Mechanical transmission surface stress laser ultrasonic detection device and automatic analysis system

Abstract

The present application relates to the technical field of nondestructive testing, and discloses a mechanical transmission part surface stress laser ultrasonic detection device and an automatic analysis system, which comprises a pulsed laser excitation module, a laser interference receiving module, a multi-axis scanning mechanism and a processing unit; the processing unit extracts the local relative curvature radius of a target point through a low-frequency channel, a phase pre-distortion is generated by feedforward control of a spatial light modulator, and a pure stress phase difference is calculated by using a joint decoupling equation to eliminate the influence of geometric divergence caused by topography; meanwhile, the scanning mechanism is controlled to perform microarray scanning, speckle suppression weight coefficients are generated based on low-frequency light intensity amplitude, and high-frequency ultrasonic time domain signals are spatially weighted and synthesized; finally, the decoupled phase difference is substituted into an acoustic-elastic equation to solve the surface residual stress value. The present application eliminates wavefront divergence caused by macroscopic curved surface topography and optical detection noise caused by microscopic roughness, and improves the accuracy of surface stress measurement.

Description

Technical Field

[0001] This invention relates to the field of nondestructive testing technology, specifically to a laser ultrasonic testing device and automatic analysis system for surface stress of mechanical transmission components. Background Technology

[0002] Residual stress is generated on the surface of mechanical transmission components during the manufacturing process. Measuring this stress state is an important step in evaluating the mechanical properties of the components. Laser ultrasonic technology, as a non-contact testing method, mainly calculates surface stress by measuring the sound velocity or phase change of ultrasonic surface waves propagating in the material, combined with the acoustoelastic effect.

[0003] Existing laser ultrasonic testing devices are typically designed for flat surfaces. When measuring mechanical transmission components with complex shapes, the macroscopic curvature of the workpiece surface causes geometric divergence of the excited ultrasonic wave front during propagation. This divergence phase generated along the curved surface is superimposed on the phase change caused by the actual residual stress. Existing signal processing systems lack structural curvature decoupling algorithms, making it difficult to separate the phase shift introduced by the geometric contour from the pure stress phase, resulting in deviations in the calculated stress values ​​for non-planar regions.

[0004] Furthermore, the surfaces of actual mechanical transmission components typically exhibit microscopic roughness due to machining. When a probe laser irradiates a rough surface, the reflected light forms a spatially randomly distributed speckle field. Existing equipment often employs a single-point fixed detection and fixed-gain signal extraction mode. During scanning, if the probe spot falls into the weak light reflection region of the speckle field, the light intensity amplitude at the interference receiver will randomly attenuate, leading to a decrease in the signal-to-noise ratio of the extracted high-frequency ultrasonic time-domain signal. Optical interference noise caused by fluctuations in the microscopic surface state can distort single-point measurement data and produce data jumps or gaps on the continuous scanning trajectory, making it difficult for existing systems to output spatially continuous and complete surface stress distribution maps. Summary of the Invention

[0005] The technical problem solved by this invention is that, when measuring the surface stress of mechanical transmission components, the curvature change of the workpiece's macroscopic morphology in existing laser ultrasonic testing technology causes distortion of the ultrasonic surface wave propagation path and generates geometric divergence phase, introducing phase error; at the same time, the micro-roughness of the surface under test causes random phase modulation of the speckle field, resulting in a decrease in the signal-to-noise ratio of the optical detection position, which in turn affects the accuracy of the stress calculation data.

[0006] To address the above problems, the present invention provides the following technical solution:

[0007] This invention provides a laser ultrasonic testing device for surface stress of mechanical transmission components, comprising:

[0008] A pulsed laser excitation module, comprising a pulsed laser, a pre-beam expander, a spatial light modulator, and a lens group;

[0009] A laser interferometer receiving module, including an interferometer and a data acquisition circuit with high and low frequency channels;

[0010] A multi-axis automatic scanning mechanism, equipped with the lens group and optical probe;

[0011] The automatic analysis system, electrically connected to the acquisition circuit, the spatial light modulator, and the multi-axis automatic scanning mechanism, includes a processing unit and a memory storing the following program:

[0012] The curvature extraction program extracts low-frequency interference voltage signals from the low-frequency channel to determine the local relative radius of curvature.

[0013] The speckle suppression weight generation program generates speckle suppression weight coefficients from local light intensity amplitudes.

[0014] A phase diagram matrix generation procedure, wherein the local relative radius of curvature generates the phase diagram matrix that drives the spatial light modulator;

[0015] A high-frequency time-domain signal weighted synthesis program combines the high-frequency time-domain signal of the high-frequency channel with the speckle suppression weighting coefficient to obtain a synthesized ultrasonic time-domain signal;

[0016] The frequency-wavenumber domain analysis program converts the synthesized ultrasonic time-domain signal into the frequency-wavenumber domain and takes the measured total phase difference.

[0017] The program for joint decoupling and solving the acoustoelastic equation uses the measured total phase difference, initial phase boundary condition values, and geometric divergence phase values ​​to calculate the pure stress phase difference and solve for the surface stress value.

[0018] This invention also provides an automated analysis method for laser ultrasonic testing of surface stress in mechanical transmission components, comprising the following steps:

[0019] The multi-axis automatic scanning mechanism is controlled to scan along the surface of the mechanical transmission component. The low-frequency interference voltage signal is extracted through the low-frequency channel and the local relative radius of curvature at each target point on the continuous scanning trajectory is calculated.

[0020] The multi-axis automatic scanning mechanism is controlled to perform discrete point microarray scanning at the target point, and the local light intensity amplitude of each dwelling point of the microarray is extracted to generate the corresponding speckle suppression weight coefficient.

[0021] The local relative radius of curvature is used as a feedforward control parameter to generate a phase diagram matrix, and the spatial light modulator is driven to load the phase diagram matrix to excite broadband ultrasonic waves on the surface of the mechanical transmission component.

[0022] The high-frequency channel is triggered to acquire high-frequency time-domain signals from each station, and the speckle suppression weighting coefficient is used to spatially weight and synthesize the high-frequency time-domain signals to obtain a synthesized ultrasound time-domain signal.

[0023] The synthetic ultrasonic time-domain signal is converted to the frequency-wavenumber domain to extract the measured total phase difference. The joint decoupling equation is then used to calculate the pure stress phase difference by combining the measured total phase difference, the obtained initial phase boundary condition value, and the geometric divergence phase value calculated based on the local relative curvature radius.

[0024] The relative rate of change of sound velocity is calculated based on the pure stress phase difference, and the surface stress value is obtained by combining the pre-calibrated acoustoelastic coefficient.

[0025] Furthermore, the specific steps for extracting the low-frequency interference voltage signal through the low-frequency channel and calculating the local relative radius of curvature at each target point on the continuous scanning trajectory are as follows:

[0026] Phase demodulation and phase expansion of the low-frequency interference voltage signal yield a spatial bias signal that reflects the changes in the macroscopic contour of the surface.

[0027] The spatial bias signal is subjected to spatial low-pass filtering to remove high-frequency spatial phase noise;

[0028] The second spatial directional derivative of the filtered spatial bias signal along the preset propagation direction is calculated, and the reciprocal of the second spatial directional derivative is used as the equivalent local relative radius of curvature.

[0029] Its innovative principle lies in:

[0030] By isolating high-frequency dynamic ultrasonic components through a low-frequency signal channel, the macroscopic curvature is directly calculated using the second-order spatial directional derivative of the steady-state spatial bias signal, thus providing an accurate input reference for subsequent acoustic dispersion stripping caused by geometric effects.

[0031] Furthermore, the specific steps for extracting the local light intensity amplitude of each dwelling point of the microarray to generate the corresponding speckle suppression weighting coefficients are as follows:

[0032] The multi-axis automatic scanning mechanism is aligned with the target point to perform a cross-shaped discrete point micro-array scan;

[0033] The average value, effective value, or peak-to-peak value of the low-frequency interference voltage signal at each dwell point within a preset time window is calculated as the local light intensity amplitude;

[0034] Normalization calculation is performed by dividing the local light intensity amplitude of a single dwelling point by the sum of the local light intensity amplitudes of all dwelling points in the same microarray, and a higher synthesis weight is assigned to the corresponding dwelling point to obtain the speckle suppression weight coefficient.

[0035] Its innovative principle lies in:

[0036] Multi-point micro-displacement scanning is used to induce spatial decorrelation of the probe spot in the random speckle field caused by the micro-roughness of the workpiece surface. Normalized arithmetic is used to evaluate the real-time optical reflection quality of each dwelling point, thereby adaptively measuring the signal-to-noise ratio contribution of the detection position in spatial distribution.

[0037] Furthermore, the specific steps for generating the phase diagram matrix using the local relative radius of curvature as a feedforward control parameter are as follows:

[0038] Construct a spatial phase analytical model superimposed with periodic grating phase function terms and curvature compensation terms;

[0039] The curvature compensation equivalent focal length is obtained by multiplying the local relative curvature radius of the feedforward input by the optical magnification of the lens group, and the curvature compensation equivalent focal length is substituted into the curvature compensation term.

[0040] The theoretical phase deflection at each coordinate point on the pixel target surface of the spatial light modulator is calculated. The theoretical phase deflection is then linearly quantized and phase-folded, and mapped to the grayscale level supported by the spatial light modulator to generate the two-dimensional phase diagram matrix.

[0041] Its innovative principle lies in:

[0042] Introducing a phase pre-distortion opposite to the bending direction of the workpiece surface into the feedforward control layer modulates the initial excitation wavefront of the pulsed laser, thereby counteracting the geometric divergence effect when ultrasonic surface waves propagate along non-planar paths.

[0043] Furthermore, the step of driving the spatial light modulator to load the phase diagram matrix to excite broadband ultrasonic waves on the surface of the mechanical transmission component also includes a protective step performed in conjunction with the above:

[0044] The pulsed laser beam emitted from the pulsed laser is controlled, and the cross-sectional area of ​​the emitted beam is increased by the lens combination inside the pre-beam expander. The peak energy density of the pulsed laser incident on the liquid crystal target surface of the spatial light modulator is reduced to below the nominal damage threshold of the spatial light modulator before entering the spatial light modulator for wavefront deflection.

[0045] Its innovative principle lies in:

[0046] By precisely controlling the energy density of the single-pulse optical field at a specific target surface in space using a beam expander, physical damage protection can be achieved for the core optical deflection layer hardware under test without losing the overall pulse energy output.

[0047] Furthermore, the specific steps for spatially weighting and synthesizing the high-frequency time-domain signal by invoking the speckle suppression weighting coefficients to obtain the synthesized ultrasound time-domain signal are as follows:

[0048] Using the high-frequency time-domain signal of the central target point of the microarray as the timing reference, the time delay of other stationary point signals relative to the timing reference is calculated by the cross-correlation algorithm.

[0049] Based on the time delay, perform a data sequence index shift operation or a fractional delay compensation operation on the high-frequency time-domain signals of the adjacent dwelling points to complete the time-domain phase alignment of the multi-point detection signals in the microarray.

[0050] The aligned high-frequency time-domain signal is multiplied by the corresponding speckle suppression weight coefficient, and the product is then spatially weighted in the digital domain to synthesize the synthesized ultrasonic time-domain signal with a high signal-to-noise ratio.

[0051] Its innovative principle lies in:

[0052] First, the acoustic path difference delay introduced by the physical displacement of the sub-spot is eliminated by the cross-correlation algorithm to prevent interference cancellation caused by the phase misalignment of the high-frequency waveform when the signal is directly accumulated. Then, the interference of dark speckle noise on the high-frequency signal is suppressed by the light intensity weight extracted in advance.

[0053] Furthermore, the specific steps for calculating the pure stress phase difference by calling the joint decoupling equation to use the measured total phase difference, the obtained initial phase boundary condition value, and the geometric divergence phase value calculated based on the local relative radius of curvature are as follows:

[0054] The synthetic ultrasonic time-domain signals of adjacent target points are recombined into a two-dimensional spatiotemporal data matrix and a two-dimensional discrete Fourier transform is performed. The center frequency is located in the two-dimensional spectrum diagram in the frequency-wavenumber domain and the measured wavenumber is extracted.

[0055] The measured total phase difference is obtained by multiplying the difference between the measured wavenumber and the pre-stored stress-free reference wavenumber by the ultrasonic surface wave propagation distance parameter.

[0056] The dimensionless geometric curvature factor is obtained by dividing the ultrasonic surface wave propagation distance parameter by the local relative radius of curvature, and the dimensionless geometric curvature factor is multiplied by the theoretical basis phase under the stress-free flat surface state to obtain the geometric divergence phase value.

[0057] The initial phase boundary condition values ​​are calculated by analyzing the phase diagram matrix and combining it with the pre-established calibration relationship.

[0058] By adding the initial phase boundary condition value to the measured total phase difference and subtracting the geometric divergence phase value using the joint decoupling equation, the pure stress phase difference, which eliminates the influence of macroscopic deformation and optical field pre-distortion, is separated.

[0059] Its innovative principle lies in:

[0060] A unified decoupled algebraic operation model was established, which discretized the total phase shift obtained from spatial measurements into material stress effect terms, geometric surface sound propagation path change terms, and spatial modulation initial phase terms. Multi-field coupling interference artifacts were eliminated through pure digital algebraic logic, and pure stress characteristic variables were extracted.

[0061] Furthermore, the specific steps for calculating the relative rate of change of sound speed based on the pure stress phase difference are as follows:

[0062] Divide the calculated pure stress phase difference by the product of the preset stress-free reference wavenumber and the preset ultrasonic surface wave propagation distance parameter, and take the opposite number to obtain the dimensionless relative sound velocity change rate.

[0063] Then, by utilizing the quantization conversion of acoustic characteristic quantities to mechanical parameters, the relative sound velocity change rate is divided by the pre-calibrated acoustoelastic coefficient to calculate the surface stress value.

[0064] Its innovative principle lies in:

[0065] Based on the principle of acoustoelasticity of solids, a deterministic correspondence between the relative deviation of dimensionless wave velocity and the residual stress (tensile stress and compressive stress) on the solid surface is determined, thus enabling precise quantification of stress values.

[0066] Furthermore, after calculating the surface stress value using the pre-calibrated acoustoelastic coefficient, a data completeness verification step is also performed:

[0067] When the surface stress value of any target point exceeds the preset physical calibration range and the signal quality does not meet the preset signal-to-noise ratio condition, the current test point is determined to be an abnormal data point and forced removal is performed. Then, a spatial pixel neighborhood window is constructed with the abnormal data point as the center, and the arithmetic mean of the surrounding valid data points is calculated to perform neighborhood numerical interpolation and reconstruction to generate a residual stress spatial distribution map.

[0068] This invention provides a laser ultrasonic testing device and automatic analysis system for surface stress of mechanical transmission components. It has the following beneficial effects:

[0069] 1. This invention extracts the local relative radius of curvature of the workpiece surface through a low-frequency channel, superimposes phase pre-distortion on the laser excitation end of the feedforward control spatial light modulator, and eliminates geometric divergence phase values ​​at the back end using a joint decoupling equation. This scheme, through the combination of physical compensation in the early stage of the optical path and algorithmic decoupling in the later stage, eliminates the interference of wavefront divergence and phase distortion caused by the macroscopic curved surface structure of mechanical transmission components on the propagation of ultrasonic surface waves, extracts the pure stress phase difference, and thus improves the accuracy of stress measurement on complex curved surfaces.

[0070] 2. This invention controls a multi-axis mechanism to perform micro-array scanning in the neighborhood of the target point, calculates the local light intensity amplitude using the low-frequency DC component of the interference signal and generates speckle suppression weight coefficients, and then performs temporal alignment and spatial weighted synthesis on the high-frequency ultrasonic time-domain signal. This processing method dynamically allocates the calculated weights according to the actual reflected light intensity of the detection point, reduces the proportion of random speckle and light signal attenuation caused by surface micro-roughness in the total signal, suppresses optical detection noise, and ensures the signal-to-noise ratio during continuous scanning.

[0071] 3. The automatic analysis system of the present invention is equipped with data completeness verification and neighborhood interpolation logic. By setting the calibration range and signal-to-noise ratio threshold, it automatically judges and removes abnormal data points, and then fills them with the arithmetic mean of the effective data in the spatial pixel neighborhood window. This mechanism eliminates the measurement dead zone interference caused by local processing defects or severe morphological changes in the workpiece. It can output a global surface residual stress spatial distribution map with continuous physical coordinates and complete data structure without manual intervention, thus improving the reliability of automated detection. Attached Figure Description

[0072] Figure 1 This is a system framework diagram of the present invention;

[0073] Figure 2 This is a flowchart of the automatic analysis system of the present invention;

[0074] Figure 3 This is a flowchart of the local relative radius of curvature extraction process of the present invention;

[0075] Figure 4 This is a schematic diagram illustrating the principle of microarray scanning and speckle weight extraction of the present invention.

[0076] Figure 5 This is a schematic diagram of the physical layer adaptive modulation optical path and control principle of the present invention;

[0077] Figure 6 This is a schematic diagram illustrating the principle of ultrasonic time-domain signal acquisition and spatial weighted synthesis according to the present invention.

[0078] Figure 7 This is a schematic diagram illustrating the principle of ultrasonic phase decoupling and stress feature extraction of the present invention.

[0079] Figure 8 This is a schematic diagram illustrating the principle of quantitative assessment and spatial mapping of residual stress in this invention.

[0080] Figure 9 This is a comparison diagram of the local light intensity amplitude and speckle suppression weight distribution of the single-target microarray of the present invention;

[0081] Figure 10 This is a comparative line graph showing the change of pure stress phase difference with spatial curvature in this invention;

[0082] Figure 11 This is a line graph comparing the signal-to-noise ratio of the high-frequency ultrasonic synthesized signal according to the present invention.

[0083] Figure 12 This is a bar chart comparing the residual stress detection errors of the present invention. Detailed Implementation

[0084] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0085] See attached document Figure 1 The present invention provides a laser ultrasonic testing device and automatic analysis system for surface stress of mechanical transmission components, including a pulsed laser excitation module, a laser interference receiving module, a multi-axis automatic scanning mechanism and an automatic analysis system.

[0086] The pulsed laser excitation module is used to excite broadband ultrasonic waves on the surface of mechanical transmission components. Its optical path includes a pulsed laser, a pre-expander, a spatial light modulator, and a lens group.

[0087] The output of the pulsed laser is optically coupled to a pre-amplifier. The output beam of the pre-amplifier is aligned with the liquid crystal target of the spatial light modulator. The reflected light from the spatial light modulator is projected onto the surface of the transmission component via a lens group. The pre-amplifier is used to reduce the energy density of the pulsed laser to below the damage threshold of the spatial light modulator. The beam expansion ratio of the pre-amplifier is determined based on the single-pulse energy, pulse width, beam diameter of the pulsed laser, and the nominal damage threshold of the spatial light modulator, ensuring that the peak energy density incident on the liquid crystal target of the spatial light modulator is lower than the damage threshold of the spatial light modulator.

[0088] The laser interferometer receiving module is used to acquire the ultrasonic vibration response of the surface of a transmission component. The module includes a probe laser and an interferometer; the output of the photodetector of the interferometer is electrically connected to the acquisition circuit.

[0089] The acquisition circuit is configured with low-frequency and high-frequency channels in parallel. The cutoff frequency range of the low-frequency channel is DC to 1kHz, and the bandwidth range of the high-frequency channel is 1MHz to 20MHz. The low-frequency channel is used to extract low-frequency interference voltage signals and provide data for surface macroscopic profile extraction and speckle reflection intensity assessment; the high-frequency channel is used to acquire broadband ultrasonic vibration time-domain signals.

[0090] The end effector of the multi-axis automatic scanning mechanism is equipped with a lens group of a pulsed laser excitation module and an optical probe of a laser interference receiving module. Through spatial displacement control, the multi-axis automatic scanning mechanism enables the detection optical path to perform continuous scanning, target point return, and microarray positioning on the surface of the transmission component. The pulsed laser, pre-expander, and spatial light modulator are fixedly mounted on the equipment base or integrated into the end effector. When the pulsed laser, pre-expander, and spatial light modulator are fixedly mounted on the equipment base, the excitation beam modulated by the spatial light modulator is transmitted through a relay reflector group, an articulated light guide arm, or an equivalent light guide assembly to the lens group on the end effector, and then focused and projected onto the surface of the mechanical transmission component.

[0091] The signal input terminal of the automatic analysis system is electrically connected to the acquisition circuit of the laser interferometer receiving module, and the control output terminal is connected to the servo controllers of the spatial light modulator and the multi-axis automatic scanning mechanism, respectively. The automatic analysis system is internally equipped with a processing unit. It also contains a memory that stores programs executable by the processing unit, including curvature extraction, speckle suppression weight generation, phase diagram matrix generation, high-frequency time-domain signal weighted synthesis, frequency-wavenumber domain analysis, joint decoupling equation solving, and acoustoelastic equation solving programs.

[0092] Before performing the test, the automated analysis system performs system calibration. System calibration includes: calibration of the conversion relationship between the interference voltage signal output by the interferometer and the equivalent optical path difference or displacement on the surface; calibration of the correspondence between the grayscale value of the spatial light modulator and the actual optical phase delay; calibration of the correspondence between the phase diagram matrix loaded by the spatial light modulator and the initial phase boundary conditions of the acoustic field; calibration of the stress-free reference wavenumber kref; and calibration of the acoustoelastic coefficient K in the acoustoelastic equation corresponding to the material under test. These calibration parameters are stored in the automated analysis system and are invoked by the processing unit during the processes of local relative radius of curvature extraction, phase diagram matrix generation, joint decoupling equation solving, and acoustoelastic equation solving.

[0093] See attached document Figure 2 The automatic analysis method based on the above-mentioned detection device provided by the present invention may include the following steps:

[0094] S100, the automatic analysis system controls the multi-axis automatic scanning mechanism to perform continuous scanning along the surface of the transmission component. During the scanning process, the automatic analysis system controls the low-frequency channel of the laser interferometer receiving module to extract the low-frequency interference voltage signal. The automatic analysis system performs phase demodulation, phase expansion, and calibration conversion on the low-frequency interference voltage signal through the processing unit to obtain a spatial bias signal reflecting the macroscopic contour changes of the surface under test. Subsequently, the automatic analysis system performs spatial low-pass filtering on the spatial bias signal and calculates the second-order spatial derivative of the spatial bias signal to obtain the local relative radius of curvature at each target point on the continuous scanning trajectory.

[0095] S200: The automatic analysis system controls the multi-axis automatic scanning mechanism to return to the target position and performs a sub-spot-scale cross-shaped discrete point micro-array scan within the neighborhood of the target. The automatic analysis system extracts the local reflected light intensity amplitude of each dwelling point of the micro-array through the low-frequency channel, and then normalizes each local reflected light intensity amplitude to generate the speckle suppression weight coefficient corresponding to each micro-node.

[0096] In the S300, the automatic analysis system uses the local relative radius of curvature as the feedforward control parameter. Based on this radius, the system calculates the geometrically conjugate phase diagram matrix and sends it to the drive circuit of the spatial light modulator. Under the control of the drive signal output from the automatic analysis system, the spatial light modulator loads the phase diagram matrix, then modulates the excitation light field into a phased array shape, and excites broadband ultrasonic waves on the surface of the transmission component.

[0097] In S400, at each dwell point of the microarray scan, the automatic analysis system triggers the high-frequency channel of the laser interferometer receiving module to acquire broadband ultrasonic vibration time-domain signals. Subsequently, the automatic analysis system calls the speckle suppression weighting coefficient to perform digital domain spatial weighted synthesis of the high-frequency time-domain signals at each dwell point, obtaining a synthesized ultrasonic time-domain signal.

[0098] The S500 automatic analysis system converts the synthetic ultrasound time-domain signal to the frequency-wavenumber domain, extracts the measured total phase difference, and calls the joint decoupling equation. The automatic analysis system substitutes the geometric divergence phase term corresponding to the equivalent local radius of curvature in the propagation direction and the initial phase boundary condition corresponding to the phase diagram matrix into the joint decoupling equation. The joint decoupling equation adds the initial phase boundary condition value to the measured total phase difference and subtracts the geometric divergence phase value to calculate the pure stress phase difference. Specifically, the measured total phase difference is obtained from the synthetic ultrasound time-domain signal acquired by the high-frequency channel through frequency-wavenumber domain analysis; the initial phase boundary condition is obtained from the phase diagram matrix loaded by the spatial light modulator combined with the system calibration relationship; and the geometric divergence phase value is obtained by converting the local relative radius of curvature calculated from the spatial offset signal extracted by the low-frequency channel.

[0099] The S600 automatic analysis system calculates the relative rate of change of sound velocity based on the pure stress phase difference and the ultrasonic surface wave propagation distance parameters. It then substitutes this rate of change of sound velocity into the acoustoelastic equation to solve for the surface stress value at the target location. After the scanning trajectory is traversed, the automatic analysis system generates global stress distribution data based on the surface stress values ​​at each point.

[0100] See attached document Figure 3 The step S100 for obtaining the local relative radius of curvature provided by the present invention may specifically include:

[0101] S110, the automatic analysis system sends a trajectory tracking command to the servo controller of the multi-axis automatic scanning mechanism. Upon receiving the command, the multi-axis automatic scanning mechanism drives the optical probe to perform a continuous sliding scan along the surface of the mechanical transmission component to be measured. The scanning trajectory covers the macroscopic contour feature area of ​​the mechanical transmission component, which can be a gear involute tooth surface path or a shaft shoulder arc path. In one embodiment, to acquire two-dimensional surface topography data, the multi-axis automatic scanning mechanism employs a raster scanning or helical scanning strategy to cover the surface area to be measured.

[0102] S120: During the continuous movement of the optical probe, the detection laser emits a continuous laser beam that illuminates the surface of the transmission component. The interferometer receives the reflected light signal carrying surface topography information and converts it into an interference voltage signal. The automatic analysis system controls the low-frequency channel of the laser interferometer receiving module to acquire this interference voltage signal in real time. The cutoff frequency of the low-frequency channel is set between DC and 1kHz, with its upper limit determined based on the moving speed of the multi-axis automatic scanning mechanism and the expected rate of change of the workpiece's macroscopic contour. This frequency band is used to filter out high-frequency ultrasonic vibration components, retaining the low-frequency interference voltage signal that reflects the change in the optical path difference of the macroscopic surface. The automatic analysis system performs phase demodulation, phase expansion, and calibration conversion on the low-frequency interference voltage signal through the processing unit to obtain a spatial offset signal reflecting the change in the macroscopic contour of the surface under test. .

[0103] S130, the automatic analysis system acquires the spatial offset signal amplitude corresponding to each coordinate point on the continuous scanning trajectory and, combined with the two-dimensional position coordinates fed back by the multi-axis automatic scanning mechanism, constructs a two-dimensional spatial distribution matrix. When the continuous scanning trajectory is an irregular curved path, the automatic analysis system interpolates and maps the sampled data to a preset regular two-dimensional grid or surface parameter coordinate grid based on the real-time coordinates fed back by the multi-axis automatic scanning mechanism, and then constructs the two-dimensional spatial distribution matrix. Machining surfaces typically have microscopic roughness, and high-frequency spatial phase noise caused by surface scattering will be superimposed on the spatial offset signal. The automatic analysis system performs spatial low-pass filtering on the two-dimensional spatial distribution matrix through the processing unit to filter out high-frequency spatial phase noise and retain the spatial offset signal reflecting the macroscopic geometry of the workpiece. Spatial low-pass filtering can be implemented using a two-dimensional Gaussian filter or a sliding window averaging filter algorithm.

[0104] S140, the automatic analysis system calculates the second-order spatial derivative of the filtered spatial bias signal to analyze the local relative radius of curvature at each target point on the continuous scanning trajectory. The automatic analysis system converts the spatial bias signal into a discrete function of two-dimensional plane coordinates according to a preset spatial sampling step size. The preset spatial sampling step size is determined based on the spot radius of the probe laser, ranging from one-quarter to one-half of the spot radius. The first derivative of the surface profile function reflects the surface's tilt state, while the second derivative reflects the rate of change of the tilt slope, i.e., the macroscopic curvature of the surface. For gently changing surfaces with small curvature, the radius of curvature can be approximated as the reciprocal of the sum of the second-order partial derivatives in the corresponding spatial direction. When the ultrasonic surface wave propagates along a preset propagation direction, the automatic analysis system preferentially selects the equivalent local radius of curvature along that propagation direction as the subsequent geometric phase compensation parameter. The specific calculation formula for the local relative radius of curvature is as follows:

[0105] ;

[0106] in, Current two-dimensional spatial coordinates The local relative radius of curvature at that location; The amplitude of the spatial offset signal at the current two-dimensional spatial coordinates (x, y); The second-order partial derivative of the spatial bias signal along the x-axis; This represents the second-order partial derivative of the spatial bias signal along the y-axis. When the absolute value of the second-order spatial derivative is lower than a preset threshold, the automatic analysis system determines that the current region is an approximately planar region and sets the local relative radius of curvature to a preset upper limit or an infinity equivalent value to avoid numerical divergence.

[0107] Since the collected data is a discrete two-dimensional spatial distribution matrix, the automatic analysis system uses the central difference method to approximate the above-mentioned second-order partial derivatives. Taking the x-axis direction as an example, the central difference operation uses the amplitude of data points at adjacent sampling steps before and after the current target point to complete the difference calculation, thereby converting the continuous analytical model into a discrete data-executable calculation process.

[0108] The automatic analysis system associates and stores the local relative radius of curvature data obtained along the scanning trajectory with the corresponding physical space coordinates, establishing a curvature mapping matrix in the global coordinate system. This curvature mapping matrix serves as the feedforward control parameter for the subsequent driving pulsed laser excitation module to perform physical modulation.

[0109] See attached document Figure 4 The step S200 of generating speckle suppression weighting coefficients provided by the present invention may specifically include:

[0110] S210, the automatic analysis system sends an absolute coordinate positioning command to the multi-axis automatic scanning mechanism, controlling the center of the probe laser spot to align with the current target position obtained by continuous scanning. After reaching the target position, the multi-axis automatic scanning mechanism performs a cross-shaped discrete point micro-array scan centered on the target point. During the cross-shaped discrete point micro-array scan, the multi-axis automatic scanning mechanism keeps the center of the excitation spot array formed by the pulsed laser excitation module at the preset excitation position corresponding to the current macroscopic target point, and drives the probe spot of the laser interferometer receiving module to perform micro-displacement scanning in the neighborhood of the target point. In another embodiment, if the excitation spot and the probe spot synchronously displace with the end effector, the automatic analysis system records the center coordinates of the excitation spot and the center coordinates of the probe spot corresponding to each dwell point, and incorporates the equivalent acoustic path difference caused by the change of the two into the subsequent time-domain alignment calculation.

[0111] In this embodiment, the cross-shaped microarray includes a central target point and four neighboring points distributed in mutually perpendicular directions, for a total of five dwelling points. The spatial displacement step size between adjacent dwelling points is set to one-quarter to one-third of the radius of the focused spot of the probe laser. This step size ensures that the illumination areas of the spot at adjacent dwelling points partially overlap, while utilizing the small spatial differences in surface micro-roughness to generate a spatial decorrelation effect in the speckle field.

[0112] S220, the multi-axis automatic scanning mechanism briefly pauses at each dwell point of the microarray. When the probe is stationary, the reflected light received by the interferometer is randomly phase-modulated by the surface micro-roughness, forming a speckle field in space. The probe light at different dwell points falls into the bright or dark speckle areas, causing differences in the steady-state light intensity level output by the low-frequency channel.

[0113] To quantify the optical signal-to-noise ratio at the current detection point, the automatic analysis system extracts the low-frequency interference voltage signal of each dwelling point within a preset time window and calculates the average, effective, or peak-to-peak value of the signal within that time window as the local light intensity amplitude reflecting the speckle reflection intensity at the current dwelling point. The local light intensity amplitude is positively correlated with the optical sensitivity of subsequent high-frequency ultrasonic vibration detection at that point. In one embodiment, the length of the preset time window is determined based on the response time of the low-frequency channel and the system sampling rate, and a steady-state data segment of 10 to 50 milliseconds can be selected.

[0114] S230: After obtaining the local light intensity amplitudes of the five dwelling points of the current microarray, the automatic analysis system performs normalization calculations through the processing unit to construct a speckle suppression weight matrix. This calculation is used to evaluate the local optical reflection quality of each dwelling point, assign higher composite weights to detection positions with higher signal-to-noise ratios, and reduce the noise contribution of weak light intensity detection positions in the dark speckle region to the overall signal. The specific normalization calculation formula for generating the speckle suppression weight coefficients is as follows:

[0115] ;

[0116] in, This represents the speckle suppression weight coefficient corresponding to the i-th dwell point in the current microarray. and The index variable is the serial number of the station, and in this embodiment, its value ranges from 1 to 5. For the first The local light intensity amplitude was calculated and extracted from each dwelling point; This is the sum of the local light intensity amplitudes of the five dwelling points in the same microarray. When the sum of the local light intensity amplitudes of all dwelling points in the same cross-shaped microarray is lower than the preset light intensity threshold, the automatic analysis system determines through the processing unit that the current target point has low reflectivity or dark speckle anomalies, and controls the multi-axis automatic scanning mechanism to expand the microarray search range, reselect dwelling points, or use a preset equal weighting coefficient as a downgrade process.

[0117] The automatic analysis system associates the five speckle suppression weight coefficients with the corresponding microscopic coordinates of the dwell points and caches them in the internal memory for spatial weighted synthesis after the subsequent acquisition of time-domain data from the high-frequency channel.

[0118] See attached document Figure 5 The step S300 of the feedforward driven pulsed laser excitation module provided by the present invention may specifically include:

[0119] S310, the automatic analysis system retrieves the local relative radius of curvature at the target point on the scanning trajectory and generates a geometrically conjugate phase diagram matrix by combining it with preset ultrasonic long parameters. To excite spatially phase-consistent and energy-concentrated array ultrasonic waves on a surface with macroscopic curvature, the automatic analysis system constructs a spatial phase analytical model superimposed with array modulation terms and curvature compensation terms. This model, based on the generation of multi-line or multi-point distributed array spots, incorporates a phase pre-distortion opposite to the workpiece surface curvature to counteract the geometric divergence generated when ultrasonic surface waves propagate along the curved surface. The specific calculation formula for the theoretical phase deflection at each coordinate point on the pixel target surface of the spatial light modulator by the automatic analysis system is as follows:

[0120] ;

[0121] in, Two-dimensional coordinates of the pixel plane of the spatial light modulator The theoretical phase deflection at that point. , , and These are integers representing the horizontal and vertical pixel indices of the spatial light modulator. The physical spacing between pixels; The periodic grating phase function is calculated based on the preset ultrasonic array spacing. The center wavelength of the laser output from the pulsed laser; The equivalent focal length for curvature compensation is determined by the local relative radius of curvature of the feedforward input. Obtained by multiplying by the optical magnification of the lens group; This is the fundamental phase constant of the optical path system, set based on the background aberration measurements during system calibration; The modulo operator is used to fold continuously calculated phase values ​​into the interval between 0 and 2π. The automatic analysis system linearly quantizes the theoretical phase deflection, mapping it to the grayscale levels supported by the spatial light modulator hardware, and generates a two-dimensional phase diagram matrix. Based on the pre-calibrated optical magnification, imaging direction, and phase modulation symbol, the automatic analysis system converts the local relative radius of curvature into a curvature-compensated equivalent focal length. ;when When the phase exceeds the adjustable phase range of the spatial light modulator, the automatic analysis system performs phase folding, amplitude limiting, or partitioning approximation on the phase diagram matrix.

[0122] S320: The automatic analysis system sends the digital frame of the two-dimensional phase diagram matrix to the pulsed laser excitation module and loads it into the liquid crystal driving circuit of the spatial light modulator. After sending the two-dimensional phase diagram matrix to the driving circuit of the spatial light modulator, the automatic analysis system waits for the spatial light modulator to reach a preset stable response state before outputting a synchronization trigger signal to the pulsed laser, so that the pulse emission time of the pulsed laser matches the stable display time of the phase diagram of the spatial light modulator.

[0123] Under system timing-triggered control, the pulsed laser emits a pulsed laser beam. The initial laser beam enters a pre-expander, which increases the cross-sectional area of ​​the emitted beam through an internal lens assembly, attenuating the peak energy density of the laser pulse below the damage threshold of the spatial light modulator. As one implementation, the beam expansion ratio of the pre-expander ensures that the laser energy density at the output surface is within a preset safe range adapted to the selected spatial light modulator, thus preventing high-energy pulses from causing optical damage to the liquid crystal deflection layer and dielectric reflective film of the spatial light modulator.

[0124] The S330 beam, after its energy density is attenuated by a pre-expander, is incident on the liquid crystal target surface of the spatial light modulator. The wavefront of the optical field undergoes phase shift deflection after spatial phase modulation by a preset two-dimensional phase diagram matrix, and the reflected beam carrying the modulated phase enters the lens group. The lens group completes optical relay transmission and focal convergence, and re-compresses the beam diameter. The refocused beam, with a safe energy density to meet the thermoelastic effect of the transmission component material, is projected onto the surface to be measured, forming multiple equally spaced linear array spots or dot matrix spots on the workpiece surface. The surface material of the transmission component absorbs the laser energy and undergoes transient thermal expansion, thereby exciting broadband ultrasonic waves that pre-converge towards the specified focal direction.

[0125] See attached document Figure 6The step S400 of generating a synthetic ultrasound time-domain signal provided by the present invention may specifically include:

[0126] The S410 multi-axis automatic scanning mechanism drives the optical probe to sequentially position itself at each dwell point of the cross-shaped microarray. After the probe reaches each dwell point and remains stationary, the automatic analysis system sends a synchronous trigger pulse to the pulsed laser excitation module, exciting broadband ultrasonic waves on the workpiece surface. The ultrasonic surface waves propagate to the area covered by the probe spot, causing nanoscale high-frequency vibrations on the surface. This mechanical vibration modulates the phase of the probe laser wavefront emitted by the probe laser. The interferometer receives the reflected light carrying the vibration information and converts it into a voltage signal. The automatic analysis system triggers the high-frequency channel of the laser interferometer receiving module to synchronously acquire the AC component of this voltage signal. The bandwidth of the high-frequency channel is set from 1MHz to 20MHz to isolate baseline drift introduced by environmental vibrations and low-frequency thermal fluctuations, while preserving the ultrasonic surface wave waveform that reflects the material stress characteristics.

[0127] S420, the automatic analysis system caches the discrete high-frequency time-domain signals collected from each station in the processing unit's memory. Due to the sub-spot-scale physical displacement difference between the stations in the microarray, the ultrasonic waves travel different paths from the excitation source to different detection stations, resulting in slight time delays between the high-frequency time-domain signals. Direct arithmetic superposition in the digital domain would cause high-frequency phase misalignment, leading to waveform interference cancellation. Therefore, the automatic analysis system uses the high-frequency time-domain signal from the center target point of the microarray as a timing reference and calculates the time delay of the signals from the other four neighboring points relative to this reference using a cross-correlation algorithm. This time delay can be an integer sampling delay or a fractional sampling delay; for fractional sampling delays, the processing unit uses interpolation delay filtering or frequency-domain phase compensation to complete time-domain alignment. After obtaining this time delay, the automatic analysis system performs data sequence index shifting or fractional delay compensation operations on the high-frequency time-domain signals of the neighboring points to complete the time-domain phase alignment of the multi-point detection signals.

[0128] S430: After completing the time-domain alignment of the multi-point signals, the automatic analysis system calls the current microarray speckle suppression weight coefficients stored in step S200. The automatic analysis system multiplies the high-frequency time-domain signal of each dwelling point with the corresponding speckle suppression weight coefficient, and performs digital domain spatial weighted synthesis on the product result. This process assigns the optical signal-to-noise ratio weight obtained from the low-frequency channel light intensity assessment to the ultrasonic signal acquired in the high-frequency channel, thereby increasing the synthesis weight of the bright speckle area detection data and reducing the noise contribution of the dark speckle area detection data. The specific calculation formula for generating the synthesized ultrasonic time-domain signal is as follows:

[0129] ;

[0130] in, This is the discrete synthesized ultrasonic time-domain signal output after time-domain alignment and spatial weighting processing; For discrete sampling points, use the time index variable. This is the index variable for the stationary points in the cross-shaped microarray, and its value ranges from 1 to 5. For the first The speckle suppression weight coefficient corresponding to each residence point; For the high-frequency channel in the first The original high-frequency time-domain signal sequence collected at each station; For the first The number of lag sampling points of the high-frequency time-domain signal at each dwell point relative to the signal at the center target point, where the number of lag sampling points is equal to the physical time delay multiplied by the analog-to-digital conversion sampling rate of the high-frequency channel. When When the corresponding residence point is the central target point itself, It is 0.

[0131] The automatic analysis system uses the synthesized ultrasonic time-domain signal as the representative ultrasonic waveform of the current macroscopic target location and transmits it to the subsequent dispersion feature extraction and decoupling algorithm module.

[0132] See attached document Figure 7 The step S500 for obtaining the pure stress phase difference provided by the present invention may specifically include:

[0133] S510, the automatic analysis system retrieves the synthesized ultrasonic time-domain signals from adjacent target point positions along the continuous scanning trajectory. The adjacent target points are multiple targets arranged along the ultrasonic surface wave propagation direction. Based on the spatial coordinates recorded by the multi-axis automatic scanning mechanism, the automatic analysis system resamples the synthesized ultrasonic time-domain signals corresponding to the multiple target points into an equally spaced spatial sequence. Combining the physical spatial step size of the adjacent target points recorded by the multi-axis automatic scanning mechanism, the automatic analysis system reassembles the one-dimensional time series into a two-dimensional spatiotemporal data matrix. Subsequently, the automatic analysis system performs a two-dimensional discrete Fourier transform on the two-dimensional spatiotemporal data matrix, converting the ultrasonic signal from the time-space domain to the frequency-wavenumber domain.

[0134] In the generated two-dimensional frequency-wavenumber spectrum, the automatic analysis system uses a two-dimensional peak-finding algorithm to locate the center frequency of the concentrated ultrasonic energy and extracts the measured wavenumber corresponding to that center frequency. The automatic analysis system subtracts the measured wavenumber from the pre-stored stress-free reference wavenumber and multiplies the difference by the physical propagation distance of the ultrasonic surface wave to obtain the measured total phase difference. The stress-free reference wavenumber kref is obtained by calibration using a stress-free standard test block with the same material, surface condition, and center frequency, or by measurement in the stress-free reference region during the system initialization phase.

[0135] S520, the automatic analysis system retrieves the curvature mapping matrix cached in step S100 and reads the local relative radius of curvature corresponding to the current macroscopic analysis region. When the ultrasonic surface wave propagates along the preset propagation direction, the automatic analysis system determines the propagation direction based on the line connecting the center of the excitation spot array formed by the pulsed laser excitation module and the center of the detection spot of the laser interferometer receiving module, and calculates the second-order directional derivative along the propagation direction in the spatial bias signal L(x,y). For gently changing surfaces with small curvature, the reciprocal of this second-order directional derivative is used as the equivalent local radius of curvature. .

[0136] In the non-planar region of a mechanical transmission component, when an ultrasonic surface wave propagates along a curved surface, the actual propagation path and wavefront shape are distorted due to the surface's geometric curvature, resulting in a phase drift independent of material stress. The automatic analysis system approximates the curved surface in the neighborhood of the current propagation path as having an equivalent local radius of curvature. The arc segment. The equivalent distance along which the ultrasonic surface wave propagates is... The geometric deflection angle corresponding to the arc segment is approximately expressed as: This ratio is a dimensionless geometric curvature factor, used to characterize the influence of surface curvature on the propagation path and wavefront development of ultrasonic surface waves. The automatic analysis system correlates this dimensionless geometric curvature factor with the theoretical basis of the stress-free flat surface state. Multiplying these yields the geometric divergence phase value. The automatic analysis system calculates the geometric divergence phase value based on the local circular arc approximation and the ray acoustic propagation relationship. The specific calculation formula is as follows:

[0137] ;

[0138] in, The geometric divergence phase value is calculated based on the surface macroscopic curvature; The equivalent ultrasonic surface wave propagation distance parameter used in the current phase assessment; Let be the equivalent local radius of curvature in the current propagation direction; The system pre-stores the stress-free reference wavenumber at the corresponding center frequency; Characterizes the dimensionless geometric curvature factor introduced by the surface curvature morphology; The theoretical basis for characterizing the stress-free, flat surface state. When a fixed excitation and probe point structure is used, The propagation distance along the workpiece surface between the center of the excitation spot array formed by the pulsed laser excitation module and the center of the detection spot of the laser interferometer receiving module; when an adjacent target point differential structure is used, The equivalent distance between adjacent target points along the propagation direction for participating in phase difference calculation.

[0139] S530, the automatic analysis system parses the two-dimensional phase diagram matrix used to drive the spatial light modulator in step S300, and calculates the initial phase boundary conditions at the current center frequency and propagation direction by combining the pre-established calibration relationship between the optical phase modulation amount, the spatial distribution of the excitation spot array, and the initial phase of the sound field. During the laser thermoelastic conversion process, the optical phase pre-deflection set by the spatial light modulator changes the relative position of the array spot on the workpiece surface in terms of spatial distribution; this change in spatial position is equivalent to the spatial initial phase shift of the synthesized ultrasonic wave array when the ultrasonic wave propagates to the far field.

[0140] To separate the phase component reflecting the residual stress state of the material from the total phase shift, the automatic analysis system invokes the joint decoupling equation. The system substitutes the geometrically divergent phase value corresponding to the local relative radius of curvature and the initial phase boundary condition value corresponding to the phase diagram matrix into the joint decoupling equation. The joint decoupling equation adds the initial phase boundary condition value and subtracts the geometrically divergent phase value from the measured total phase difference to calculate the pure stress phase difference. The plus and minus signs are determined according to the positive phase direction convention established during system calibration; when the optical path or propagation direction changes, the automatic analysis system corrects the signs of the initial phase boundary condition and geometrically divergent phase values ​​according to the calibration parameters. The specific calculation formula for the joint decoupling equation is as follows:

[0141] ;

[0142] in, This is the pure stress phase difference output after physical field decoupling calculation; This refers to the measured total phase difference extracted and calculated by the automatic analysis system in the frequency-wavenumber domain. The initial phase boundary condition values ​​introduced when loading the phase diagram matrix into the spatial light modulator are obtained by converting the phase diagram matrix, the spatial distribution of the excitation spot array, and the initial phase calibration relationship of the acoustic field. The value of the geometric divergence phase is calculated based on the local relative radius of curvature. Based on the equivalent local radius of curvature in the propagation direction The calculated geometric divergence phase value.

[0143] In the above process, the joint decoupling equations are used to separate geometric acoustic effects, optical phase modulation effects, and acoustoelastic effects in the digital domain. The automatic analysis system caches the pure stress phase difference in an internal register and uses it as an input parameter for the next computational stage to calculate the relative rate of change of sound speed.

[0144] See attached document Figure 8 The step S600 of obtaining the decoupled target physical quantity and constructing the spatial distribution of residual stress provided by the present invention may specifically include:

[0145] S610, The automatic analysis system retrieves the pure stress phase difference calculated and cached in step S500. Based on the pre-defined theoretical foundation of surface acoustic waves (SAWs), the phase shift caused by the stress field is converted into a dimensionless relative rate of change of sound velocity. Residual stress alters the material's elastic constants through acoustoelastic effects, thereby causing a change in the propagation velocity of SAWs. This velocity change manifests as a pure stress phase difference in the wavenumber domain. The specific formula for calculating the relative rate of change of sound velocity is as follows:

[0146] ;

[0147] in, To calculate the rate of change of the relative speed of sound in the output; The output is the pure stress phase difference after physical field decoupling; The pre-stored stress-free reference wavenumber for the system; The equivalent ultrasonic surface wave propagation distance parameter used in the current phase assessment; the negative sign in the formula is used to characterize the corresponding relationship between the decrease in sound velocity of surface acoustic waves under tensile stress and the increase in sound velocity under compressive stress.

[0148] The S620 automatic analysis system inputs the relative rate of change of sound velocity into its integrated stress analysis module. This module introduces an acoustoelastic coefficient matched to the material of the transmission component under test, enabling a quantitative conversion from acoustic characteristics to mechanical residual stress values. To ensure the calculation results match the measured object, the acoustoelastic coefficient in the acoustoelastic equation... Calibration coefficients are used to match the material grade, heat treatment state, ultrasonic surface wave mode, center frequency, propagation direction, and testing temperature of the mechanical transmission component under test. Before system testing, uniaxial or biaxial stress calibration experiments are performed on standard test blocks of the same material to obtain the correspondence between the relative rate of change of sound velocity and the known stress, and the fitted acoustoelastic coefficients are then used. Write it into the parameter library of the automatic analysis system.

[0149] In this embodiment, for high-strength alloy steel commonly used in mechanical transmission components, the range of the acoustoelastic coefficient is set to -1.2 × 10⁻⁶. -5 MPa -1 to -3.8×10 -5 MPa -1 The specific formula for calculating the residual stress value at the target point on the surface of the mechanical transmission component is as follows:

[0150] ;

[0151] in, This represents the residual stress value at the current target point, in MPa. When... When the value is positive, it indicates that the workpiece surface is under tensile stress; when... When the value is negative, it indicates that the surface of the workpiece is under compressive stress. The acoustic elastic coefficient is determined through calibration experiments. The relative rate of change of sound is the input.

[0152] In step S630, the automatic analysis system synchronously correlates the residual stress values ​​calculated at each target point with the two-dimensional position coordinates fed back by the multi-axis automatic scanning mechanism in step S100, establishing a stress mapping matrix in the global coordinate system. The stress analysis module performs reconstruction processing on the stress mapping matrix to generate a continuous spatial distribution map of residual stress on the workpiece surface.

[0153] To reduce the impact of anomalous data caused by edge scattering from complex curved surfaces and abrupt noise, the stress analysis module performs data completeness verification before reconstruction. The system reads the yield strength, calibrated range, and signal quality threshold of the material under test as anomaly criteria. When the residual stress value calculated for a target point exceeds the preset physical allowable range or calibrated range, and the signal-to-noise ratio, coherence coefficient, or phase residual corresponding to that target point does not meet preset conditions, the automatic analysis system determines that the point is an anomalous data point through the processing unit and performs removal and neighborhood interpolation filling.

[0154] After removing outlier data points, the automatic analysis system constructs a 3×3 spatial pixel neighborhood window centered on the outlier point. It calculates the arithmetic mean of the valid residual stress data points within this window and uses it for numerical interpolation. If there are no valid data points within the 3×3 window, the window size is progressively expanded by pixel increments until at least one valid data point is obtained for averaging. When an outlier point is located in a boundary region, the automatic analysis system uses existing valid neighboring points for calculation. If no valid data points are found after expanding the window to a preset maximum size, the point is marked as an invalid region and a prompt message is output.

[0155] The automatic analysis system outputs the generated spatial distribution map of residual stress to the display terminal for visualization, or writes the spatial distribution map of residual stress as a control parameter for predicting the fatigue life of transmission components and evaluating the quality of processing technology into the local non-volatile memory.

[0156] To further aid in understanding the technical solutions of the embodiments of the present invention, the above steps are described in detail below with reference to a specific industrial application scenario and numerical calculation process, and relevant experimental verification and effect comparison data are provided.

[0157] See attached document Figure 9The test object is a 20CrMnTi alloy steel gear that has undergone surface shot peening. Due to the involute morphology of the gear tooth surface, its local relative radius of curvature continuously changes with spatial position. In the macroscopic morphology extraction stage of step S100, the automatic analysis system controls the probe to move to a target point near the pitch circle of the tooth surface. After extracting the spatial bias signal using the low-frequency channel, the equivalent local radius of curvature at that point is obtained through second-order spatial derivative calculation. =25.0mm.

[0158] In the microarray speckle weight extraction stage of step S200, the automatic analysis system performs a cross-shaped microarray scan at the target point. The local light intensity amplitudes of the five extracted dwell points are 1.2, 0.8, 0.4, 1.6, and 1.0, respectively. The dwell point corresponding to 0.4 falls precisely into the dark speckle region. The automatic analysis system performs normalization calculations to obtain the speckle suppression weight coefficients. The values ​​are 0.24, 0.16, 0.08, 0.32, and 0.20, respectively. Figure 9 The positive correlation between the light intensity amplitude at each dwell point and the assigned weight coefficient is visually demonstrated, which gives high signal-to-noise ratio points a higher synthesis weight. In steps S300 and S400, the spatial light modulator loads a phase diagram matrix based on a curvature radius of 25.0 mm for feedforward compensation. The high-frequency channel acquires five time-domain signals and aligns them in the time domain. Then, they are weighted and synthesized according to the above weights to output a high signal-to-noise ratio synthesized ultrasonic time-domain signal.

[0159] In the decoupling calculation stage of step S500, the system preset detection parameters are as follows: ultrasonic surface wave propagation distance parameter. =5.0mm, stress-free reference wavenumber obtained during calibration =2.0 rad / mm. The theoretical basis phase calculated from this is... =10.0 rad. Initial phase boundary condition value obtained by phase diagram calculation using a spatial light modulator. =0.10rad. The automatic analysis system performs frequency-to-wavenumber domain conversion on the synthesized ultrasound time-domain signal and extracts the measured total phase difference. =1.84 rad. Simultaneously, the geometric divergence phase value is calculated based on the radius of curvature:

[0160] ;

[0161] Substituting into the joint decoupling equation:

[0162] ;

[0163] In the stress transformation stage of step S600, the automatic analysis system first calculates the relative rate of change of sound velocity:

[0164] ;

[0165] The acoustic elastic modulus of 20CrMnTi steel stored internally in the system =-2.0×10 -5 MPa -1 Substitute the solutions into the acoustic elastic equation to calculate the residual stress:

[0166] ;

[0167] The calculation results accurately reflect the presence of a residual compressive stress of 300 MPa on the surface of the shot-peened area.

[0168] To verify the effectiveness and accuracy of the decoupling algorithm and signal synthesis scheme proposed in this invention, a systematic comparative experiment was conducted using X-ray diffraction (XRD) as a benchmark.

[0169] See attached document Figure 10 In the experiment, 20 test target points were selected at equal intervals along the tooth root to the tooth tip on the involute tooth surface. The actual residual stress distribution of the material along this path was basically uniform, but the local relative radius of curvature on the surface showed a significant nonlinear monotonic decrease. Figure 10 In the figure, the horizontal axis represents the spatial position index of the test target point along the tooth surface, and the vertical axis represents the calculated pure stress phase difference. The dashed broken line in the figure represents the data of the traditional direct measurement method without incorporating the curvature decoupling and optical compensation of this invention. It can be seen that, due to the superposition of geometric dispersion effects, the phase difference measured by the traditional method undergoes severe nonlinear baseline drift as the radius of curvature decreases, completely masking the true stress characteristics. The solid broken line in the figure represents the pure stress phase difference data after processing with the joint decoupling equation of this invention. The decoupled curve maintains high horizontal stability throughout the entire space range, greatly eliminating the artifact fluctuations introduced by macroscopic geometric deformation, proving the mathematical self-consistency and robustness of the multiphysics separation logic of this invention in an engineering environment.

[0170] See attached document Figure 11 To verify the effectiveness of the speckle suppression weighting mechanism, the signal quality of 100 consecutive macroscopic target points of the microarray was recorded in the experiment. The horizontal axis represents the target point number, and the vertical axis represents the signal-to-noise ratio (SNR) of the signal, in dB. Figure 11 The dotted line trajectory in the lower middle represents the signal-to-noise ratio distribution of the ultrasonic time-domain signal obtained by direct detection at a single point. Affected by the random modulation of the speckle field caused by the micro-roughness of the surface under test, its signal-to-noise ratio is extremely unstable. When it falls into the dark speckle region, it drops multiple times (signal-to-noise ratio is below 5dB), causing the subsequent peak finding algorithm to fail. Figure 11The solid line trajectory at the top center represents the signal-to-noise ratio distribution after synthesis using the cross-shaped microarray and local intensity adaptive weighted synthesis of this invention. The overall signal-to-noise ratio of the synthesized signal is improved by 12dB to 18dB, and the fluctuation amplitude of the curve is significantly narrowed, eliminating dead zones caused by random drops and ensuring high fidelity of the time-domain waveform during continuous scanning.

[0171] See attached document Figure 12 The data was extracted from the absolute residual stress test results of three samples with different combinations of roughness and curvature (sample A is a smooth plane, sample B is a rough plane, and sample C is a rough curved surface). The XRD calibration value was used as the zero-point reference. The horizontal axis is divided into three groups, corresponding to the three types of samples; the vertical axis represents the absolute value of the absolute measurement error, in MPa. For the smooth plane sample A, the errors of both the traditional method and the method of this invention are within 15 MPa, showing comparable performance. For the rough plane sample B, which has superimposed microscopic speckle noise, the error of the traditional method expands to approximately 45 MPa, while the error of this invention is stabilized at around 18 MPa due to the weighting mechanism introduced in step S200. For the complex sample C, which simultaneously exhibits speckle noise and surface dispersion, the calculation results of the traditional method show numerical divergence, with errors exceeding 120 MPa, rendering it unusable for quantitative engineering evaluation. The method of this invention, benefiting from the physical-level optical field pre-distortion and the digital-level joint decoupling equation, strictly controls the measurement error within 22 MPa, significantly approaching the measurement accuracy boundary of XRD. The experimental data directly confirm the anti-interference ability and accurate quantification ability of the method of the present invention under complex working conditions.

Claims

1. A laser ultrasonic testing device for surface stress of mechanical transmission components, characterized in that, include: A pulsed laser excitation module, comprising a pulsed laser, a pre-beam expander, a spatial light modulator, and a lens group; A laser interferometer receiving module, including an interferometer and a data acquisition circuit with high and low frequency channels; A multi-axis automatic scanning mechanism, equipped with the lens group and optical probe; The automatic analysis system, electrically connected to the acquisition circuit, the spatial light modulator, and the multi-axis automatic scanning mechanism, includes a processing unit and a memory storing the following program: The curvature extraction program extracts low-frequency interference voltage signals from the low-frequency channel to determine the local relative radius of curvature. The speckle suppression weight generation program generates speckle suppression weight coefficients from local light intensity amplitudes. A phase diagram matrix generation procedure, wherein the local relative radius of curvature generates the phase diagram matrix that drives the spatial light modulator; A high-frequency time-domain signal weighted synthesis program combines the high-frequency time-domain signal of the high-frequency channel with the speckle suppression weighting coefficient to obtain a synthesized ultrasonic time-domain signal; The frequency-wavenumber domain analysis program converts the synthesized ultrasonic time-domain signal into the frequency-wavenumber domain and takes the measured total phase difference. The program for joint decoupling and solving the acoustoelastic equation uses the measured total phase difference, initial phase boundary condition values, and geometric divergence phase values ​​to calculate the pure stress phase difference and solve for the surface stress value.

2. An automated analysis method for laser ultrasonic testing of surface stress in mechanical transmission components, characterized in that, The laser ultrasonic testing device for surface stress of mechanical transmission components as described in claim 1 comprises the following steps: The multi-axis automatic scanning mechanism is controlled to scan along the surface of the mechanical transmission component. The low-frequency interference voltage signal is extracted through the low-frequency channel and the local relative radius of curvature at each target point on the continuous scanning trajectory is calculated. The multi-axis automatic scanning mechanism is controlled to perform discrete point microarray scanning at the target point, and the local light intensity amplitude of each dwelling point of the microarray is extracted to generate the corresponding speckle suppression weight coefficient. The local relative radius of curvature is used as a feedforward control parameter to generate a phase diagram matrix, and the spatial light modulator is driven to load the phase diagram matrix to excite broadband ultrasonic waves on the surface of the mechanical transmission component. The high-frequency channel is triggered to acquire high-frequency time-domain signals from each station, and the speckle suppression weighting coefficient is used to spatially weight and synthesize the high-frequency time-domain signals to obtain a synthesized ultrasound time-domain signal. The synthetic ultrasonic time-domain signal is converted to the frequency-wavenumber domain to extract the measured total phase difference. The joint decoupling equation is then used to calculate the pure stress phase difference by combining the measured total phase difference, the obtained initial phase boundary condition value, and the geometric divergence phase value calculated based on the local relative curvature radius. The relative rate of change of sound velocity is calculated based on the pure stress phase difference, and the surface stress value is obtained by combining the pre-calibrated acoustoelastic coefficient.

3. The automatic analysis method according to claim 2, characterized in that, The specific steps for extracting the low-frequency interference voltage signal through the low-frequency channel and calculating the local relative radius of curvature at each target point on the continuous scanning trajectory are as follows: Phase demodulation and phase expansion of the low-frequency interference voltage signal yield a spatial bias signal that reflects the changes in the macroscopic contour of the surface. The spatial bias signal is subjected to spatial low-pass filtering to remove high-frequency spatial phase noise; The second spatial directional derivative of the filtered spatial bias signal along the preset propagation direction is calculated, and the reciprocal of the second spatial directional derivative is used as the equivalent local relative radius of curvature.

4. The automatic analysis method according to claim 2, characterized in that, The specific steps for extracting the local light intensity amplitude of each dwelling point of the microarray to generate the corresponding speckle suppression weighting coefficients are as follows: The multi-axis automatic scanning mechanism is aligned with the target point to perform a cross-shaped discrete point micro-array scan; The average value, effective value, or peak-to-peak value of the low-frequency interference voltage signal at each dwell point within a preset time window is calculated as the local light intensity amplitude; Normalization calculation is performed by dividing the local light intensity amplitude of a single dwelling point by the sum of the local light intensity amplitudes of all dwelling points in the same microarray, and a higher synthesis weight is assigned to the corresponding dwelling point to obtain the speckle suppression weight coefficient.

5. The automatic analysis method according to claim 2, characterized in that, The specific steps for generating the phase diagram matrix by using the local relative radius of curvature as a feedforward control parameter are as follows: Construct a spatial phase analytical model superimposed with periodic grating phase function terms and curvature compensation terms; The curvature compensation equivalent focal length is obtained by multiplying the local relative curvature radius of the feedforward input by the optical magnification of the lens group, and the curvature compensation equivalent focal length is substituted into the curvature compensation term. The theoretical phase deflection at each coordinate point on the pixel target surface of the spatial light modulator is calculated. The theoretical phase deflection is then linearly quantized and phase-folded, and mapped to the grayscale level supported by the spatial light modulator to generate the two-dimensional phase diagram matrix.

6. The automatic analysis method according to claim 2, characterized in that, The step of driving the spatial light modulator to load the phase diagram matrix to excite broadband ultrasonic waves on the surface of the mechanical transmission component also includes a protective step performed in conjunction with the above: The pulsed laser beam emitted from the pulsed laser is controlled, and the cross-sectional area of ​​the emitted beam is increased by the lens combination inside the pre-beam expander. The peak energy density of the pulsed laser incident on the liquid crystal target surface of the spatial light modulator is reduced to below the nominal damage threshold of the spatial light modulator before entering the spatial light modulator for wavefront deflection.

7. The automatic analysis method according to claim 2, characterized in that, The specific steps for spatially weighting and synthesizing the high-frequency time-domain signal by calling the speckle suppression weighting coefficient to obtain the synthesized ultrasound time-domain signal are as follows: Using the high-frequency time-domain signal of the central target point of the microarray as the timing reference, the time delay of other stationary point signals relative to the timing reference is calculated by the cross-correlation algorithm. Based on the time delay, perform a data sequence index shift operation or a fractional delay compensation operation on the high-frequency time-domain signals of the adjacent dwelling points to complete the time-domain phase alignment of the multi-point detection signals in the microarray. The aligned high-frequency time-domain signal is multiplied by the corresponding speckle suppression weight coefficient, and the product is then spatially weighted in the digital domain to synthesize the synthesized ultrasonic time-domain signal with a high signal-to-noise ratio.

8. The automatic analysis method according to claim 2, characterized in that, The specific steps for calculating the pure stress phase difference by calling the joint decoupling equation to use the measured total phase difference, the obtained initial phase boundary condition value, and the geometric divergence phase value calculated based on the local relative curvature radius are as follows: The synthetic ultrasonic time-domain signals of adjacent target points are recombined into a two-dimensional spatiotemporal data matrix and a two-dimensional discrete Fourier transform is performed. The center frequency is located in the two-dimensional spectrum diagram in the frequency-wavenumber domain and the measured wavenumber is extracted. The measured total phase difference is obtained by multiplying the difference between the measured wavenumber and the pre-stored stress-free reference wavenumber by the ultrasonic surface wave propagation distance parameter. The dimensionless geometric curvature factor is obtained by dividing the ultrasonic surface wave propagation distance parameter by the local relative radius of curvature, and the dimensionless geometric curvature factor is multiplied by the theoretical basis phase under the stress-free flat surface state to obtain the geometric divergence phase value. The initial phase boundary condition values ​​are calculated by analyzing the phase diagram matrix and combining it with the pre-established calibration relationship. By adding the initial phase boundary condition value to the measured total phase difference and subtracting the geometric divergence phase value using the joint decoupling equation, the pure stress phase difference, which eliminates the influence of macroscopic deformation and optical field pre-distortion, is separated.

9. The automatic analysis method according to claim 2, characterized in that, The specific steps for calculating the relative rate of change of sound speed based on the pure stress phase difference are as follows: Divide the calculated pure stress phase difference by the product of the preset stress-free reference wavenumber and the preset ultrasonic surface wave propagation distance parameter, and take the opposite number to obtain the dimensionless relative sound velocity change rate. Then, by utilizing the quantization conversion of acoustic characteristic quantities to mechanical parameters, the relative sound velocity change rate is divided by the pre-calibrated acoustoelastic coefficient to calculate the surface stress value.

10. The automatic analysis method according to claim 2, characterized in that, After calculating the surface stress value using the pre-calibrated acoustoelastic coefficient, a data completeness verification step is also performed: When the surface stress value of any target point exceeds the preset physical calibration range and the signal quality does not meet the preset signal-to-noise ratio condition, the current test point is determined to be an abnormal data point and forced removal is performed. Then, a spatial pixel neighborhood window is constructed with the abnormal data point as the center, and the arithmetic mean of the surrounding valid data points is calculated to perform neighborhood numerical interpolation and reconstruction to generate a residual stress spatial distribution map.

Images