Intelligent control method for water-cooling heat dissipation ultra-small integrated handheld laser welding machine

Abstract

The invention relates to the technical field of laser welding control, and discloses an intelligent control method for a water-cooling heat dissipation ultra-small integrated handheld laser welding machine. The method comprises the steps that a thermal dynamic regulation and control model is trained and generated according to real-time temperature monitoring data of a water-cooling heat dissipation unit and power operation parameters of a laser transmitter; the model is applied to predict the heat distribution change in the welding operation period in real time, and the flow speed of cooling liquid of a water cooling unit is dynamically adjusted based on the prediction result; then, in combination with the adjusted cooling liquid flow velocity data, virtual analogue simulation is conducted on the movement track of the laser welding joint, and quality characteristic parameters of the simulated welding joint are evaluated; and finally, aiming at a predefined welding quality optimization target, executing optimization operation on the simulation quality characteristic parameter data set, and generating and outputting an intelligent control execution scheme. According to the method, intelligent control over the welding process is achieved, the heat dissipation accuracy and the welding quality stability are improved, and the method is suitable for the efficient operation requirement of the ultra-small integrated handheld laser welding machine.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of laser welding control, in particular to an intelligent control method for a water-cooled and heat-dissipated ultra-small integrated handheld laser welding machine. BACKGROUND

[0002] With the development of manufacturing industry towards precision and portability, handheld laser welding machines are increasingly widely used in metal processing, automobile repair, medical devices and other fields due to their flexible operation and high welding precision. However, the ultra-small integrated design of handheld devices poses a serious challenge to heat dissipation and stable control. During laser welding, the laser emitter generates a large amount of heat. If heat dissipation is not timely or uniform, it will not only affect the stability of the laser output power, but also may cause damage to the core components of the device due to overheating, thereby shortening the service life of the device. Most handheld laser welding machines use traditional water-cooled heat dissipation methods, which control heat dissipation through a pre-set cooling liquid flow rate. This method does not take into account the dynamic changes of laser power and real-time temperature distribution differences during welding operations, making it difficult to achieve precise control of heat dissipation. When the laser power suddenly increases, the pre-set flow rate may not be able to remove the excess heat in time, causing local temperature to be too high. When the power decreases, the high flow rate of the cooling liquid will result in energy waste, and may affect the thermal stability of the welding area due to sudden temperature drop.

[0003] The motion trajectory of the laser welding head directly affects the forming quality of the welded joint. In the prior art, the motion trajectory planning relies on the experience of the operator and lacks linkage with the real-time heat dissipation state. During welding, changes in heat distribution will change the melting and solidification characteristics of the material. If the motion trajectory is not dynamically adjusted according to the heat distribution, defects such as weld deviation, pores and cracks may occur. In addition, the evaluation of the quality of the welded joint often needs to be realized through offline detection after welding, which cannot provide real-time feedback and optimize the control parameters during welding, resulting in insufficient stability of the welding quality and a high rate of defective products. In the existing control method, there is a lack of real-time prediction capability for changes in heat distribution, and a dynamic correlation model between heat dissipation parameters, motion trajectory and welding quality has not been established, making it difficult to achieve intelligent closed-loop control of the welding process. These problems seriously restrict the application of ultra-small integrated handheld laser welding machines in high-precision and high-stability welding scenarios, and an intelligent control method that combines real-time heat dissipation state and welding parameters is urgently needed. SUMMARY

[0004] The present application aims to provide an intelligent control method for a water-cooled and heat-dissipated ultra-small integrated handheld laser welding machine to solve the problems raised in the background.

[0005] To achieve the above-mentioned purpose, the present application provides an intelligent control method for a water-cooled and heat-dissipated ultra-small integrated handheld laser welding machine, which comprises: According to the real-time temperature monitoring data of the water-cooled heat dissipation unit and the power operation parameters of the laser emitter, a thermal dynamic regulation model is trained; The trained thermal dynamic regulation model is applied to real-time prediction of thermal distribution changes during welding operation, and the cooling liquid flow rate of the water-cooled unit is dynamically adjusted based on the predicted thermal distribution change results; The motion trajectory of the laser welded joint is virtually simulated in combination with the adjusted cooling liquid flow rate data, and the quality characteristic parameters of the simulated welded joint are evaluated; The quality characteristic parameter data set of the simulated welded joint is optimized for the pre-defined welding quality optimization target, and an intelligent welding control execution scheme is generated and output.

[0006] Preferably, wherein the thermal dynamic regulation model is trained, including: collecting real-time temperature monitoring data of the water-cooled heat dissipation unit and power operation parameter data of the laser emitter, the real-time temperature monitoring data including temperature sensor time series readings and heat dissipation efficiency indicators, and the power operation parameter data including laser power setting value and pulse action duration; Based on the real-time temperature monitoring data and the power operation parameter data, the thermal behavior correlation characteristics are analyzed, and a thermal dynamic training data set is constructed; The thermal dynamic training data set is used for model training iteration process until the model convergence condition is reached, and the thermal dynamic regulation model is obtained.

[0007] Preferably, wherein based on the real-time temperature monitoring data and the power operation parameter data, the thermal behavior correlation characteristics are analyzed, and a thermal dynamic training data set is constructed, including: analyzing the feature change law of the welding heat affected zone through the real-time temperature monitoring data, and determining the interaction coefficient of temperature and power; Integrating the power operation parameter data and the interaction coefficient of temperature and power for collaborative regulation influence evaluation, and obtaining the collaborative influence factor of power adjustment on temperature stability; Fusing the interaction coefficient of temperature and power and the collaborative influence factor of power adjustment, constructing a thermal dynamic training data set, and the data label of the thermal dynamic training data set is the predicted temperature value.

[0008] Preferably, wherein the interaction coefficient of temperature and power is determined by analyzing the feature change law of the welding heat affected zone through the real-time temperature monitoring data, including: using the real-time temperature monitoring data to analyze the temperature change mode under different power configuration conditions; The actual temperature steady-state value under different power configuration conditions is obtained through experimental test; Aligning the power configuration conditions, temperature change mode and actual temperature steady-state value, determining the interaction coefficient of temperature and power.

[0009] Preferably, the power adjustment synergistic influence factor is obtained by integrating the power operation parameter data and the temperature-power interaction coefficient, including: referring to the power operation parameter data, creating a comparison temperature characteristic data set, and the comparison temperature characteristic data set covering temperature record information of the same power configuration and the same welding material under multiple heat dissipation conditions; According to the comparison temperature characteristic data set, a temperature stability experiment test is performed to obtain a temperature deviation amount of the comparison temperature characteristic data set; The temperature deviation amount and the temperature-power interaction coefficient under the no-power adjustment reference are combined to calculate the power adjustment synergistic influence factor.

[0010] Preferably, the model training iteration process using the thermal dynamic training data set includes: proportionally dividing the thermal dynamic training data set to form a model training subset and a model verification subset; An adaptive deep learning algorithm is selected for thermal dynamic regulation model training, wherein the model training subset is used for weight optimization and hyperparameter adjustment of the thermal dynamic regulation model, and the model verification subset is used for error checking and performance tuning of the optimized model until the preset accuracy standard is met, and the final thermal dynamic regulation model is confirmed.

[0011] Preferably, the trained thermal dynamic regulation model is applied to real-time prediction of thermal distribution changes during welding operations, and the cooling liquid flow rate of the water cooling unit is dynamically adjusted based on the predicted thermal distribution change results, including: obtaining the predicted thermal distribution change results output by the thermal dynamic regulation model; The cooling liquid flow rate adjustment strategy of the water cooling unit is searched according to the predicted thermal distribution change results to generate the optimal cooling liquid flow rate configuration with the goal of minimizing thermal fluctuations.

[0012] Preferably, the adjusted cooling liquid flow rate data is combined to virtually simulate the motion trajectory of the laser welding head, and the quality characteristic parameters of the simulated welding joint are evaluated, including: receiving the adjusted cooling liquid flow rate data, simulating the motion trajectory of the laser welding head, and outputting the quality characteristic parameter data of the simulated welding joint, including weld depth, molten pool width, and thermal deformation amount; The welding path is explored in terms of geometric shape, moving speed, and energy input scheme based on the quality characteristic parameter data of the simulated welding joint to optimize the welding control channel with the welding quality optimization target as the evaluation reference.

[0013] Preferably, the quality characteristic parameter data set of the simulated welding joint is optimized for the predefined welding quality optimization target, including: performing multi-dimensional quality characteristic quantization integration processing according to the quality characteristic parameter data of the simulated welding joint to generate a multi-quality index quantization matrix; The environment compensation gradient is analyzed based on the multi-quality index quantization matrix, and a multi-level compensation control network is constructed according to the gradient level division, and the multi-level compensation control network includes multi-level control parameter items. The environment dynamic collaborative control parameters are obtained, and the multi-level compensation control network is calibrated step by step by applying the environment dynamic collaborative control parameters to generate an environment collaborative control strategy. The evaluation benchmark is adjusted according to the environment collaborative control strategy. According to the adjusted evaluation benchmark, the geometric shape, movement rate and energy input scheme of the welding path are searched and optimized.

[0014] Preferably, the predefined welding quality optimization target includes a welding structure shape parameter target and a quality tolerance target.

[0015] Compared with the prior art, the present application has the following advantages: The water-cooled heat dissipation super-small integrated handheld laser welding machine intelligent control method realizes accurate grasp of the heat distribution change in the welding process by constructing a thermal dynamic control model. The model is trained based on the real-time temperature monitoring data of the water-cooled heat dissipation unit and the power operation parameters of the laser emitter, so that the model can fully learn the heat transfer law under different working conditions, thereby predicting the heat distribution change in real time during welding operation. This prediction method based on actual operation data breaks the limitation of traditional preset cooling parameters, and makes the cooling control more suitable for actual welding requirements. Based on the predicted heat distribution change result, the cooling liquid flow rate of the water cooling unit is dynamically adjusted, so that the heat dissipation intensity and the heat generation rate can be matched in real time. When the laser power increases and heat is concentrated, the cooling liquid flow rate increases correspondingly to quickly remove excess heat and avoid the influence of local overheating on equipment and welding quality; when the laser power decreases, the flow rate decreases in time to reduce unnecessary energy consumption, while maintaining the relative stability of the welding area temperature. This dynamic adjustment mechanism not only improves the cooling efficiency, but also guarantees the stability of the thermal environment, creating favorable conditions for the smooth progress of the welding process. The motion trajectory of the laser welding joint is virtually simulated combined with the adjusted cooling liquid flow rate data, and the heat dissipation state is included in the trajectory evaluation system, so that the simulation result is closer to the actual welding scene. By evaluating the quality characteristic parameters of the simulated welding joint, possible quality problems such as poor joint forming caused by uneven heat distribution can be predicted in advance. This early prediction mechanism reduces the blindness of relying on experience to plan the trajectory, and provides a reliable basis for optimizing the motion trajectory. The optimization operation is performed on the simulation quality characteristic parameter dataset for a predefined welding quality optimization target, various factors such as heat dissipation state and motion trajectory can be comprehensively considered, and an intelligent control scheme suitable for specific welding requirements is generated. The scheme not only covers the dynamic regulation and control strategy of the cooling liquid flow rate, but also includes the optimized path of the welding head motion trajectory, so that the heat dissipation, power output and trajectory planning form an organic synergy. This synergy can adapt to the welding requirements of workpieces of different materials and different thicknesses, reduce welding defects caused by parameter mismatch, and improve the adaptability and operation stability of the handheld laser welding machine under complex working conditions. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 A timing diagram of the intelligent control method of the water-cooled heat dissipation ultrasmall integrated handheld laser welding machine; Figure 2 A flowchart of the thermal dynamic regulation and control model training; Figure 3 A flowchart of the thermal dynamic training dataset construction; Figure 4 A flowchart of the welding quality characteristic parameter optimization operation. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0018] Please refer to Figure 1 The present application provides an intelligent control method of a water-cooled heat dissipation ultrasmall integrated handheld laser welding machine, which comprises: According to the real-time temperature monitoring data of the water-cooled heat dissipation unit and the power operation parameters of the laser emitter, a thermal dynamic regulation and control model is generated through data acquisition, feature analysis and model training iteration process. The model is applied to predict the thermal distribution change during welding operation in real time, and the cooling liquid flow rate of the water-cooled unit is dynamically adjusted based on the prediction result. The motion trajectory of the laser welding head is virtually simulated combined with the adjusted cooling liquid flow rate data, and the quality characteristic parameters of the simulated welding joint are evaluated. The optimization operation is performed on the quality characteristic parameter dataset for a predefined welding quality optimization target, and an intelligent welding control execution scheme is generated and output.

[0019] Embodiment 1: Please refer to Figure 2The training process of the thermal dynamic regulation model mainly includes three main stages: data collection, feature analysis, and model training iteration. In the data collection stage, real-time temperature monitoring data of the water-cooled heat dissipation unit and power operating parameter data of the laser emitter are synchronously acquired. Real-time temperature monitoring data is sourced from temperature sensors installed at key positions of the water-cooled heat dissipation unit, which are usually distributed at the water inlet, water outlet, and heat dissipation fin surface, etc. to collect temperature values at a frequency of once per second, forming continuous time series readings. The heat dissipation efficiency index is also included in the collection range, which represents the heat dissipation efficiency of the water-cooled unit by calculating the ratio of the heat carried away by the cooling liquid to the input laser energy per unit time. Power operating parameter data mainly come from the control system of the laser emitter, including the set value of laser power, which can be adjusted from 100 watts to 5000 watts in continuous mode or set peak power and base power in pulse mode according to welding requirements; in addition, it also includes pulse action time, which specifically covers the duration of a single pulse ranging from 0.1 milliseconds to 10 milliseconds, and the interval time between adjacent pulses ranging from 0.5 milliseconds to 20 milliseconds. The collection process of these data needs to ensure time synchronization to facilitate the subsequent analysis of the corresponding relationship between temperature changes and power adjustments.

[0020] The task of the feature analysis stage is to analyze the correlation characteristics between thermal behaviors based on the collected temperature monitoring data and power operating parameter data, and to construct a data set for model training. It is necessary to analyze the temperature variation law of the welding heat affected zone under different power configuration conditions. For example, in low power mode, the temperature rise may be relatively slow, the peak temperature is low, and the cooling rate is also relatively slow; while in high power mode, the temperature rise rate will significantly accelerate, the peak temperature will rise, and the cooling rate will also accelerate. This process needs to be completed by analyzing the change trend of temperature time series data to identify the characteristics of temperature change with time, such as rise rate, peak temperature value, and cooling rate, etc. In order to more accurately understand the relationship between temperature and power, it is also necessary to obtain the actual temperature steady-state value under different power configurations through experimental tests. Experimental tests usually use the same materials as actual welding to make standard test pieces, and welding operations are carried out in a controlled temperature environment. The temperature curve is recorded after welding until the temperature stabilizes in a range close to the ambient temperature, and the temperature value at this time is the actual temperature steady-state value. For example, welding a stainless steel test piece under fixed power and pulse parameters, recording the temperature from the peak to the stable state of 10 degrees Celsius above the ambient temperature within 10 seconds after welding, and the stable temperature is the steady-state value obtained by experiment.

[0021] After obtaining the temperature change pattern and the actual temperature steady-state value, it is necessary to further determine the interaction coefficient between temperature and power. This coefficient reflects the degree of influence of power change on temperature, which is obtained by aligning and analyzing the temperature change pattern under different power configurations and the corresponding steady-state value. For example, when the power setting value increases, observe the change of temperature rising rate and peak temperature, and find the quantitative relationship between power change and temperature change by comparing multiple sets of data under different power configurations.

[0022] To evaluate the synergistic effect of power adjustment under different heat dissipation conditions, it is also necessary to integrate power operation parameter data and temperature and power interaction coefficient for synergistic regulation and influence evaluation. This process requires creating a comparison temperature feature data set that covers temperature record information under multiple heat dissipation conditions, such as natural cooling, forced air cooling, and different flow rate water cooling, for the same power configuration and the same welding material. By performing temperature stability experiment test, changing the heat dissipation condition (such as adjusting the water cooling flow rate), recording the temperature fluctuation range under different conditions, and obtaining the temperature deviation amount. For example, when the water cooling flow rate increases from 2 liters per minute to 5 liters per minute, observe the change of temperature fluctuation range, calculate the temperature deviation amount under different flow rates. Combined with the temperature and power interaction coefficient under the benchmark of no power adjustment, the synergistic influence factor of power adjustment can be calculated, which reflects the additional temperature influence of power adjustment under different heat dissipation conditions.

[0023] The model training iteration phase needs to use the constructed thermal dynamic training data set to train and optimize the model. The thermal dynamic training data set is divided into model training subset and model validation subset according to certain proportion, common division proportion is 7:3 or 8:2. The training subset is mainly used for model weight optimization and hyperparameter adjustment, and the validation subset is used for error checking and performance tuning of the optimized model. The selected model algorithm is usually adaptive deep learning algorithm, such as recurrent neural network model based on gated recurrent unit, which can effectively process time series data and capture the dynamic correlation between temperature change and power adjustment. In the training process, the training subset is input into the model, the connection weight of the internal neurons of the model is adjusted through the back propagation algorithm, and the hyperparameters (such as learning rate, number of hidden layer neurons, etc.) of the model are adjusted to improve the fitting ability of the model to data.

[0024] The optimized model is tested by the validation subset, the prediction error (such as root mean square error) of the model on the validation subset is calculated, and the model parameters are further adjusted according to the error result until the prediction error of the model on the validation subset is less than the preset accuracy standard, at this time the model training is completed, and the final thermal dynamic regulation model is determined. The pre-defined welding quality optimization target includes a welding structure morphology parameter target and a quality tolerance target, wherein the structure morphology parameter target specifies the allowable deviation range of the welding seam width, penetration and other parameters, and the quality tolerance target sets the maximum allowable value of the thermal deformation amount, the length of the undercut defect and other quality indicators, which provides a clear reference standard for subsequent welding process control and optimization.

[0025] Example 2: refer to Figure 3 The construction of the thermal dynamic training data set is the basis for model training, and the core is to establish a data set that can reflect the correlation between temperature change and power adjustment, and heat dissipation conditions through the determination of temperature and power interaction coefficients and the evaluation of synergistic factors. This process needs to combine real-time monitoring data and experimental test results and be completed in stages.

[0026] The determination of the temperature and power interaction coefficient first relies on the analysis of real-time temperature monitoring data. Real-time temperature monitoring data is collected by temperature sensors installed at key positions of the water-cooled heat dissipation unit. These sensors are usually distributed at the water inlet, water outlet, and surface of the heat dissipation fins, and record temperature values at a fixed frequency (such as once per second) to form a continuous time series data sequence. The heat dissipation efficiency index is used to characterize the heat dissipation efficiency of the water-cooled unit by calculating the ratio of the heat carried away by the cooling liquid per unit time to the total energy input by the laser. This index is directly related to the rate of temperature change. When analyzing the temperature change pattern, the temperature time series data needs to be classified according to the power configuration conditions, for example, the data of different power segments such as 2000W, 3000W, and 4000W in continuous welding mode are extracted respectively, and the temperature change trend with time in the same power segment is observed. Specifically, the rate of temperature rise (such as the ratio of the time from room temperature to peak temperature to the amount of temperature change), the specific value of peak temperature (such as the highest temperature reached during welding), and the rate of cooling (such as the ratio of the time from peak temperature to ambient temperature to the amount of temperature change) need to be identified. These characteristics reflect the dynamic balance relationship between heat accumulation and dissipation under different power inputs.

[0027] To determine the quantitative correlation between temperature and power more accurately, it is necessary to obtain the actual temperature steady-state values under different power configurations through experimental testing. Experimental testing usually uses standard specimens made of the same materials as actual welding (such as 304 stainless steel or aluminum alloy) in a controlled temperature experimental environment. During the experiment, the laser emitter works according to the preset power setting value (such as 2000 watts in continuous mode) and pulse parameters (such as pulse frequency, duty cycle), and after completing the welding, the specimen temperature is continuously monitored until the temperature stabilizes in a range close to the ambient temperature (such as when the ambient temperature is 25℃, the steady-state temperature stabilizes between 35℃ and 40℃). The temperature value at this time is recorded as the actual temperature steady-state value under this power configuration. For example, in the 2000-watt continuous welding mode, the average steady-state temperature obtained after multiple repeated experiments is the typical steady-state value corresponding to this power. By comparing the temperature change patterns (rising rate, peak temperature, cooling rate) and corresponding steady-state values under different power configurations (such as 2000 watts, 3000 watts, 4000 watts), a correlation between power changes and temperature changes can be established. For example, when the power increases from 2000 watts to 3000 watts, the temperature rising rate is observed to accelerate, the peak temperature rises, and the cooling rate also accelerates accordingly. Through statistical analysis of multiple groups of data, the approximate proportional relationship between temperature and power can be determined, i.e. the interaction coefficient of temperature and power.

[0028] The determination of the synergistic influence factor requires the integration of power operating parameter data and the interaction coefficient of temperature and power to evaluate the additional impact of power adjustment under different heat dissipation conditions. Create a comparison temperature feature dataset that covers temperature record information under various heat dissipation conditions for the same power configuration and the same welding material. For example, for the 3000-watt continuous welding mode, different heat dissipation conditions are set in the experiment: natural cooling (no active cooling), forced air cooling (wind speed 0.5 m / s), low flow rate water cooling (1 L / min), high flow rate water cooling (5 L / min), etc. The temperature change curve during welding and the final steady-state temperature under each heat dissipation condition are recorded. Through this dataset, the differences in temperature changes under different heat dissipation conditions under the same power input can be observed, for example, the steady-state temperature under high flow rate water cooling is significantly lower than that under natural cooling.

[0029] Temperature stability experiment tests are performed to obtain temperature deviation amounts. In the experiment, parameters such as laser power, welding speed, material type, etc. are kept constant, and only the heat dissipation working conditions are adjusted (such as changing the water cooling flow rate), and the temperature fluctuation range under different working conditions is recorded. For example, under the condition of 3000W continuous welding mode, when the water cooling flow rate is increased from 1L / min to 3L / min, the peak value change of temperature during welding (such as the peak temperature from 400℃ to 350℃) and the temperature fluctuation amplitude during the cooling stage (such as from ±20℃ to ±10℃) are observed, and the temperature deviation amount under different flow rates (such as the temperature deviation of flow rate 1L / min and reference flow rate 2L / min is +30℃) is calculated.

[0030] Combined with the temperature and power interaction coefficient under the reference of power adjustment (i.e. the power-temperature correlation ratio without considering the difference of heat dissipation working conditions) and the temperature deviation amount, the synergistic influence factor of power adjustment is calculated. For example, under the reference flow rate of 2L / min, the temperature rises by about 200℃ for every 1000W increase in power (interaction coefficient); while under the flow rate of 1L / min, the same power increase causes the temperature to rise by 230℃, so the synergistic influence factor is 30℃ / (kW·(L / min)), which reflects the additional temperature influence of power adjustment under different heat dissipation conditions. Through this process, a thermal dynamic training data set containing the temperature and power interaction coefficient and the synergistic influence factor can be constructed, providing comprehensive data support for subsequent model training.

[0031] Example 3: The construction of the thermal dynamic training data set needs to be realized through the determination of the temperature and power interaction coefficient and the evaluation of the synergistic influence factor. This process needs to combine real-time monitoring data and experimental test results, and systematically analyze the correlation between temperature change, power adjustment and heat dissipation conditions.

[0032] The determination of the temperature and power interaction coefficient first relies on the collection and analysis of real-time temperature monitoring data. Real-time temperature monitoring data is obtained by temperature sensors installed at key positions of the water cooling heat dissipation unit. These sensors are usually distributed at the water inlet, water outlet, heat dissipation fin surface, etc. to record temperature values at a fixed frequency (such as once per second), forming a continuous time series data sequence.

[0033] The heat dissipation efficiency index characterizes the heat dissipation efficiency of a water-cooled unit by calculating the ratio of the heat removed by the coolant per unit time to the total laser input energy. This index is directly related to the rate of temperature change. When analyzing temperature change patterns, the temperature time-series data needs to be categorized according to power configuration conditions. For example, data from different power ranges such as 2000W, 3000W, and 4000W in continuous welding mode should be extracted separately to observe the temperature change trend over time within the same power range. Specifically, it is necessary to identify the rate of temperature rise (e.g., the ratio of time to temperature change from room temperature to peak temperature), the specific value of the peak temperature (e.g., the highest temperature reached during welding), and the rate of cooling (e.g., the ratio of time to temperature change from peak temperature to near ambient temperature). These characteristics reflect the dynamic balance between heat accumulation and dissipation under different power inputs.

[0034] To more accurately determine the quantitative correlation between temperature and power, it is necessary to obtain the actual steady-state temperature values ​​under different power configurations through experimental testing. Experimental testing typically uses standard specimens made of the same materials as the actual welding (such as 304 stainless steel or aluminum alloy) and is conducted in a temperature-controlled experimental environment. During the experiment, the laser emitter operates according to a preset power setting (e.g., 2000 watts in continuous mode) and pulse parameters (e.g., pulse frequency, duty cycle). After welding, the specimen temperature is continuously monitored until it stabilizes within a range close to the ambient temperature (e.g., at an ambient temperature of 25°C, the steady-state temperature stabilizes between 35°C and 40°C). This temperature value is recorded as the actual steady-state temperature value under that power configuration. For example, in 2000-watt continuous welding mode, the average steady-state temperature obtained after multiple repetitions of the experiment is the typical steady-state value corresponding to that power. By comparing the temperature change patterns (rise rate, peak temperature, cooling rate) with the corresponding steady-state values ​​under different power configurations (e.g., 2000 watts, 3000 watts, 4000 watts), a correlation between power changes and temperature changes can be established.

[0035] In this process, the relationship between temperature and power can be quantitatively described by the following formula:

[0036] in: Indicates time Temperature value at time (unit: °C). This indicates the laser power setting value (unit: W). The coefficient representing the effect of power on the rate of temperature change (unit: ℃ / (W·s)) reflects the temperature change caused by a unit power per unit time. The coefficient representing the effect of the power square term on temperature (unit: ℃ / (W²)) reflects the accelerating effect of temperature change when power increases; This represents the ambient temperature baseline (unit: °C), i.e., the ambient temperature when there is no power input. By fitting real-time temperature monitoring data into this formula, we can obtain... , , The specific values ​​of the three parameters are used to determine the interaction coefficient between temperature and power. For example, in a 2000-watt continuous welding mode, if the fitted result is... ℃ / (W·s) ℃ / (W²) ℃, then when time The predicted temperature value is [value] per second. ℃.

[0037] Determining the synergistic influencing factors requires integrating power operating parameter data with the interaction coefficient between temperature and power to assess the additional impact of power adjustments under different heat dissipation conditions. A comparative temperature feature dataset was created, encompassing temperature records for the same power configuration and welding materials under various heat dissipation conditions. For example, for a 3000-watt continuous welding mode, different heat dissipation conditions were set in the experiment: natural cooling (no active cooling), forced air cooling (wind speed 0.5 m / s), low-flow-rate water cooling (1 L / min), and high-flow-rate water cooling (5 L / min), etc., recording the temperature change curves and final steady-state temperature during the welding process under each heat dissipation condition. This dataset allows observation of the differences in the impact of different heat dissipation conditions on temperature changes under the same power input; for example, the steady-state temperature under high-flow-rate water cooling is significantly lower than that under natural cooling.

[0038] Temperature stability tests were conducted to obtain the temperature deviation. During the experiment, parameters such as laser power, welding speed, and material type were kept constant, and only the heat dissipation conditions (e.g., changing the water cooling flow rate) were adjusted. The temperature fluctuation range under different conditions was recorded. For example, in 3000-watt continuous welding mode, when the water cooling flow rate increased from 1 L / min to 3 L / min, the peak temperature change during welding (e.g., peak temperature decreasing from 600℃ to 550℃) and the temperature fluctuation amplitude during the cooling stage (e.g., decreasing from ±30℃ to ±15℃) were observed. The temperature deviation at different flow rates was calculated (e.g., the temperature deviation between a flow rate of 1 L / min and a reference flow rate of 2 L / min is +50℃).

[0039] By combining the temperature-power interaction coefficient (i.e., the power-temperature correlation ratio without considering differences in heat dissipation conditions) and the temperature deviation under a no-power-adjustment baseline, the synergistic influence factor of power adjustment is calculated. For example, at a baseline flow rate of 2 L / min, for every 1000 watts increase in power, the temperature rises by approximately 200°C (as shown in the formula). and Parameter determination); while at the flow rate of 1 L / min, the same power increase leads to the temperature rise of 250 °C, and the synergistic effect factor is 50 °C / (kW·(L / min)), which reflects the additional temperature influence of power adjustment under different heat dissipation conditions. Through this process, a thermal dynamic training data set containing the temperature and power interaction coefficient and the synergistic effect factor can be constructed to provide comprehensive data support for subsequent model training.

[0040] Example 4: In the scenario of handheld laser welding of aluminum alloy sheet, the process of applying the trained thermal dynamic regulation model to realize intelligent control can be divided into four continuous links: real-time thermal distribution prediction, cooling liquid flow rate adjustment, virtual simulation simulation, and welding control optimization. The implementation details of each link are described in detail below with specific examples.

[0041] In a certain aluminum alloy sheet welding task, the workpiece to be welded is 5052 aluminum alloy with a thickness of 1 mm, the weld length is 200 mm, the weld depth is required to be 0.8 mm ± 0.1 mm, the weld width is 1.0 mm ± 0.1 mm, and the thermal deformation is ≤0.2 mm. The welding equipment uses an ultra-small integrated water-cooled heat dissipation system, the laser emitter has a maximum output power of 3000 W, and supports continuous mode (power set value 1000 W-3000 W) and pulse mode (peak power 2000 W-3000 W, base power 500 W-1000 W, pulse frequency 10 Hz-50 Hz).

[0042] Real-time thermal distribution prediction: During the welding process, temperature sensors deployed at key positions (inlet, outlet, and surface of heat dissipation fins) of the water-cooled heat dissipation unit collect temperature data at a frequency of 10 Hz, while the laser emitter control system synchronously outputs power operation parameters (current power set value 2500 W, continuous mode). The temperature sensor data (partial time points) are shown in Table 1.

[0043] Table 1: Temperature sensor data (partial time points).

[0044] Time (s) Inlet water temperature (°C) Outlet water temperature (°C) Fin surface temperature (°C) Laser power setpoint (W) 5 22.1 25.3 28.7 2500 10 23.4 26.8 30.2 2500 15 24.7 28.1 32.5 2500 20 26.2 29.5 35.1 2500 The thermal dynamic regulation model receives the above real-time data with a time window of every 2 seconds and outputs the predicted thermal distribution change results. The prediction results are presented in the form of a three-dimensional temperature field, covering the welding area (weld center and both sides within a range of 5 mm) and the key positions of the water-cooled heat dissipation unit, with a time resolution of 0.1 seconds. For example, at 15 seconds after the start of welding, the model predicts that the maximum temperature in the welding area will rise to 480 °C, and reach a peak of 520 °C at 20 seconds, and the outlet temperature of the water-cooled heat dissipation unit will rise to 31 °C at 20 seconds.

[0045] Coolant flow rate adjustment: Based on the predicted thermal distribution results, the system searches for the optimal coolant flow rate configuration with the goal of minimizing thermal fluctuations. The initial flow rate of the water cooling unit is 3 L / min. The prediction shows that if this flow rate is maintained, the temperature of the welding area at 20 seconds will exceed the maximum temperature allowed by the material (550°C), and the outlet water temperature will continue to rise to 33°C, which may lead to a decrease in heat dissipation efficiency. The system calculates through an optimization algorithm that increasing the flow rate to 5 L / min will increase the heat carried away by the coolant by about 30%, effectively inhibiting the rapid rise in temperature. After adjustment, the coolant flow rate is switched to 5 L / min at 18 seconds. At this time, the predicted maximum temperature of the welding area at 20 seconds is reduced to 500°C, and the outlet water temperature stabilizes at 29°C, with a thermal fluctuation variance that is 25% lower than before the adjustment.

[0046] Virtual simulation: Based on the adjusted coolant flow rate (5 L / min), the system simulates the motion trajectory of the laser welding joint in the virtual simulation platform. The simulation input parameters include: laser power 2500W (continuous mode), welding speed 5m / min, material 5052 aluminum alloy (thermal conductivity 190W / (m·K), specific heat capacity 900J / (kg·K)), coolant flow rate 5L / min (inlet temperature 20°C). The simulation model couples the laser heat source (Gaussian distribution, power density linearly decays with scanning position), molten pool flow, heat conduction (considering material anisotropy), and coolant convection (surface heat transfer coefficient calculated based on Newton's cooling law), etc. Physical fields, output the quality characteristic parameters of the simulated welded joint. Simulation results (after welding is completed), refer to Table 2.

[0047] Table 2: Simulation results (after welding is completed).

[0048] Quality characteristic Simulated value Predefined quality target Weld depth 0.82 mm 0.8 mm ± 0.1 mm Weld pool width 1.05 mm 1.0 mm ± 0.1 mm Thermal distortion 0.18 mm ≤ 0.2 mm Weld pool tail undercut length 0.5 mm ≤ 1 mm Welding control optimization: Based on the predefined welding quality optimization goals (structural form parameter goals and quality tolerance goals), the quality characteristic parameters of the simulation results are evaluated. The weld depth 0.82mm (target range 0.7mm-0.9mm), molten pool width 1.05mm (target range 0.9mm-1.1mm) all meet the requirements; the thermal deformation amount 0.18mm (target ≤0.2mm) meets the tolerance; the molten pool tail undercut length 0.5mm (target ≤1mm) is also within the allowed range.

[0049] To further optimize the control channel, the system explores the geometry of the welding path, the moving speed and the energy input scheme. For example, in the simulation, it is found that when the welding speed is reduced to 4.5 m / min at the middle section of the weld (100-150 mm position) based on 5 m / min, the molten pool width in this area can be stabilized at 1.0-1.05 mm, and the thermal deformation is further reduced to 0.16 mm; the laser power is dynamically adjusted to 2400 W (the original set value is 2500 W) in this area, which can avoid the overburning of the molten pool caused by local overheating. The final optimized welding control execution scheme includes: the cooling liquid flow rate is maintained at 3 L / min for the first 10 seconds before welding, and then increased to 5 L / min and maintained; the moving speed of the laser welding head trajectory is reduced from 5 m / min to 4.5 m / min at the middle section of the weld (100-150 mm); the laser power of the corresponding area is adjusted from 2500 W to 2400 W.

[0050] This scheme ensures the thermal stability of the welding process through real-time thermal distribution prediction, dynamic adjustment of cooling liquid flow rate and virtual simulation optimization, while meeting the predefined welding quality targets, providing specific control guidance for actual welding operations.

[0051] Example 5: see Figure 4 Through multi-dimensional quality feature integration, environmental compensation analysis, multi-level control network construction and dynamic parameter calibration, etc., the intelligent welding control execution scheme that meets the quality requirements is finally generated. The following will combine the aluminum alloy sheet welding scene to explain the implementation details of each link.

[0052] In the aluminum alloy sheet welding task, the predefined welding quality optimization targets include structural morphology parameter targets (such as the weld depth needs to be controlled between 0.7 mm and 0.9 mm, and the molten pool width needs to be maintained within the range of 0.9 mm to 1.1 mm) and quality tolerance targets (such as the thermal deformation is not more than 0.2 mm, and the length of the undercut defect at the tail of the molten pool is less than 1 mm). The quality characteristic parameter data set of the simulated welding joint includes weld depth, molten pool width, thermal deformation, undercut defect length, etc. These parameters are obtained by simulating the results under different welding conditions through the virtual simulation platform.

[0053] The first step of optimization is to quantify and integrate multiple quality characteristics. This process converts multiple independent quality parameters into comparable numerical indicators and considers their impact on the overall welding quality. Each quality parameter is normalized: for example, the target range for weld depth is 0.7mm to 0.9mm. If the simulation value is 0.82mm, the normalized value is (0.82-0.7) / (0.9-0.7)=0.6; if the simulation value is 0.68mm (below the lower limit), the normalized value is (0.68-0.7) / (0.9-0.7)=-0.1. Similarly, the target range for pool width is 0.9mm to 1.1mm, and the normalized value for a simulation value of 1.05mm is (1.05-0.9) / (1.1-0.9)=0.75; the target upper limit for thermal deformation is 0.2mm, and the normalized value for a simulation value of 0.18mm is (0.2-0.18) / 0.2=0.1 (assuming reverse normalization, with smaller values indicating closer to the target); the target upper limit for undercut defect length is 1mm, and the normalized value for a simulation value of 0.5mm is (1-0.5) / 1=0.5.

[0054] The normalized parameters are weighted and summed based on their impact on welding quality to obtain a comprehensive quality indicator. The weights are determined based on historical welding experience and industry standards: for example, weld depth has a greater impact on structural strength, with a weight of 0.3; thermal deformation directly affects the assembly accuracy of the workpiece, with a weight of 0.4; pool width affects the density of the weld, with a weight of 0.2; and undercut defect length reflects the surface quality, with a weight of 0.1. Taking a simulation result as an example, the normalized weld depth is 0.6 (weight 0.3), the pool width is 0.75 (weight 0.2), the thermal deformation is 0.1 (weight 0.4), and the undercut defect length is 0.5 (weight 0.1), so the comprehensive quality indicator is 0.6x0.3+0.75x0.2+0.1x0.4+0.5x0.1=0.18+0.15+0.04+0.05=0.42. This indicator is compared with the pre-defined target comprehensive indicator (usually set to 1 or a specific threshold) to determine whether the current simulation result meets the optimization requirements.

[0055] The purpose of analyzing the environmental compensation gradient is to analyze the influence of environmental factors on the comprehensive quality indicator and establish a quantitative relationship between environmental parameters and quality indicators. Environmental factors include environmental temperature (e.g. 20℃ to 30℃), relative humidity (e.g. 40% to 60%), wind speed (e.g. 0.1m / s to 0.5m / s), etc. These factors indirectly change the heat distribution during welding by affecting heat dissipation efficiency and material thermal conductivity. For example, an increase in environmental temperature increases the initial temperature of the material, slows down the cooling rate of the molten pool under the same welding conditions, and may increase the weld depth and thermal deformation.

[0056] The analysis process needs to collect multiple sets of environmental parameters and corresponding comprehensive quality indicators. For example, under the same welding parameters (laser power 2500W, welding speed 5m / min), record the corresponding comprehensive quality indicators under different environmental temperatures (20°C, 25°C, 30°C), humidity (40%, 50%, 60%), and wind speed (0.1m / s, 0.3m / s, 0.5m / s). Through statistical analysis, it is found that for every 5°C increase in environmental temperature, the comprehensive quality indicator decreases by 0.05; for every 10% increase in humidity, the comprehensive quality indicator decreases by 0.03; for every 0.2m / s increase in wind speed, the comprehensive quality indicator increases by 0.02 (assuming that increasing wind speed accelerates heat dissipation, which may improve thermal deformation but affects the stability of the molten pool). These quantitative relationships constitute the environmental compensation gradient, which is used for subsequent regulation.

[0057] The construction of the multi-level compensation regulation network is based on the environmental compensation gradient, which divides the adjustment range of the control parameters according to the gradient level. The regulation network usually includes three levels of parameters: the primary parameter is the cooling liquid flow rate (adjustment range 3L / min to 8L / min), the secondary parameter is the laser power (adjustment range 2000W to 3000W), and the tertiary parameter is the welding speed (adjustment range 4m / min to 6m / min). Each level of parameter corresponds to a different compensation gradient level, for example, when the environmental temperature deviates from the reference temperature (25°C) within ±2°C, only the primary parameter needs to be adjusted; when the deviation exceeds ±2°C but does not exceed ±5°C, the primary and secondary parameters need to be adjusted simultaneously; when the deviation exceeds ±5°C, the three parameters need to be adjusted simultaneously.

[0058] The acquisition of environmental dynamic collaborative regulation parameters requires real-time environmental sensors to collect current environmental data and input the environmental compensation gradient model to calculate the compensation values of each parameter. For example, the environmental temperature is 28°C, which is 3°C higher than the reference temperature of 25°C. According to the gradient relationship, the comprehensive quality indicator needs to be compensated by 0.03 (for every 5°C increase in temperature, the compensation is 0.05, and 3°C corresponds to 0.03); the humidity is 55%, which is 5% higher than the reference humidity of 50%, and the compensation is 0.015 (for every 10% increase, the compensation is 0.03, and 5% corresponds to 0.015); the wind speed is 0.4m / s, which is 0.1m / s higher than the reference wind speed of 0.3m / s, and the compensation is -0.01 (for every 0.2m / s increase, the compensation is 0.02, and 0.1m / s corresponds to -0.01). The comprehensive compensation value is 0.03+0.015-0.01=0.035, which needs to be adjusted by adjusting the control parameters to make the comprehensive quality indicator increase by 0.035 to reach the target value.

[0059] The parameter calibration process adjusts the parameters of the multi-stage control network according to the comprehensive compensation value. The primary parameter, the cooling liquid flow rate, is adjusted: the reference flow rate is 5 L / min, and the flow rate needs to be increased to improve the heat dissipation efficiency when the ambient temperature rises by 3°C. Assuming that the comprehensive quality index increases by 0.02 for each 1 L / min increase in flow rate, the flow rate needs to be increased from 5 L / min to 6 L / min. At this time, the comprehensive index increases to 0.42+0.02=0.44, which is still lower than the target value (assuming the target value is 0.45), and the secondary parameter, the laser power, needs to be further adjusted. The reference power is 2500W, and reducing the power by 50W can accelerate the cooling rate of the molten pool, and the comprehensive index increases by 0.015, so the adjusted power is 2450W (compensation 0.015), and the comprehensive index increases to 0.44+0.015=0.455, which is close to the target value. If it still does not fully meet the requirements, the tertiary parameter, the welding speed, needs to be adjusted. The reference speed is 5 m / min, and reducing the speed by 0.2 m / min to 4.8 m / min can reduce the heat input, and the comprehensive index increases by 0.005, and the final comprehensive index reaches 0.46, which meets the optimization requirements.

[0060] The generated intelligent welding control execution scheme includes a specific parameter adjustment sequence: the cooling liquid flow rate is maintained at 5 L / min for 5 seconds before welding, and then increased to 6 L / min and maintained; the laser power is linearly reduced from 2500W to 2450W during the welding process; the welding speed is maintained at 5 m / min in the first half of the weld (0mm-100mm), and reduced to 4.8 m / min in the second half (100mm-200mm). The scheme integrates multi-dimensional quality characteristics, analyzes environmental compensation, and adjusts multiple parameters to ensure the thermal stability of the welding process, while meeting the pre-defined structural form parameters and quality tolerance targets, providing clear control guidance for actual welding operations.

[0061] It should be noted that, in this document, relational terms such as first and second and the like can be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0062] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, replacements and variations can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent control of a water-cooled, ultra-compact, integrated handheld laser welding machine, comprising the following steps: Based on the real-time temperature monitoring data of the water-cooled heat dissipation unit and the power operation parameters of the laser emitter, a thermal dynamic control model is trained and generated. The trained thermal dynamic control model is used to predict the heat distribution changes during welding operations in real time, and the coolant flow rate of the water cooling unit is dynamically adjusted based on the predicted heat distribution changes. By combining the adjusted coolant flow rate data, the motion trajectory of the laser welding head is virtually simulated to evaluate the quality characteristic parameters of the simulated welded joint. The system performs optimization calculations on the dataset of quality characteristic parameters of the simulated welded joint based on a predefined welding quality optimization objective, and generates and outputs an intelligent welding control execution scheme.

2. The intelligent control method for a water-cooled, ultra-small integrated handheld laser welding machine according to claim 1, wherein training and generating the thermal dynamic control model includes: The system collects real-time temperature monitoring data of the water-cooled heat dissipation unit and power operation parameter data of the laser emitter. The real-time temperature monitoring data includes the time-series readings of the temperature sensor and the heat dissipation performance index, while the power operation parameter data includes the laser power setpoint and pulse duration. Thermal behavior correlation feature analysis is performed based on real-time temperature monitoring data and power operation parameter data to construct a thermal dynamic training dataset. The model is trained iteratively using a thermal dynamic training dataset until the model convergence condition is met, thus obtaining a thermal dynamic control model.

3. The intelligent control method for a water-cooled, ultra-small integrated handheld laser welding machine according to claim 2, wherein thermal behavior correlation feature analysis is performed based on real-time temperature monitoring data and power operation parameter data to construct a thermal dynamic training dataset, including: By analyzing the characteristic variation law of the welding heat-affected zone through real-time temperature monitoring data, the interaction coefficient between temperature and power is determined. By integrating power operation parameter data with the interaction coefficient between temperature and power, the synergistic effect of power adjustment on temperature stability is assessed, and the synergistic influence factor of power adjustment on temperature stability is derived. A thermal dynamic training dataset is constructed by integrating the interaction coefficient between temperature and power and the synergistic influence factor of power adjustment. The data label of the thermal dynamic training dataset is the predicted temperature value.

4. The intelligent control method for the ultra-small integrated handheld laser welding machine with water cooling heat dissipation according to claim 3, wherein determining the interaction coefficient between temperature and power by analyzing the characteristic change law of the welding heat-affected zone through real-time temperature monitoring data includes: By utilizing real-time temperature monitoring data, we can analyze temperature change patterns under different power configurations. The actual steady-state temperature values ​​under different power configurations were obtained through experimental testing. Align the power configuration conditions, temperature variation patterns, and actual steady-state temperature values ​​to determine the interaction coefficient between temperature and power.

5. The intelligent control method for a water-cooled, ultra-small integrated handheld laser welding machine according to claim 3, wherein the integrated power operation parameter data and the interaction coefficient between temperature and power are used for synergistic regulation impact assessment, and the synergistic impact factors for power adjustment are derived as follows: Based on the power operating parameter data, a comparison temperature feature dataset is created, which covers temperature record information under various heat dissipation conditions with the same power configuration and the same welding materials. Based on the comparison temperature feature dataset, a temperature stability test was performed to obtain the temperature deviation of the comparison temperature feature dataset. By combining the temperature deviation with the interaction coefficient between temperature and power under the no-power-adjustment baseline, the synergistic influence factor of power adjustment is calculated.

6. The intelligent control method for a water-cooled, ultra-small integrated handheld laser welding machine according to claim 2, wherein the model training iteration process using a thermal dynamic training dataset includes: The dynamic training dataset is divided proportionally to form a model training subset and a model validation subset. An adaptive deep learning algorithm was selected for training the thermal dynamic control model. The training subset was used to optimize the weights and adjust the hyperparameters of the thermal dynamic control model, while the validation subset was used to verify the error and optimize the performance of the optimized model until the preset accuracy standard was met, thus confirming the final thermal dynamic control model.

7. The intelligent control method for a water-cooled, ultra-small integrated handheld laser welding machine according to claim 1, wherein the application of a trained thermal dynamic control model to predict heat distribution changes during welding operations in real time, and the dynamic adjustment of the coolant flow rate of the water-cooling unit based on the predicted heat distribution change results, includes: Obtain the predicted heat distribution changes output by the thermal dynamic control model; With the goal of minimizing thermal fluctuations, the system searches for a coolant flow rate adjustment strategy for the water-cooled unit based on the predicted changes in heat distribution, and generates the optimal coolant flow rate configuration.

8. The intelligent control method for a water-cooled, ultra-small integrated handheld laser welding machine according to claim 1, wherein the motion trajectory of the laser welding head is virtually simulated based on the adjusted coolant flow rate data, and the quality characteristic parameters of the simulated welded joint are evaluated, including: Receive the adjusted coolant flow rate data, simulate the motion trajectory of the laser welding head, and output the quality characteristic parameter data of the simulated welded joint, including weld depth, molten pool width and thermal deformation. Using the welding quality optimization target as the evaluation benchmark, the welding path geometry, movement speed and energy input schemes are explored based on the quality characteristic parameter data of the simulated welded joint, and the welding control channel is optimized.

9. The intelligent control method for a water-cooled, ultra-small integrated handheld laser welding machine according to claim 8, wherein performing optimization calculations on the simulated weld joint quality characteristic parameter dataset for a predefined welding quality optimization target includes: Based on the quality characteristic parameter data of the simulated welded joint, multi-dimensional quality feature quantification and integration processing is performed to generate a multi-quality index quantification matrix. Environmental compensation gradient analysis is performed based on a multi-quality index quantification matrix. A multi-level compensation and control network is constructed according to the gradient level. The multi-level compensation and control network contains multi-level control parameter items. Acquire dynamic and coordinated environmental control parameters; By applying dynamic and coordinated control parameters of the application environment, parameter calibration is performed step by step on a multi-level compensation and control network to generate an environmental coordinated control strategy. The assessment benchmarks were adjusted in accordance with the environmental coordinated regulation strategy. Based on the adjusted evaluation criteria, a search for geometric shape, movement speed, and energy input schemes is performed on the welding path to optimize the welding control channel.

10. The intelligent control method for a water-cooled, ultra-small integrated handheld laser welding machine according to claim 1, wherein the predefined welding quality optimization targets include welding structure morphology parameter targets and quality tolerance targets.

Images