transaction: refers to the sequential execution of six consecutive and causally constrained stages for each measurement node: READY, MOVING, SETTLING, ARMED, ACQUIRING, and COMMITTED. The successful completion of the previous stage is a necessary condition for entering the next stage.

[0029] (2) Stability condition: refers to the preset standard for judging whether the movement of the two-dimensional angle positioning mechanism has stopped and reached the state of being collectable. Specifically, it may include: the position error of the two axes of the positioning mechanism is continuously less than the first threshold within a preset time; or the jitter index (such as variance or rate of change) of the two-axis position feedback signal is continuously less than the second threshold within a preset time.

[0030] (3) Data version information: refers to a set of unique identifiers and metadata generated for each measurement, used for data traceability. The core feature is that repeated measurements (re-measurements) of the same physical measurement point will generate a new data version (Revision) that does not overwrite the original record, ensuring the integrity of the measurement history.

[0031] (4) Normal compensation derived data version: refers to a new set of data generated after geometric projection correction of the original data based on the local normal direction of the surface of the measured object, while retaining the original line-of-sight (LOS) measurement data. This derived data coexists with the original data and can be traced.

[0032] (5) Device Adaptation Layer: This refers to an intermediate layer in the control terminal software architecture, which decouples the drivers (SDKs) of specific hardware devices from the upper-layer application business logic. It defines a unified interface (capability contract) so that the upper-layer logic can call hardware with the same function but different brands or models in the same way.

[0033] Please see Figures 1 to 3This application provides a pan-tilt-zoom scanning vibration measurement and modal analysis method and system, aiming to solve problems in existing technologies such as asynchronous motion and acquisition, uncontrollable data quality, unreliable analysis results, and untraceable processes. This invention ensures the accuracy and reliability of measurement data through a closed-loop quality control scheme covering the entire process from physical acquisition to data analysis.

[0034] The method provided in this application first executes a transaction, including movement, stabilization, and acquisition phases, sequentially on multiple measurement points. This transactional processing replaces simple cyclic scanning, and its design aims to establish strict quality control for the measurement process of each measurement point. By decomposing the operation of each measurement point into independent, sequentially dependent phases, it ensures that the next phase can only proceed after the previous phase is successfully completed. This guarantees the standardization of each operation and the validity of the data from a process perspective, solving the problem of inconsistent data quality caused by loose process control in traditional scanning methods.

[0035] During the stabilization phase, a core step of this method is to determine whether the positioning mechanism meets preset stabilization conditions based on motion feedback, and use this determination as a gating condition for the synchronous data acquisition device to enter the acquisition phase. Specifically, after the pan-tilt unit points to a new measurement point, the control system continuously monitors its position feedback signal until the signal error or jitter remains within a very small range for a certain period of time. This design aims to address the common problem of asynchronous motion and acquisition in the prior art, where the measuring device may begin acquiring data before the pan-tilt unit is completely stationary, causing the measurement position to deviate from the target point. By introducing the stable state of physical motion as a "hardware lock" for data acquisition, this invention ensures that each frame of data is acquired at a precise, stationary target position, fundamentally improving the spatial accuracy of the data.

[0036] During the acquisition phase, the system synchronously acquires the vibration response and reference signal of the measurement points. The purpose of synchronous acquisition is to establish a unified phase reference among all measurement point data. In modal analysis, the relative phase of vibration at each point is crucial information necessary for constructing mode shapes. Traditional sequential scanning makes it difficult to guarantee phase consistency due to the different acquisition times at each point. This invention introduces one or more reference signals (such as the acceleration response or excitation force signal at a fixed point) acquired synchronously with all measurement points, providing the necessary conditions for subsequent phase correction and global modal analysis. In this way, the technical challenge of accurate modal analysis with sequential scanning is solved, making it possible to perform operational modal analysis (OMA) or experimental modal analysis (EMA) using single-point scanning equipment.

[0037] To achieve automated scanning, this method utilizes a vision device to acquire geometric information associated with the measurement points. The vision device (such as a camera) can capture the shape, contour, and three-dimensional coordinates of the object being measured, thus providing a basis for the automatic planning and positioning of the measurement points. This step is designed to free the operator from tedious manual positioning and to handle objects with complex shapes. By establishing a precise calibration relationship between the vision coordinate system and the pan-tilt-zoom (PTZ) motion coordinate system, the system can automatically convert any point on the image into a driving command for the PTG, achieving highly efficient and high-precision automated scanning.

[0038] To achieve end-to-end data traceability, this method generates data version information containing a version number for each measurement and stores the associated geometric information, measurement data, and data version information related to the measurement point. Re-measurement of the same measurement point generates a new data version that does not overwrite the original record. This "versioned storage" mechanism is the core of the data management in this invention. Its design aims to solve the problems of data being easily overwritten and modification history being untraceable in traditional testing. When a measurement fails and needs to be retested, the system creates a new data version instead of overwriting the old record. In this way, all successful, failed, original, and retested data are completely preserved, forming a clear audit chain. This approach ensures data integrity and historical traceability, providing a solid foundation for subsequent quality audits and problem investigation.

[0039] After data storage, this method also includes quality verification of the associated stored data versions. This step is designed to perform a "health check" on the collected raw data, intercepting obviously substandard data before it enters the analysis process. By checking indicators such as sample size, timestamp continuity, signal overload, and backlight intensity, bad data caused by hardware failure, environmental interference, or operational errors can be effectively identified, thus ensuring that the dataset entering subsequent analysis stages is clean and valid.

[0040] Finally, this method constrains the type of subsequent modal analysis based on the type of the reference signal or whether its signal-to-noise ratio meets preset conditions, and only processes the data versions that have passed the quality verification. This design of "analysis constraint" and "quality-gated post-processing" aims to solidify the expert's analytical experience into the software process and prevent data misuse. For example, the system automatically determines whether a valid force reference signal exists; if not, experimental modal analysis is prohibited. Simultaneously, only data versions that have passed the aforementioned quality verification are allowed for use in the final modal parameter calculation. Through this dual constraint, the present invention ensures that the final output analysis results are based on correct data and correct analytical methods, greatly improving the scientific rigor and credibility of the final conclusions.

[0041] Furthermore, in a preferred embodiment, the mechanical structure of the synchronous belt two-dimensional angle positioning mechanism is specifically defined. Its yaw transmission component includes an annular synchronous belt and an arc-shaped synchronous gear ring; and its pitch transmission component includes a pitch synchronous belt and a synchronous pulley pair. This design utilizes the inherent advantages of low backlash and smooth transmission of synchronous belt drives, combined with the arc-shaped synchronous gear ring to achieve a wide range of angle coverage. Compared to traditional gear trains or worm gear structures, this solution, while ensuring positioning accuracy, has a simpler structure, lower cost, and a larger achievable range of motion, providing a high-performance and economical hardware foundation for scanning and measuring large structures.

[0042] In another preferred embodiment, the stability condition is specifically defined, including: the position error of the two axes of the synchronous belt two-dimensional angle positioning mechanism is continuously less than a first threshold within a preset time; or the jitter index of the position feedback signal of the two axes of the synchronous belt two-dimensional angle positioning mechanism is continuously less than a second threshold within a preset time. This specific definition transforms the abstract concept of "stability" into a quantifiable and programmable judgment criterion. For example, the position error can be set to be continuously less than 0.02 degrees within 100 milliseconds. This makes the control logic of "stability gating" more explicit and robust, ensuring the consistency of judgment criteria under different equipment and different operating conditions, and improving the reliability and repeatability of the method.

[0043] In one optional implementation, data version quality verification specifically includes checking at least one of the following: sample count, timestamp, signal overload, backlight intensity, and trigger quality. This provides a detailed checklist for data quality verification. For example, the system can check whether the acquisition duration meets requirements, whether there are timestamp jumps between data blocks, whether the signal amplitude exceeds the acquisition card's range, whether the laser backlight signal intensity is too low (potentially leading to a poor signal-to-noise ratio), and whether the hammer excitation experiences "double-clicks," etc. Through these meticulous checks, data quality can be more comprehensively assessed, improving the accuracy of bad data identification.

[0044] Furthermore, the constraint on the subsequent modal analysis type based on the type of reference signal specifically refers to the following: when the reference signal is a fixed response reference, run modal analysis is selected; when the reference signal is a force reference, experimental modal analysis is selected; when the reference signal is missing, or its signal-to-noise ratio is lower than a preset threshold, only run deformation analysis is allowed. This implementation provides clear execution rules for the "analysis constraint," constituting an intelligent decision-making system. It simulates the judgment logic of a test engineer, automatically selecting the most suitable analysis method based on the available data types, thereby avoiding typical errors such as forcibly performing experimental modal analysis without an excitation signal, and ensuring the correct application of the analysis method.

[0045] In another preferred embodiment, the associated storage further includes: retaining the original line-of-sight response data and generating a traceable normal compensation derived data version based on a preset calibration matrix characterizing the transformation relationship between the measurement coordinate system and the global coordinate system, and the surface normal of the measured object extracted from the geometric information. This design is primarily aimed at measurement scenarios of complex three-dimensional curved surfaces. Laser vibrometers measure vibration along their line-of-sight direction, but for curved surfaces, the normal vibration perpendicular to their surface is more meaningful. This solution, while retaining the original measurement facts (line-of-sight data), utilizes three-dimensional geometric information for coordinate transformation to calculate the normal vibration and stores it as a new "derived data version." In this way, it provides analysis results that are more in line with physical intuition, while ensuring that all processing is traceable and reproducible, enhancing the accuracy and rigor of the analysis of complex objects.

[0046] Please see Figure 1 and Figure 3 This application also provides a pan-tilt-zoom (PTZ) scanning vibration measurement and modal analysis system. This system serves as the physical carrier for implementing the aforementioned methods, and includes a vision device 113, a synchronous two-dimensional angle positioning mechanism, a non-contact vibration measurement device 112, a synchronous data acquisition device 117, and a control terminal 118. The control terminal 118 acts as the "brain" of the system, and its internal software is configured to execute the methods described in any of the preceding embodiments. By combining the functional limitations of the method with the physical structure of the system, protection of the product form is achieved.

[0047] Furthermore, such as Figure 1 As shown, the hardware structure of the synchronous belt two-dimensional angle positioning mechanism is preferably defined. It includes a yaw drive assembly and a pitch drive assembly. The yaw drive assembly includes an arc-shaped synchronous gear ring 102 connected to a rotatable frame 106 about a yaw axis, and an annular synchronous belt 103 for driving the arc-shaped synchronous gear ring 102. The pitch drive assembly includes a pitch synchronous belt 108 and a synchronous pulley pair 110 for driving an instrument mounting plate 111 for mounting the non-contact vibration measuring device 112. This structure corresponds to the preferred embodiment in the method claims and is a clear description of the preferred physical form of the system, protecting this specific mechanical design with wide-ranging and low-cost advantages.

[0048] In a preferred embodiment, such as Figure 3As shown, the software architecture of the control terminal 118 is defined, including a device adaptation layer 310. The device adaptation layer 310 decouples the hardware driver from the upper-layer application logic, allowing different hardware to be called through a unified interface. For example, whether it's a camera from brand A or a capture card from brand B, the device adaptation layer 310 will encapsulate it into a unified "visual acquisition module 301" or "synchronous DAQ module 304" interface for upper-layer calls. This design greatly improves the modularity and scalability of the system. When it's necessary to replace or upgrade some hardware, only a new adapter needs to be developed, without modifying the core business logic (such as the scan transaction layer 321, data audit layer 322, etc.), thus protecting the advanced ideas of this invention in software engineering implementation.

[0049] Furthermore, such as Figure 3 As shown, the control terminal 118 also includes a simulation or playback branch 311 that shares the same business logic as the real hardware. This branch can simulate the behavior of the hardware (e.g., return preset data or simulate a motion process) without being connected to the physical hardware. Upper-layer application logic, such as the calibration layer 320 and the scan transaction layer 321, can interact seamlessly with the real hardware or the simulation branch. This design allows software development, testing, and verification to be performed independently of expensive and complex hardware environments, significantly accelerating development iteration speed and facilitating testing for various anomalies, thus improving software robustness.

[0050] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0051] In one specific embodiment, this invention provides a complete implementation scheme for pan-tilt-zoom (PTZ) scanning vibration measurement and modal analysis. Its technical principle lies in establishing a precise mapping between the measurement point and the pan-tilt-zoom angle through visual calibration, then using a six-stage acquisition process based on "stability gating" to scan the object under test point by point, while simultaneously acquiring synchronous reference signals. All acquired data undergoes rigorous quality verification and is stored in a versioned manner. Finally, based on manual review and approval, an appropriate modal analysis algorithm is intelligently selected according to the validity of the reference signals, thereby achieving a closed-loop quality control system from physical acquisition to data analysis.

[0052] Please see Figure 1The system hardware platform utilized in this embodiment includes a two-dimensional gimbal mechanical structure, a synchronous data acquisition device 117, and an industrial control terminal 118 as the control core. This two-dimensional gimbal mechanical structure is designed for wide-range, high-precision angle positioning. It is fixedly supported by a stable gimbal base 101. Its yaw motion is driven by a yaw drive motor 104 through a small yaw synchronous pulley 105, which drives an annular synchronous belt 103. This annular synchronous belt 103 meshes with a large-diameter yaw arc-shaped synchronous gear ring 102, causing the yaw gyroscope 106 to rotate horizontally. Its pitch motion is driven by a pitch drive motor 109 mounted on the yaw gyroscope 106, which drives the instrument mounting plate 111 to rotate around the pitch axis (i.e., the pitch shaft 107) through a pitch synchronous belt 108 and a pitch synchronous pulley pair 110. This transmission method, which combines a synchronous belt with an arc-shaped gear ring, offers a wider transmission ratio range, smoother operation, and lower manufacturing costs compared to traditional gear drives. It enables large-scale scanning, such as horizontal scanning exceeding ±170 degrees. A non-contact laser vibration measurement front-end 112 and an RGB-D vision camera 113 for positioning are mounted on the instrument mounting plate 111. The laser vibration measurement front-end 112 moves along with the gimbal at the bottom, while the RGB-D vision camera 113 remains stationary at the top. To achieve closed-loop control, both axes of the gimbal are equipped with encoders and limit components 114 to provide real-time feedback of precise angular position to the industrial control terminal 118. In the data acquisition link, the vibration signal output from the laser vibration measurement front-end 112, along with reference signals from one or more synchronous reference sensors 119 (e.g., accelerometers fixed to the structure or force hammers used for excitation) used to establish a phase reference, are simultaneously connected to the synchronous DAQ device 117 for high-precision synchronous digitization. The industrial control terminal 118 sends control commands to the yaw drive motor 104 and the pitch drive motor 109 via the two-axis driver 120 to achieve precise angle positioning. To ensure the safety and smooth movement of the cable during the movement, the system is also equipped with a cable guide 115.

[0053] Reference Figure 2The workflow shown in this embodiment can be divided into three main stages: preparation, scanning, and data processing. In the preparation stage, the operator first creates a project on the industrial control terminal 118 (S101), and the system automatically establishes a corresponding directory structure to store various types of data. Subsequently, the system connects to and self-tests all hardware devices. The operator uses the RGB-D vision camera 113 to acquire images of the object under test 116, interactively defines the measurement contour of interest on the software interface, and then the system performs visual calibration (S104). By acquiring the correspondence between the coordinates of several sets of image points and the actual angles of the gimbal, the system calculates the transformation matrix between the two, thereby accurately converting each measurement point in the topology into the yaw and pitch target angles of the gimbal and setting the grid parameters (such as generating measurement points according to a 10x10 matrix). The system then generates a measurement point grid, forming a "frozen" topology (S105).

[0054] During the scanning phase, the system initiates automatic scanning and strictly executes a six-stage transaction (S107) for each measurement point in the topology. Taking the measurement of the Nth point as an example: First, it enters the "READY" state; then, the industrial control terminal 118 issues a command to drive the pan-tilt unit to the target angle of the Nth point, at which point the state switches to "MOVING"; when the encoder of the pan-tilt unit reports that it is close to the target position, the state enters "SETTLING". In this stage, the industrial control terminal 118 continuously reads the feedback value of the encoder 114 at a high frequency (e.g., 1kHz) and calculates the position error and signal jitter in real time. Only when the position error is continuously less than 0.02 degrees within 100 milliseconds, and the jitter index is also lower than the preset threshold, does the system determine that the pan-tilt unit has truly "stabilized" and enter the next stage. This process is called "stabilization gating". If stabilization fails, the measurement at that point is paused and an alarm is triggered. After successful stabilization, the state switches to "ARMED", and the industrial control terminal 118 sends an arming command to the synchronous DAQ device 117 to prepare for data acquisition. Immediately afterwards, the status changes to "ACQUIRING," and the DAQ device 117 synchronously acquires the response signal of the laser vibration front-end 112 and the reference signal of the synchronous reference sensor 119 under a unified clock. After acquisition, the data is first written to a temporary transaction, and then the system performs quality verification on the data block (S108), such as checking whether the sample count is complete, whether there is signal overload, and whether the laser backlight intensity meets the standard. After all checks pass, the data block is written to the HDF5 engineering data 330 and is given a "COMMITTED" flag, indicating that the data at this point has been validly submitted. At this point, the transaction for one measurement point is considered complete, and the system then begins the transaction for the next measurement point.

[0055] During the data processing phase, after all measurement points have been scanned, the operator can, as follows: Figure 5Data review is performed on the data review and retest decision interface 500. The measurement point data version list 501 in the interface clearly displays the quality status ("pass" or "fail") of each measurement point and each version (such as data version entries 511, 512, 513). For measurement points that fail quality verification (such as 512), for "hard failure" verification, the system will directly stop the measurement and require the operator to retest. For "warning" verification, the operator can select it and view the specific reason for failure in the quality verification details area 502 (such as verification result 522 showing "signal overload"). Then, the operator can decide in the decision operation area 503 whether to accept the version (Accept Version button 531), delete the version (Delete Version button 533), or add it to the retest queue (Retest This Point button 532). If a retest is selected, the system will re-execute the six-phase transaction for that point, generating a new data version (e.g., test point P13 - version 2, as shown in 513), while the original failed version (test point P13 - version 1, as shown in 512) is still retained, achieving versioned data traceability. In an optional implementation, the process of generating a new data version can be considered as a separate step S109. Finally, the operator initiates modal analysis. The system will automatically activate the corresponding analysis module (S111) based on whether a valid force reference or fixed response reference exists in the channel configuration. The analysis module will only read the latest valid data versions marked "COMMITTED" and manually reviewed, perform modal parameter identification, and as follows... Figure 6 As shown, the results are presented in the form of a modal parameter list 602 and a mode shape cloud map 631 on the modal analysis results and mode shape cloud map display interface 600. Through this embodiment, high-precision and high-reliability automated scanning vibration measurement of the object under test can be achieved. Due to the existence of the "stability gating" mechanism, the data quality of each measurement point is guaranteed; due to synchronous reference and version management, the final modal analysis results are accurate and traceable.

[0056] In another embodiment, for test objects with complex three-dimensional curvature (e.g., turbine blades), the present invention provides a more advanced calibration and compensation method. In this embodiment, the system structure is basically the same as in the previous embodiments, but the depth information provided by the RGB-D vision camera 113 is fully utilized. During the calibration phase, the system no longer uses a simple two-dimensional planar model, but performs three-dimensional calibration. The operator selects several calibration points on the surface of the test object, and the system acquires the three-dimensional coordinates P of these points in the camera coordinate system. c i The system simultaneously records the actual pitch angle θ and yaw angle ψ when the gimbal points to these points. Based on (θ, ψ), the system can construct a unit direction vector d. i The measurement ray. Ideally, point P in the camera coordinate system.c i After a rigid body transformation with a rotation matrix R and a translation vector T, the object should fall exactly on this ray. Therefore, the calibration problem is transformed into solving for the optimal R and T that satisfy the following constraints: P g i = RP c i + T = λ i d i R T R = I, det(R) = 1 d(θ,ψ) = [cosθ·sinψ, -sinθ, cosθ·cosψ] T r i = w i ¹ / ² · {[(RP c i + T) / ‖RP c i + T‖2] × d i} (R ,T ) = arg min Σ i ρ(‖r i ||2²) Among them, P g i Let P be a 3D point in the global coordinate system, R be the rotation matrix, and P be the rotation matrix. c i Let T be a 3D point in the camera coordinate system, and λ be the translation vector. i d is the ray distance. i To measure the ray direction, a robust estimation algorithm (such as the Soft-L1 loss function) is used in the solution process to improve the robustness of the calibration results and resist individual point selection errors or depth measurement noise. This algorithm minimizes the angular deviation between the predicted ray and the actual ray. After calibration, the system uses a set of independent, uncomputed points for verification to ensure, for example, that the maximum angular error is within 0.2 degrees, thus confirming the accuracy and applicability of the calibration results.

[0057] In the data processing stage, this embodiment also introduces normal compensation. When generating the measurement point topology, the system uses the depth map to calculate the local unit normal vector n of the surface of each measurement point. After scanning and acquisition, the laser vibration front-end 112 obtains the vibration component x along the line-of-sight direction d. LOs (t). To obtain a more physically meaningful normal vibration perpendicular to the surface, the system performs compensation calculations according to the following formula: x n (t) = x LOs (t) / (d T n) = x LOs (t) / cosα,|d T n| ≥ c min Where, x n (t) represents the derivation relation of the normal response, x LOs (t) represents the original LOS value, d is the unit line-of-sight vector, n is the unit normal of the measurement point, α is the angle between d and n, and c min This is the cosine threshold used to determine the effective projection. This compensation is only performed when the incident angle α is less than a preset threshold (e.g., 60 degrees) to avoid amplifying errors due to an excessively small denominator caused by grazing. Crucially, the system will completely preserve the original x... LOs (t) data, and calculate x n (t) is stored separately as a traceable "normal compensation derived data version". This allows users to view the corrected results and also to revert to the original measurement data at any time, ensuring the transparency and rigor of data processing. Compared to the basic embodiment, this embodiment can handle the measured object with complex curved surfaces more accurately, significantly improving the accuracy of 3D object modal analysis.

[0058] In yet another embodiment, the method of the present invention further includes intelligently selecting an analysis algorithm based on the configuration of the reference signal to ensure the scientific validity of the analysis. The implementation of this function relies on the definition of sensor roles during the channel configuration phase and the internal decision-making logic of the modal analysis module. This is illustrated below through three typical scenarios.

[0059] Scenario 1: Performing Operational Modal Analysis (OMA). Assume the object under test is a running compressor whose vibrations are generated by its own operation and cannot be stimulated by external forces. Before testing, in addition to setting up the scanning measurement points, the operator needs to install an accelerometer at a fixed location on the compressor (outside the scanning area) and define the sensor's channel role as "Fixed Response Reference" in the channel configuration interface of the industrial control terminal 118. After all measurement points have been scanned, the operator starts the analysis. The modal analysis layer 323 first checks if a valid "Fixed Response Reference" channel exists in the data version, covering the acquisition process of all scanning measurement points. Upon detecting the existence of this channel, the system determines that the conditions for OMA are met and automatically activates an OMA analysis algorithm, such as Frequency Domain Decomposition (EFDD). This algorithm constructs the power spectral density matrix G between all response channels (including the reference channel). γγ (ω k ), and perform singular value decomposition on it at each frequency point: G γγ (ω k ) = U(ω k ) Σ(ω k ) U H (ω k ) r ≈ u1(ω r ), σ1(ω r ) σ2(ω r ) Among them, G γγ (ω k ) is the output spectral density matrix, U(ω) k ) is a vector containing singular vector u i unitary matrix, Σ(ω k ) contains singular values ​​σ i diagonal matrix, superscript H This represents the conjugate transpose. The principal singularity σ1(ω) is identified near the resonance peak. r ) is much larger than the subsingular value σ2(ω r The frequency points of ω are identified, and the corresponding principal singular vectors u1(ω) are extracted. r As a mode shape r This allows us to identify the modal parameters of the system. Their relationship can be expressed as follows: r ≈ u1(ω r ).

[0060] Scenario 2: Performing Experimental Modal Analysis (EMA). Assume the test object is a stationary car brake disc, and the test uses a force hammer to excite it. The operator connects a force hammer with a force sensor to the synchronous DAQ device 117 and defines its channel role as "force reference". During the ACQUIRING phase of the scan, the system waits for the pulse signal generated by the force hammer to trigger data acquisition. During the analysis phase, the modal analysis layer 323 detects a valid "force reference" channel in the data version, thus determining that the conditions for EMA are met, and automatically activates the EMA analysis algorithm, such as the multi-reference least squares complex frequency domain method (P-LSCF). The system calculates the frequency response function (FRF) H1(ω) between each response point and the force reference point, using the following formula: H1(ω) = G γ f(ω) · Gff†(ω) Where H1(ω) is the FRF estimator, G γf(ω) is the cross-spectral matrix, Gff(ω) is the autospectral matrix, and † is the Moore-Penrose generalized inverse. Then, the P-LSCF algorithm fits the entire FRF matrix with a polynomial matrix in a common denominator, thereby identifying the common poles of the system (i.e., modal frequencies and damping).

[0061] Scenario 3: No valid reference signal. Suppose that during testing, the operator neither sets up a fixed reference point nor uses force excitation, or the signal-to-noise ratio of the reference signal is extremely low. During the analysis phase, modal analysis layer 323 checks and finds that there is no valid synchronous reference that can be used to establish global phase. In this case, the system will determine that OMA or EMA analysis cannot be performed because the phase relationship between different measurement points cannot be guaranteed to be correct. In this situation, the system will automatically disable the OMA and EMA modules, allowing only running deformation analysis (ODS). ODS analysis only displays the vibration amplitude of each measurement point at a specific frequency, presented in animation form, but the system will clearly inform the user that this is only working deformation, not a modal shape with a globally consistent phase. Through this intelligent constraint, erroneous analysis of unqualified data is effectively avoided, ensuring the scientific validity of the final conclusion.

[0062] In the system embodiments of the present invention, the software architecture of the control terminal 118 embodies a highly modular and scalable design philosophy. For example... Figure 3 As shown, the entire software system is clearly divided into different layers. The bottom layer is the device adaptation layer 310, which is responsible for communicating with specific hardware, such as the vision acquisition module 301, the synchronous 2D pan-tilt unit 302, the laser vibration measurement front-end 303, and the synchronous DAQ module 304. The core function of the device adaptation layer 310 is to encapsulate hardware drivers (SDKs) from different manufacturers and models, providing a unified and standardized interface to the upper layers. For example, whether it's a camera from brand A or brand B, the calibration layer 320 calls the same "image acquisition" function. This design isolates hardware heterogeneity at the bottom layer, allowing the core business logic of the upper layers, including the calibration layer 320, the scanning transaction layer 321, the data review layer 322, and the modal analysis layer 323, to be developed and evolved completely independently of specific hardware. When support for a new piece of hardware is needed, only a new adapter plugin needs to be developed for it, without modifying the upper-layer code, greatly improving the maintainability and scalability of the system.

[0063] Furthermore, to facilitate development and testing, the software architecture includes a Mock / Replay emulation branch 311 parallel to the real hardware driver. This emulation branch also follows the standard interface defined by the device adaptation layer 310, but it does not communicate with the real hardware internally. Instead, it responds to upper-layer calls by returning preset data or simulating hardware behavior. For example, when the upper layer calls the "Mobile Gimbal" interface, the emulation branch can wait for a period of time and then directly return a "Motion Complete" success status. Because the emulation branch shares the same business logic path as the real driver, developers can perform complete testing and verification of core functions such as scanning transactions, data auditing, and modal analysis without a physical hardware connection. This not only significantly reduces reliance on expensive testing equipment and accelerates the development iteration cycle, but also makes the simulation and testing of various complex abnormal scenarios (such as hardware failures and communication interruptions) simple and feasible, thereby effectively improving the stability and reliability of the entire system.

Claims

1. A pan-tilt-zoom (PTZ) scanning vibration measurement and modal analysis method, the system of which includes a vision device, a two-dimensional angle or position positioning mechanism, a non-contact vibration measurement device, a synchronous data acquisition device, and a control terminal, characterized in that, include: The transaction, which includes the movement, stabilization, and acquisition phases, is executed sequentially for multiple measurement points. During the stabilization phase, the motion feedback of the two-dimensional angle or position positioning mechanism is used to determine whether it meets the preset stabilization conditions, and the determination result is used as the gate condition for the synchronous data acquisition device to enter the acquisition phase. During the acquisition phase, the vibration response and reference signal of the measuring point are acquired simultaneously. Geometric information associated with the measurement point is acquired using a vision device; For each measurement, a data version information containing a version number is generated, and the geometric information associated with the measurement point, the measurement data, and the data version information are stored together. For repeated measurements of the same measurement point, a new data version is generated that does not overwrite the original record. Perform quality verification on the data version of the associated storage; Based on the type of the reference signal or whether its signal-to-noise ratio meets preset conditions, the subsequent modal analysis type is constrained, and only the data version that has passed the quality verification is processed.

2. The method according to claim 1, characterized in that, The yaw transmission component of the synchronous belt two-dimensional angle positioning mechanism includes an annular synchronous belt and an arc-shaped synchronous gear ring; and its pitch transmission component includes a pitch synchronous belt and a synchronous pulley pair.

3. The method according to claim 1 or claim 2, characterized in that, The stability conditions include: The position error of the two axes of the two-dimensional angle or position positioning mechanism is continuously less than a first threshold within a preset time; or the jitter index of the position feedback signal of the two axes of the two-dimensional angle or position positioning mechanism is continuously less than a second threshold within a preset time.

4. The method according to claim 1, characterized in that, The data version quality verification includes "hard failure" and "warning". Hard failure includes incomplete records, channel / length mismatch, NaN / Inf, all zeros, no true trigger edge, severe clipping, and lack of specified common references, which can prevent release. Warnings include low backlight, pre-trigger offset, multiple threshold overruns / suspected double-clicks, approaching the range, and insufficient attenuation / repeatability / coherence. These should generally be issued as warnings and subject to manual review, rather than being automatically judged as failures. The relationship between "quality verification passed", "approval released", and "can enter analysis" is defined accordingly.

5. The method according to claim 1, characterized in that, The specific modal analysis type constrained by the type of the reference signal is as follows: When the reference signal is a fixed response reference, select to run deformation analysis and run modal analysis; When the reference signal is a force reference, select to run deformation analysis and experimental modal analysis; When the reference signal is missing, or its signal-to-noise ratio is lower than a preset threshold, only deformation analysis is allowed, and only amplitude or relative values ​​can be displayed.

6. The method according to claim 1, characterized in that, The associated storage includes: The original line-of-sight response data is retained, and a traceable normal compensation derived data version is generated based on the preset calibration matrix that characterizes the transformation relationship between the measurement coordinate system and the global coordinate system and the surface normal of the measured object extracted from the geometric information; in the calibration matrix, the planar calibration matrix is ​​the homography matrix H, and the surface is the camera-gimbal extrinsic parameter R / T.

7. A pan-tilt-zoom (PTZ) scanning vibration measurement and modal analysis system, comprising a vision device, a synchronous two-dimensional angle positioning mechanism, a non-contact vibration measurement device, a synchronous data acquisition device, and a control terminal, characterized in that, The control terminal is configured to perform the method as described in any one of claims 1 to 6.

8. The system according to claim 7, characterized in that, The synchronous belt two-dimensional angle positioning mechanism includes: a yaw transmission assembly and a pitch transmission assembly; The yaw transmission assembly includes an arc-shaped synchronous gear ring connected to a slewing frame rotatable about a yaw axis, and an annular synchronous belt for driving the arc-shaped synchronous gear ring. The pitch transmission assembly includes a pitch timing belt and timing pulley for driving an instrument mounting plate for mounting the non-contact vibration measuring device.

9. The system according to claim 7 or claim 8, characterized in that, The software architecture of the control terminal includes a device adaptation layer, which is used to decouple the hardware driver from the upper-layer application logic so as to call different hardware through a unified interface.

10. The system according to claim 9, characterized in that, The control terminal also includes a simulation or playback branch that shares the same business logic as the real hardware, used to test and verify the upper-layer application logic when there is no physical hardware connection.

Images