Real-time monitoring method for laser cleaning of carbon fiber composite material

Abstract

The invention relates to the technical field of laser cleaning processing, and particularly discloses a real-time monitoring method for laser cleaning of a carbon fiber composite material. An optical coherence tomography (OCT) detection unit and a sensor module are innovatively adopted to form a real-time monitoring system, and according to the surface treatment requirement of the carbon fiber composite material, a temperature sensor in the sensor module is used for collecting a heat radiation signal of a laser irradiation area in real time, and the surface layer temperature of a base material during machining is monitored in real time; the distance between a machining head and a base material is dynamically tracked through a laser ranging module in the sensor module, the OCT detection unit obtains micron-sized surface topography characteristics through frequency domain interference signal analysis, and surface roughness parameters are synchronously calculated. The data of the three channels are fused through the edge calculation module and then fed back in real time through the displayer, and the temperature field, the roughness and the focal length are effectively monitored in the cleaning process. According to the method, damage to the base material due to factors such as laser ablation can be avoided, the operation time is effectively shortened, and the operation efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of laser cleaning processing, and particularly relates to a real-time monitoring method for laser cleaning carbon fiber composites. Background Technique

[0002] Due to its non-contact, high-precision, and environmentally friendly characteristics, laser cleaning technology has gradually replaced traditional mechanical or chemical cleaning processes in fields such as aerospace and automotive manufacturing. For high-value substrates such as carbon fiber composites (CFRP), laser cleaning needs to balance the dual requirements of efficiently removing contaminants (such as mold release agents and oxide layers) and avoiding thermal damage to the substrate. However, the following bottleneck problems still exist in the prior art: (1) Lag in roughness detection The prior art needs to evaluate the surface roughness (Ra value) through a contact probe (such as a profilometer) or an off-line optical device (such as a confocal microscope) after cleaning, resulting in a lag in process parameter optimization and a reduction in efficiency.

[0003] (2) Risk of thermal damage due to lack of real-time parameters Carbon fiber has poor thermal conductivity (~5 W / m·K), and the resin matrix has a low heat resistance threshold (usually <300°C). Existing composite laser cleaning lacks real-time monitoring of the substrate surface temperature and only relies on preset power parameters, which easily causes local overheating and leads to fiber breakage or resin carbonization.

[0004] (3) Focal length offset affects cleaning accuracy The carbon fiber surface often has micron-level undulations (such as weaving textures or damage depressions), and the offset of the laser focus position (>±50μm) will cause fluctuations in the energy density (the fluctuation amplitude reaches ±30%), resulting in insufficient cleaning or over-etching. Traditional devices use a fixed focal length or an off-line focusing mode and cannot adapt to the dynamically changing surface topography. Summary of the Invention

[0005] In order to make up for the deficiencies of the prior art, the present invention provides a collaborative system that combines an OCT detection head and a sensor module to synchronously obtain carbon fiber surface, temperature, and focal length data during the laser cleaning process, and solves the core problems of thermal damage, insufficient accuracy, and low efficiency in the prior art. A real-time monitoring method for laser cleaning carbon fiber composites.

[0006] The present invention is realized through the following technical solutions: A real-time monitoring method for laser cleaning carbon fiber composites, using carbon fiber composites with dirt on the surface as the material to be processed, including the following steps: Step S1: Place the material to be processed on a three-dimensional displacement platform. A laser cleaning head and an OCT optoelectronic balance detector are arranged above the three-dimensional displacement platform. Adjust the material to be processed to be located on the focal plane of the focusing lens in the laser cleaning head. Step S2: Control the sensor module and its control system to turn on and establish a data transmission connection with the computer; control the OCT optoelectronic balance detector and its control system to turn on and establish a data transmission connection with the computer; set the parameter ranges of the laser cleaning head, the sensor module, and the OCT optoelectronic balance detector. Step S3: The laser cleaning head, the scanning galvanometer of the sensor module, and the OCT optoelectronic balance detector achieve hardware-level synchronization through the FPGA in the control system. The laser pulse triggers the OCT acquisition window, and the laser pulse frequency is the same as the OCT acquisition frequency. Step S4: Obtain the image of the material to be processed by combining two-dimensional cross-sectional scanning and three-dimensional surface scanning. Step S5: The OCT optoelectronic balance detector collects the reference spectrum I ref (k) without a sample, and deducts the background noise in real time; calibrate the wavenumber k space through the fixed reflection peak of the reference mirror to eliminate the influence of spectrometer nonlinearity. Apply a Hanning window to I(k) to suppress the FFT sidelobes, collect data and process it. Step S6: Compare the above processed data with the specified parameter range and perform real-time control through the control system. Step S7: According to the set parameter range, emit laser to process the surface of the carbon fiber composite material. At the same time, the laser cleaning head and the OCT optoelectronic balance detector move synchronously to achieve the removal of surface dirt.

[0007] The present invention innovatively adopts an optical coherence tomography (OCT) detection unit and a sensor module to form a real-time monitoring system. Aiming at the surface treatment requirements of carbon fiber composite materials, the temperature sensor in the sensor module is used to collect the thermal radiation signal in the laser irradiation area in real time to monitor the temperature of the substrate surface during processing; the laser ranging module in the sensor module is used to dynamically track the distance between the processing head and the substrate. The OCT detection unit analyzes the frequency-domain interference signal to obtain micron-level surface topography features and synchronously calculates the surface roughness parameters. The above three-channel data are fused and processed by the edge computing module and then fed back in real time through the display to achieve effective monitoring of the temperature field, roughness, and focal length during the cleaning process.

[0008] A more optimal technical solution of the present invention is: In step S2, the scanning speed of the laser cleaning head is 3000 - 4000 mm / s, the cleaning speed is 0.0036 - 0.004 m / s, the laser connected to the laser cleaning head is an ultraviolet laser, its power range is 1 - 3 W, and the laser repetition frequency is 30 - 35 Hz. The sensor module includes a temperature sensor and a laser ranging module. Among them, the temperature range of the temperature sensor is 200 - 800 °C, the temperature detection accuracy is ±10 °C, the thermal radiation response time is ≤0.5 ms, and the dynamic range of the laser ranging module is 0 - 10 mm; The arithmetic mean roughness detection range of the OCT photoelectric balance detector is 0.1 - 10 μm, the detection range of the maximum height difference is 1 - 50 μm, the OCT axial resolution is ≤5 μm, and the lateral scan coverage is ≥10 mm.

[0009] In step S3, the laser pulse triggers the OCT acquisition window. The laser pulse frequency is the same as the OCT acquisition frequency, and the delay time is <0.1 μs.

[0010] In step S4, the two-dimensional cross-sectional scan uses the single-line scan and depth-direction scan methods to analyze the surface topography, and the three-dimensional surface scan uses the XY-plane grid scan for high-precision surface reconstruction.

[0011] In step S5, the OCT photoelectric balance detector uses a 2048-pixel linear InGaAs camera. Its spectral range is 1200 - 1400 nm, the resolution is Δλ≈0.1 nm, the data is transmitted to the FPGA through the Camera Link interface, and the single-frame acquisition time is 10 μs; Its specific data acquisition and processing include the steps: 1) Interference signal calibration: Calibrate the wavenumber k space through the fixed reflection peak of the reference mirror to eliminate the nonlinear response of the spectrometer; Collect the reference spectrum I ref (k) without a sample, and calculate the correction signal in real time ; Fit the nonlinear kk-space through the reflection peak position of the calibration mirror, and use cubic spline interpolation to correct: k linear =a0 + a1λ + a2λ 2 + a3λ 3 where λ is the spectrometer pixel index, and the coefficients a0 - a3 are obtained through calibration; Apply the Hanning window to I(k) to suppress the FFT sidelobes: where N = 2048; Calculate the DC signal mean value and subtract it: ; Correct the non-uniformity of the temperature data: where Toffset and G(x,y) are obtained through blackbody furnace calibration; Perform moving average filtering on the displacement data to suppress high-frequency vibration noise: ; 2) Depth profile A-scan reconstruction: Perform FFT on the 2048-point interference signal of each A-scan to generate a complex depth profile A(z), extract the amplitude information |A(z)|, and record the phase φ(z)=arg(A(z)): A(z)=Γ -1 {I DC-free (k)}, Search for the position Z of the maximum value of the surface reflection peak amplitude peak : ; 3) Surface topography reconstruction: A single B-scan line contains 512 A-scan points, which is reconstructed into a two-dimensional depth profile in the X-Z plane. Take the Z of each A-scan along the X direction peak , generate a one-dimensional height curve h(x); The XY galvanometer scans step by step to generate a three-dimensional point cloud matrix Z(x, y) for off-line high-precision roughness analysis.

[0012] Further preferably, the process of roughness calculation includes the steps: a) Denoise the obtained h(x) using Gaussian low-pass filtering with a cut-off wavelength λc = 50 μm to obtain , Then fit the reference plane h base (x)=ax + b using the least squares method for tilt removal, and calculate the residual height h corrected (x)=h filtered (x)-h base (x); b) Calculate the roughness parameters according to the formula from the obtained residual height to obtain the arithmetic mean roughness , Root mean square roughness , The maximum height difference Rz = max(h i ) - min(h i ).

[0013] Step S6 includes the steps: 1) Establish a roughness-power dynamic control model: Assume that the target arithmetic mean roughness is 3 μm, with an allowable deviation of ±0.5, set the proportionality coefficient K p = 0.2 W / μm, set the integral coefficient K i= 0.05 W / (μm·s), the set constraint conditions are that the lower limit of power limit is 1 W and the upper limit of power is 5 W, and then the real-time power is adjusted ; The real-time control process is to obtain the roughness Ra every 2 ms and perform deviation calculation: ΔRa = Ra - Ra target ; When |ΔRa| ≤ 0.5 μm, keep the current power, If ΔRa > 0.5 μm, increase the power according to the formula, If ΔRa < 0.5 μm, decrease the power according to the formula; Send a 16-bit PWM signal to control the laser through the RS422 interface, with a resolution of 0.1 W; 2) Set the temperature over-limit protection mechanism: Since the glass transition temperature Tmax of carbon fiber epoxy resin is 300 °C, set the dynamic threshold T threshold (x, y), 250 °C < T threshold (x, y) < 275 °C.

[0014] The infrared thermal imager in the temperature sensor outputs the temperature matrix T(x, y) at a frequency of 60 Hz, and set the emergency stop condition: ∃(x, y) s.t. T(x, y) > T threshold (x, y) + 10 °C; When the set temperature threshold is exceeded, the FPGA hardware emergency stop signal is triggered, the light source is immediately turned off, the galvanometer controls the laser focus to move out of the processing area, and the over-temperature position (x, y) and temperature value are recorded.

[0015] There is a safety redundancy designed in the parameter interval in step S7: 1) The FPGA hardware watchdog continuously monitors the heartbeat signal of the main control program, and if it times out, it automatically triggers the emergency stop signal; 2) Dual-channel temperature verification, the confocal displacement sensor synchronously measures the surface reflectivity R. If R drops suddenly and the temperature is not over-limit, a secondary protection is still triggered: the ablation determination condition is R < 0.3 and dR / dt < -0.1 / μs; 3) Power gradient ramp, any power adjustment goes through a ramp transition. In addition, the time constant τ = 0.2 ms to ensure that the 10 - 90% rise time < 1 ms: P(t) = P initial +(P target -P initial )·(1 - e -t / τ ).

[0016] The present invention also discloses a device for laser cleaning carbon fiber composites based on the above real-time monitoring method, which includes a laser control system, a sensor control system, an optical detection head control system, and a three-dimensional displacement platform connected to a computer; the laser control system is connected to a laser cleaning head, the sensor control system is connected to a sensor module, the optical detection head control system is connected to an OCT optoelectronic balance detector, the carbon fiber composites to be cleaned are placed on the three-dimensional displacement platform, and the positions of the laser cleaning head, the sensor module, and the OCT optoelectronic balance detector correspond to the carbon fiber composites.

[0017] Compared with the prior art, the method of the present invention has the following beneficial technical effects because a sensor module and an optical detection head are added to the device during the cleaning process: (1) Improve the cleaning quality: For carbon fiber composites, when the temperature is too high (T > 650 °C), the strength decreases by more than 30%. When the laser cleaning head is defocused during processing, bubbles will be generated on the surface layer of the carbon fiber composites due to plasma and impact effects. Due to the presence of the sensor module and the optical detection head, the surface temperature of the material to be processed and the real-time focal length of the laser cleaning head can be monitored in real time, improving the cleaning quality; (2) Improve the efficiency: Since a sensor module and an optical detection head are added to the cleaning device, the real-time focal length of the laser cleaning head, the surface temperature of the carbon fiber composites, and the roughness can be obtained in real time only by setting through a computer during the processing, without the need for a detection step after processing, improving the efficiency.

[0018] The present invention can monitor the laser cleaning process of carbon fiber composites in real time, avoid damage to the substrate due to factors such as laser ablation, and at the same time, due to the synchronous monitoring of roughness, the operation time can be effectively shortened and the operation efficiency can be improved. Brief Description of the Drawings

[0019] The present invention will be further described below with reference to the drawings.

[0020] Figure 1 It is a schematic diagram of the detection process of the method of the present invention; Figure 2 It is a schematic diagram of the structure of the device of the present invention.

[0021] In the figure, 1 is a computer, 2 is a laser control system, 3 is a sensor control system, 4 is an optical detection head control system, 5 is a sensor module, 6 is a laser cleaning head, 7 is an OCT optoelectronic balance detector, 8 is the material to be processed, and 9 is a three-dimensional displacement platform. Detailed Embodiments

[0022] To make the above objects, features, and advantages of the present invention more apparent and understandable, the following provides a detailed description of the specific embodiments of the present invention in conjunction with the accompanying drawings. A lot of specific details are set forth in the following description to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this invention belongs. The terms used in the description of the present invention herein are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0024] The following provides a detailed description of the present invention in conjunction with the accompanying drawings.

[0025] Embodiment 1: A real-time monitoring method for laser cleaning of carbon fiber composites Using carbon fiber composites with dirt on the surface as the material to be processed, the method includes the following steps: Step S1: Place the material to be processed on a three-dimensional displacement platform. Above the three-dimensional displacement platform, a laser cleaning head and an OCT optoelectronic balance detector are provided. Adjust the material to be processed to be located on the focal plane of the focusing lens in the laser cleaning head. Among them, the OCT optoelectronic balance detector consists of five parts, namely a light source module, an interferometer module, a detection module, a scanning module, and a control and communication module.

[0026] (1) Light source module: Provide a beam with a wide spectral range and generate an interference signal. A broadband light source is selected to provide a light source with a wide spectral range. An ultrashort pulse laser is used, and the near-infrared band (1300 nm) is selected to balance the penetration depth and resolution. (2) Interferometer module: Divide the light source into a reference light and a sample light and generate an interference signal. It consists of a reference arm, a sample arm, and a beam splitter. Reference arm: Includes a mirror and an adjustable delay line for adjusting the optical path of the reference light. Sample arm: Includes a focusing lens and a scanning galvanometer for focusing the beam on the sample surface and performing lateral scanning. Beam splitter: Divide the light source into a reference light and a sample light and recombine the reflected light to form an interference signal. (3) Detection module: Receive the interference signal and perform frequency-domain analysis to extract surface topography information. It consists of a spectrometer, a detector, and a signal processing unit. Spectrometer: Receives interference signals and decomposes them into spectral components of different wavelengths; Detector: Converts spectral signals into electrical signals; Signal processing unit: Performs a fast Fourier transform (FFT) on the signals output by the detector to extract surface topography information; (4) Scanning module: Achieves lateral scanning of the sample surface. Composed of a lateral scanning galvanometer and a focusing lens; Lateral scanning galvanometer: Achieves lateral scanning of the light beam on the sample surface by quickly swinging the reflecting mirror; Focusing lens: Focuses the light beam onto the sample surface to ensure high-resolution imaging; (5) Control and communication module: Controls each module of the OCT photodetector and enables communication with an external system. Composed of a control unit and a communication interface; Control unit: Controls the synchronous operation of the light source, scanning galvanometer, and detection module; Communication interface: Communicates with a computer or edge computing module to transmit detection data in real time.

[0027] Step S2: Control the sensor module and its control system to turn on and establish a data transmission connection with the computer; control the OCT photoelectric balance detector and its control system to turn on and establish a data transmission connection with the computer; set the parameter ranges of the laser cleaning head, sensor module, and OCT photoelectric balance detector. The specific steps are as follows: (1) Parameter setting of the laser cleaning head The scanning speed of the laser cleaning head is 3000 - 4000 mm / s, the cleaning speed is 0.0036 - 0.004 m / s, the laser connected to the laser cleaning head is an ultraviolet laser, its power range is 1 - 3 W, and the laser repetition frequency is 30 - 35 Hz; (2) Parameter setting of the sensor module The sensor module includes a temperature sensor and a laser ranging module.

[0028] The temperature sensor monitors the temperature on the surface of the carbon fiber composite material in real time through non-contact thermal radiation measurement technology. Its core is based on the Stefan - Boltzmann law, that is, the radiation power per unit area of an object is proportional to the fourth power of its absolute temperature: P = εσT 4 , where P is the radiation power, ε is the material emissivity, σ is the Stefan - Boltzmann constant, and T is the surface temperature (unit: K); The sensor infers the surface temperature by receiving the thermal radiation signal in the laser irradiation area and combining the preset emissivity of the carbon fiber composite material. The temperature sensor used in this embodiment is an infrared thermopile sensor.

[0029] The laser ranging module uses the triangulation method to achieve dynamic focal length tracking. Its core formula is: d = L·sinθ / sin(θ + ψ) Where d is the distance between the processing head and the substrate, L is the baseline distance between the emitter and the receiver, θ is the laser emission angle, and ψ is the reflected light reception angle (measured by the change in the position of the photoelectric sensor).

[0030] In this embodiment, the dynamic range of the laser ranging module is 0 - 10 mm, and the focus offset is compensated in real time in combination with the galvanometer scanning parameters. When defocus occurs due to surface undulation, the system adjusts the galvanometer deflection angle α through the following formula: Δz = f·tanα, Where Δz is the focus offset amount and f is the focal length of the focusing lens.

[0031] In this embodiment, the range of the temperature sensor is 200 - 800 °C, the temperature detection accuracy is ±10 °C, the thermal radiation response time is ≤0.5 ms, and the dynamic range of the laser ranging module is 0 - 10 mm; (3) OCT optoelectronic balance detector parameter setting This embodiment uses frequency-domain OCT (FD-OCT). Due to considering the plasma and laser wavelength interference during the laser cleaning process, the optical detection head and the laser cleaning head are not integrated in the device of this embodiment. The working principle of frequency-domain OCT can be roughly divided into three parts: a) Low-coherence interferometric measurement: The light source of the OCT optoelectronic balance detector selects a swept laser source with a central wavelength of 1310 nm and a bandwidth of 100 nm. Its short coherence length (Lc = λ0² / Δλ ≈ 17 μm) can achieve high axial resolution (Δz ≈ Lc / 2 ≈ 8.5 μm); b) Interference signal generation: The reflected lights of the reference arm and the sample arm interfere at the beam splitter, and the intensity of the interference signal is: , Where k = 2π / λ is the wave number, R R and R S are the reflectivities of the reference arm and the sample arm, and Δz is the optical path difference; c) Frequency-domain signal processing: The interference spectrum is obtained through a spectrometer or a swept light source, and the Fourier transform (FFT) is performed on I(k) to obtain the depth-resolved reflectivity distribution: , The surface topography is determined by detecting the position of the reflection peak.

[0032] In this embodiment, the arithmetic mean roughness detection range of the OCT optoelectronic balance detector is 0.1 - 10 μm, the detection range of the maximum height difference is 1 - 50 μm, the OCT axial resolution ≤ 5 μm, and the transverse scan coverage ≥ 10 mm Step S3: The laser cleaning head, the scanning galvanometer of the sensor module, and the OCT optoelectronic balance detector achieve hardware-level synchronization through the FPGA in the control system. The laser pulse triggers the OCT acquisition window, and the laser pulse frequency is the same as the OCT acquisition frequency, with a delay time < 0.1 μs.

[0033] Step S4: Obtain the image of the material to be processed by combining two-dimensional cross-sectional scanning and three-dimensional surface scanning; Among them, two-dimensional cross-sectional scanning uses single-line scanning and depth-direction scanning to analyze the surface topography, and three-dimensional surface scanning uses XY-plane grid scanning for high-precision surface reconstruction.

[0034] Step S5: The OCT optoelectronic balance detector collects the reference spectrum I ref (k) without a sample, and deducts the background noise in real time; calibrates the wavenumber k space through the fixed reflection peak of the reference mirror to eliminate the influence of spectrometer nonlinearity, applies a Hanning window to I(k) to suppress the FFT sidelobes, and collects and processes the data; Parameter setting of the data acquisition device: (1) Hardware synchronization and triggering Main clock source: The FPGA board provides a global clock (100 MHz); Trigger synchronization mechanism: The laser pulse (1 kHz) is used as the main trigger signal, and the FPGA generates a frequency division / delay signal; The OCT camera, galvanometer, thermal imager, and displacement sensor are all synchronized through PWM or TTL signals; Timing error compensation: The trigger delay of each sensor is corrected through a pre-calibrated table (LUT), with an accuracy < 1 μs; (2) Data acquisition OCT optoelectronic detector data: The laser cleaning pulse (1 kHz) generates a synchronous trigger signal through the FPGA to control the OCT light source, detector, and galvanometer. The detector acquisition window starts 10 μs after the end of the laser pulse to avoid plasma flash interference; The single-frame data format is 2048 pixel points (corresponding to wavenumber k spatial sampling), and each pixel value is the intensity I(k). The single-frame size is 4 KB; the real-time bandwidth is 200 MB / s; Temperature data: The thermal imager collects the temperature field during the laser pulse interval (900 μs) and transmits it to the computer via GigE (16 ms per frame); Displacement data: The confocal sensor continuously measures the distance and transmits ASCII format data via RS485 (baud rate 1 Mbps, 100 μs per point).

[0035] Thus, in this embodiment, the OCT optoelectronic balanced detector uses a linear array InGaAs camera with 2048 pixels. Its spectral range is 1200 - 1400 nm, the resolution Δλ ≈ 0.1 nm, and the data is transmitted to the FPGA via the Camera Link interface. The single-frame acquisition time is 10 μs; Specific data acquisition and processing include the steps: 1) Interference signal calibration Calibrate the wavenumber k space through the fixed reflection peak of the reference mirror to eliminate the non - linear response of the spectrometer; collect the reference spectrum I ref (k) without the sample and calculate the correction signal in real time ; Fit the non - linear kk - space through the reflection peak position of the calibration mirror and correct it using cubic spline interpolation: k linear = a0 + a1λ + a2λ 2 + a3λ 3 ,[[]] where λ is the pixel index of the spectrometer, and the coefficients a0 - a3 are obtained through calibration; Apply the Hanning window to I(k) to suppress the FFT sidelobes: , where N = 2048; Calculate the mean value of the DC signal and subtract it: ; Correct the non - uniformity of the temperature data: , where Toffset and G(x, y) are obtained through blackbody furnace calibration; Perform moving average filtering on the displacement data to suppress high - frequency vibration noise: ; 2) Depth profile A - scan reconstruction Perform FFT on the 2048 - point interference signals of each A - scan to generate the complex depth profile A(z), extract the amplitude information |A(z)|, and record the phase φ(z)=arg(A(z)): A(z)=Γ -1{I DC-free (k)} Search for the position Z of the maximum value of the surface reflection peak amplitude peak : ; 3) Surface topography reconstruction A single B-scan line contains 512 A-scan points, which are reconstructed into a two-dimensional depth profile in the X-Z plane. Take the Z of each A-scan along the X direction peak , generating a one-dimensional height curve h(x); The XY galvanometer scans step by step to generate a three-dimensional point cloud matrix Z(x, y) for offline high-precision roughness analysis. The newly added personalized design for carbon fiber composites is mainly reflected in the data processing for carbon fiber substrates, after extracting information from the profile A-scan reconstruction.

[0036] 4) Adopt the fiber texture feature separation algorithm - dual-scale Gaussian filtering Short-wavelength filtering (λc = 20μm): Use Gaussian low-pass filtering for denoising, with a cut-off wavelength of 20μm, to extract the surface high-frequency components (including fiber weaving texture and micro-defects): h filtered_short (x) = h(x) * G(x, σ short ) where G(x, σ short ) is a Gaussian kernel with a standard deviation of ; Long-wavelength filtering (λc = 50μm): Use another Gaussian low-pass filter with a cut-off wavelength of 50μm to separate the low-frequency components (mainly including macroscopic defects such as residual pollutants or matrix damage): h filtered_long (x) = h(x) * G(x, σ long ); Texture and defect separation: h texture (x) = h filtered_short (x); h defect (x) = h filtered_long (x) - h filtered_short (x); Through the differential operation, retain the defect features in the long-wavelength filtering result and remove the fiber texture interference at the same time.

[0037] The above roughness calculation process includes the steps: a) Denoise the obtained h(x) (at this time h(x) is h defect (x) after the above processing) using Gaussian low-pass filtering with a cut-off wavelength λc = 50μm to obtain , Then, the least squares method is used to fit the reference plane h base (x)=ax + b for detrending and calculate the residual height h corrected (x)=h filtered (x)-h base (x); b) Calculate the roughness parameters by combining the obtained residual height with the formula to obtain the arithmetic mean roughness , Root mean square roughness: , Maximum height difference: Rz = max(h i ) - min(h i ); Step S6: Compare the above - processed data with the specified parameter range and perform real - time control through the control system. The specific steps are as follows: 1) Establish a roughness - power dynamic control model a) Roughness feedback control logic Preset the target roughness range: Assume the target arithmetic mean roughness is 3μm (set according to requirements), and the allowable deviation is ±0.5 (i.e., Ra ∈ [2.5, 3.5]μm); Set the proportionality coefficient K p = 0.2W / μm (for every 1μm deviation from the target, the power is adjusted by 0.2W), and set the integral coefficient K i = 0.05W / (μm·s) (cumulative error compensation to prevent steady - state deviation); Set the constraint conditions that the lower limit of power is 1W and the upper limit of power is 5W, and then adjust the real - time power ; b) Control process The real - time control process is to obtain the roughness Ra every 2ms, Perform deviation calculation: ΔRa = Ra - Ra target ; Control decision: When |ΔRa| ≤ 0.5μm, keep the current power, If ΔRa > 0.5μm, increase the power according to the formula, If ΔRa < 0.5μm, decrease the power according to the formula; Instruction issuance: Send a 16 - bit PWM signal to control the laser through the RS422 interface, with a resolution of 0.1W; 2) Set the temperature over - limit protection mechanism a) Temperature safety threshold The glass transition temperature of the carbon fiber epoxy resin is Tmax = 300 °C, and the dynamic threshold is set as 250 °C < T threshold (x,y) < 275 °C; b) Protection trigger logic The infrared thermal imager in the temperature sensor outputs the temperature matrix T(x,y) at a frequency of 60 Hz, and the emergency stop condition is set as: ∃(x,y) s.t. T(x,y) > T threshold (x,y) + 10 °C; When the set temperature threshold is exceeded, the FPGA hardware emergency stop signal is triggered, the light source is immediately turned off, the galvanometer controls the laser focus to move out of the processing area, and the over-temperature position (x,y) and temperature value are recorded.

[0038] Step S7: According to the set parameter range, emit laser to process the surface of the carbon fiber composite material, and at the same time, the laser cleaning head and the OCT optoelectronic balance detector move synchronously to achieve the removal of surface dirt.

[0039] There is a safety redundancy designed in the parameter range: 1) Hardware watchdog: The FPGA continuously monitors the heartbeat signal of the main control program. If it times out (e.g., no response for > 1 ms), it automatically triggers the emergency stop signal (ESTOP); 2) Dual-channel temperature verification: The confocal displacement sensor synchronously measures the surface reflectivity R. If R drops suddenly (a sign of carbon fiber ablation) and the temperature is not exceeded, a secondary protection is still triggered: The ablation determination condition is R < 0.3 and dR / dt < -0.1 / μs; 3) Power gradient ramp: Any power adjustment goes through a ramp transition. In addition, the time constant τ = 0.2 ms to ensure that the 10 - 90% rise time < 1 ms: P(t) = P initial +(P target -P initial )·(1 - e -t / τ ).

[0040] The method disclosed in this embodiment is used to clean the natural fiber composite material, and the specific cleaning result parameters are as follows.

[0041] In the above experiment, in the five given experiments, the types of lasers, laser powers, laser repetition frequencies, galvanometer scanning speeds, and cleaning speeds used are the same. The traditional method refers to manually measuring the focal length, temperature, and using a contact roughness measurement device to measure after the cleaning is completed.

[0042] The above content is only part of the experimental data. During another part of the experiment, due to the design of a dynamic temperature threshold, the method of the present invention stops emitting light when exceeding the threshold compared with the traditional method, achieving the purpose of protecting the substrate, while the traditional method causes damage to the substrate.

[0043] Example 2: Device for Laser Cleaning Carbon Fiber Composite Materials Based on Optical Coherence Tomography As shown in the appendix Figure 2 The device includes a laser control system 2, a sensor control system 3, an optical detection head control system 4, and a three-dimensional displacement platform 9 connected to a computer 1; the laser control system 2 is connected to a laser cleaning head 6, the sensor control system 3 is connected to a sensor module 5, the optical detection head control system 4 is connected to an OCT optoelectronic balance detector 7, and a carbon fiber composite material to be cleaned (material to be processed 8) is placed on the three-dimensional displacement platform 9. The positions of the laser cleaning head 6, the sensor module 5, and the OCT optoelectronic balance detector 7 correspond to the above carbon fiber composite material.

[0044] Place the material to be processed 8 on the three-dimensional displacement platform 9. Above the three-dimensional displacement platform 9, there are a laser cleaning head 6 and an OCT optoelectronic balance detector 7. Adjust the material to be processed 8 to be located on the focal plane of the focusing lens in the laser cleaning head 6; Control the sensor module 5 and its control system to start and establish a data transmission connection with the computer 1; control the OCT optoelectronic balance detector 7 and its control system to start and establish a data transmission connection with the computer 1; set the parameter ranges of the laser cleaning head 6, the sensor module 5, and the OCT optoelectronic balance detector 7; The scanning galvanometers of the laser cleaning head 6, the sensor module 5, and the OCT optoelectronic balance detector 7 achieve hardware-level synchronization through the FPGA in the control system. The laser pulse triggers the OCT acquisition window, and the laser pulse frequency is the same as the OCT acquisition frequency; Adopt a combination of two-dimensional cross-sectional scanning and three-dimensional surface scanning to obtain the image of the material to be processed 8; The OCT optoelectronic balance detector 7 acquires the reference spectrum I ref (k) of no sample and deducts the background noise in real time; calibrates the wavenumber k space through the fixed reflection peak of the reference mirror to eliminate the influence of spectrometer nonlinearity, applies a Hanning window to I(k) to suppress the FFT sidelobes, and acquires and processes the data; Compare the above processed data with the specified parameter range and perform real-time control through the control system; According to the set parameter range, emit laser to process the surface of the carbon fiber composite material, and at the same time, the laser cleaning head 6 and the OCT optoelectronic balance detector 7 move synchronously to achieve the removal of surface dirt.

[0045] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and the specification of the present invention.

Claims

1. A real-time monitoring method for laser cleaning of carbon fiber composites, using carbon fiber composites with dirt on the surface as the material to be processed, characterized in that, It includes the following steps: Step S1: Place the material to be processed on a three-dimensional displacement platform. There is a laser cleaning head and an OCT optoelectronic balance detector above the three-dimensional displacement platform. Adjust the material to be processed to be on the focal plane of the focusing lens in the laser cleaning head; Step S2: Control the sensor module and its control system to turn on and establish a data transmission connection with the computer; control the OCT optoelectronic balance detector and its control system to turn on and establish a data transmission connection with the computer; set the parameter ranges of the laser cleaning head, the sensor module, and the OCT optoelectronic balance detector; Step S3: The laser cleaning head, the scanning galvanometer of the sensor module, and the OCT optoelectronic balance detector achieve hardware-level synchronization through the FPGA in the control system. The laser pulse triggers the OCT acquisition window, and the laser pulse frequency is the same as the OCT acquisition frequency; Step S4: Obtain the image of the material to be processed by combining two-dimensional cross-sectional scanning and three-dimensional surface scanning; Step S5: The OCT optoelectronic balance detector collects the reference spectrum I ref (k) without a sample, and deducts the background noise in real time; calibrates the wavenumber k space through the fixed reflection peak of the reference mirror to eliminate the influence of spectrometer nonlinearity, applies a Hanning window to I(k) to suppress the FFT sidelobes, collects data and processes it; Step S6: Compare the above processed data with the specified parameter range and perform real-time control through the control system; Step S7: According to the set parameter range, emit laser to process the surface of the carbon fiber composite material, and at the same time, the laser cleaning head and the OCT optoelectronic balance detector move synchronously to achieve the removal of surface dirt.

2. The real-time monitoring method for laser cleaning carbon fiber composites according to claim 1, characterized in that: In step S2, the scanning speed of the laser cleaning head is 3000 - 4000 mm / s, the cleaning speed is 0.0036 - 0.004 m / s, the laser connected to the laser cleaning head is an ultraviolet laser, its power range is 1 - 3 W, and the laser repetition frequency is 30 - 35 Hz; the sensor module includes a temperature sensor and a laser ranging module. Among them, the measuring range of the temperature sensor is 200 - 800 °C, the temperature detection accuracy is ±10 °C, the thermal radiation response time ≤ 0.5 ms, and the dynamic range of the laser ranging module is 0 - 10 mm; the arithmetic mean roughness detection range of the OCT optoelectronic balance detector is 0.1 - 10 μm, the detection range of the maximum height difference is 1 - 50 μm, the OCT axial resolution ≤ 5 μm, and the transverse scanning coverage rate ≥ 10 mm.

3. The real-time monitoring method for laser cleaning of carbon fiber composites according to claim 1, characterized in that: In step S3, the laser pulse triggers the OCT acquisition window, the laser pulse frequency is the same as the OCT acquisition frequency, and the delay time < 0.1 μs.

4. The real-time monitoring method for laser cleaning of carbon fiber composites according to claim 1, characterized in that: In step S4, the two-dimensional cross-sectional scan uses single-line scanning and depth-direction scanning to analyze the surface topography, and the three-dimensional surface scan uses XY-plane grid scanning for high-precision surface reconstruction.

5. The real-time monitoring method for laser cleaning carbon fiber composites according to claim 1, characterized in that: In step S5, the OCT optoelectronic balanced detector uses a linear array InGaAs camera with 2048 pixels. Its spectral range is 1200 - 1400 nm, the resolution Δλ≈0.1 nm, and the data is transmitted to the FPGA through the Camera Link interface. The single-frame acquisition time is 10 μs. The specific data acquisition and processing include the following steps: 1) Interference signal calibration: Calibrate the wavenumber k space through the fixed reflection peak of the reference mirror to eliminate the non-linear response of the spectrometer; collect the reference spectrum I ref (k) without a sample, and calculate the correction signal in real time ; Fit the non-linear kk-space through the reflection peak position of the calibration mirror and use cubic spline interpolation to correct: k linear =a0 + a1λ + a2λ 2 + a3λ 3 , where λ is the pixel index of the spectrometer, and the coefficients a0 - a3 are obtained through calibration; apply the Hanning window to I(k) to suppress the FFT sidelobes: , where N = 2048; calculate the mean value of the DC signal and subtract it: ; Correct the temperature data non-uniformity: , where Toffset and G(x,y) are obtained through blackbody furnace calibration; perform moving average filtering on the displacement data to suppress high-frequency vibration noise: ; 2) Depth profile A-scan reconstruction: Perform FFT on the 2048-point interference signal of each A-scan to generate a complex depth profile A(z), extract the amplitude information |A(z)|, and record the phase φ(z) = arg(A(z)): A(z) = Γ -1 {I DC-free (k)}, search for the position Z peak of the maximum amplitude of the surface reflection peak: ; 3) Surface topography reconstruction: A single B-scan line contains 512 A-scan points and is reconstructed into a two-dimensional depth profile in the X-Z plane. Take the Z peak of each A-scan along the X direction to generate a one-dimensional height curve h(x); the XY galvanometer scans step by step to generate a three-dimensional point cloud matrix Z(x,y) for offline high-precision roughness analysis.

6. The real-time monitoring method for laser cleaning of carbon fiber composites according to claim 5, characterized in that: In step 3), the process of roughness calculation includes step a): denoise the obtained h(x) by Gaussian low-pass filtering with a cut-off wavelength λc = 50 μm to obtain , and then fit the reference plane h base (x) = ax + b by the least squares method, perform detrending, and calculate the residual height h corrected (x) = h filtered (x) - h base (x); b) calculate the roughness parameters by combining the obtained residual height with the formula to obtain the arithmetic mean roughness , the root mean square roughness , and the maximum height difference Rz = max(h i ) - min(h i ).

7. The real-time monitoring method for laser cleaning of carbon fiber composites according to claim 1, characterized in that: Step S6 includes step 1) establishing a roughness-power dynamic control model: assuming the target arithmetic mean roughness is 3 μm, with an allowable deviation of ±0.5, setting the proportionality coefficient K p = 0.2 W / μm, setting the integral coefficient K i = 0.05 W / (μm·s), setting the constraint conditions that the lower limit of power limit is 1 W and the upper limit of power is 5 W, and then adjusting the real-time power ; The real-time control process is to obtain the roughness Ra every 2 ms, and perform deviation calculation ΔRa = Ra - Ra target ; When |ΔRa| ≤ 0.5 μm, keep the current power. If ΔRa > 0.5 μm, increase the power according to the formula. If ΔRa < 0.5 μm, decrease the power according to the formula; send a 16-bit PWM signal to control the laser through the RS422 interface, with a resolution of 0.1 W; 2) setting a temperature over-limit protection mechanism: the glass transition temperature Tmax of carbon fiber epoxy resin is 300 °C, setting the dynamic threshold 250 °C < T threshold (x,y) < 275 °C; the infrared thermal imager in the temperature sensor outputs the temperature matrix T(x,y) at a frequency of 60 Hz, setting the emergency stop condition: ∃(x,y) s.t. T(x,y) > T threshold (x,y) + 10 °C; when the set temperature threshold is exceeded, the FPGA hardware emergency stop signal is sent, immediately turning off the light source, the galvanometer controls the laser focus to move out of the processing area, and records the over-temperature position (x,y) and the temperature value.

8. The real-time monitoring method for laser cleaning carbon fiber composite materials according to claim 1, characterized in that: There is a safety redundancy designed in the parameter range in step S7: 1) The FPGA hardware watchdog continuously monitors the heartbeat signal of the main control program, and automatically triggers an emergency stop signal if it times out; 2) Dual-channel temperature verification, the confocal displacement sensor synchronously measures the surface reflectivity R. If R drops suddenly and the temperature is not exceeded, a secondary protection is still triggered: the ablation determination condition is R < 0.3 and dR / dt < -0.1 / μs; 3) Power gradient ramp, any power adjustment goes through a ramp transition. Additionally, the time constant τ = 0.2 ms, ensuring that the 10 - 90% rise time < 1 ms: P(t) = P initial +(P target -P initial )·(1 - e -t / τ ).

9. An apparatus for laser cleaning carbon fiber composite materials based on the real-time monitoring method according to claim 1, characterized in that: It includes a laser control system, a sensor control system, an optical detection head control system, and a three-dimensional displacement platform connected to a computer; the laser control system is connected to the laser cleaning head, the sensor control system is connected to the sensor module, the optical detection head control system is connected to the OCT optoelectronic balance detector, the carbon fiber composite material to be cleaned is placed on the three-dimensional displacement platform, and the positions of the laser cleaning head, the sensor module, and the OCT optoelectronic balance detector correspond to the carbon fiber composite material.

Images