An adaptive thermally excited infrared thermal wave nondestructive testing method and equipment

Abstract

This invention discloses an adaptive thermally excited infrared thermal wave nondestructive testing method and device. The method includes: querying a pre-stored database based on the scanned position signal to obtain the theoretical standard energy distribution spectrum under the current pose; triggering pre-aiming sensing; acquiring and generating the actual pre-aiming energy distribution spectrum; comparing the two to calculate the actual energy non-uniformity, and comparing it with a preset threshold to decide whether to initiate compensation; driving the actuator to perform physical adjustments and confirming the adjustments; and finally synchronously triggering the main thermal excitation and infrared thermal image acquisition. The device includes a scanning motion and pose feedback module, a high-precision multispectral pre-aiming sensing module, an intelligent decision-making and synchronous control module, a finely adjustable main excitation execution module, and a data management and communication interface module. This invention can proactively correct the uneven excitation energy distribution caused by scanning geometric changes from the physical source, significantly improving the quality and reliability of the detected image.

Description

Technical Field

[0001] This invention relates to the field of technology, specifically to an adaptive thermally excited infrared thermal wave nondestructive testing method and equipment. Background Technology

[0002] Infrared thermal nondestructive testing (IDT) is a method that detects internal defects by actively heating a workpiece and observing changes in its surface temperature field. Its accuracy and reliability are highly dependent on the uniformity of the thermal excitation applied to the workpiece surface. For small workpieces, good excitation uniformity can be achieved through a fixed excitation source and precise optical design.

[0003] However, when dealing with large-sized plate-shaped or curved workpieces commonly found in aerospace, wind power, and other fields, a "moving scan, zoned inspection" working mode must be adopted. In this mode, the position and angle of the excitation source integrated on the scanning head relative to the workpiece surface are constantly changing, leading to a fundamental technical challenge: due to changes in scanning geometry, the distribution of excitation energy on the workpiece surface undergoes a systematic, predictable, but non-uniform change with position. Specifically, the heat flux density received is higher in the central region of the workpiece or in the region directly facing the excitation source; while in the edge regions or regions with larger tilt angles, the received heat flux density is significantly attenuated. This excitation non-uniformity determined by the scanning pose is not random noise, but a fixed physical law. It directly results in different temperature change baselines in the thermal image sequences acquired from different sub-detection regions. During defect signal extraction and image stitching, this non-uniform base evolves into severe background gradients or striped artifacts, not only masking weak defect signals but also potentially generating false defect indications, greatly reducing the signal-to-noise ratio, quantitative evaluation capability, and overall reliability of the detection.

[0004] To address the aforementioned problems, existing technical solutions mainly employ two approaches: one is to use a fixed excitation source with an ultra-large area, and the other is to test... Figure one One approach is to cover the entire workpiece at once, but this would result in a huge waste of energy, bulky equipment, and unsatisfactory edge uniformity. The second approach is to try complex digital algorithms for background subtraction and correction in post-processing image processing. This is a "post-processing remedy" that is not only computationally complex but also difficult to fully restore the true physical state, which may introduce secondary errors and damage the original objectivity of the data.

[0005] Therefore, in view of the above problems, this application proposes an adaptive thermally excited infrared thermal wave nondestructive testing method and equipment. Summary of the Invention

[0006] The purpose of this invention is to provide an adaptive thermally excited infrared thermal wave nondestructive testing method and device to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: An adaptive thermally excited infrared thermal wave nondestructive testing method includes the following steps: S1: Determine the current scanning pose based on the scanning position signal, and query the pre-stored pose-standard energy distribution mapping database according to the current scanning pose to obtain the corresponding theoretical standard energy distribution map. S2: Based on the current scanning pose, trigger the pre-aiming sensing operation, synchronously collect the light signal reflected by the workpiece surface, and generate the actual pre-aiming energy distribution map; S3: Compare the actual target energy distribution map with the theoretical standard energy distribution map, calculate the actual energy non-uniformity, and compare it with a preset threshold to decide whether to initiate compensation: if no compensation is needed, proceed to S6; if compensation is needed, proceed to S4. S4: Calculate the energy deviation map based on the actual pre-aiming energy distribution map and the theoretical standard energy distribution map, perform spatial feature pattern recognition on the energy deviation map to determine the dominant non-uniformity mode, and generate a corresponding physical compensation instruction set according to the dominant non-uniformity mode; S5: Drive the actuator to perform physical adjustments based on the physical compensation instruction set, and generate an execution completion confirmation signal after confirming that the adjustments are completed; S6: Upon receiving a determination that no compensation is required or an execution completion confirmation signal, the main thermal excitation is triggered synchronously and the infrared thermal imager is controlled to acquire a thermal image sequence. At the same time, the detection process data is associated with and stored with the thermal image sequence.

[0008] As a further aspect of the present invention: In step S1, the pose-standard energy distribution mapping database is established by the following offline calibration method: an ideal diffuse reflection standard plate is placed at the detection station, and the scanning device is controlled to carry the thermal excitation module to traverse the grid points arranged in the work space according to the preset density. At each grid point, the precise pose of the scanning head is recorded, and the raw data of the light intensity reflected from the standard plate is acquired using the pre-aiming sensing module under that pose. The collected raw data is normalized to eliminate the inhomogeneity of the guide light source itself; the mapping relationship between all pose points and the corresponding normalized energy distribution maps is stored to form the pose-standard energy distribution mapping database. As a further aspect of the present invention: step S2 includes: triggering a millisecond-level fast pre-aiming operation by the intelligent decision and synchronization control module; controlling the low-power guiding light source to emit a weak light pulse with a duration of T1 according to the strict synchronization timing generated by the field programmable gate array, and simultaneously controlling the high-sensitivity photoelectric sensor array to start synchronous integration; After the light pulse ends, the analog-to-digital conversion value of the sensor array is read, and a two-dimensional actual pre-aiming energy distribution map is generated based on its physical arrangement mapping relationship; The guiding light source is an LED array, the center wavelength of which is similar to the effective thermal excitation band of the main thermal excitation source.

[0009] As a further aspect of the present invention: step S3 includes: performing data preprocessing on the actual pre-aimed energy distribution map to obtain a corrected actual energy distribution map; Calculate the average signal intensity of all pixels in the corrected actual energy distribution map. m and standard deviation s ; Based on formula = ( s / m ) * 100%, calculate the actual energy non-uniformity. ; Will With the preset threshold of uniformity Compare the results and determine the next process branch based on the comparison.

[0010] As a further aspect of the present invention: in step S4, spatial feature pattern recognition is performed on the energy deviation spectrum, and the determined dominant non-uniformity mode includes at least one of the following: The overall gradient pattern is characterized by a monotonic gradient change in energy deviation in space; Local high-frequency modes are characterized by isolated bright spots, dark spots, or regular stripes in energy deviations. The overall translation mode is characterized by an energy deviation that is close to a non-zero constant value.

[0011] As a further aspect of the present invention: the dominant non-uniformity mode generates a corresponding physical compensation instruction set, specifically as follows: When the dominant non-uniformity mode is the overall gradient mode, an instruction set is generated for performing reflector deformation compensation. The reflector deformation compensation controls the extension and retraction of multiple independent piezoelectric actuators on the back of the reflector to produce controllable elastic micro-deformation of the local curved surface of the reflector, thereby changing the distribution of light reflection angle. When the dominant non-uniform mode is a local high-frequency mode, an instruction set is generated for performing micro-translation compensation of the lamp array. The micro-translation compensation of the lamp array changes the relative spatial position between multiple lamps by driving each excitation lamp or lamp group to perform independent micron-level translation, thereby adjusting the interference superposition mode of light waves on the workpiece surface. When the dominant non-uniform mode is the overall translation mode, an instruction set is generated for performing independent energy fine-tuning. The independent energy fine-tuning directly controls the total energy released by each unit by independently adjusting the energy storage capacitor capacity or charging voltage of each thermal excitation unit.

[0012] As a further aspect of the present invention: the "action-effect" physical model upon which the reflector deformation compensation, lamp array micro-translation compensation, and independent energy fine adjustment are based is established by the following method: For the reflector deformation model, an optical model including the reflector, thermal excitation light source and workpiece surface is established using optical simulation software. Through parametric simulation and data fitting, the relationship between the displacement of the piezoelectric actuator and the change in the illuminance gradient of the workpiece surface is established. For the lamp array micro-translation model and the independent energy fine adjustment model, through experimental calibration, the input quantity is systematically changed and the change in energy distribution on the workpiece surface is measured. Based on the "input-output" data, a lookup table or fitting model is established.

[0013] A further aspect of the present invention: In step S4, when the analysis of the energy deviation spectrum indicates the existence of a non-uniform mixing mode, a multi-objective collaborative optimization method is used to solve for the compensation command, the method comprising: Preliminary compensation instructions are calculated for each of the identified non-uniformity modes. Under a unified optimization framework, considering the coupling effects between the actuators, a set of comprehensive compensation instructions is solved to simultaneously improve the predicted energy distribution after compensation for multiple abnormal modes.

[0014] As a further aspect of the present invention: in step S5, the condition for confirming the completion of the adjustment is that the deviation between the actual state feedback value of all actuators and the target value in the physical compensation instruction set is less than their respective preset allowable error tolerance.

[0015] As a further aspect of the present invention: an adaptive thermally excited infrared thermal wave nondestructive testing device, used to implement an adaptive thermally excited infrared thermal wave nondestructive testing method, comprising: The scanning motion and pose feedback module is used to provide scanning positioning signals and accurate spatial pose data; A high-precision multispectral pre-aiming sensing module is used to perform pre-aiming sensing operations and generate actual pre-aiming energy distribution maps; The intelligent decision-making and synchronization control module is used to perform data query, calculation, decision-making, instruction generation, and synchronization trigger control. The fine-tunable main excitation execution module includes at least a fine-tunable mechanism for performing physical compensation, a main thermal excitation source, and a state feedback sensor; The data management and communication interface module is used for data transmission and associated storage.

[0016] Compared with existing technologies, the beneficial effects of this invention are as follows: By constructing a "perception-decision-execution" closed loop, this invention utilizes pre-aiming perception to acquire the actual energy distribution in real time after each scan positioning and before formal excitation, and performs intelligent comparison and pattern recognition with theoretical benchmarks. This drives a precision mechanism to actively compensate for the thermal excitation source, fundamentally eliminating the systematic energy distribution gradient or anomaly introduced by pose changes in large-size scanning detection, thus ensuring excitation uniformity from the source. This technical solution avoids the limitations and secondary errors of traditional post-processing digital correction algorithms, significantly improving the background uniformity of infrared thermal image sequences, defect signal-to-noise ratio, comparability of data from different regions, and the quantitative evaluation capability and reliability of overall detection. Attached Figure Description

[0017] Figure 1 This is a flowchart of an adaptive thermally excited infrared thermal wave nondestructive testing method.

[0018] Figure 2 This is an architecture diagram of an adaptive thermally excited infrared thermal wave non-destructive testing device. Detailed Implementation

[0019] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0020] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0021] Please see Figure 1-Figure 2 An adaptive thermally excited infrared thermal wave nondestructive testing method, comprising: S1: Determine the current scanning pose based on the scanning position signal, and query the pre-stored pose-standard energy distribution mapping database according to the current scanning pose to obtain the corresponding theoretical standard energy distribution map.

[0022] The scanning motion and pose feedback module utilizes high-precision displacement sensors and angle encoders to monitor the spatial pose of the scanning head. Once the scanning head reaches the predetermined detection position and stabilizes, a "positioning complete" signal is acquired. Based on this "positioning complete" signal, the precise spatial coordinates of the scanning head in the workpiece coordinate system are captured and locked in real-time from the motion control bus. X, Y, Z ) and attitude angle Determine the current scanning pose Based on the scanning pose The system queries a "pose-standard energy distribution" mapping database pre-stored in the device's non-volatile memory. This mapping database, established through offline calibration, records the theoretically expected standard energy distribution pattern when the device thermally excites an ideal, uniform, highly reflective surface at various discrete standard positions in space. Through querying and interpolation calculations, the system determines the pose relative to the current continuous scanning position. The only corresponding theoretical standard energy distribution map The theoretical standard energy distribution map is described. This will serve as the benchmark for all subsequent analyses and comparisons.

[0023] Among them, the device's non-volatile memory is a hardware chip or storage unit integrated in the intelligent decision-making and synchronization control module for long-term data storage, such as eMMC chip, SSD solid-state drive or FLASH memory. It is used to permanently save key data such as "pose-standard energy distribution" mapping database and system calibration parameters, so that the data will not be lost even if the device is powered off.

[0024] It should be noted that the method for establishing the "pose-standard energy distribution" mapping database is as follows: Before the system is put into use, an ideal diffuse reflection standard plate is placed at the detection station, and the scanning equipment carrying the thermal excitation module is controlled to traverse a series of dense grid points in the workspace; at each grid point, its precise pose is recorded, and under this pose, a high-precision multispectral pre-aiming sensing module is used to collect the light intensity distribution data reflected by the standard plate; the data collected from all pose points are normalized to eliminate the non-uniformity of the guiding light source itself, and a correspondence table between pose coordinates and standardized energy distribution map is established and stored as a mapping database.

[0025] It is important to note that the grid density is determined based on the size of the workspace and the desired interpolation accuracy. Typically, within the effective working volume of the scanning head, the grid points are arranged according to the principle that the spatial step size is no greater than 1 / 3 of the characteristic size of the main thermal excitation spot.

[0026] As one possible embodiment, if the spot feature size is 30mm, then the grid point spacing is no greater than 10mm. For scanning pose... If it does not fall directly on a grid point, a combination of bilinear interpolation for position coordinates and spherical linear interpolation for attitude angles is used to calculate the corresponding value from the theoretical standard energy distribution map stored at neighboring grid points. Corresponding continuous theoretical standard energy distribution map .

[0027] The specific process of normalizing the data collected from all pose points is as follows: During the offline calibration phase, at each grid point pose, a high-precision multispectral pre-aiming sensing module is used to acquire the raw light intensity data matrix reflected by the ideal diffuse reflection standard plate. Under the same light source driving parameters, the standard plate was replaced with a reference plate with a known reflectivity close to 100% and absolutely uniform spatial reflection characteristics. The original light intensity distribution matrix of the light source itself was acquired under the same pose. Normalization formula: ; in,( i, j ) represents the pixel coordinates of the sensor array. C This is a global calibration constant used to map the results to a standardized energy intensity range; that is, through this operation, This eliminates the inherent spatial intensity non-uniformity of the guide light source and the vignetting effect of the optical system, reflecting the theoretical light intensity distribution that an ideal uniform surface should exhibit under the current geometric pose. (All pose points) After denoising and smoothing, the matrix is ​​stored as a theoretical standard energy distribution map. Together, they constitute a "pose-standard energy distribution" mapping database.

[0028] S2: Based on the current scanning pose, trigger the pre-aiming sensing operation, synchronously collect the light signal reflected by the workpiece surface, and generate the actual pre-aiming energy distribution map.

[0029] Based on the current scanning pose determined in step S1 And found the theoretical standard energy distribution map The intelligent decision-making and synchronization control module determines and triggers a millisecond-level rapid pre-aiming operation.

[0030] It should be noted that the intelligent decision-making and synchronization control module includes a central controller, a synchronization timing module, etc. The central controller is set as a high-performance microprocessor, which is responsible for running the upper-level decision-making, calculation, and management algorithms. The synchronization timing module is a field-programmable gate array (FPGA) or a dedicated timer / counter circuit, which is responsible for generating hardware timing signals with nanosecond or microsecond precision. It is used to strictly synchronize the two actions of lighting up the light source and starting / ending the sensor integration, ensuring the spatiotemporal consistency of the acquired data.

[0031] Specifically, based on a preset strict synchronization sequence, trigger commands are sent to the low-power guide light source driver and the high-sensitivity photoelectric sensor array. The guide light source is triggered, emitting a weak light pulse with a duration of T1 to simulate the spectral characteristics of the main thermal excitation. Simultaneously, the photoelectric sensor array is triggered, begins synchronous integration, and collects the light intensity signal reflected back from the current workpiece area. After the light pulse ends, the intelligent decision-making and synchronization control module reads the analog-to-digital conversion values ​​of all channels of the sensor array. Based on the synchronously acquired raw light intensity data and according to the physical arrangement mapping relationship of the photoelectric sensor array, a two-dimensional actual pre-aiming energy distribution map is determined. The actual pre-aiming energy distribution map A "preview snapshot" that directly reflects the distribution of thermal excitation energy that will be formed under the current real geometric conditions and workpiece surface characteristics.

[0032] It should be noted that the power guide light source driver is an electronic circuit that drives a low-power guide light source, such as an LED array. After receiving the trigger pulse from the FPGA, it can control the light source to emit light pulses of a specific width and intensity. The spectral characteristics of the main thermal excitation refer to the light emitted by the guide light source, whose wavelength range is similar to that of the main thermal excitation source, such as a xenon lamp, which produces an effective thermal excitation band. This ensures that the optical response characteristics of the pre-aiming perception can truly reflect the effect of the subsequent main thermal excitation.

[0033] Among them, the guiding light source can be an LED array of a specific wavelength, and its center wavelength is selected to be similar to the main absorption band of the main thermal excitation source (pulsed xenon lamp) that generates effective thermal excitation on the surface of the material under test. As a possible embodiment, for most polymer composites and coatings, the effective thermal excitation band for thermal wave detection is mostly in the near-infrared (0.8-1.5 μm) to short-wave infrared (1.5-2.5 μm) range. The guiding light source can be an LED with a center wavelength of 0.95 μm or 1.55 μm, with a typical half-width at half-maximum (FWHM) of ±20 nm. The pulse duration T1 is set to 100-200 microseconds, etc., mainly to ensure that the photoelectric sensor array obtains a sufficient signal-to-noise ratio without saturation, that is, within the typical detection distance and workpiece surface reflectivity range, the relationship between the sensor output signal strength and the light source pulse width is measured.

[0034] The photoelectric sensor array includes photodiodes arranged in a matrix.

[0035] The raw light intensity data is the digital value output by the analog-to-digital converter after each pixel in the photoelectric sensor array converts the received light signal, without any processing. It directly corresponds to the light intensity reflected by each tiny area.

[0036] S3: Compare the actual target energy distribution map with the theoretical standard energy distribution map, calculate the actual energy non-uniformity, and compare it with a preset threshold to decide whether to start compensation: if no compensation is needed, jump to S6; if compensation is needed, proceed to S4.

[0037] Based on the actual pre-aiming energy distribution map determined in step S2 Data preprocessing was performed to eliminate hardware system errors: from the actual target energy distribution map The pre-stored sensor dark noise background matrix is ​​subtracted, and gain non-uniformity correction is performed on each photoelectric sensor channel to determine the corrected actual energy distribution spectrum. Based on the corrected actual energy distribution spectrum Calculate its global statistical characteristics, including the average signal strength of all pixels. m and standard deviation s .

[0038] Based on formula Calculate and determine the actual energy non-uniformity percentage at the current detection location. .based on This is compared with a preset threshold that characterizes the acceptable uniformity of the system. (For example, 8.0%) is compared. Based on the comparison results, the subsequent decision path is determined: when ≤ If the current thermal excitation energy distribution already meets the uniformity requirement, no physical compensation is needed. Then, a "zero compensation" instruction flag is generated, and the process jumps directly to step S6.

[0039] when > When it is determined that the current thermal excitation energy distribution is uneven, an adaptive compensation mechanism must be activated. Then, proceed to the next step, S4, for in-depth analysis and strategy formulation. and As key input data.

[0040] It should be noted that the preset threshold It can be configured according to the accuracy requirements of different detection tasks; For high-precision defect detection It can be set to a lower value, such as 5%; For rapid screening It can be set to a higher value, such as 10%-15%.

[0041] S4: Calculate the energy deviation map based on the actual pre-aiming energy distribution map and the theoretical standard energy distribution map, perform spatial feature pattern recognition on the energy deviation map to determine the dominant non-uniformity mode, and generate a corresponding physical compensation instruction set according to the dominant non-uniformity mode.

[0042] Based on the corrected actual energy distribution map input in step S3 and theoretical standard energy distribution map The relative energy deviation spectrum was determined by comparing and calculating element by element. DM ,in ， Characterizes the percentage deviation of the measured energy in each local region from the theoretical geometric reference.

[0043] Based on energy deviation map ΔM, Pattern recognition analysis is performed on its spatial characteristics, such as calculating the average gradient in the row / column directions, performing spatial filtering to extract high-frequency components, or calculating the overall offset. Based on the results of the pattern recognition analysis, the current energy-dominant inhomogeneity mode type is determined, which specifically includes the following: If the analysis results show DM If a clear, monotonous spatial gradient change is observed, it is identified as the dominant mode, known as the "overall gradient mode," which is mainly caused by scanning geometry, i.e., large incident angles.

[0044] If the analysis results show DM If isolated bright spots, dark spots, or regular stripes are present, the dominant mode is identified as a "local high-frequency mode," which may originate from local abnormal reflections on the workpiece surface or interference from the light source.

[0045] If the analysis results show DM If all elements are close to a non-zero constant, then the dominant mode is determined to be the "overall translation mode," indicating the existence of a systematic reflectivity or distance deviation.

[0046] After determining the dominant uneven pattern, the optimal compensation mechanism is selected based on the preset mapping rule base of "dominant pattern → compensation strategy".

[0047] The physical compensation set includes the following types: For the "overall gradient mode", "reflector deformation compensation" is selected as the compensation strategy; Specifically, the physical process of reflector deformation compensation is as follows: The actuator is the reflector dynamic deformation mechanism in the finely adjustable main excitation execution module. The reflector dynamic deformation mechanism consists of multiple (usually ≥3) independently controlled piezoelectric actuators connected flexibly to the main reflector. Each piezoelectric actuator generates precise micron-level expansion and contraction under the drive of an electrical signal. Simultaneously, the piezoelectric actuators are installed in a ring or array between the back of the reflector and the rigid base. By coordinating the expansion and contraction of each piezoelectric actuator, controllable elastic micro-deformation is generated on the local curved surface of the rigid reflector, thereby changing the local curvature of the reflector and thus altering the distribution of the light reflection angle.

[0048] When an overall gradient in energy distribution from left to right (or from top to bottom) is detected, the system calculates a set of piezoelectric actuator displacement commands to make the left side of the reflector appropriately concave and the right side appropriately convex (or vice versa), guiding more reflected light to the area with lower energy, thereby reversing the spatial gradient and achieving uniform energy distribution.

[0049] For the "local high-frequency mode", "lamp array micro-translation compensation" is selected as the compensation strategy.

[0050] Specifically, the physical process of micro-translation compensation for the lamp array is as follows: The actuator is the thermal excitation source pose fine-tuning mechanism in the finely adjustable main excitation execution module. The thermal excitation source pose fine-tuning mechanism is driven by a precision linear motor or piezoelectric motor.

[0051] Each excitation lamp (or lamp group) is mounted on an independent micro-motion platform. The platform can perform micron-level translational movements in a plane parallel to the lamp axis (XY plane), with positioning accuracy reaching the sub-micron level. By fine-tuning the relative spatial positions between multiple lamps, the interference superposition mode of the light waves they emit on the workpiece surface is changed.

[0052] When local bright spots, dark spots, or interference fringes are detected in the energy distribution, the system calculates the small translation amount required for each lamp. For example, to eliminate the central bright spot, the lamps on both sides may be instructed to move slightly towards the center, changing the constructive / destructive interference conditions of the superimposed light fields, thereby smoothing local high-frequency fluctuations and eliminating spots or fringes.

[0053] For the "overall translation mode", "independent energy fine adjustment" is selected as the compensation strategy.

[0054] Specifically, the physical process of independent energy fine-tuning compensation is as follows: The actuator is an independent energy fine-tuning control circuit in the finely adjustable main excitation execution module. The independent energy fine-tuning control circuit equips each thermal excitation unit with an independent programmable energy storage and discharge module. Through a fast electronic switching network, the capacity of the energy storage capacitor connected to each lamp circuit is dynamically adjusted, or the charging voltage is precisely set, directly controlling the total energy released by each thermal excitation unit.

[0055] When a systematic deviation from the expected energy level across the entire area is detected (overall too bright or too dim), the system calculates the required adjustment percentage for each energy channel proportionally. For example, if the overall energy is 5% lower, the charging voltage or capacitor value of all lamps is increased proportionally to raise the overall thermal excitation intensity and compensate for the system deviation introduced by changes in distance or differences in overall reflectivity.

[0056] Based on the selected compensation strategy and the pre-stored "action-effect" physical model of the actuator, the least squares optimization algorithm is used to solve for the solution to eliminate or minimize the effect. DM The specific adjustments required for the implementing agencies; Preferably, the "action-effect" physical model includes a reflector deformation model, a lamp array micro-translation model, and an independent energy fine adjustment model; The specific method for establishing the above model is as follows: A precise optical model of the reflector deformation model, including the reflector surface, the thermal excitation light source, and the workpiece surface, is established using optical simulation software. In the simulation, the simulated displacements of each piezoelectric actuator on the back of the reflector are parametrically changed to calculate the change in illuminance distribution on the workpiece surface. Through a large number of simulation data points, a mapping function is fitted to correlate the actuator displacement vector with the illuminance gradient change vector on the workpiece surface.

[0057] The micro-translation model of the lamp array and the fine-tuning model of independent energy were established based on experimental calibration. Under the calibration fixture, the micro-translation amount of a single lamp was systematically changed, or the output of a certain energy channel was independently adjusted. The change in energy distribution on the workpiece surface was measured using a high-precision area array radiometer. The data pairs of input (translation / adjustment amount) and output (energy distribution change) were recorded, and a lookup table or a simple linear / nonlinear fitting model was established.

[0058] Specifically, this includes the target displacement of each piezoelectric actuator, the micro-translation of each lamp, and the adjustment percentage of each energy channel. Based on the calculated adjustment amounts and the physical constraints of each actuator, such as stroke limits and speed limits, feasibility verification and optimization are performed. The final feasible adjustment parameters are then converted into specific drive instructions that can be directly executed by the hardware, such as analog voltage values ​​and digital pulse counts, and packaged into a complete set of compensation control instructions. .

[0059] It should be noted that element-wise comparison refers to comparing two matrices with the same dimensions, such as a corrected actual energy distribution map. and theoretical standard energy distribution map In this process, two values ​​in the same row and column position undergo one-to-one mathematical operations, such as division and subtraction, to generate a new matrix of the same dimension, namely the energy deviation map. DM .

[0060] As a possible embodiment of the application, when the analysis determines that there is a mixed non-uniform mode, the compensation strategy is determined by a multi-objective collaborative optimization method, that is, the preliminary compensation instructions are calculated for the "overall gradient mode" and the "local high frequency mode" respectively, and then the coupling effect between the various actuators is considered. Under a unified optimization framework, a set of comprehensive compensation instructions are solved so that the predicted energy distribution after compensation simultaneously satisfies the objectives of overall gradient smoothness and local anomaly smoothness.

[0061] S5: Drives the actuator to perform physical adjustments based on the physical compensation instruction set, and generates an execution completion confirmation signal after confirming that the adjustment is completed.

[0062] Based on the compensation control instruction set generated in step S4 The intelligent decision-making and synchronous control module sends instructions to the precision actuators integrated on the excitation device through the corresponding hardware interface; Specifically, multiple piezoelectric actuators on the back of the drive reflector generate a specified displacement, drive the lamp tube micro-motion platform to move to a specified position, and adjust the output parameters of the independent energy control circuit.

[0063] Hardware interfaces include: analog output cards, digital I / O cards, dedicated motion control buses, etc.

[0064] While each actuator is in motion, the actual state vector of each actuator is determined based on the physical state data read in real time by the integrated state feedback sensors integrated within these actuators. The intelligent decision-making and synchronous control module continuously... and The target value of the instruction is compared in real time. When the deviation between the actual state and the target state of all actuators is less than their respective preset allowable error tolerance ε (e.g., displacement error < 0.1 micrometers), the physical compensation adjustment is determined to be completed, and a high-level "execution completion confirmation signal" is generated. The execution completion confirmation signal This is a necessary condition to allow subsequent main thermal excitation to be triggered.

[0065] Preferably, the integrated state feedback sensor includes a strain gauge built into the piezoelectric actuator, a grating ruler integrated into the micro-motion platform, and a voltage / current monitoring module for the energy circuit.

[0066] S6: Upon receiving a determination that no compensation is required or an execution completion confirmation signal, the main thermal excitation is triggered synchronously and the infrared thermal imager is controlled to acquire a thermal image sequence. At the same time, the detection process data is associated with and stored with the thermal image sequence.

[0067] The intelligent decision-making and synchronization control module continuously monitors two key input signals: one is the "no adjustment needed, jump to S6" judgment flag from step S3; the other is the execution completion confirmation signal from step S5. Once any of the above valid signals is detected, it is determined that the timing for synchronous triggering of the main thermal excitation is ripe.

[0068] The intelligent decision-making and synchronization control module initiates a very short fixed delay. This is used to ensure that any mechanical micro-motions have been stabilized. During the delay... After completion, based on the precise time reference at this moment, the synchronization timing generator synchronously outputs two trigger pulses with a strict time alignment relationship: one pulse is sent to the trigger terminal of the high-energy pulse xenon lamp power supply to detonate and release the main thermal excitation flash used for detection; the other pulse is sent to the external trigger input port of the infrared thermal imager to command it to start acquiring the infrared thermal image sequence of the workpiece surface.

[0069] Meanwhile, all key data objects generated, processed, and latched throughout the entire process from steps S1 to S5, including but not limited to: current pose Actual target energy distribution map Corrected actual energy distribution spectrum Theoretical standard energy distribution map Actual energy non-uniformity percentage Energy deviation spectrum DM Compensation control instruction set Actual state vector In addition to precise timestamps, the intelligent decision-making and synchronization control module encapsulates and defines this data into a complete detection process metadata data package containing rich contextual information. Subsequently, the data management and communication interface module will... The thermal image data stream acquired and returned by the infrared thermal imager is logically associated and bound for storage through global serial numbers or high-precision timestamps, thereby establishing a traceable and complete data archive for each detection.

[0070] The detection process metadata generated in step S6 Its data can be used for offline analysis to optimize system parameters, for example, by statistically analyzing a large number of detection points. The value can be used to evaluate and adjust the preset threshold for uniformity. By analyzing commonly used compensation command patterns under different poses, the control model of the actuator can be optimized, and the effectiveness of the entire adaptive homogenization decision chain can be verified and optimized by associating metadata with the quality of the final detected image.

[0071] In other words, by constructing and executing the aforementioned rigorous data-driven decision-making chain based on a closed loop of "perception-decision-execution," proactive correction of uneven thermal excitation energy distribution is achieved from the physical source. Based on the determination of the current precise pose, a theoretical energy distribution benchmark is obtained; then, the actual energy distribution is determined by pre-aiming perception; by comparing the two, the non-uniformity is determined and a decision is made on whether to compensate; when compensation is required, the optimal compensation strategy and instructions are determined based on intelligent pattern analysis of the deviation spectrum; by driving the precision actuator and confirming its state, precise adjustments to the physical world are achieved; finally, after confirming that homogenization has been achieved, effective main thermal excitation and data acquisition are triggered simultaneously, and metadata from the entire process is associated.

[0072] This chain-like processing method ensures that the thermal excitation conditions are highly consistent and controllable at every site in large-size scanning detection, fundamentally eliminating systematic errors introduced by scanning geometry changes, and significantly improving the background uniformity, defect signal-to-noise ratio, and reliability of quantitative assessment of infrared thermal wave detection images.

[0073] Secondly, the present invention provides an adaptive thermally excited infrared thermal wave nondestructive testing device, applied to the aforementioned adaptive thermally excited infrared thermal wave nondestructive testing method, specifically including: The system includes a scanning motion and pose feedback module, a high-precision multispectral pre-aiming sensing module, an intelligent decision-making and synchronous control module, a fine-tunable main excitation execution module, and a data management and communication interface module.

[0074] The scanning motion and pose feedback module is integrated with the external scanning mechanism to provide the "positioning complete" signal and high-precision pose data stream in step S1. .

[0075] The external scanning mechanism includes robots, gantry cranes, etc. The high-precision multispectral aiming sensing module includes a low-power guiding light source and a two-dimensional photoelectric sensor array, used to execute step S2, completing the illumination and synchronous data acquisition of the target area within milliseconds, and generating an actual aiming energy distribution map. .

[0076] The intelligent decision-making and synchronization control module is the core computing and control unit of the device, typically employing an architecture that combines a Field-Programmable Gate Array (FPGA) with a high-performance microprocessor. This module is configured to execute the baseline query in step S1, data processing and decision branch judgment in step S3, deviation pattern analysis and compensation instruction calculation in step S4, and synchronization trigger logic and metadata generation in step S6. Internally, it stores the "pose-standard energy distribution" mapping database and a uniformity preset threshold. The compensation strategy mapping rule base and the physical model parameters of each execution agency.

[0077] The fine-tunable master stimulus execution module comprises three key sub-parts: Adjustable mechanism: includes at least one set of reflector dynamic deformation mechanism, one set of thermal excitation source pose fine adjustment mechanism, and independent energy fine control circuit; The dynamic deformation mechanism of the reflector consists of multiple (usually ≥3) independently controlled piezoelectric actuators flexibly connected to the main reflector, which can generate controllable micron-level elastic deformation of the local curved surface of the reflector under the drive of electrical signals.

[0078] The thermal excitation source pose fine-tuning mechanism is driven by a precision linear motor or piezoelectric motor, which can drive the excitation lamp tube or lamp group to perform two-dimensional or three-dimensional micron-level translation or rotation.

[0079] Each thermal excitation unit is equipped with an independent programmable energy storage and discharge module, featuring a precise energy control circuit. Through a fast electronic switching network, the capacity of the energy storage capacitor connected to each lamp circuit is dynamically adjusted, or the charging voltage is precisely set, directly controlling the total energy released by each thermal excitation unit.

[0080] Main thermal excitation source: usually a high-energy pulsed xenon lamp array or a high-power laser, used to respond to the trigger pulse issued by the intelligent decision and synchronization control module in the final step S6 and release the main thermal excitation flash required for detection.

[0081] Integrated status feedback sensors: Embedded in the fine-tunable mechanism, the strain sensor integrated in the piezoelectric actuator and the grating ruler or encoder integrated in the micro-motion platform are used to monitor the physical adjustment status in real time and provide feedback signals to the control module. This is to achieve closed-loop control and completion confirmation in step S5.

[0082] The data management and communication interface module is responsible for processing the process metadata data generated in step S6. The infrared thermal imager transmits the image data it collects to the host computer or cloud server and receives configuration parameters from the host computer.

[0083] Other features and advantages will be set forth in the following description, and the objects and other advantages of the invention are realized and obtained through the structures particularly pointed out in the description and the drawings.

[0084] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and non-volatile computer storage media are basically similar to the method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0085] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0086] The above description is merely one or more embodiments of this specification and is not intended to limit this specification. Various modifications and variations can be made to the one or more embodiments of this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of one or more embodiments of this specification should be included within the scope of the claims of this specification.

Claims

1. An adaptive thermally excited infrared thermal wave nondestructive testing method, characterized in that, Includes the following steps: S1: Determine the current scanning pose based on the scanning position signal, and query the pre-stored pose-standard energy distribution mapping database according to the current scanning pose to obtain the corresponding theoretical standard energy distribution map. S2: Based on the current scanning pose, trigger the pre-aiming sensing operation, synchronously collect the light signal reflected by the workpiece surface, and generate the actual pre-aiming energy distribution map; S3: Compare the actual target energy distribution map with the theoretical standard energy distribution map, calculate the actual energy non-uniformity, and compare it with a preset threshold to decide whether to start compensation: if no compensation is needed, proceed to S6; If compensation is required, proceed to S4; S4: Calculate the energy deviation map based on the actual pre-aiming energy distribution map and the theoretical standard energy distribution map, perform spatial feature pattern recognition on the energy deviation map to determine the dominant non-uniformity mode, and generate a corresponding physical compensation instruction set according to the dominant non-uniformity mode; S5: Drive the actuator to perform physical adjustments based on the physical compensation instruction set, and generate an execution completion confirmation signal after confirming that the adjustments are completed; S6: Upon receiving a determination that no compensation is required or an execution completion confirmation signal, the main thermal excitation is triggered synchronously and the infrared thermal imager is controlled to acquire a thermal image sequence. At the same time, the detection process data is associated with and stored with the thermal image sequence.

2. The adaptive thermally excited infrared thermal wave nondestructive testing method according to claim 1, characterized in that, In step S1, the pose-standard energy distribution mapping database is established by the following offline calibration method: an ideal diffuse reflection standard plate is placed at the detection station, and the scanning equipment is controlled to carry the thermal excitation module to traverse the grid points arranged according to the preset density in the work space. At each grid point, the precise pose of the scanning head is recorded, and the raw data of the light intensity reflected from the standard plate is acquired using the pre-aiming sensing module under that pose. The collected raw data is normalized to eliminate the non-uniformity of the guide light source itself; the mapping relationship between all pose points and the corresponding normalized energy distribution map is stored to form the pose-standard energy distribution mapping database.

3. The adaptive thermally excited infrared thermal wave nondestructive testing method according to claim 1, characterized in that, Step S2 includes: triggering a millisecond-level fast pre-aiming operation by the intelligent decision and synchronization control module; controlling the low-power guiding light source to emit a weak light pulse with a duration of T1 according to the strict synchronization timing generated by the field programmable gate array, and simultaneously controlling the high-sensitivity photoelectric sensor array to start synchronous integration. After the light pulse ends, the analog-to-digital conversion value of the sensor array is read, and a two-dimensional actual pre-aiming energy distribution map is generated based on its physical arrangement mapping relationship; The guiding light source is an LED array, the center wavelength of which is similar to the effective thermal excitation band of the main thermal excitation source.

4. The adaptive thermally excited infrared thermal wave nondestructive testing method according to claim 1, characterized in that, Step S3 includes: performing data preprocessing on the actual pre-aimed energy distribution map to obtain a corrected actual energy distribution map; Calculate the average signal intensity of all pixels in the corrected actual energy distribution map. μ and standard deviation σ ; Based on formula = ( σ / μ ) * 100%, calculate the actual energy non-uniformity. ; Will With the preset threshold of uniformity Compare the results and determine the next process branch based on the comparison.

5. The adaptive thermally excited infrared thermal wave nondestructive testing method according to claim 1, characterized in that, In step S4, spatial feature pattern recognition is performed on the energy deviation spectrum, and the determined dominant inhomogeneity mode includes at least one of the following: The overall gradient pattern is characterized by a monotonic gradient change in energy deviation in space; Local high-frequency modes are characterized by isolated bright spots, dark spots, or regular stripes in energy deviations. The overall translation mode is characterized by an energy deviation that is close to a non-zero constant value.

6. The adaptive thermally excited infrared thermal wave nondestructive testing method according to claim 5, characterized in that, The dominant non-uniformity mode generates a corresponding physical compensation instruction set, specifically: When the dominant non-uniformity mode is the overall gradient mode, an instruction set is generated for performing reflector deformation compensation. The reflector deformation compensation controls the extension and retraction of multiple independent piezoelectric actuators on the back of the reflector to produce controllable elastic micro-deformation of the local curved surface of the reflector, thereby changing the distribution of light reflection angle. When the dominant non-uniform mode is a local high-frequency mode, an instruction set is generated for performing micro-translation compensation of the lamp array. The micro-translation compensation of the lamp array changes the relative spatial position between multiple lamps by driving each excitation lamp or lamp group to perform independent micron-level translation, thereby adjusting the interference superposition mode of light waves on the workpiece surface. When the dominant non-uniform mode is the overall translation mode, an instruction set is generated for performing independent energy fine-tuning. The independent energy fine-tuning directly controls the total energy released by each unit by independently adjusting the energy storage capacitor capacity or charging voltage of each thermal excitation unit.

7. The adaptive thermally excited infrared thermal wave nondestructive testing method according to claim 6, characterized in that, The "action-effect" physical model upon which the reflector deformation compensation, lamp array micro-translation compensation, and independent energy fine adjustment are based is established by the following methods: For the reflector deformation model, an optical model including the reflector, thermal excitation light source and workpiece surface is established using optical simulation software. Through parametric simulation and data fitting, the relationship between the displacement of the piezoelectric actuator and the change in the illuminance gradient of the workpiece surface is established. For the lamp array micro-translation model and the independent energy fine adjustment model, through experimental calibration, the input quantity is systematically changed and the change in energy distribution on the workpiece surface is measured. Based on the "input-output" data, a lookup table or fitting model is established.

8. The adaptive thermally excited infrared thermal wave nondestructive testing method according to claim 1 or 5, characterized in that, In step S4, when the analysis of the energy deviation spectrum indicates the existence of a non-uniform mixing mode, a multi-objective collaborative optimization method is used to solve for the compensation command. The method includes: Preliminary compensation instructions are calculated for each of the identified non-uniformity modes. Under a unified optimization framework, considering the coupling effects between the actuators, a set of comprehensive compensation instructions is solved to simultaneously improve the predicted energy distribution after compensation for multiple abnormal modes.

9. The adaptive thermally excited infrared thermal wave nondestructive testing method according to claim 1, characterized in that, In step S5, the condition for confirming the completion of the adjustment is that the deviation between the actual state feedback value of all actuators and the target value in the physical compensation instruction set is less than their respective preset allowable error tolerance.

10. An adaptive thermally excited infrared thermal wave nondestructive testing device, used to implement the adaptive thermally excited infrared thermal wave nondestructive testing method according to any one of claims 1-9, characterized in that, include: The scanning motion and pose feedback module is used to provide scanning positioning signals and accurate spatial pose data; A high-precision multispectral pre-aiming sensing module is used to perform pre-aiming sensing operations and generate actual pre-aiming energy distribution maps; The intelligent decision-making and synchronization control module is used to perform data query, calculation, decision-making, instruction generation, and synchronization trigger control. The fine-tunable main excitation execution module includes at least a fine-tunable mechanism for performing physical compensation, a main thermal excitation source, and a state feedback sensor; The data management and communication interface module is used for data transmission and associated storage.

Images