Laser welding parameter optimization method and system for aerosol can welding seam

Abstract

The embodiment of the invention provides a laser welding parameter optimization method and system for an aerosol can welding seam, and relates to the technical field of aerosol can welding parameter optimization. The method comprises the steps that thermodynamic state data of a weld pool of a weld joint in the aerosol can production welding process are collected in real time; dynamically quantifying a deviation signal of a welding state based on a preset reference thermodynamic template and the thermodynamic state data acquired in real time; and based on the deviation signal, closed-loop feedback adjustment is conducted on the laser welding parameters in the production welding process. By means of the method and device, the problem that welding parameters cannot be dynamically adjusted is solved, and then the effect of improving the welding quality is achieved.

Description

Technical Field

[0001] This invention relates to the field of optimizing welding parameters for insecticide aerosol cans, specifically to a method and system for optimizing laser welding parameters for aerosol can welds. Background Technology

[0002] Insecticide aerosol cans are widely used pressure vessels in daily life. Their interiors typically contain corrosive chemical solvents, propellants, and active ingredients. These corrosive contents place extremely stringent requirements on the long-term sealing and chemical stability of the can's welds. Any minute defects in the welds, such as pinholes, incomplete penetration, or microcracks, can not only cause product failure but may also become the starting point for corrosion during long-term storage, ultimately leading to leakage and posing safety and environmental risks. Laser welding, due to its high speed and efficiency, is the preferred process for the large-scale production of these aerosol cans.

[0003] However, existing laser welding processes generally employ open-loop control, meaning welding parameters are set once before production. These static parameters cannot effectively handle dynamic disturbances during production, such as batch-to-batch material variations and surface condition fluctuations, easily leading to unstable welding quality. The microscopic defects resulting from these inconsistencies are precisely the fatal hidden dangers affecting the long-term reliability of insecticide aerosol cans. Summary of the Invention

[0004] This invention provides a method and system for optimizing laser welding parameters for aerosol can welds, thereby at least solving the problem of the inability to dynamically adjust parameters in related technologies.

[0005] According to an embodiment of the present invention, a method for optimizing laser welding parameters of aerosol can welds is provided, comprising:

[0006] Real-time acquisition of thermodynamic state data of the weld pool during the welding process of aerosol can production;

[0007] Based on a preset reference thermodynamic template and the real-time collected thermodynamic state data, the deviation signal of the welding state is dynamically quantified, wherein the reference thermodynamic template corresponds to the standard welding process of the aerosol can.

[0008] Based on the deviation signal, the laser welding parameters in the production welding process are adjusted using closed-loop feedback.

[0009] In one exemplary embodiment, the deviation signal of the dynamically quantified welding state includes:

[0010] Based on the real-time position of the laser welding head on the weld, the corresponding reference thermodynamic state is extracted from the reference thermodynamic template.

[0011] The real-time acquired thermodynamic state data is compared with the reference thermodynamic state to generate the deviation signal.

[0012] In an exemplary embodiment, the step of adjusting the laser welding parameters using closed-loop feedback based on the deviation signal includes:

[0013] The deviation signal is input to a preset PID controller to generate a feedback adjustment amount to compensate for the deviation signal;

[0014] The laser welding parameters are adjusted based on the feedback adjustment amount.

[0015] In one exemplary embodiment, the method further includes:

[0016] Obtain the preset geometric feature information of the aerosol can;

[0017] Before the laser welding head reaches the weld position corresponding to the geometric feature information, a feedforward compensation amount is generated;

[0018] The feedforward compensation is superimposed on the adjustment process of the laser welding parameters to suppress welding disturbances caused by the geometric feature information in advance.

[0019] In one exemplary embodiment, the method further includes:

[0020] Obtain a sequence of real-time temperature field distribution images of the molten pool;

[0021] The real-time temperature field distribution image sequence is analyzed to extract a first feature of the molten pool, the first feature including the area of ​​the molten pool, the sloshing frequency of the molten pool, and at least one of the welding spatter events;

[0022] Based on the first feature, welding abnormalities during the welding process are determined.

[0023] In one exemplary embodiment, identifying welding abnormalities during the welding process includes:

[0024] Based on the area of ​​the molten pool, determine the area stability factor;

[0025] Based on the sloshing frequency of the molten pool, the oscillation stability factor is determined;

[0026] Based on the aforementioned welding spatter events, a spatter stability factor is determined;

[0027] The welding stability index is determined based on the area stability factor, the oscillation stability factor, and the spatter stability factor.

[0028] The welding abnormality is determined based on the welding stability index.

[0029] In one exemplary embodiment, the method further includes:

[0030] Based on the type of welding abnormality, a laser parameter correction sequence is determined from a preset correction strategy library;

[0031] The laser parameter correction sequence is superimposed on the laser welding parameter adjustment process to suppress anomalies.

[0032] According to another embodiment of the present invention, a laser welding parameter optimization system for aerosol can welds is provided, comprising:

[0033] The state perception module is used to collect real-time thermodynamic state data of the weld pool during the welding process of aerosol can production.

[0034] The control module is used to dynamically quantify the deviation signal of the welding state based on a preset reference thermodynamic template and the real-time acquired thermodynamic state data, wherein the reference thermodynamic template corresponds to the standard welding process of the aerosol can.

[0035] The welding execution module is used to perform closed-loop feedback adjustment of the laser welding parameters in the production welding process based on the deviation signal.

[0036] According to yet another embodiment of the present invention, a computer-readable storage medium is also provided, wherein a computer program is stored therein, wherein the computer program is configured to perform the steps in any of the above method embodiments when executed.

[0037] According to yet another embodiment of the present invention, an electronic device is also provided, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0038] This invention achieves instantaneous adjustment of the welding process by real-time monitoring of the thermodynamic state of the weld pool and comparing it with a reference template. This fundamentally solves the problem that static parameters cannot cope with dynamic disturbances. Therefore, it can solve the problem of not being able to adaptively and dynamically adjust welding parameters, thereby improving welding quality. Attached Figure Description

[0039] Figure 1 This is a flowchart of a method for optimizing laser welding parameters for aerosol can welds according to an embodiment of the present invention;

[0040] Figure 2 This is a structural block diagram of a laser welding parameter optimization system for aerosol can welds according to an embodiment of the present invention;

[0041] Figure 3 This is a simulation diagram of the stability factor and welding stability index according to an embodiment of the present invention;

[0042] Figure 4 This is a schematic diagram of the light spot oscillation according to an embodiment of the present invention;

[0043] Figure 5 This is a high-frequency acoustic signal diagram of the welding process according to an embodiment of the present invention. Detailed Implementation

[0044] The technical solutions of the embodiments of this application will be described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0045] In the following description, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0046] Furthermore, in this application, directional terms such as "upper," "lower," "left," and "right" may be defined relative to the orientation of the components shown in the accompanying drawings. It should be understood that these directional terms can be relative concepts, used for relative description and clarification, and may change accordingly depending on the orientation of the components in the accompanying drawings.

[0047] In this application, unless otherwise expressly specified and limited, the term "connection" should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral part; it can be a direct connection or an indirect connection through an intermediate medium. Furthermore, the term "coupled" can refer to an electrical connection that enables signal transmission.

[0048] As used herein, “about,” “approximately,” or “approximately” includes the stated value and the average value within an acceptable range of deviation from the given value, wherein the acceptable range of deviation is determined by a person skilled in the art taking into account the measurement under discussion and the error associated with the measurement of the given quantity (i.e., the limitations of the measurement system).

[0049] Example 1

[0050] This embodiment provides a method for optimizing laser welding parameters for aerosol can welds. This method constructs a closed-loop feedback control loop based on the real-time thermodynamic state of the molten pool, which can dynamically adjust the laser welding parameters during the production process. This solves the technical problem in the prior art where static welding parameters cannot adapt to production disturbances such as material batch fluctuations and equipment status changes, thus leading to inconsistent welding quality. This method achieves the beneficial effect of significantly improving the yield and reliability of welds.

[0051] Reference Figure 1 The method described in this embodiment may include the following steps:

[0052] S100: Construct a reference thermodynamic template corresponding to the standard welding process for aerosol cans.

[0053] In this embodiment, the reference thermodynamic template RTP is a dynamic dataset associated with the spatial location of the weld, which essentially depicts the evolution parameters of key thermodynamic characteristics (such as temperature distribution and geometry) of the molten pool as it moves along the entire weld path under optimal welding conditions.

[0054] Specifically, the first step is sample preparation and calibration. This involves selecting standard aerosol cans that meet the highest quality standards in terms of size, material grade (e.g., DX51D galvanized steel sheet, thickness 0.2 mm), and surface condition as standard samples. Subsequently, these standard samples are calibrated using preset standard welding parameters that have been experimentally verified and can stably produce high-quality welds. These standard parameters can be set as follows: the laser uses continuous wave mode with an average power of 800 watts (W), the laser spot diameter is 0.3 mm, and the welding scanning speed is 100 mm / s.

[0055] Throughout the calibration welding process, thermodynamic data acquisition is required. Specifically, a high-speed, high-resolution infrared thermal imager, mounted coaxially or off-axis with the laser welding head, is strictly synchronized with the position encoder of the laser scanning galvanometer via hardware triggering. This thermal imager continuously captures two-dimensional temperature field distribution images of the weld pool at a rate of, for example, 2000 frames per second (fps). Each frame is associated with a weld position coordinate. Correspondingly, this yields one or more datasets representing the full thermodynamic state of the molten pool throughout an ideal or standard welding process. Each element of this dataset is a data pair. ,in It is in position The size collected at the location is, for example Temperature matrix of pixels.

[0056] Finally, the reference template is generated and stored. This process involves complex signal processing of the large amount of raw thermal image data to extract stable and representative features. To eliminate random noise in a single sample or single measurement, moving average filtering in the time series or Gaussian filtering in the spatial domain can be used. Furthermore, to accurately quantify the geometry of the molten pool, this embodiment employs a three-dimensional Gaussian surface fitting algorithm, which converts the discrete temperature matrix... Fit to a continuous two-dimensional Gaussian function:

[0057]

[0058] in, It is the peak amplitude of the Gaussian function. These are the subpixel coordinates of the center of the molten pool. and These represent the width (standard deviation) of the molten pool in the two principal axis directions, respectively. This is the background temperature; by solving the parameters of this model using the least squares method, a more accurate peak temperature of the molten pool can be obtained than by directly searching for the maximum value in the pixel matrix. ), center position and geometric dimensions.

[0059] The template is stored in the non-volatile memory of the data processing and control module in the form of an efficient lookup table (LUT). The data structure of the lookup table is designed as follows:

[0060] struct RtpPoint { float position; float peak_temp; float area_ref;...}

[0061] This array of structures records each discrete location point along the entire weld path at intervals of, for example, 0.1 mm. The corresponding ideal molten pool state parameter vector.

[0062] S200: Real-time acquisition of thermodynamic state data of the weld pool during the production and welding process of aerosol cans.

[0063] On an automated production line, once an aerosol can to be welded is precisely positioned at the welding station, the system starts the welding process; at the same time, the status sensing module deployed on the welding head also starts working simultaneously.

[0064] In this embodiment, the state perception module is a high-speed infrared thermal imager in the near-infrared (NIR) band specifically designed for high-temperature measurement. Its temperature measurement range covers 800 to 2000 degrees Celsius, and its sampling frequency is set to a very high level, such as 1000 Hz. This means that every 1 millisecond (ms), the system can acquire a new snapshot of the molten pool state. The data acquisition process is strictly synchronized with the movement of the laser welding head. This high-frequency, synchronous acquisition method ensures that the control system can respond to any tiny and rapid changes in the welding molten pool with extremely high spatiotemporal resolution.

[0065] S300: Based on a reference thermodynamic template and real-time acquired thermodynamic state data, dynamically quantifies the deviation signal of the welding state.

[0066] In this embodiment, the process is executed cyclically at a very high frequency within the data processing and control module. For each real-time data packet received from S200, the controller first utilizes the real-time location information... It queries and interpolates within its internally stored reference thermodynamic template (RTP) to calculate the position. The corresponding reference thermodynamic state .

[0067] After acquiring the reference state, the controller immediately calculates the deviation signal; the deviation signal is defined as the difference between the real-time peak temperature of the molten pool and the reference peak temperature of the molten pool, to intuitively reflect the degree of deviation of the current welding energy input, i.e.:

[0068]

[0069] S400: Based on the deviation signal, it performs closed-loop feedback adjustment of laser welding parameters during the production welding process.

[0070] In this embodiment, the quantized deviation signal is converted into specific control actions for the physical device (laser) to complete the closed loop. This embodiment adopts a composite adjustment strategy that combines a proportional-integral-derivative PID controller with feedforward control to achieve fast and stable control.

[0071] PID feedback control generates a deviation signal in each control cycle. The input is fed into a digital PID controller, which calculates the feedback adjustment by executing a discrete-time positional PID control algorithm. To ensure optimal performance of the PID controller, its three key parameters... The Ziegler-Nichols tuning method is used to perform systematic calibration offline.

[0072] Building upon this, a feedforward compensation mechanism is introduced to address predictable disturbances. In this case, the controller obtains the preset geometric feature information of the aerosol can from the geometric information input unit. When the controller anticipates that the laser welding head is about to reach a reinforcing rib structure, it calculates a positive power compensation amount in advance. .

[0073] Ultimately, the controller will use the feedback adjustment calculated by the PID controller. and feedforward compensation By performing algebraic superposition, the total power adjustment is obtained, and the new laser power command is calculated as follows:

[0074]

[0075] This new power command is immediately sent to the laser's power controller, completing one closed-loop adjustment, and so on.

[0076] Example 2

[0077] Reference Figure 2 This application also provides a laser welding parameter optimization system for aerosol can welds. This system is designed to perform all the steps in the foregoing method embodiments, and the system can be physically or logically divided into the following modules:

[0078] Execution module 10: contains a high-power fiber laser and its high-speed two-dimensional scanning galvanometer system.

[0079] Status perception module 20: The core is a high-speed infrared thermal imager.

[0080] Template building module 30: Logically responsible for offline construction of benchmark thermodynamic templates.

[0081] Control module 40: Implemented by a field-programmable gate array (FPGA); the parallel processing architecture of the FPGA enables it to process pixel data streams, perform RTP lookup and interpolation operations, and run PID and feedforward control algorithms simultaneously with nanosecond-level latency; in addition, the module also includes a geometric information input unit 41 to provide input for feedforward control.

[0082] Example 3

[0083] Traditional PID control can only maintain macroscopic stability, but it responds poorly to specific, transient defect patterns (such as porosity and spatter). To solve this problem, unlike Embodiment 1, this embodiment integrates a parallel second control loop to actively diagnose and correct microscopic welding defects. Specifically, it includes the following steps executed in parallel with PID control to jointly influence the laser welding parameters:

[0084] S500: Real-time extraction of molten pool morphology and dynamic features

[0085] In this embodiment, multi-dimensional features that characterize the health of the welding process are extracted from the acquired real-time temperature field distribution image sequence. These features go beyond a single peak temperature and aim to capture the geometry and dynamic behavior of the molten pool. This step is implemented in the welding quality diagnosis module of the control module 40 through a dedicated image processing pipeline.

[0086] Specifically, such as Figure 3 As shown, this module calculates the following features in real time:

[0087] Molten pool area By segmenting the temperature image according to a preset temperature threshold (e.g., the melting point of steel is 1650K), the number of pixels above the threshold is calculated, which is the area of ​​the molten pool; this feature directly reflects the size of the molten pool.

[0088] Spectral characteristics of molten pool sloshing : A sequence of the melt pool area over a recent period (e.g., the last 128 frames, or 128 milliseconds). Treating it as a time signal, performing a Fast Fourier Transform (FFT) on it yields a spectrum that can reveal whether the molten pool experiences periodic oscillations; the features extracted here... It refers to the maximum spectral amplitude in the critical frequency band (e.g., 100 Hz to 200 Hz) associated with keyhole instability.

[0089] Welding spatter incident rate Spatter manifests as small, bright, isolated hot spots that separate from the main molten pool region. The diagnostic module identifies these events using image differencing and morphological analysis algorithms; the features extracted here... It refers to the average number of splash events detected per unit of time within a relatively short time window (e.g., 100 milliseconds) in the past, measured in "events per second".

[0090] S600: Identification and Correction Strategy Selection for Welding Abnormalities

[0091] Based on the extracted features, a nonlinear comprehensive evaluation model is used to quantitatively assess the current welding state and match corresponding correction strategies. This model treats the stability of the welding process as a joint probability that multiple independent physical processes do not fail; therefore, severe instability in any dimension will lead to a sharp decrease in overall stability. Specifically:

[0092] S610: Calculation of Welding Stability Index (WPSI)

[0093] like Figure 3As shown, WPSI is a dimensionless exponent between 0 (completely unstable) and 1 (absolutely stable), and its calculation formula is the product of three independent stability factors:

[0094]

[0095] in, These are the area stability factor, oscillation stability factor, and splash stability factor, respectively.

[0096] 1. Area stability factor Used to characterize the deviation of the weld pool area from its ideal value; since the ideal welding process should maintain the reference area defined by the reference template. Since it is nearby, this factor can be constructed as a Gaussian function centered on the reference area:

[0097]

[0098] in, It is an area tolerance parameter, which characterizes the system's tolerance to area fluctuations.

[0099] For example, suppose at location Reference area Area tolerance parameter in units of 250 pixels It is set to 10% of the reference area, that is Pixel unit; if the area being measured in real time If the unit is 270 pixels, then the area stability factor is calculated as follows:

[0100] .

[0101] 2. Oscillation stability factor Used to characterize the dynamic stability of the weld pool; since a healthy welding process should not have violent periodic oscillations, this factor can be constructed as an inverse S-shaped function (inverse Sigmoid function). When the oscillation characteristic value is below a critical threshold, the factor is close to 1, and drops rapidly to 0 after exceeding the threshold.

[0102]

[0103] in, It is the critical threshold for the spectral amplitude within the key frequency band. It is the slope parameter that controls the steepness of the curve.

[0104] For example, the critical spectral amplitude of keyhole instability was determined experimentally. The slope parameter is set to 0.8 (in any unit). Set to 10; if the currently measured peak spectral value If the value is 0.85, then the oscillation stability factor is calculated as follows:

[0105] .

[0106] 3. Splash stability factor This factor is used to characterize the mass loss and process instability caused by splashing; since the higher the incidence of splashing events, the worse the stability, this factor can be constructed as an exponential decay function:

[0107]

[0108] in, It is the splash sensitivity coefficient.

[0109] For example, the splash sensitivity coefficient It is set to 0.5. If the currently measured splash event rate is... The splash stability factor is calculated as follows: (number of splashes per second)

[0110] .

[0111] Therefore, by combining the WPSI calculation and anomaly identification, the three independently calculated factors mentioned above are multiplied together to obtain the final comprehensive welding stability index.

[0112] At this time, the system presets the WPSI trigger threshold. Since the currently calculated WPSI value (0.101) is lower than the threshold, the system identifies that the current welding process is in a serious "abnormal state" and so on.

[0113] Furthermore, by analyzing which of the three factors has the lowest contribution (in this case, it is...) and The system can infer the main causes of the abnormal state (keyhole instability and splashing) and select an optimal correction strategy accordingly.

[0114] S700: Execution of laser parameter correction sequence

[0115] In this embodiment, once the calculated WPSI index is lower than a preset threshold, the system selects a laser parameter correction sequence CAP corresponding to the abnormal state type from the preset correction strategy library and immediately superimposes it onto the current welding parameters. It should be noted that CAP is not a simple numerical adjustment, but a pre-arranged parameter dynamic modulation waveform that lasts for several milliseconds to tens of milliseconds.

[0116] For example, to address the excessively low WPSI index caused by a combination of keyhole instability and splashing, the system may select a composite CAP whose parameter modulation includes:

[0117] Laser power: A sinusoidal modulation with a frequency of 500 Hz and an amplitude of 15% of the base power is superimposed on the base power (determined by the PID circuit) to stabilize the keyhole.

[0118] Scanning galvanometer (spot oscillation): such as Figure 4 As shown, a circular oscillation with a radius of 0.1 mm and a frequency of 800 Hz is superimposed on the original straight scanning path to improve the fluidity of the molten pool and suppress splashing.

[0119] After the CAP is completed, the system returns to a control mode dominated by PID and continues to monitor the welding status through the WPSI model.

[0120] At the system level, the internal logic of the control module 40 in this embodiment is more complex. It integrates a welding quality diagnosis module and a correction strategy execution module in parallel. The welding quality diagnosis module is responsible for implementing the functions of S500 and S600. It has an independent image processing pipeline, an FFT calculation core, and a WPSI calculation submodule. The correction strategy execution module stores multiple CAPs waveform data and, based on the trigger signal from the diagnosis module, performs real-time hardware-level superposition of the corresponding modulation signal with the signal of the main control path, ultimately outputting a composite control command.

[0121] Example 4

[0122] Laser welding (especially the keyhole effect in deep penetration welding) is a violent physical process that generates intense plasma and emits broadband acoustic signals. Normal and stable welding processes produce acoustic signals with relatively stable spectral characteristics. However, when a burn-through defect occurs, the laser energy penetrates the workpiece, opening the bottom of the keyhole, and high-temperature, high-pressure metal vapor and plasma are instantly ejected towards the back of the workpiece. Traditional post-weld inspection methods, such as water bath and differential pressure methods, cannot detect burn-through defects in real time; inspection can only be performed after welding is complete, impacting production efficiency.

[0123] In response, this embodiment, based on all the aforementioned embodiments, achieves real-time online detection and prediction of fatal defects such as weld burn-through that directly lead to leakage by performing in-depth analysis of the physical signals generated during the welding process.

[0124] Specifically, the following steps are included:

[0125] S800: Synchronous acquisition of high-frequency acoustic signals during welding process

[0126] In this embodiment, one or more highly sensitive acoustic sensors are installed near the welding head, in a location not directly impacted by spatter. Preferably, a broadband acoustic emission sensor AE is used, made of piezoelectric ceramic (PZT), with a response frequency range covering 100 kHz to 1 MHz. This sensor is connected to a high-speed data acquisition card (DAQ) via a low-noise preamplifier. The DAQ's analog-to-digital converter (ADC) has a resolution of at least 12 bits and a sampling rate of at least 2 megabits per second (MS / s) to ensure distortion-free capture of microsecond-level acoustic impact events. The trigger input of the DAQ is synchronized with the laser's output signal to ensure that the acoustic signal acquisition strictly corresponds to the welding process. The acquired high-frequency acoustic signal time series... It is transmitted in real time to the dedicated signal processing module in the control module 40.

[0127] For example, such as Figure 5 As shown, assuming the ADC sampling rate is 2MS / s, the system will collect 2000 voltage sampling points of acoustic signals in each millisecond (ms) control cycle. This data block containing 2000 points forms the basis for subsequent spectrum analysis.

[0128] S900: Online airtightness diagnosis and intervention based on real-time spectrum analysis

[0129] In this embodiment, decisive information related to the weld airtightness is extracted from the acquired, noisy raw acoustic signal, and corresponding intervention actions are performed accordingly. This step is implemented in the dedicated digital signal processing (DSP) logic of the control module 40 or in the parallel processing pipeline of the FPGA, and includes a two-level diagnostic logic; wherein:

[0130] S910: First-level diagnosis: Real-time detection of weld penetration defects based on feature spectrum matching

[0131] The goal of this level is to capture solder burn-through events with zero latency, specifically:

[0132] S911: Offline feature spectrum library construction.

[0133] During the offline calibration phase, a series of typical burn-through defects are created on the standard sample by deliberately using excessively high laser power. While creating these defects, the high-frequency acoustic signals generated are simultaneously acquired. Subsequently, these signals are analyzed by short-time Fourier transform (STFT) or wavelet transform to extract the acoustic spectral features that are statistically consistent at the moment of burn-through. This characteristic spectrum is the acoustic leakage characteristic spectrum ALSS. ALSS usually exhibits a sudden and sharp increase in energy in a specific high-frequency band (e.g., 300kHz-500kHz). The ALSS is quantized into an eigenvector and stored in the system's non-volatile memory.

[0134] S912: Online real-time spectrum analysis and pattern matching.

[0135] During the production welding process, the system performs a Fast Fourier Transform (FFT) (e.g., a 1024-point FFT) on the real-time acquired acoustic signal data blocks (e.g., 2000 sampling points every 1ms) to generate a real-time acoustic spectrum. Then, the system uses a fast pattern matching algorithm to calculate the normalized cross-correlation coefficient or Euclidean distance between the real-time spectrum and the pre-stored ALSS feature vectors.

[0136] S913: Instant intervention action trigger.

[0137] The system has a preset matching threshold (e.g., a cross-correlation coefficient greater than 0.9); once at a certain moment... If the calculated matching degree exceeds the threshold, the system immediately determines the corresponding position in the weld. If a burn-through defect occurs, this determination will immediately trigger a top-priority immediate intervention action, which includes:

[0138] S1, within microseconds, forcibly shuts down the laser output via a hardware interrupt.

[0139] S2, stop the feeding motion of the aerosol can.

[0140] S3, through the production line control system (PLC), marks the aerosol can with the clearly identified defect location as a defective product and automatically removes it from the subsequent workstation.

[0141] For example, in At 10:00, the real-time spectrum calculated by FFT showed a narrowband peak near 420kHz with energy 30 dB higher than normal. This characteristic closely matched the stored ALSS, with a calculated cross-correlation coefficient of 0.95. The system immediately detected a solder burn-through. At 10:00 a.m. (with a delay of only 500 microseconds), the laser is turned off, the tank is marked as scrap, and so on.

[0142] S920: Level 2 Diagnostics: Predictive Maintenance of Leakage Trends Based on Acoustic Health Indicators

[0143] The goal of this level is to predict trends in airtightness degradation and implement preventative interventions by monitoring subtle changes in the acoustic characteristics of the welding process before catastrophic weld burn-through events occur; specifically:

[0144] S921: Construct a healthy acoustic template.

[0145] During the offline calibration phase, a large number of acoustic signals of the standard under ideal welding conditions are collected. The spectrum of these signals is averaged to obtain the healthy acoustic template (HAT), which contains the spectral envelope of normal energy distribution in each frequency band.

[0146] S922: Calculate the Acoustic Health Index (AHI) of welds.

[0147] During the production welding process, the system will display the acoustic spectrum in real time. Compared with HAT, a dimensionless weld acoustic health index between 0 and 1 is calculated. The index can be calculated based on the similarity of the spectra:

[0148]

[0149] The formula calculates the normalized area difference between the real-time spectrum and the healthy template. When the real-time spectrum and the template perfectly overlap, the exponent is 1; as the deviation increases, the exponent approaches 0.

[0150] S923: Preventive fine-tuning.

[0151] The system has a more lenient warning threshold than immediate intervention (e.g., AHI index below 0.8). When the AHI index remains below this warning threshold for a period of time (e.g., 10 consecutive milliseconds), the system determines that although the welding process has not yet achieved burn-through, its stability is deteriorating and there is a risk of future leakage. At this time, the system will not shut down abruptly, but will trigger a preventive fine-tuning action. For example, it can send a small negative target value bias to the PID control loop, causing the PID controller to automatically reduce the peak temperature of the molten pool by 5-10 Kelvin, thereby pulling the keyhole process, which is on the verge of instability, back to a stable state, thus eliminating the defect before it forms.

[0152] At the system level, the control module 40 in this embodiment also integrates an acoustic diagnostic module. This module has an independent high-speed ADC interface and a dedicated DSP processing core, which can perform FFT, pattern matching and AHI exponent calculation in parallel. The output of this module (an instant intervention trigger signal and a preventive fine-tuning request signal) is connected to the main control logic of the system, realizing deep coupling between the acoustic diagnostic loop and the thermodynamic control loop.

[0153] Example 5

[0154] After obtaining the aforementioned parameters, this embodiment uses a multi-level cognitive control arbitrator (HCCA) to make intelligent decisions on the three relatively independent control / diagnostic loops.

[0155] Specifically, the following steps are included:

[0156] S1000: Real-time fusion and comprehensive evaluation of multi-dimensional states during the welding process

[0157] In each control cycle (e.g., 1 millisecond), the arbitrator receives all critical status information from the three underlying loops in real time, specifically including:

[0158] Thermodynamic steady-state loop information: Real-time temperature deviation signal .

[0159] WPSI circuit information: Comprehensive welding stability index and its three sub-factors .

[0160] Acoustic diagnostic circuit information: Weld acoustic health index and a Boolean-type weld burn-through alarm indicator. .

[0161] HCCA fuses these multidimensional heterogeneous data to obtain a comprehensive state vector describing the overall health of the current welding process. .

[0162] S1010: Dynamic Partitioning of System Operating States Based on Comprehensive State Vector

[0163] Based on the generated integrated state vector, HCCA dynamically divides the current working state of the system into one of the following four hierarchical states according to preset rules:

[0164] State 1: Stable Optimization State

[0165] Entry requirements: All core metrics are within a highly healthy range; for example, and and and .

[0166] Arbitration strategy: In this state, the system operates smoothly; the HCCA-authorized PID loop is fully responsible for fine-tuning the laser power to maintain temperature stability; the WPSI loop and acoustic loop only monitor and do not generate intervention commands.

[0167] State 2: Prediction and Early Warning State

[0168] Entry criteria: Macroscopic thermodynamic stability, but slight deterioration in microscopic or acoustic parameters, indicating potential risks; for example, but or .

[0169] Arbitration Strategy: In this state, HCCA adopts a "conflict resolution" strategy. For example, an acoustic circuit may request a mild power reduction command to enhance stability due to a decrease in the AHI index; while a PID circuit may request a small power increase command due to a slight temperature drop. According to the meta-rule of "safety first," HCCA will prioritize approving the acoustic circuit's power reduction request, but may attenuate the magnitude of its request (e.g., -20W) (e.g., ultimately approving -10W), while temporarily suppressing the PID circuit's power increase request to avoid command conflicts between the two circuits that could cause system oscillation.

[0170] State 3: Active Correction State

[0171] Entry criteria: Identification of microscopic defects requiring proactive intervention; for example, .

[0172] Arbitration Strategy: In this state, HCCA adopts a "cooperative correction" strategy. The WPSI loop generates a specific laser parameter correction sequence (CAP) (e.g., CAP-01). While approving the execution of this CAP, HCCA anticipates that the CAP (e.g., containing high-frequency power modulation) will affect the average temperature of the molten pool (e.g., potentially causing a 15K drop in average temperature). Therefore, HCCA generates a feedforward compensation instruction, adjusting the target temperature setpoint of the PID loop within the 20-millisecond cycle of CAP-01 execution. By temporarily increasing the power by 15K, the PID loop will actively coordinate to increase the base power to counteract the disturbance caused by CAP, ensuring that the overall thermodynamic stability is not compromised when performing fine micro-defect correction operations.

[0173] State 4: Emergency Intervention State

[0174] Entry conditions: A catastrophic or impending catastrophic defect is detected; for example, or It remains below a very low threshold for an extended period.

[0175] Arbitration Strategy: In this state, HCCA has the highest authority and executes the "absolute priority" strategy, which means immediately ignoring any instructions from all other loops and directly generating and issuing a highest priority emergency intervention instruction, such as immediately shutting down the laser and marking the defective product.

[0176] S1100: Unified generation and issuance of final control commands

[0177] After S1010 determines the current operating state and executes the corresponding arbitration logic, HCCA integrates all approved and modified instruction components to generate a unified, co-optimized final control instruction vector. This vector may include:

[0178] 1) The final laser power setting value (may be synthesized from the PID base value + CAP modulation value + HCCA compensation value).

[0179] 2) Final beam oscillation mode instructions (possibly from CAP).

[0180] 3) System status flags (e.g., shutdown alarm flags).

[0181] This final control command vector is sent to the underlying execution module to complete a full intelligent control cycle.

[0182] At the system level, the control module 40 in this embodiment includes a thermodynamic control module, a WPSI diagnostic module, and an acoustic diagnostic module as parallel low-level processing units. Above these modules, there is also a core cognitive control arbitrator module. This module receives state information from all low-level modules, implements a state machine to execute the state switching logic defined in S1010, and includes a rule engine and a collaborative model to execute the arbitration strategy described in S1010. The output of the arbitrator module is the final control command, which directly drives the welding execution module.

[0183] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0184] It should be noted that the above modules can be implemented by software or hardware. For the latter, they can be implemented in the following ways, but are not limited to: all the above modules are located in the same processor; or, the above modules are located in different processors in any combination.

[0185] Embodiments of the present invention also provide a computer-readable storage medium storing a computer program, wherein the computer program is configured to perform the steps in any of the above method embodiments when executed.

[0186] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard disk, magnetic disk, or optical disk.

[0187] Embodiments of the present invention also provide an electronic device including a memory and a processor, the memory storing a computer program and the processor being configured to run the computer program to perform the steps in any of the above method embodiments.

[0188] In one exemplary embodiment, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor and the input / output device is connected to the processor.

[0189] Through the above description of the embodiments, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0190] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another apparatus, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0191] The units described as separate components may or may not be physically separate. A component shown as a unit can be one or more physical units; that is, it can be located in one place or distributed in multiple different locations. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0192] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0193] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, in essence, or the parts that contribute to the prior art, or all or part of the technical solutions, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0194] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method of laser welding parameter optimization for aerosol can seams, characterized by, The method comprises: real-time acquisition of thermodynamic state data of a molten pool of a weld seam in a production welding process of an aerosol can; dynamic quantification of a deviation signal of a welding state based on a preset reference thermodynamic template and the real-time acquired thermodynamic state data, wherein the reference thermodynamic template corresponds to a standard welding process of the aerosol can; closed-loop feedback adjustment of laser welding parameters in the production welding process based on the deviation signal.

2. The method of claim 1, wherein, The dynamic quantification of the deviation signal of the welding state comprises: extraction of a corresponding reference thermodynamic state from the reference thermodynamic template according to a real-time position of a current laser welding head on the weld seam; comparison of the real-time acquired thermodynamic state data with the reference thermodynamic state to generate the deviation signal.

3. The method of claim 2, wherein, The closed-loop feedback adjustment of the laser welding parameters based on the deviation signal comprises: input of the deviation signal into a preset PID controller to generate a feedback adjustment amount for compensating the deviation signal; adjustment of the laser welding parameters based on the feedback adjustment amount.

4. The method of claim 1, wherein, The method further comprises: acquisition of preset geometric feature information of the aerosol can; generation of a feedforward compensation amount before the laser welding head reaches a weld seam position corresponding to the geometric feature information; superposition of the feedforward compensation amount into the adjustment process of the laser welding parameters to suppress welding disturbances caused by the geometric feature information in advance.

5. The method of claim 1, wherein, The method further comprises: acquisition of a real-time temperature field distribution image sequence of the molten pool; analysis of the real-time temperature field distribution image sequence to extract a first feature of the molten pool, the first feature comprising at least any one of an area of the molten pool, a shaking frequency of the molten pool, and a welding spatter event; determination of a welding abnormal state in the welding process based on the first feature.

6. The method of claim 1, wherein, The determination of the welding abnormal state in the welding process comprises: determination of an area stability factor based on the area of the molten pool; determination of an oscillation stability factor based on the shaking frequency of the molten pool; determination of a spatter stability factor based on the welding spatter event; determination of a welding stability index based on the area stability factor, the oscillation stability factor, and the spatter stability factor; determination of the welding abnormal state based on the welding stability index.

7. The method of claim 1, wherein, The method further comprises: determination of a laser parameter correction sequence from a preset correction strategy library based on a type of the welding abnormal state; superposition of the laser parameter correction sequence into the adjustment process of the laser welding parameters for abnormality suppression.

8. A laser welding parameter optimization system for aerosol can seams, characterized by, The system comprises: a state perception module for real-time acquisition of thermodynamic state data of a molten pool of a weld seam in a production welding process of an aerosol can; a control module for dynamic quantification of a deviation signal of a welding state based on a preset reference thermodynamic template and the real-time acquired thermodynamic state data, wherein the reference thermodynamic template corresponds to a standard welding process of the aerosol can; a welding execution module for closed-loop feedback adjustment of laser welding parameters in the production welding process based on the deviation signal.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, wherein the computer program is configured to execute the method in any one of claims 1 to 7 when running. 10.An electronic device comprising a memory and a processor, the electronic device characterized by, The memory stores a computer program, and the processor is configured to execute the computer program to execute the method in any one of claims 1 to 7.

Images