A method of laser engraving a product surface

Abstract

This invention discloses a laser engraving method for product surfaces, comprising: during the duration of a laser pulse, alternately applying low-frequency sweep excitation signals and high-frequency fixed-frequency excitation signals through a coaxial ring induction electrode according to a preset timing sequence, and acquiring plasma impedance response; calculating and obtaining a layer feature vector F1 characterizing the material layer and a quality feature vector F2 characterizing the surface microstate; performing similarity matching between F1 and a preset fingerprint database to determine the layer, and coarsely adjusting the laser parameters when the layer changes; simultaneously comparing the deviation of F2 with a standard quality feature range, and finely adjusting the subsequent laser parameters, with the fine adjustment response being faster than the coarse adjustment; when in the interlayer transition zone, pausing the coarse parameter adjustment and only performing the fine adjustment, and synthesizing the transition zone quality standard based on the similarity distance between F1 and the fingerprints of adjacent layers. This invention can effectively eliminate interface processing defects and improve the interface flatness and batch consistency of multilayer composite material engraving.

Description

Technical Field

[0001] This application relates to the field of laser processing and online inspection, and in particular to a method for laser engraving on the surface of a product. Background Technology

[0002] In surface laser engraving of composite heterogeneous material products, existing technologies typically employ a single-wavelength, single-pulse laser with preset fixed parameters for engraving. For multilayer composite materials consisting of a surface coating and a metal substrate, multilayer printed circuit board copper foil and an insulating layer, or leather and fabric substrates, some improved solutions introduce layer detection methods based on spectral analysis or optical coherence tomography. By collecting the plasma emission spectrum generated during laser processing or measuring the processing depth, it is determined whether the current engraving has penetrated to the target layer. When a layer change is detected, the laser is triggered to stop or a simple parameter switch is performed to avoid burning through the underlying substrate.

[0003] However, the aforementioned existing technologies have the following drawbacks in practical applications: First, both spectral analysis and optical coherence tomography are "post-detection" methods, with detection delays ranging from tens of microseconds to milliseconds. During this period, the laser has already caused irreversible overheating damage to the underlying material, making true real-time layer-following control impossible. Second, existing solutions can only achieve a binary switch-type judgment of "whether the target layer has been reached," lacking the ability to monitor the engraving quality within the same material layer online. When the laser power drifts due to thermal effects or the focus shifts slightly, it cannot be actively corrected, resulting in poor consistency in engraving depth and surface roughness during batch processing. Furthermore, in the interface transition area between two material layers, existing technologies have not proposed a targeted control strategy. The pulse group effect generated by parameter switching will form step-like processing marks at the interface, affecting the product's appearance quality and coating adhesion.

[0004] To address this, the present invention proposes a laser engraving method for product surfaces. By alternately applying low-frequency and high-frequency excitation signals to a coaxial ring electrode during the duration of a laser pulse, layer feature vectors characterizing material layers and quality feature vectors characterizing surface microstates are simultaneously extracted from the same photo-induced plasma. These feature vectors drive layer coarse adjustment closed loops and quality fine adjustment closed loops, respectively. Furthermore, a dedicated collaborative control strategy is executed in the interlayer transition region to achieve high-precision, high-quality adaptive engraving of multilayer composite heterogeneous materials. Summary of the Invention

[0005] To address the aforementioned problems, this application provides a method for laser engraving on a product surface, employing the following technical solution: A method for laser engraving on a product surface includes the following steps: S1: During the duration of the laser engraving pulse acting on the surface of the composite heterogeneous material, a low-frequency sweep frequency excitation signal and a high-frequency fixed frequency excitation signal are alternately applied to the plasma region generated by the laser in a preset sequence through a ring induction electrode coaxially set at the exit of the engraving head, and the low-frequency impedance response of the plasma to the low-frequency sweep frequency excitation signal and the high-frequency impedance response to the high-frequency fixed frequency excitation signal are collected respectively. S2: Perform equivalent circuit model fitting and solution on the low-frequency impedance response to obtain the layer feature vector F1, which is used to characterize the material layer to which the current processing position belongs; perform time-domain feature extraction and solution on the high-frequency impedance response to obtain the quality feature vector F2, which is used to characterize the micro-state of the current processing surface. S3: Perform similarity matching between the layer feature vector F1 and the preset composite material impedance feature fingerprint database to determine the material layer at the current processing position; when the determined layer changes, switch the laser engraving parameters to the preset parameter group corresponding to the material layer in the first response time. S4: Compare the quality feature vector F2 with the preset standard quality feature range. When F2 deviates from the standard quality feature range, adjust and correct the subsequent laser engraving parameters in real time according to the preset quality parameter mapping model within the second response time. The second response time is less than the first response time. S5: When the layer feature vector F1 shows that the current processing position is in the interlayer transition zone between two layers of material, pause the layer driving parameter switching operation in step S3, and only perform the quality driving parameter fine-tuning correction in step S4; until F1 stably falls into the new layer region within a series of laser pulse cycles, then resume the parameter switching operation in step S3.

[0006] Preferably, in step S1: The low-frequency sweep excitation signal has a frequency range of 1MHz to 10MHz and is used to detect the bulk conductivity and dielectric properties of the plasma. The high-frequency fixed-frequency excitation signal has a frequency range of 30MHz to 50MHz and is used to detect the uniformity and stability of the plasma sheath distribution.

[0007] Preferably, in step S2: The layer feature vector F1 includes at least the plasma resistance Rp, sheath capacitance Cp, and equivalent inductance Lp in the plasma equivalent circuit model. The quality feature vector F2 includes at least the root mean square fluctuation of the impedance amplitude Δ|Z| and the transient deviation of the phase angle Δφ in the high-frequency impedance response.

[0008] Preferably, step S2 further includes an environmental parameter compensation step, specifically: The temperature (T), relative humidity (RH), and atmospheric pressure (P) of the processing environment are collected in real time by a miniature temperature, humidity, and pressure composite sensor integrated into the engraving head. Based on the pre-calibrated environmental-impedance compensation model, the impedance spectrum correction ΔZ(T,RH,P) corresponding to the current environmental parameters is calculated. The impedance spectrum correction value ΔZ is used to compensate and correct the acquired low-frequency impedance response and high-frequency impedance response respectively, so as to eliminate the influence of environmental fluctuations on the accuracy of impedance feature extraction.

[0009] Preferably, it also includes a user-defined extension step for the composite material impedance feature fingerprint library, specifically: Enter calibration mode and execute a preset calibration engraving program on a standard sample with a known layer structure; Collect F1 sample data of the layer feature vectors corresponding to each layer of material during the calibration and engraving process, and calculate the cluster center value and variance range of each layer of F1 sample data; The cluster center value and variance range are associated with the corresponding material layer information and stored as new material entries in the composite material impedance feature fingerprint database to adapt to the identification needs of different batches or niche composite materials.

[0010] Preferably, the specific determination condition for F1 to stably fall into the new layer region within multiple consecutive laser pulse cycles in step S5 is as follows: Set a counting threshold N, with N ranging from 5 to 10; when the weighted Euclidean distance between the layer feature vector F1 calculated within N consecutive laser pulse cycles and the target layer fingerprint is less than the preset threshold, it is determined that the new layer region has been fully entered.

[0011] Preferably, the quality parameter mapping model in step S4 is established through pre-calibration experiments, specifically including: Establish the mapping relationship between Δ|Z| and the surface roughness Ra in the quality feature vector F2, and the mapping relationship between Δφ and the microcrack density on the surface. Establish a PID control law for the deviation between the actual roughness value and the standard value and the laser power correction ΔP, and a proportional control law for the deviation between the actual microcrack density value and the standard value and the scanning speed correction ΔV.

[0012] Preferably, step S1 further includes an electromagnetic compatibility control step, specifically: The inner ring of the annular induction electrode is configured as a grounded shielding ring to bypass the high-voltage transient electric field generated by the laser pulse to the ground. The transmission paths of the low-frequency sweep excitation signal and the high-frequency fixed-frequency excitation signal adopt double-shielded coaxial cable, and an active low-pass filter with a cutoff frequency of 500MHz is connected in series at the signal acquisition front end.

[0013] Preferably, the composite heterogeneous material is a multilayer composite structure, specifically including at least one of the following types: Metal or alloy substrates with an organic polymer film coating on their surface; Printed circuit board with alternating layers of copper foil and insulating layers; A soft interior material composed of a leather top layer and a fabric backing layer.

[0014] Preferably, in step S5, when step S4 is performed in the interlayer transition zone, the preset standard quality feature range is obtained by weighting the standard quality feature ranges of the two adjacent layers by using the similarity distance between the current layer feature vector F1 and the layer fingerprints of the two adjacent material layers as weights.

[0015] In summary, this application includes at least one of the following beneficial technical effects: 1. By simultaneously extracting the layer feature vector F1 and the quality feature vector F2 from the same photoinduced plasma within a single laser pulse duration, a one-source two-dimensional sensing for layer identification and processing quality monitoring of composite heterogeneous materials is realized. This eliminates the need for additional independent quality monitoring sensors, simplifying the system structure while obtaining richer processing information. 2. By constructing a dual-layer collaborative control architecture of coarse adjustment closed loop and fine adjustment closed loop of quality, the large-scale switching of cross-layer parameters and the continuous fine adjustment of processing quality within the layer are adapted to each other in terms of response speed and control granularity, which solves the technical contradiction that layer determination and quality control cannot be taken into account in a single closed loop in the existing technology. 3. By adopting a collaborative control strategy in the interlayer transition zone, which involves pausing the switching of layer-driven parameters and only performing fine-tuning of quality-driven parameters, and dynamically determining the quality control benchmark of the transition zone by weighted synthesis of the standard quality characteristic range of adjacent layers, the processing defects caused by parameter switching delay and pulse group effect at the material interface in the traditional method are effectively eliminated, and the interface smoothness and interlayer bonding quality of multilayer composite heterogeneous material engraving are significantly improved. 4. Through the online compensation mechanism for environmental parameters, the impedance feature extraction can adapt to the fluctuations in workshop temperature, humidity and air pressure, avoiding misjudgment of strata and deviation in quality assessment caused by environmental changes, and improving the measurement stability and control reliability of the system during long-term operation in industrial sites. 5. Through the user-defined extension function of the composite material impedance characteristic fingerprint library, the system can quickly adapt to the impedance characteristic differences of different batches, different suppliers or niche composite materials, reduce the recalibration cost when changing production equipment, and enhance the versatility and flexibility of processing methods. Attached Figure Description

[0016] Figure 1 This is a flowchart of a product surface laser engraving method according to an embodiment of this application. Detailed Implementation

[0017] This invention proposes a laser engraving method for product surfaces. The core of this method lies in the following: during the brief duration of the laser engraving pulse applied to the surface of a composite heterogeneous material, low-frequency and high-frequency electrical excitation signals are alternately applied to the photoinduced plasma through a coaxial ring induction electrode. Two-dimensional impedance feature vectors characterizing the material layers and the micro-state of the processed surface are synchronously extracted from the same plasma source. This drives the coordinated operation of the layer coarse adjustment closed loop and the quality fine adjustment closed loop. A dedicated control strategy is executed in the interlayer transition region to achieve high-precision, high-quality adaptive engraving of multilayer composite heterogeneous materials.

[0018] In this invention, the preset timing sequence refers to the time allocation rule within the duration of a single laser engraving pulse, which divides time into multiple acquisition windows, sequentially applying different frequency excitation signals to the plasma region and correspondingly acquiring the impedance response. This timing sequence is preset based on the laser pulse width, plasma formation and evolution characteristics, and the response speed of the signal acquisition circuit to ensure complete acquisition of both low-frequency and high-frequency impedance responses within the pulse duration.

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that the following embodiments are only for explaining the invention and are not intended to limit the scope of protection of the invention.

[0020] Example 1: Logo engraving on the surface coating of a mobile phone aluminum alloy casing This embodiment uses logo engraving on the PET film layer of a mobile phone aluminum alloy casing as an application scenario. The workpiece has a multi-layered composite structure: the outer layer is a transparent PET organic polymer film layer with a thickness of approximately 20μm, the bottom layer is a 6061 aluminum alloy substrate, and the two layers are bonded together by a transparent acrylate adhesive layer with a thickness of approximately 3-5μm. The processing requirements are: the engraving depth should just penetrate the PET film layer and enter the adhesive layer, but without damaging the surface of the aluminum alloy substrate; the engraved surface should have clear lines, no carbonization at the edges, and no step marks at the interface.

[0021] Reference Figure 1 This embodiment provides a method for laser engraving on the surface of a product, including the following steps: S1. Alternating application of dual-frequency excitation signals and acquisition of impedance response: During the duration of the laser engraving pulse acting on the surface of the composite heterogeneous material, a low-frequency sweep excitation signal and a high-frequency fixed-frequency excitation signal are alternately applied to the plasma region induced by the laser through a ring induction electrode coaxially set at the exit of the engraving head, according to a preset time sequence, and the low-frequency impedance response of the plasma to the low-frequency sweep excitation signal and the high-frequency impedance response to the high-frequency fixed-frequency excitation signal are acquired respectively. S2. Dual-dimensional impedance feature calculation: The low-frequency impedance response is fitted with an equivalent circuit model to obtain a layer feature vector F1 that characterizes the material layer to which the current processing position belongs; the high-frequency impedance response is extracted and calculated using time-domain features to obtain a quality feature vector F2 that characterizes the micro-state of the current processing surface. S3. Layer determination and parameter coarse adjustment: The layer feature vector F1 is matched with the preset composite material impedance feature fingerprint database to determine the material layer at the current processing position; when the determined layer changes, the laser engraving parameters are switched to the preset parameter group corresponding to the material layer in the first response time. S4. Quality Deviation Judgment and Parameter Fine-tuning: The quality feature vector F2 is compared with the preset standard quality feature range. When F2 deviates from the standard quality feature range, the subsequent laser engraving parameters are fine-tuned in real time according to the preset quality parameter mapping model within the second response time. The second response time is less than the first response time. S5. Interlayer transition zone collaborative control: When the layer feature vector F1 shows that the current processing position is in the interlayer transition zone between two layers of material, the layer driving parameter switching operation in step S3 is paused, and only the quality driving parameter fine-tuning correction in step S4 is performed; until F1 stably falls into the new layer region within multiple consecutive laser pulse cycles, the parameter switching operation in step S3 is resumed.

[0022] By simultaneously extracting the layer feature vector F1 and the quality feature vector F2 from the same photoinduced plasma within a single laser pulse duration, a single-source dual-dimensional sensing is achieved, enabling simultaneous material layer identification and processing quality monitoring without the need for additional sensors. A dual-layer collaborative control architecture of layer coarse adjustment and quality fine adjustment is constructed, allowing for a balance between large-scale switching of cross-layer parameters and continuous fine-tuning of intra-layer processing quality in terms of response speed and control granularity. In particular, a collaborative strategy is employed in the interlayer transition zone, pausing layer-driven switching, performing only quality-driven fine adjustment, and combining this with weighted synthesis of standard quality feature ranges from adjacent layers. This effectively eliminates the step-like processing marks caused by pulse group effects at material interfaces, significantly improving the interface smoothness of multilayer composite heterogeneous material engraving. Furthermore, through online environmental parameter compensation and user-defined expansion of the composite material impedance feature fingerprint database, the system can adapt to fluctuations in workshop temperature and humidity and batch differences in materials, greatly improving the robustness and consistency of batch processing.

[0023] The specific implementation methods for each step are described in detail below.

[0024] Step S1, the specific implementation of alternating application of dual-frequency excitation signals and impedance response acquisition: In this embodiment, the laser engraving head is equipped with a coaxial annular induction electrode assembly at its front end. This assembly includes: an inner grounding shield ring made of gold-plated copper, with an inner diameter of 3.2 mm and an outer diameter of 3.8 mm, reliably connected to the engraving head housing via a metal elastic contact; and an outer induction electrode made of gold-plated copper, with an inner diameter of 4.0 mm and an outer diameter of 4.6 mm, separated from the inner ring by a polytetrafluoroethylene insulating ring with a gap of 0.2 mm. The coaxiality deviation between the axes of the inner and outer rings and the laser beam optical axis is no greater than 0.05 mm. The vertical distance between the lower end face of the induction electrode and the workpiece surface is controlled within the range of 1.5 mm to 3.0 mm.

[0025] In addition to the electrode structure described above, regarding the timing design of the dual-frequency excitation signal, this embodiment employs a nanosecond-level pulsed fiber laser with a center wavelength of 1064 nm, a pulse width of approximately 100 ns (full width at half maximum), and a pulse repetition frequency of 50 kHz, meaning the interval between adjacent pulses is 20 μs. Within the duration of each laser pulse, the excitation and acquisition timing design is as follows: Following the laser pulse leading edge (i.e., the instant of plasma formation, approximately 2-3 ns after triggering), the first acquisition window begins, lasting 30 ns. Within this window, a low-frequency sweep excitation signal of 1 MHz to 10 MHz is applied to the plasma region through the outer ring of the annular induction electrode. The sweep step is 1 MHz, with each excitation step lasting approximately 3 ns, for a total of 10 frequency points. The low-frequency impedance response of the plasma to this low-frequency sweep excitation signal is simultaneously acquired, and the acquired data includes the impedance amplitude |Z| at each frequency point. L | and phase angle φL The frequency range of the low-frequency sweep excitation signal is 1MHz to 10MHz. The reason for selecting this frequency range is that within this frequency band, the impedance response of the plasma is most sensitive to the bulk conductivity and dielectric properties of the material, which can effectively reflect the intrinsic properties of the parent material and provide highly recognizable characteristic parameters for subsequent stratigraphic identification.

[0026] The second acquisition window then begins, lasting 50 ns. During this window, a 40 MHz high-frequency fixed-frequency excitation signal is applied, with the frequency falling within the 30 MHz to 50 MHz range. The plasma's high-frequency impedance response to this excitation signal is simultaneously acquired, and the acquired data includes the impedance amplitude |Z| at that frequency. H | and phase angle φ H The frequency range of this high-frequency fixed-frequency excitation signal is 30MHz to 50MHz. The reason for selecting this frequency range is that within this frequency band, the impedance response of the plasma is most sensitive to changes in the sheath thickness, which can effectively reflect the influence of the micro-undulations of the processed surface on the uniformity of the sheath distribution, and provide a highly sensitive detection signal for subsequent quality monitoring.

[0027] The remaining approximately 20ns are used for data settlement and parameter updates, completing all processing before the next laser pulse arrives.

[0028] It should be noted that during the duration of each frequency point of the low-frequency sweep excitation, the number of oscillation cycles of the excitation signal at that frequency point is limited (for example, at a frequency of 1 MHz, a duration of 3 ns contains only about 0.003 complete cycles). Those skilled in the art can use lock-in amplification technology or IQ quadrature demodulation technology to perform phase-sensitive detection on the acquired response signal using a reference signal with the same frequency as the excitation signal, and extract the impedance amplitude and phase angle information at that frequency point from the sub-periodic signal segment. Such signal processing techniques are conventional techniques in this field and will not be elaborated here.

[0029] In terms of electromagnetic compatibility design, the inner grounding shield bypasses the high-voltage transient electric field generated during laser pulse triggering directly to ground, preventing this strong interference from coupling into the outer impedance measurement circuit. The low-frequency sweep excitation signal and the high-frequency fixed-frequency excitation signal are transmitted to the outer induction electrode via a double-shielded coaxial cable. The outer shield is connected to the engraving head housing ground, and the inner shield is connected to the signal ground. A ferrite core is filled between the inner and outer shields to suppress common-mode interference. A fourth-order Butterworth active low-pass filter with a cutoff frequency of 500MHz is connected in series before the signal acquisition front-end, i.e., before the ADC input port, to filter out high-frequency noise generated by the laser driver power switch.

[0030] Furthermore, this embodiment employs the following layout design: all analog signal traces (impedance detection signals) are arranged in the left slot of the engraving head, while all digital control signals and high-frequency power cables are arranged in the right slot, separated by a metal partition. The analog signal ground plane and the digital ground plane are connected at a single point below the ADC chip, effectively suppressing ground loop interference.

[0031] Step S2, Specific implementation of two-dimensional impedance characteristic calculation: After acquiring low-frequency impedance spectrum data (including |Z| and φ) at 10 frequency points in the first acquisition window, the data is converted into the real and imaginary parts of complex impedance, and then nonlinear least-squares fitting is performed using a plasma equivalent circuit model. This equivalent circuit model is a series-parallel hybrid topology: the plasma resistance Rp is connected in series with the equivalent inductance Lp, and then in parallel with the sheath capacitance Cp.

[0032] The fitting algorithm employed the Levenberg-Marquardt algorithm, with the objective function being the minimization of the sum of squared complex deviations between the measured impedance and the model impedance at 10 frequency points. The fitting outputs the values ​​of three model parameters, forming a layer eigenvector F1=[Rp,Cp,Lp]. Here, Rp represents the bulk resistance of the plasma, primarily dependent on the electron density and electron-neutral particle collision frequency; Cp represents the sheath capacitance formed at the plasma-material interface, dependent on the sheath thickness and electrode coupling area; and Lp represents the inertial inductance effect of electrons, dependent on the electron mass density effect.

[0033] The table below shows typical F1 values ​​for the three materials in this embodiment (under the conditions of 50kHz pulse, 100ns pulse width, and 5W average power):

[0034] As can be seen from the table, the F1 parameters of different materials exhibit significant differences in magnitude, demonstrating good layer identification capabilities. The Rp of the metal substrate is only about 1 / 20 of that of the organic film, the Cp is about 1 / 5, and the Lp is about 1 / 8. This is because the electron density of metals is much higher than that of organic materials, resulting in a correspondingly significant increase in plasma conductivity.

[0035] After obtaining the high-frequency impedance response of 40MHz in the second acquisition window, time-domain statistical analysis was performed on 32 consecutive sampling points (sampling rate of 200MHz, sampling duration of 160ns, covering the second acquisition window and the transition bands of 55ns before and after it), and the following two quality characteristic parameters were extracted: The root mean square ripple of the impedance amplitude, Δ|Z|, is calculated using the following formula:

[0036] in Let be the high-frequency impedance amplitude at the i-th sampling point. It is the arithmetic mean of 32 sampling points.

[0037] The transient deviation Δφ of the phase angle is calculated using the following formula:

[0038] in Let be the high-frequency impedance phase angle of the i-th sampling point. This is the smoothed value after median filtering (window width 5 points). Δφ is taken as the maximum absolute value deviating from the smoothed value among the 32 sampling points to capture the phase abrupt peaks caused by microcracks.

[0039] The quality feature vector F2 = [Δ|Z|, Δφ]. Experiments have verified the following physical correlation between Δ|Z| and Δφ and the microscopic state of the processed surface: When the surface roughness Ra increases, the local thickness unevenness of the plasma sheath on the micro-undulating surface intensifies, leading to scattering loss of high-frequency signals at the sheath boundary, manifested as a systematic increase in the root mean square fluctuation of Δ|Z|; when microcracks appear on the surface, local electron escape channels are formed at the cracks, causing abrupt changes in the equivalent capacitance of the plasma-material interface, manifested as an abnormal spike in the transient jump variable of Δφ.

[0040] The F1 extracted in this step will be used as the input for the subsequent step S3 layer determination, and the F2 extracted will be used as the input for the subsequent step S4 quality deviation determination.

[0041] In addition, considering that fluctuations in temperature, humidity, and air pressure in the workshop environment may affect the thickness of the plasma sheath, thereby causing the impedance measurement reference to drift, this embodiment also includes an environmental parameter compensation step in step S2, specifically including the following steps: A miniature temperature, humidity, and pressure composite sensor (Bosch BME280, sampling frequency 100Hz) is integrated into the engraving head to collect real-time data on the processing environment's temperature (T), relative humidity (RH), and atmospheric pressure (P). An environment-impedance compensation model is established through pre-calibration experiments. The calibration scheme involves placing the device in a controlled environment chamber and performing impedance spectroscopy measurements on the same standard sample (PET-coated standard sample) under various combinations of temperature (20℃-35℃, in 5℃ steps), relative humidity (30%-70%RH, in 10%RH steps), and atmospheric pressure (95kPa-105kPa, in 2.5kPa steps). The baseline values ​​of Rp, Cp, and Lp under different environmental conditions are recorded. Using the baseline values ​​under standard environmental conditions (T0=25℃, RH0=50%RH, P0=101.325kPa) as a reference, a compensation model is established using multiple linear regression.

[0042] Regression analysis yielded the following results: a1 = -2.3 Ω / ℃, a2 = -0.8 Ω / %RH, a3 = 1.4 Ω / kPa (taking Rp as an example, the compensation coefficients for Cp and Lp were obtained similarly). In actual processing, the impedance spectrum correction ΔZ is calculated based on the current environmental sensor readings. Before the impedance data enters the solution step, the impedance spectrum correction ΔZ is used to compensate and correct the acquired low-frequency and high-frequency impedance responses, respectively. This is achieved by subtracting (or adding according to the coefficient sign) ΔZ from the acquired original impedance value to eliminate the influence of environmental fluctuations on the accuracy of impedance feature extraction.

[0043] The linear assumptions of the above environmental-impedance compensation model are based on the following: In a typical industrial processing environment, the temperature range is 20℃ to 35℃, the relative humidity range is 30% to 70%RH, and the atmospheric pressure range is 95kPa to 105kPa. Within this narrow range, the plasma impedance can be approximated as linear with respect to the changes in these environmental parameters (first-order Taylor expansion approximation), with a linear fitting error of less than 1%, which is sufficient to meet engineering accuracy requirements.

[0044] The compensation coefficients a1, a2, and a3 were obtained through a three-factor, five-level orthogonal calibration experiment. Specifically: Temperature (T) was set at five levels: 20℃, 25℃, 30℃, 35℃, and 40℃; relative humidity (RH) was set at five levels: 30%, 40%, 50%, 60%, and 70%; and atmospheric pressure (P) was set at five levels: 95 kPa, 98 kPa, 101 kPa, 104 kPa, and 107 kPa. A total of 5 × 5 × 5 = 125 sets of experiments were conducted. For each set of experiments, impedance data from 100 laser pulses were collected on the same standard sample under corresponding environmental conditions, and the average value was taken as the measurement value for that set.

[0045] Using the reference impedance value Z0 under standard environmental conditions (T0=25℃, RH0=50%RH, P0=101.325kPa) as a reference, the impedance deviation ΔZ under various environmental conditions was calculated. The Levenberg-Marquardt algorithm was used to perform multiple linear regression fitting on 125 sets of data, yielding the following results: a1 = -2.3Ω / ℃, a2 = -0.8Ω / %RH, a3 = 1.4Ω / kPa (taking Rp as an example, the compensation coefficients of Cp and Lp are obtained using the same method).

[0046] The goodness-of-fit R² = 0.987 indicates that the linear model is in high agreement with the experimental data. Each experiment was independently repeated three times and the average value was taken. The cross-validation prediction error was less than 1.2Ω, which verifies the accuracy and generalization ability of the compensation model.

[0047] During routine processing, if the workshop temperature gradually increases due to continuous equipment operation (e.g., from 25°C to 32°C), without compensation, the measured Rp value will deviate by approximately 16Ω (about 3.5% deviation), potentially affecting the accuracy and robustness of the subsequent S3 layer determination. Enabling environmental compensation effectively eliminates this deviation.

[0048] It should be noted that laser-induced plasma exhibits instantaneous random fluctuations. To suppress their impact on impedance measurement accuracy, this invention employs a triple mechanism for noise reduction: First, synchronous acquisition is performed pulse by pulse. The impedance acquisition timing is locked to the laser pulse trigger signal, and the acquisition window is opened after a delay of 2ns to 3ns after the laser pulse leading edge to avoid the high-voltage transient electromagnetic noise generated at the pulse leading edge.

[0049] Second, multi-pulse statistical filtering. Impedance data acquired over three consecutive laser pulse cycles are first subjected to median filtering to remove gross errors, followed by mean filtering. The filtering window length N = 3 to 5 (N = 3 in this embodiment). This window length effectively suppresses random fluctuations while ensuring response speed.

[0050] Third, statistical identification based on eigenvectors. This invention uses plasma resistance Rp, sheath capacitance Cp, and equivalent inductance Lp obtained by fitting an equivalent circuit model to form the eigenvector F1, rather than relying on single-point impedance values. Since Rp, Cp, and Lp are statistical fitting results of impedance data at multiple frequencies, their statistical stability is much higher than the original sampled values. Experimental verification shows that the above triple mechanism can suppress the influence of random plasma fluctuations on the eigenvector to within 3%, meeting the accuracy requirements of engineering measurements.

[0051] Step S3, Layer Determination and Coarse Parameter Adjustment: Before the equipment is put into use, the target composite material is calibrated offline to establish a composite material impedance characteristic fingerprint database. For the mobile phone shell in this embodiment, the calibration steps are as follows: Take a standard sample with a known layer structure, and execute a set of engraving calibration programs on the PET film surface, the adhesive layer surface exposed after manually peeling off the PET, and the aluminum alloy surface exposed after manually peeling off all the film and adhesive layers. For each calibration, collect F1 data of no less than 100 laser pulses, and calculate the mean μ and standard deviation σ of each dimension parameter. The fingerprint database entry records the range μ±3σ as the feature space of the material at that layer. In this embodiment, the fingerprint database contains three records:

[0052] In the actual processing, for each laser pulse, F1 obtained in step S2 is used to calculate its weighted Euclidean distance D with the fingerprint of the current layer in the fingerprint database:

[0053] Among them, the weighting coefficient =0.5、 =0.3、 =0.2, determined based on the sensitivity of each parameter to material type identification. Rp has the largest interlayer difference (the Rp difference between metals and organic materials is on the order of magnitude), so it has the highest weight; Cp is next; Lp has a relatively small interlayer difference, so it has the lowest weight.

[0054] When D is consistently less than a preset threshold (set to 1.5 in this embodiment, meaning F1 is within 1.5σ of the fingerprint space), it is determined that the current processing position is still in the current layer. When D increases to exceed the threshold and the weighted Euclidean distance between F1 and the fingerprint of the new layer is less than the threshold of the new layer for multiple consecutive pulse cycles, it is determined that the new layer region has been entered.

[0055] When a change in the determined layer is detected, a coarse parameter adjustment switch is performed within the first response time. In this embodiment, the first response time is 1μs to 10μs. For this embodiment, when switching from the PET coating layer to the adhesive layer, the laser parameters are switched from the first parameter group (peak power approximately 50W, repetition frequency 80kHz, scanning speed 1200mm / s) to the second parameter group (peak power approximately 25W, repetition frequency 60kHz, scanning speed 800mm / s) to accommodate the differences in ablation thresholds of different materials.

[0056] Coarse parameter adjustment switching is achieved through direct hardware connection of the FPGA to the laser acousto-optic modulator. After determining the layer change, the impedance analysis FPGA directly reads the corresponding parameter group from the preset parameter table at the hardware level and loads it into the acousto-optic modulator driver circuit, without going through upper computer software scheduling, thus controlling the response delay to the order of microseconds. The determination result of this step will serve as the basis for the subsequent transition zone control in step S5.

[0057] To accommodate different batches of coating materials or niche composite materials (such as impedance characteristic shifts caused by differences in filler formulations in PET film materials from different suppliers), the method in this embodiment also includes a user-defined extension step for the composite material impedance characteristic fingerprint library.

[0058] The operation procedure is as follows: The operator enters calibration mode and places 3-5 standard samples with known layer structures on the worktable. The equipment automatically executes the preset calibration engraving program—engraving a 20mm long straight line on each sample, using the equipment's default processing parameters. During the calibration engraving process, the system automatically collects F1 sample data for each pulse cycle and automatically determines the layer boundary points according to the order in which the material enters the system.

[0059] After engraving, the system performs K-means clustering analysis on the F1 sample dataset collected during the calibration engraving process for each material layer (the K value is pre-input by the operator, i.e., the known number of material layers). It calculates the cluster center value and standard deviation of each dimension for each layer of F1 sample data, thereby determining the variance range (μ±3σ) for each dimension. The cluster center value and variance range are then associated with the corresponding material layer information and stored as new material entries in the composite material impedance feature fingerprint database to adapt to the identification needs of different batches or niche composite materials.

[0060] In subsequent processing, when the matching degree between F1 and the newly added entry is higher than that with the existing entries, the system prioritizes using the custom entry for hierarchical determination. If the matching of the custom entry does not meet the confidence requirement, the system automatically falls back to the main fingerprint database for matching.

[0061] Step S4, Specific Implementation of Quality Deviation Judgment and Parameter Fine-tuning: During the period of layer stability (F1 is located within the fingerprint feature space of a certain layer), quality closed-loop fine-tuning is performed. First, it is necessary to determine the standard quality characteristic range for various materials. In this embodiment, F2 data is collected through multiple engraving experiments on standard samples, and the mean ± 3σ is taken as the standard range. When engraving the PET coating layer under standard parameters (peak power 50W, repetition frequency 80kHz), the standard F2 range is: Δ|Z|=3.0-3.8Ω, Δφ=1.0-2.5°.

[0062] This embodiment establishes a complete mass parameter mapping model through pre-calibration experiments, specifically including: The mapping relationship between Δ|Z| and the surface roughness Ra was established: Five standard PET-coated samples with varying roughness were prepared (each polished with sandpaper of different grits). The surface Ra of each sample was measured using a roughness meter (Mitutoyo SJ-210), yielding Ra values ​​of 0.15 μm, 0.28 μm, 0.42 μm, 0.55 μm, and 0.68 μm, respectively. Each sample was then laser-engraved, and the Δ|Z| data in F2 were collected, yielding the following corresponding relationship:

[0063] Linear fitting yielded the mapping relationship: Ra = 0.088 × Δ|Z| + 0.003 (correlation coefficient R² = 0.994). The standard Ra for the PET coating layer is set to 0.30 μm (to meet the requirements for electronic product appearance parts), corresponding to a standard Δ|Z| of 3.37 Ω.

[0064] The mapping relationship between Δφ and the microcrack density of the processed surface was established: Microcracks were observed in the above samples using a scanning electron microscope (SEM, 500x magnification). The number of independent microcracks with a length exceeding 5μm within a unit area (100μm × 100μm field of view) was used as the microcrack density index. The calibration data are as follows:

[0065] A mapping relationship was established through polynomial fitting. The standard microcrack density was set to no more than 2 cracks / 10000μm², corresponding to a standard value of Δφ of no more than 4°.

[0066] The physical mechanism of the mass parameter mapping model is as follows: (1) Linear relationship between Δ|Z| and surface roughness Ra: When the surface roughness Ra increases, the uniformity of the plasma sheath thickness distribution on the micro-undulating surface decreases, resulting in scattering loss of the high-frequency excitation signal at the sheath boundary. Experimental calibration shows that under the processing conditions of this application (Ra range 0.1μm~0.7μm), Δ|Z| and Ra exhibit a good linear relationship. By performing least squares linear fitting on the measurement data of five sets of calibration samples, the following empirical formula is obtained: Ra=0.088×Δ|Z|+0.003, goodness of fit R²=0.994.

[0067] (2) Polynomial relationship between Δφ and microcrack density: When microcracks appear on the processed surface, local plasma quenching channels are formed at the cracks, causing a sudden change in the equivalent capacitance of the plasma-material interface, which is reflected as a spike in the transient deviation Δφ of the high-frequency impedance phase angle. Experiments show that the microcrack density x (unit: cracks / 10000μm²) has a quadratic polynomial relationship with Δφ. By fitting the data of five sets of calibration samples with a quadratic polynomial, the following empirical formula is obtained: Δφ = 0.32x² + 0.85x + 1.2 The constant term 1.2 corresponds to the inherent phase noise at zero microcracks. The goodness of fit R² = 0.991, and the effective range of x is 0–8 cracks / 10000 μm². The prediction error of this polynomial was determined using leave-one-out cross-validation, with a mean absolute error of less than 0.3°, accurately characterizing the relationship between microcrack density and phase angle deviation. The coefficients in the above empirical formula were obtained using the PET coating material and corresponding processing parameters in Example 1 of this application. For different materials or processing parameters, the same calibration method should be used for refitting.

[0068] When the real-time acquired F2 deviates from the standard quality characteristic range, parameter fine-tuning correction is performed within the second response time, and the correction is applied to subsequent laser pulses. The second response time is less than the period of a single laser pulse and does not exceed 100ns. This value range is determined based on the data throughput rate of the impedance spectrum acquisition circuit and the parallel processing capability of the FPGA, ensuring that the quality detection result of each pulse is processed and the parameters are updated before the trigger signal of the next pulse arrives.

[0069] The roughness deviation is corrected using a PID control law:

[0070] in, =Ra actual(t) - Ra standard, which is the deviation between the Ra value calculated for the current pulse and the standard Ra value (0.30μm). Δt is taken as the pulse repetition period of 20μs. After experimental tuning, Kp=15W / μm, Ki=3W / (μm·ms), Kd=2W·ms / μm. The correction amount ΔP is directly added to the current preset laser peak power. When Ra exceeds the standard value, ΔP is negative (power reduction), and vice versa.

[0071] The correction for microcrack density uses a proportional control law: ΔV=K·e 密度 Where K = 50 mm / s·(strips / 10000 μm²) -1 When the microcrack density exceeds the standard, ΔV is a positive value (increasing the scanning speed to reduce the energy deposition density); conversely, the speed can be appropriately reduced to improve processing efficiency.

[0072] To ensure processing stability, the fine-tuning correction amount is limited as follows: |ΔP|≤5W (not exceeding 10% of the preset power), |ΔV|≤200mm / s (not exceeding 25% of the preset scanning speed).

[0073] Step S5, Specific Implementation of Coordinated Control of Interlayer Transition Zone: When the determination result of F1 in step S3 shows that the current processing position has entered the interlayer transition zone, that is, the weighted Euclidean distance between F1 and the fingerprint of the current layer begins to increase continuously (exceeding 80% of the threshold as a warning line), but the determination conditions of the new layer have not yet been met, the transition zone control strategy is triggered, and the collaborative control mode of step S5 is entered.

[0074] The core logic of transition zone control is to pause the layer-driven parameter switching operation in step S3 and only perform the quality-driven parameter fine-tuning correction in step S4. Specifically, although the F1 determination result in step S3 may indicate that "the layer is changing," the system does not perform cross-layer jump switching of the preset parameter group at this stage. Instead, it continuously fine-tunes the laser parameters based on the quality feedback from F2 in step S4 to maintain the processing quality in the transition zone. The layer determination program in step S3 continues to run in the background, but the determination result does not trigger parameter switching; it is only used to determine whether the transition zone has been left.

[0075] When performing step S4 in the interlayer transition zone, the preset standard quality feature range in step S4 is dynamically determined as follows: the standard quality feature range of each of the two adjacent layers is obtained by weighting and synthesizing based on the similarity distance between the current layer feature vector F1 and the fingerprint of the adjacent two layers.

[0076] The specific weighting formula is as follows: Standard range 过渡 =Standard Range A +(1-α)·Standard range B ; Among them, the standard range A The standard quality characteristic range for material A (current layer), standard range B The standard quality characteristic range of material B (target layer). The weighting coefficient α = D_B / (D_A+D_B), where D_A and D_B are the weighted Euclidean distances between the current F1 and the layer fingerprints of material A and material B, respectively.

[0077] The physical meaning of the above weighted synthesis method is that when the layer feature vector F1 is in the transition zone between material A and material B, its similarity distance with the fingerprints of the two layers reflects the relative position of the current processing position in the material space. The weight coefficient α = D_B / (D_A + D_B) satisfies the distance decay principle—the closer F1 is to material A (the smaller D_A is), the closer α is to 1, and the standard range is closer to the standard quality feature range of material A; the closer it is to material B (the smaller D_B is), the closer α is to 0, and the standard range is closer to the standard quality feature range of material B. This linear weighting method ensures that the quality benchmark in the transition zone transitions continuously and smoothly with the layer changes.

[0078] To verify the smoothness of this weighting method, the applicant conducted a comparative experiment: on the same multilayer composite board (PET film layer + adhesive layer + aluminum alloy substrate), engraving was performed using both the unweighted fixed standard range (i.e., directly using the standard range of material A until completely switching to material B) and the weighted synthesis method of this invention, and the surface roughness fluctuation of the transition zone was measured. The results show that: Unweighted method: The roughness Ra of the transition zone fluctuates within ±0.15 μm, and obvious step marks are visible at the interface; The weighting method of this invention: the roughness Ra of the transition zone fluctuates within a range of ±0.03 μm, and scanning electron microscopy (SEM) observation shows that the interface is smooth with no visible steps, and the smoothness is improved by about 80% ((0.15-0.03) / 0.15=0.8).

[0079] The above experiment was repeated on five different batches of samples, and the results were consistent, verifying that the weighted synthesis method can effectively ensure the smoothness of the processing quality in the transition zone.

[0080] As F1 gradually migrates from the feature space of material A to the feature space of material B, D_A gradually increases, D_B gradually decreases, and the weight α smoothly transitions from approaching 1 to approaching 0. The standard quality feature range also smoothly transitions from the range of material A to the range of material B. This calculation is updated once per pulse cycle, achieving dynamic tracking.

[0081] When F1 stably falls into the new layer region within multiple consecutive laser pulse cycles, the parameter switching operation in step S3 is resumed. In this embodiment, the specific implementation of this determination condition is as follows: a counting threshold N=5 is set. When the weighted Euclidean distance between the layer feature vector F1 calculated within 5 consecutive laser pulse cycles and the target layer fingerprint is less than a preset threshold (1.5), it is determined that the new layer region has been fully entered. The value of N ranges from 5 to 10. If the value of N is too small, it may lead to misjudgment due to plasma noise. If the value of N is too large, it will increase the dwell time in the transition zone.

[0082] After confirming entry into the new layer, resume the parameter switching operation in step S3, completely switch the laser parameters to the preset parameter group of the new layer, and end the transition zone control strategy.

[0083] Finally, to verify the beneficial effects of the technical solution in this embodiment, the following comparative experiment was conducted.

[0084] Experiment 1: Comparison of interlayer overburn width On the same batch of mobile phone casing samples (20μm PET coating + aluminum alloy substrate), logo engraving was performed using the traditional fixed parameter method, the layer identification method only (without quality closed loop and transition zone strategy), and the method of this invention, with 50 pieces engraved in each group. After engraving, metallographic sections were made along the edge of the logo, and the overburning width at the coating-adhesive layer interface and the adhesive layer-aluminum alloy interface was observed under an optical microscope (200x magnification). The results showed that the traditional fixed parameter method resulted in an overburning width of 35-65μm at the coating-adhesive layer interface and 28-52μm at the adhesive layer-aluminum alloy interface; the layer identification method only (without closed loop) resulted in overburning widths of 12-20μm and 8-18μm, respectively; the method of this invention controlled the overburning width at both interfaces to below 5μm, significantly outperforming the comparative methods. The transition zone collaborative control strategy effectively eliminated the step-like processing marks caused by the pulse group effect, and the interface transition was smooth with no visible steps under SEM observation.

[0085] Experiment 2: Comparison of surface roughness consistency in batch processing 200 samples were continuously engraved (running without interruption, simulating an actual production line). Every 20 samples, one sample was randomly selected to measure the surface roughness Ra of the logo's bottom surface. The Ra fluctuation range of the traditional fixed-parameter method is 0.22-0.48 μm (standard deviation ±0.09 μm), while the Ra fluctuation range of the method of this invention is 0.28-0.33 μm (standard deviation ±0.02 μm). The closed-loop quality tuning effectively compensates for the power deviation caused by thermal drift during long-term laser operation, achieving high-quality consistency.

[0086] Experiment 3: Verification of Environmental Compensation Effect Tests were conducted in an environmental test chamber, where impedance measurements were performed on standard samples under three environmental conditions: 20℃ / 30%RH, 25℃ / 50%RH, and 32℃ / 70%RH. Without environmental compensation, the maximum deviation of the Rp measurement reached 28Ω (approximately 6% deviation); with environmental compensation enabled, the maximum deviation of the Rp measurement decreased to 3Ω (approximately 0.7% deviation), effectively eliminating the adverse effects of environmental fluctuations on the accuracy of stratigraphic determination.

[0087] Example 2: Blind via fabrication in multilayer PCBs This embodiment uses laser engraving of blind vias on a four-layer printed circuit board as an example. The workpiece has a multi-layer composite structure: from top to bottom, it consists of a surface copper foil layer (18μm thick), an FR4 epoxy glass cloth insulating layer (100μm thick), an inner copper foil layer (35μm thick), and an FR4 insulating layer (100μm thick). The processing requirement is to start engraving from the surface copper foil, just penetrating the FR4 insulating layer and stopping at the surface of the inner copper foil, forming blind vias for interlayer interconnection.

[0088] The method steps in this embodiment are the same as those in Embodiment 1, with the main difference being the material type and the corresponding parameter settings.

[0089] In step S1, the engraving head structure is the same as in Example 1, and the excitation and acquisition timing are also the same as in Example 1.

[0090] In step S2, impedance fingerprints of the copper foil layer and the FR4 insulating layer are established through offline calibration. The Rp range of the copper foil layer F1 is 15-40Ω, the Cp range is 0.3-0.8pF, and the Lp range is 5-12nH; the Rp range of the FR4 insulating layer F1 is 350-480Ω, the Cp range is 2.0-3.2pF, and the Lp range is 65-85nH. The differences between the two are significant, demonstrating good layer location identification capability. The environmental parameter compensation steps are the same as in Example 1.

[0091] In step S3, the fingerprint database contains four records, corresponding to the surface copper foil, FR4 insulating layer, inner copper foil, and bottom FR4, respectively. During processing, the laser parameters are initially set to the copper foil layer parameters (peak power approximately 80W to penetrate the copper foil). When F1 indicates that it has penetrated the copper foil and entered the FR4 insulating layer, it switches to the insulating layer parameters (peak power approximately 40W to ablate FR4) within the first response time. When F1 again indicates that it has penetrated the FR4 and entered the inner copper foil, it switches to the stop processing or low-energy cleaning parameters within the first response time. The fingerprint database user-defined expansion steps are the same as in Example 1.

[0092] In step S4, quality parameter mapping models are established for the copper foil layer and the FR4 insulating layer, respectively. The standard Ra of the copper foil layer is set to 0.4 μm, and the standard microcrack density is set to no more than 3 cracks / 10000 μm²; the standard Ra of the FR4 insulating layer is set to 1.0 μm, and the standard microcrack density is set to no more than 5 cracks / 10000 μm². The parameter fine-tuning and correction are performed in the same way as in Example 1.

[0093] In step S5, in the interface transition zone of FR4-inner copper foil, a collaborative control strategy is initiated: layer drive parameter switching is paused, and only quality fine-tuning is performed. The quality reference standard for the transition zone is a weighted synthesis of the standard quality characteristic ranges of FR4 and the inner copper foil, maintaining interface flatness. Once F1 stably falls into the inner copper foil region within N=6 consecutive pulse cycles, parameter switching is resumed.

[0094] After processing, the bottom of the blind hole was inspected by slicing. There were no overheated depressions on the copper foil surface, and the sidewalls of the blind hole were smooth.

[0095] Example 3: Engraving decorative patterns on the surface of automotive interior leather This embodiment uses the surface decorative texture engraving of an automotive door panel interior component as an example. The workpiece is a soft, multi-layered composite structure: the surface layer is 0.8mm thick PU (polyurethane) leather, and the bottom layer is a polyester fabric substrate. The processing requirements are: the engraving depth should just penetrate the PU leather layer to form the decorative texture, but without damaging the polyester fabric substrate, ensuring a soft feel and no deformation.

[0096] The method steps in this embodiment are the same as those in Embodiment 1, with the main difference being the material type and the corresponding parameter settings.

[0097] In step S1, the engraving head structure is basically the same as in Example 1. To accommodate slight deformation of soft materials, the vertical distance between the lower end face of the annular induction electrode and the workpiece surface is adjusted to 2.5 mm to 4.0 mm to accommodate the flexible undulations of the material. The excitation and acquisition timing sequence is the same as in Example 1.

[0098] In step S2, impedance fingerprints for the PU leather layer and the polyester fabric layer are established through offline calibration. The Rp range of the PU leather layer F1 is 380-510Ω, the Cp range is 2.2-3.0pF, and the Lp range is 72-88nH; the Rp range of the polyester fabric layer F1 is 220-310Ω, the Cp range is 1.2-1.9pF, and the Lp range is 35-50nH. The environmental parameter compensation steps are the same as in Example 1.

[0099] In step S3, the fingerprint database contains two records, corresponding to the PU leather layer and the polyester fabric layer, respectively. During processing, the initial parameters are set to the parameters specific to the PU leather layer (peak power approximately 15W to prevent the leather from scorching). When F1 indicates that it has entered the polyester fabric substrate layer, the parameters are switched to the fabric layer parameters (peak power approximately 8W, producing only visual texture without cutting the fibers) within the first response time. The user-defined expansion steps for the fingerprint database are the same as in Example 1.

[0100] In step S4, quality parameter mapping models are established for the PU leather layer and the polyester fabric layer, respectively. The standard Ra of the PU leather layer is set to 0.8 μm, and the standard microcrack density is set to no more than 4 cracks / 10000 μm². The parameter fine-tuning and correction are performed in the same way as in Example 1.

[0101] In step S5, a collaborative control strategy is initiated in the PU leather-polyester fabric interface transition zone. The quality reference standard for the transition zone is a weighted synthesis of the standard quality characteristic ranges of the PU leather and polyester fabric respectively. Once F1 stably falls within the polyester fabric layer region for N=5 consecutive pulse cycles, the execution parameter switching resumes.

[0102] After processing, the decorative patterns have clear edges, the underlying fabric fibers are intact, and the fabric is soft to the touch and does not deform.

[0103] In one embodiment of the invention, the laser engraving pulse is generated by a nanosecond or picosecond pulsed fiber laser. The first response time in step S3 is 1 μs to 10 μs, a range determined by combining typical parameters of current industrial-grade pulsed fiber lasers with the response speed of the acousto-optic modulator. The second response time in step S4 is less than the period of a single laser pulse and does not exceed 100 ns. This range is determined by combining the data throughput rate of the impedance spectrum acquisition circuit and the parallel processing capability of the FPGA, ensuring that the quality detection result of each pulse is processed and parameters are updated before the trigger signal of the next pulse arrives.

[0104] The above numerical ranges can be adjusted according to the specific hardware configuration in practical applications, but they should all meet the timing constraint that the second response time is less than the first response time, so as to ensure that the response of the quality fine-tuning closed loop takes precedence over the response of the layer coarse-tuning closed loop, thereby ensuring that the quality fine-tuning has made the necessary corrections to the current processing state before a large parameter change occurs during layer switching.

[0105] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit the scope of protection of the invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on these embodiments, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art can still combine, add, delete, or otherwise adjust the features of the various embodiments of the present invention according to the circumstances without conflict or creative effort, thereby obtaining different technical solutions that do not fundamentally depart from the concept of the present invention. These technical solutions also fall within the scope of protection of the present invention.

Claims

1. A method for laser engraving on the surface of a product, characterized in that, Includes the following steps: S1: During the duration of the laser engraving pulse acting on the surface of the composite heterogeneous material, a low-frequency sweep frequency excitation signal and a high-frequency fixed frequency excitation signal are alternately applied to the plasma region generated by the laser in a preset sequence through a ring induction electrode coaxially set at the exit of the engraving head, and the low-frequency impedance response of the plasma to the low-frequency sweep frequency excitation signal and the high-frequency impedance response to the high-frequency fixed frequency excitation signal are collected respectively. S2: Perform equivalent circuit model fitting and solution on the low-frequency impedance response to obtain the layer feature vector F1, which is used to characterize the material layer to which the current processing position belongs; perform time-domain feature extraction and solution on the high-frequency impedance response to obtain the quality feature vector F2, which is used to characterize the micro-state of the current processing surface. S3: Perform similarity matching between the layer feature vector F1 and the preset composite material impedance feature fingerprint database to determine the material layer at the current processing position; when the determined layer changes, switch the laser engraving parameters to the preset parameter group corresponding to the material layer in the first response time. S4: Compare the quality feature vector F2 with the preset standard quality feature range. When F2 deviates from the standard quality feature range, adjust and correct the subsequent laser engraving parameters in real time according to the preset quality parameter mapping model within the second response time. The second response time is less than the first response time. S5: When the layer feature vector F1 shows that the current processing position is in the interlayer transition zone between two layers of material, pause the layer driving parameter switching operation in step S3, and only perform the quality driving parameter fine-tuning correction in step S4. The parameter switching operation in step S3 is resumed only after F1 has stably fallen into the new layer region within several consecutive laser pulse cycles.

2. The laser engraving method for a product surface according to claim 1, characterized in that, In step S1: The low-frequency sweep excitation signal has a frequency range of 1MHz to 10MHz and is used to detect the bulk conductivity and dielectric properties of the plasma. The high-frequency fixed-frequency excitation signal has a frequency range of 30MHz to 50MHz and is used to detect the uniformity and stability of the plasma sheath distribution.

3. The laser engraving method for a product surface according to claim 1, characterized in that, In step S2: The layer feature vector F1 includes at least the plasma resistance Rp, sheath capacitance Cp, and equivalent inductance Lp in the plasma equivalent circuit model. The quality feature vector F2 includes at least the root mean square fluctuation of the impedance amplitude Δ|Z| and the transient deviation of the phase angle Δφ in the high-frequency impedance response.

4. The laser engraving method for a product surface according to claim 1, characterized in that, Step S2 further includes an environmental parameter compensation step, specifically: The temperature (T), relative humidity (RH), and atmospheric pressure (P) of the processing environment are collected in real time by a miniature temperature, humidity, and pressure composite sensor integrated into the engraving head. Based on the pre-calibrated environmental-impedance compensation model, the impedance spectrum correction ΔZ(T,RH,P) corresponding to the current environmental parameters is calculated. The impedance spectrum correction amount ΔZ is used to compensate and correct the acquired low-frequency impedance response and high-frequency impedance response respectively, so as to eliminate the influence of environmental fluctuations on the accuracy of impedance feature extraction.

5. The laser engraving method for a product surface according to claim 1, characterized in that, It also includes user-defined extension steps for the composite material impedance feature fingerprint library, specifically: Enter calibration mode and execute a preset calibration engraving program on a standard sample with a known layer structure; Collect F1 sample data of the layer feature vectors corresponding to each layer of material during the calibration and engraving process, and calculate the cluster center value and variance range of each layer of F1 sample data; The cluster center value and variance range are associated with the corresponding material layer information and stored as new material entries in the composite material impedance feature fingerprint database to adapt to the identification needs of different batches or niche composite materials.

6. The laser engraving method for a product surface according to claim 1, characterized in that, The specific criteria for determining whether F1 stably falls into the new layer region within multiple consecutive laser pulse cycles in step S5 are as follows: Set a counting threshold N, with N ranging from 5 to 10; when the weighted Euclidean distance between the layer feature vector F1 calculated within N consecutive laser pulse cycles and the target layer fingerprint is less than the preset threshold, it is determined that the new layer region has been fully entered.

7. The laser engraving method for a product surface according to claim 1, characterized in that, The quality parameter mapping model in step S4 is established through pre-calibration experiments, specifically including: Establish the mapping relationship between Δ|Z| and the surface roughness Ra in the quality feature vector F2, and the mapping relationship between Δφ and the microcrack density on the surface. Establish a PID control law for the deviation between the actual roughness value and the standard value and the laser power correction ΔP, and a proportional control law for the deviation between the actual microcrack density value and the standard value and the scanning speed correction ΔV.

8. The laser engraving method for a product surface according to claim 1, characterized in that, Step S1 also includes an electromagnetic compatibility control step, specifically: The inner ring of the annular induction electrode is configured as a grounded shielding ring to bypass the high-voltage transient electric field generated by the laser pulse to the ground. The transmission paths of the low-frequency sweep excitation signal and the high-frequency fixed-frequency excitation signal adopt double-shielded coaxial cable, and an active low-pass filter with a cutoff frequency of 500MHz is connected in series at the signal acquisition front end.

9. The laser engraving method for a product surface according to claim 1, characterized in that, The composite heterogeneous material is a multilayer composite structure, specifically including at least one of the following types: Metal or alloy substrates with an organic polymer film coating on their surface; Printed circuit board with alternating layers of copper foil and insulating layers; A soft interior material composed of a leather top layer and a fabric backing layer.

10. A method for laser engraving on a product surface according to claim 1, characterized in that, In step S5, when step S4 is executed in the interlayer transition zone, the preset standard quality feature range is obtained by weighting the standard quality feature ranges of the two adjacent layers by using the similarity distance between the current layer feature vector F1 and the fingerprints of the adjacent two material layers as weights.

Images