An ultrafast laser precision cutting method with parallel burr removal

Abstract

The present invention discloses an ultrafast laser precision cutting method for removing burrs in parallel, comprising: conveying materials to the cutting positions of the first galvanometer system and the second galvanometer system, the materials comprising metal foil and a coating coated on the metal foil; determining the total path of the tab cutting and the straight cutting path in the total path of the tab cutting according to the tab specification parameters; pre-cutting the metal foil along the straight cutting path by the first galvanometer system; then cutting the metal foil along the total path of the tab cutting by the second galvanometer system, and cleaning the pre-cutting path of the metal foil; the laser light sources of the first galvanometer system and the second galvanometer system are nanosecond lasers and ultrafast lasers respectively; and the straight cutting path is located at the coating part of the metal foil. The method can reduce the investment cost of the tab cutting equipment, reduce the burrs and melt beads generated during the tab cutting process, achieve high-quality tab cutting at a lower cost, and improve battery performance and yield.

Description

Technical Field

[0001] The invention relates to the technical field of laser precision cutting processing, and in particular to an ultrafast laser precision cutting method for concurrently removing burrs. Background Art

[0002] As the scale of the new energy vehicle market continues to expand, the demand for power batteries is also increasing, and power lithium batteries have also received more attention. At present, there are two ways to die-cut the tabs: metal die-cutting and laser die-cutting. Among them, metal die-cutting has the problem of tool wear, which easily leads to unstable cutting process, resulting in poor quality of pole piece die-cutting. Laser die-cutting has the characteristics of high production efficiency and good process stability, and has been widely used in the die-cutting of the tabs of power lithium batteries.

[0003] The principle of laser die-cutting of the tabs is that the laser beam acts on the battery pole piece according to a certain trajectory under the action of the galvanometer system, thereby completing the tab die-cutting. Since there are battery material coating areas in the positive and negative electrode areas of the battery pole piece, for example, there is a graphite coating layer more than 100 microns thick on the negative electrode copper foil, and there is a lithium iron phosphate coating layer more than 100 microns thick on the positive electrode aluminum foil; due to the existence of this coating layer, a high-power nanosecond laser is often required for cutting. The strong thermal effect during the cutting process of this high-power laser energy can easily produce burrs and melt beads, which has an important impact on the quality of the cutting edge and the performance and quality of the battery.

[0004] Patent No. CN202111152468.9 discloses a segmented laser cutting method, which determines the tab cutting path and linear cutting path according to the tab specification parameters, uses the first galvanometer system to cut the metal foil material along the tab cutting path, and the second galvanometer system to cut the coating along the linear cutting path, and alternately cuts out the formed tabs in sequence. This method can avoid the situation where the laser cutting path of the galvanometer system is too long, resulting in laser energy attenuation, so that the material jitter has a certain error rate when the equipment runs at high speed, avoids jitter problems caused by waste materials, and improves the cutting accuracy and cutting efficiency of laser cutting. However, the first galvanometer system and the second galvanometer system in this patent both use high-energy nanosecond lasers, which will inevitably lead to a large number of burrs and melt beads in the tab cutting process, affecting the cutting edge quality and reducing the product yield. Summary of the invention

[0005] In view of the above-mentioned technical problems existing in the prior art, the present invention provides an ultrafast laser precision cutting method for parallel burr removal, which can reduce the investment cost of the pole lug cutting equipment, reduce the burrs and melt beads generated during the pole lug cutting process, achieve high-quality pole lug cutting at a lower cost, and improve battery performance and yield.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] The present invention provides an ultrafast laser precision cutting method for concurrently removing burrs, comprising the following steps:

[0008] Conveying materials to cutting positions of the first galvanometer system and the second galvanometer system, the materials comprising metal foil and a coating applied on the metal foil;

[0009] Determine the total path for cutting the tab and the straight-line cutting path in the total path for cutting the tab according to the specification parameters of the tab;

[0010] First, the first galvanometer system pre-cuts the metal foil along a straight cutting path;

[0011] The second galvanometer system then cuts the metal foil along the total path of the tab cutting and cleans the pre-cut path of the metal foil;

[0012] The first galvanometer system uses a nanosecond laser as a laser light source, and the second galvanometer system uses an ultrafast laser as a laser light source;

[0013] The straight cutting path is located at the coating portion of the metal foil, and the end of the straight cutting path forms a circular arc corner and extends through the circular arc corner to the uncoated coating portion of the metal foil.

[0014] The first galvanometer system and the second galvanometer system can generate a laser beam to the laser head through a laser, and the galvanometer controls the deflection of the laser beam. The field lens focuses the laser beam and then emits it to the surface of the material. The laser beam cuts the surface of the material along a specified scanning path to remove excess waste, thereby forming the required power lithium battery tabs.

[0015] Using the above technical solution:

[0016] At present, the lasers used in laser tab cutting are mainly nanosecond lasers. In the process of cutting the pole piece, the surface of the material will be greatly affected by heat, resulting in the generation of molten beads and burrs, which has an important impact on the quality of the cutting edge and the performance and quality of the battery. Ultrafast lasers have the characteristics of ultra-short pulse width and ultra-high peak power density, and can achieve the effect of "cold processing", which can greatly reduce the thermal effect of the processing process, effectively avoid contamination of the cutting edge surface, and obtain better cutting quality. However, when using picosecond and femtosecond lasers for tab cutting, it is often necessary to use high-power lasers due to the presence of the coating layer, and ultrafast lasers often increase in price as the power increases, which increases the research and development cost of the equipment.

[0017] The present invention utilizes nanosecond laser and ultrafast laser for parallel processing to realize high-quality tab electrode cutting. First, the nanosecond laser in the first galvanometer system is used to pre-cut the thicker coated portion of the metal foil along a straight cutting path. The cost of the nanosecond laser is much lower than that of the ultrafast laser. Selecting a nanosecond laser with appropriate power can meet the cutting requirements of metal foil coatings of different types and thicknesses. Further, a low-power ultrafast laser (femtosecond or picosecond laser) is used in the second galvanometer system to perform secondary cutting of the metal foil according to the total tab cutting path, and the pre-cutting path of the metal foil is cleaned to clean the burrs and melt beads generated by the nanosecond laser cutting. At the same time, the ultrafast laser is used to complete the cutting of the thinner portion of the metal foil, thereby realizing complete cutting of the tab, effectively improving the processing quality and reducing the cost of the equipment.

[0018] Furthermore, the power of the nanosecond laser is 10-200W, and the power of the ultrafast laser is 10-30W.

[0019] The wavelength of the nanosecond laser light source can be selected based on common positive and negative electrodes including but not limited to copper foil, aluminum foil, etc. According to different material types and coating thicknesses, a nanosecond laser light source with a power range of 10 to 200W can be selected to meet the cutting requirements of various coating battery types.

[0020] The wavelength of the ultrafast laser light source can be selected based on common positive and negative electrodes including but not limited to copper foil, aluminum foil, etc. According to different material types and the thickness of the uncoated layer, an ultrafast laser light source with a power range of 10 to 30W can be selected to meet the overall cutting requirements of the tab.

[0021] Furthermore, during the process of cutting the metal foil by the second galvanometer system along the total path of the tab cutting, the cutting path of the second galvanometer system is monitored and automatically adjusted in real time by the CCD image sensor and the host computer, so that the path for cleaning the pre-cutting path of the metal foil by the second galvanometer system coincides with the pre-cutting path of the first galvanometer system.

[0022] During the cutting process, the nanosecond laser in the first galvanometer system pre-cuts the thicker coated part of the metal foil along the straight cutting path, the cutting path is short, and the cutting is along a straight line, and the cutting accuracy of this process is high. However, when the metal foil is further cut by the low-power ultrafast laser in the second galvanometer system according to the total path of the ear cutting, the metal foil moves rapidly, and the laser beam in the second galvanometer system moves rapidly. Cutting off the waste during the secondary cutting will cause the problem of material jitter. The galvanometer system is prone to jitter when performing precision cutting according to the preset flight cutting algorithm, resulting in deviation between the actual cutting path and the set path, and the cutting shape changes. In order to solve this technical problem, the present invention realizes real-time monitoring and automatic control of the secondary cutting path through the cooperation of the CCD image sensor facing the cutting area and the host computer, and the cleaning path of the ultrafast laser in the second optical path is overlapped with the pre-cutting path of the nanosecond laser in the first optical path with high precision. When the actual cutting path deviates from the set path, the host computer can automatically correct the secondary cutting path through the algorithm to reduce the error generated in the actual product cutting process.

[0023] Furthermore, the process of real-time monitoring and automatic control of the cutting path of the second galvanometer system by the CCD image sensor and the host computer is as follows:

[0024] The instantaneous motion state of the galvanometer in the second galvanometer system relative to the metal foil is obtained in real time by using a CCD image sensor;

[0025] The host computer processes the instantaneous motion state of the galvanometer in the second galvanometer system relative to the metal foil in real time to obtain the instantaneous speed of the galvanometer relative to the metal foil, and substitutes the instantaneous speed into the relationship model between the instantaneous speed and the jitter parameters of the galvanometer that is automatically established through a machine learning method. The jitter parameters of the galvanometer are dynamically adjusted in real time, and a new metal foil cutting path is synchronously generated according to the jitter parameters of the galvanometer adjusted in real time, and the scanning speed of the galvanometer is adjusted according to the adjusted jitter parameters. The galvanometer is adjusted with the new cutting path and scanning speed, and the output of the laser in the second galvanometer system is controlled to continue to run to complete the metal foil cutting operation.

[0026] Through the above technical scheme, real-time adaptive adjustment is introduced, the vibration compensation capability of the galvanometer in non-uniform motion is enhanced, and the anti-shake design of the galvanometer is improved. Based on the instantaneous motion speed of the galvanometer in the second galvanometer system, the vibration parameters are adaptively adjusted in real time, and the flying cutting trajectory is corrected to ensure the stability and consistency of the cutting shape, thereby improving the overall quality of the final product.

[0027] Furthermore, a multi-layer perceptron is used to automatically establish a relationship model between speed and jitter parameters.

[0028] In the present invention, a relationship model between the instantaneous speed and the jitter parameter of the galvanometer in the second galvanometer system is automatically established by a machine learning method, thereby replacing the manually set speed-jitter parameter comparison table, which can continuously update the model and optimize the adjustment strategy of the galvanometer jitter parameter according to different batches of metal foil cutting data (data sets of instantaneous speed V and jitter parameter d of different galvanometers relative to metal foil). The relationship model between speed and jitter parameters is automatically established using a multilayer perceptron, which solves the problem of nonlinear mapping relationship between galvanometer motion speed and jitter parameters.

[0029] Multilayer perceptron is a deep learning model that consists of multiple layers, including an input layer, one or more hidden layers, and an output layer. The input layer inputs the instantaneous velocity V of the galvanometer relative to the metal foil. The input layer contains one node and normalizes the input data. The hidden layer is used to learn the nonlinear relationship between the speed and the jitter parameter. There are three hidden layers in total. The number of neurons in the first layer is 64 (the specific number can be adjusted according to the complexity of the data), and the ReLU activation function is used; the number of neurons in the second layer is 64, and the ReLU activation function is also used; the number of neurons in the third layer is 32, and the ReLU activation function is also used. The number of these hidden layers and the number of neurons can be adjusted according to the specific data and performance requirements. The output layer outputs the jitter parameter d, and the output node is 1 node, corresponding to the jitter parameter d generated by the instantaneous velocity V of each input, and the corresponding activation function is a linear activation function. The loss function uses the mean square error (MSE), which is defined as:

[0030] ;

[0031] in, is the jitter parameter predicted by the model, is the actual jitter parameter. The neural network is optimized using the Adam optimizer.

[0032] Furthermore, the method of automatically establishing a relationship model between speed and jitter parameters using a multilayer perceptron includes:

[0033] Firstly, a batch of data sets of different instantaneous speeds V and jitter parameters d of the galvanometer relative to the metal foil during the metal foil cutting process are obtained, the data sets are divided into a training set and a validation set, and the input parameters are normalized;

[0034] Secondly, construct a multi-layer perceptron network structure and randomly initialize the weights;

[0035] Then, the instantaneous speed V in the training set is input through each layer of the neural network to obtain the predicted jitter parameters , calculate the loss function, and use the Adam optimizer to update the model weights so that the loss function gradually decreases;

[0036] Finally, after multiple rounds of training iterations, the model training is completed, the prediction ability of the model is evaluated by the validation set, and the trained model is saved for subsequent operations of real-time adaptive extraction of galvanometer jitter parameters based on motion speed for regulation.

[0037] In practical applications, when the cutting process involves multiple variables, the multilayer perceptron can be expanded to a multi-input network structure instead of relying solely on a single speed variable, which allows the model to more comprehensively reflect changes in working conditions.

[0038] In actual operation, the neural network structure can be more complex to optimize model performance, and machine learning methods such as convolutional neural networks, recurrent neural networks, and deep reinforcement learning algorithms can be applied to the present invention.

[0039] In the present invention, the instantaneous motion state of the galvanometer in the second galvanometer system relative to the metal foil is acquired in real time by a CCD image sensor, and the instantaneous velocity V of the galvanometer relative to the metal foil is obtained through processing.

[0040] Furthermore, during the cutting process, the material is in a static or uniformly transported state.

[0041] In specific applications, a bearing-type structure can be used to unwind the metal foil, and the material conveyed by the bearing is stationary or moves at a constant speed.

[0042] Furthermore, the first galvanometer system and the second galvanometer system cut the material synchronously and in parallel, and the cut material forms a tab.

[0043] In the processing of the tab motor, a dual optical path parallel cutting method is adopted, which can realize fully automatic synchronous cutting through the parameters set in advance, and has good adaptability and convenience for the tab cutting parameters required by various models and specifications.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] 1. The first galvanometer system of the present invention can perform laser processing pre-cutting on the metal foil with the coating layer along a straight cutting path by nanosecond laser. The cost of nanosecond laser is much lower than that of ultrafast laser. Selecting nanosecond laser with appropriate power can meet the requirements of cutting the coating layer of metal foil of different types and thicknesses;

[0046] 2. The second galvanometer system uses a low-power ultrafast laser (femtosecond or picosecond laser) to perform secondary cutting of the metal foil along the total path of the tab cutting, and cleans the pre-cutting path of the metal foil to clean the burrs and melt beads generated by nanosecond laser cutting. At the same time, the ultrafast laser is used to cut the thinner part of the metal foil, achieving complete cutting of the tab. The specific function can be achieved without the need for expensive high-power ultrafast lasers, which effectively improves the processing quality and reduces the cost of the equipment.

[0047] 3. The present invention can be applied to the processing of battery tab electrodes of various models and coating thicknesses. The first optical path nanosecond laser and the second optical path ultrafast laser can perform laser processing simultaneously and in parallel. After determining the electrode parameters and tab parameters, high-precision tab cutting and tiny burr removal can be achieved on a single processing line. The final products can directly meet the high-standard industrial application requirements.

[0048] 4. In the cutting process of the present invention, the nanosecond laser in the first galvanometer system pre-cuts the thicker coated part of the metal foil along a straight cutting path. The cutting path is short and cuts along a straight line. This process has high cutting accuracy. When the low-power ultrafast laser in the second galvanometer system performs a secondary cutting on the metal foil according to the total path of the tab cutting, the automatic monitoring and automatic control of the path can be achieved through the cooperation of the CCD image sensor facing the cutting area and the host computer. The cleaning path of the ultrafast laser in the second optical path is highly precisely overlapped with the pre-cutting path of the nanosecond laser in the first optical path. When the actual cutting path deviates from the set path, the host computer can automatically correct the secondary cutting path through an algorithm to reduce the error generated in the actual product cutting process. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0050] Figure 1 This is a workflow diagram of ultrafast laser precision cutting operation for parallel burr removal in the present invention;

[0051] Figure 2 A straight-line cutting path diagram of the first galvanometer system in the present invention;

[0052] Figure 3 It is a total path diagram of the tab cutting of the second galvanometer system in the present invention;

[0053] Figure 4 The basic structure of the neural network model between speed and jitter parameters is automatically established by using a machine learning method in the present invention;

[0054] Figure 5 A conventional lithium battery tab flying cutting trajectory in metal foil cutting generated without real-time adaptive adjustment of jitter parameters in the prior art;

[0055] Figure 6 A conventional tab flight cutting trajectory of a lithium battery in metal foil cutting generated by real-time adaptive adjustment of the jitter parameter of the motion speed in the present invention;

[0056] Figure 7 It is a schematic diagram of the structure of the ultrafast laser precision cutting system for parallel burr removal in the present invention.

[0057] Among them, the specific drawings are marked as:

[0058] Nanosecond laser 1, ultrafast laser 2, nanosecond laser transmission optical path 3, laser fixing frame 4, nanosecond laser adaptation two-dimensional galvanometer 5, nanosecond laser adaptation field mirror 6, nanosecond laser galvanometer signal input port 7, nanosecond laser processing beam 8, ultrafast laser adaptation two-dimensional galvanometer 9, ultrafast laser processing beam 10, ultrafast laser transmission optical path 11, ultrafast laser galvanometer signal input port 12, laser fixing frame base frame 13, coated metal foil part 14, bearing structure 15, ultrafast laser adaptation field mirror 16, uncoated metal foil part 17, CCD image sensor 18, host computer 19, straight line cutting path 20, tab cutting total path 21. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] Example 1

[0061] This embodiment provides an ultrafast laser precision cutting method for simultaneously removing burrs, such as Figure 1 As shown, the following steps are included:

[0062] Conveying materials to cutting positions of the first galvanometer system and the second galvanometer system, the materials comprising metal foil and a coating applied on the metal foil;

[0063] Determine the total path for cutting the tab and the straight-line cutting path in the total path for cutting the tab according to the specification parameters of the tab;

[0064] First, the first galvanometer system pre-cuts the metal foil along a straight cutting path. The straight cutting path is as follows: Figure 2 As shown;

[0065] The second galvanometer system then cuts the metal foil along the total path of the tab cutting and cleans the pre-cut path of the metal foil. The total path of the tab cutting is as follows: Figure 3 As shown;

[0066] The first galvanometer system uses a nanosecond laser as a laser light source, and the second galvanometer system uses an ultrafast laser as a laser light source;

[0067] The straight cutting path is located at the coating portion of the metal foil, the end of the straight cutting path forms a circular arc corner, and the straight cutting path extends to the uncoated coating portion of the metal foil through the circular arc corner.

[0068] The first galvanometer system and the second galvanometer system can generate a laser beam to the laser head through a laser, and the galvanometer controls the deflection of the laser beam. The field lens focuses the laser beam and then emits it to the surface of the material. The laser beam cuts the surface of the material along a specified scanning path to remove excess waste, thereby forming the required power lithium battery tabs.

[0069] At present, the lasers used in laser tab cutting are mainly nanosecond lasers. In the process of cutting the pole piece, the surface of the material will be greatly affected by heat, resulting in the generation of molten beads and burrs, which has an important impact on the quality of the cutting edge and the performance and quality of the battery. Ultrafast lasers have the characteristics of ultra-short pulse width and ultra-high peak power density, and can achieve the effect of "cold processing", which can greatly reduce the thermal effect of the processing process, effectively avoid contamination of the cutting edge surface, and obtain better cutting quality. However, when using picosecond and femtosecond lasers for tab cutting, it is often necessary to use high-power lasers due to the presence of the coating layer, and ultrafast lasers often increase in price as the power increases, which increases the research and development cost of the equipment.

[0070] The present invention utilizes nanosecond laser and ultrafast laser for parallel processing to realize high-quality tab electrode cutting. First, the nanosecond laser in the first galvanometer system is used to pre-cut the thicker coated portion of the metal foil along a straight cutting path. The cost of the nanosecond laser is much lower than that of the ultrafast laser. Selecting a nanosecond laser with appropriate power can meet the cutting requirements of metal foil coatings of different types and thicknesses. Further, a low-power ultrafast laser (femtosecond or picosecond laser) is used in the second galvanometer system to perform secondary cutting of the metal foil according to the total tab cutting path, and the pre-cutting path of the metal foil is cleaned to clean the burrs and melt beads generated by the nanosecond laser cutting. At the same time, the ultrafast laser is used to complete the cutting of the thinner portion of the metal foil, thereby realizing complete cutting of the tab, effectively improving the processing quality and reducing the cost of the equipment.

[0071] Among them, the selection of the pulse width of the ultrafast laser light source can be based on the industrial design requirements, and ultrafast picosecond lasers or ultrafast femtosecond lasers can be selected. The power of nanosecond lasers is 10 to 200W, and the power of ultrafast lasers is 10 to 30W.

[0072] The wavelength of the nanosecond laser light source can be selected based on common positive and negative electrodes including but not limited to copper foil, aluminum foil, etc. According to different material types and coating thicknesses, a nanosecond laser light source with a power range of 10 to 200W can be selected to meet the cutting requirements of various coating battery types.

[0073] The wavelength of the ultrafast laser light source can be selected based on common positive and negative electrodes including but not limited to copper foil, aluminum foil, etc. According to different material types and the thickness of the uncoated layer, an ultrafast laser light source with a power range of 10 to 30W can be selected to meet the overall cutting requirements of the tab.

[0074] During the cutting process, the material is in a static state or is transported at a constant speed. In specific applications, a bearing-type structure can be used to unwind the metal foil, and the material transported by the bearing is static or moves at a constant speed.

[0075] The first galvanometer system and the second galvanometer system perform synchronous and parallel cutting on the material, and the cut material forms the tab. In the processing of the tab motor, the dual optical path parallel cutting method is adopted, which can realize full-automatic synchronous cutting through the parameters set in advance, and has good adaptability and convenience for the tab cutting parameters required by various models and specifications.

[0076] Among them, during the process of cutting the metal foil by the second galvanometer system along the total path of the tab cutting, the cutting path of the second galvanometer system is monitored and automatically adjusted in real time through the CCD image sensor and the host computer, so that the path for cleaning the pre-cutting path of the metal foil by the second galvanometer system coincides with the pre-cutting path of the first galvanometer system.

[0077] During the cutting process, the nanosecond laser in the first galvanometer system pre-cuts the thicker coated part of the metal foil along the straight cutting path, the cutting path is short, and the cutting is along a straight line, and the cutting accuracy of this process is high. However, when the metal foil is further cut by the low-power ultrafast laser in the second galvanometer system according to the total path of the ear cutting, the metal foil moves rapidly, and the laser beam in the second galvanometer system moves rapidly. Cutting off the waste during the secondary cutting will cause the problem of material jitter. The galvanometer system is prone to jitter when performing precision cutting according to the preset flight cutting algorithm, resulting in deviation between the actual cutting path and the set path, and the cutting shape changes. In order to solve this technical problem, the present invention realizes real-time monitoring and automatic control of the secondary cutting path through the cooperation of the CCD image sensor facing the cutting area and the host computer, and the cleaning path of the ultrafast laser in the second optical path is overlapped with the pre-cutting path of the nanosecond laser in the first optical path with high precision. When the actual cutting path deviates from the set path, the host computer can automatically correct the secondary cutting path through the algorithm to reduce the error generated in the actual product cutting process.

[0078] The specific process of real-time monitoring and automatic control of the cutting path of the second galvanometer system by using the CCD image sensor and the host computer is as follows:

[0079] The instantaneous motion state of the galvanometer in the second galvanometer system relative to the metal foil is obtained in real time by using a CCD image sensor;

[0080] The host computer processes the instantaneous motion state of the galvanometer in the second galvanometer system relative to the metal foil in real time to obtain the instantaneous speed of the galvanometer relative to the metal foil, and substitutes the instantaneous speed into the relationship model between the instantaneous speed and the jitter parameters of the galvanometer that is automatically established through a machine learning method. The jitter parameters of the galvanometer are dynamically adjusted in real time, and a new metal foil cutting path is synchronously generated according to the jitter parameters of the galvanometer adjusted in real time, and the scanning speed of the galvanometer is adjusted according to the adjusted jitter parameters. The galvanometer is adjusted with the new cutting path and scanning speed, and the output of the laser in the second galvanometer system is controlled to continue to run to complete the metal foil cutting operation.

[0081] In the present invention, a relationship model between the instantaneous speed and the jitter parameter of the galvanometer is automatically established by a machine learning method, thereby replacing the manually set speed-jitter parameter comparison table, which can continuously update the model and optimize the adjustment strategy of the galvanometer jitter parameter according to different batches of metal foil cutting data (data sets of instantaneous speed V and jitter parameter d of different galvanometers relative to metal foil). The relationship model between the speed and the jitter parameter is automatically established using a multilayer perceptron, which solves the problem of the nonlinear mapping relationship between the galvanometer motion speed and the jitter parameter.

[0082] A multilayer perceptron is a deep learning model that consists of multiple layers, including an input layer, one or more hidden layers, and an output layer. Figure 4 As shown, the input layer inputs the instantaneous velocity V of the galvanometer relative to the metal foil. The input layer contains one node and normalizes the input data. The hidden layer is used to learn the nonlinear relationship between the speed and the jitter parameter. There are three hidden layers in total. The number of neurons in the first layer is 64 (which can be adjusted according to the complexity of the data), and the ReLU activation function is used; the number of neurons in the second layer is 64, and the ReLU activation function is also used; the number of neurons in the third layer is 32, and the ReLU activation function is also used. The number of these hidden layers and the number of neurons can be adjusted according to the specific data and performance requirements. The output layer outputs the jitter parameter d, and the output node is 1 node, corresponding to the jitter parameter d generated by the instantaneous velocity V of each input, and the corresponding activation function is a linear activation function. The loss function uses the mean square error (MSE), which is defined as:

[0083] ;

[0084] in, is the jitter parameter predicted by the model, is the actual jitter parameter. The neural network is optimized using the Adam optimizer.

[0085] In a specific embodiment, a multi-layer perceptron is used to automatically establish a relationship model between speed and jitter parameters, which specifically includes the following steps:

[0086] Firstly, a batch of data sets of different instantaneous speeds V and jitter parameters d of the galvanometer relative to the metal foil during the metal foil cutting process are obtained, the data sets are divided into a training set and a validation set, and the input parameters are normalized;

[0087] Secondly, construct a multi-layer perceptron network structure and randomly initialize the weights;

[0088] Then, the instantaneous speed V in the training set is input through each layer of the neural network to obtain the predicted jitter parameters , calculate the loss function, use the Adam optimizer to update the model weights, and gradually reduce the loss function. When the loss function value is less than the preset threshold ε, if the loss function is less than this threshold in the next few rounds and the change of the loss function is very small, it can be considered that the loss function has converged. Further training of the model will lead to overfitting, so the training needs to be terminated;

[0089] Finally, after multiple rounds of training iterations, the training is terminated when the preset upper limit of training rounds is reached or the loss function converges, completing the model training. The prediction ability of the model is evaluated by the validation set, and the trained model is saved for subsequent operations based on real-time adaptive extraction of the galvanometer jitter parameters for regulation based on motion speed.

[0090] If the training process reaches the preset upper limit of the number of training rounds, the loss function has not converged and the model's generalization ability is insufficient, consider adjusting the model structure, optimizing hyperparameters, or adding training data for retraining to improve model performance.

[0091] Use the validation set to predict the current model and calculate the loss function to help monitor overfitting and underfitting during training. During training, use the validation set to predict the model regularly, and the validation set loss function should continue to decrease until convergence. In order to prevent overfitting, an early stopping mechanism is introduced, which means that training can be stopped early when the loss function on the validation set no longer decreases. When the model accuracy on the validation set is high, the loss function is small, and the performance of the validation set is close to the performance of the model on the training set, it can be considered that the model has good generalization ability and the trained model can be saved.

[0092] In practical applications, when the cutting process involves multiple variables, the multilayer perceptron can be expanded to a multi-input network structure instead of relying solely on a single speed variable, which allows the model to more comprehensively reflect changes in working conditions.

[0093] In actual operation, the neural network structure can be more complex to optimize model performance, and machine learning methods such as convolutional neural networks, recurrent neural networks, and deep reinforcement learning algorithms can be applied to the present invention.

[0094] Figure 5 Shown in the figure is the conventional lithium battery tab flight cutting trajectory in the metal foil cutting generated by the prior art without real-time adaptive adjustment of the jitter parameters of the movement speed. Since the material strip moves rapidly in the lateral direction, when the tab and the galvanometer move relative to each other, the jitter of the galvanometer in the lateral direction is much greater than the jitter of the galvanometer in the longitudinal direction. The shape of the tab formed on the tab based on the flight cutting trajectory will change, thereby affecting the battery quality. Figure 6 What is shown in the figure is the flight cutting trajectory of a conventional lithium battery tab in metal foil cutting generated by real-time adaptive adjustment of the jitter parameters of the motion speed using the method in the present invention, and the generated trajectory is close to the theoretical simulation. It should be noted that the present invention only takes the trajectory within the cutting cycle of a single tab as an example for illustration, but is not limited to this, and can also be a bipolar tab, a multi-pole tab, an unconventional-shaped tab, a tab with a positioning hole, or the cutting of different materials such as circuit boards, flexible materials, film materials, and plastics. The scanning speed of the galvanometer is the speed at which the galvanometer guides the laser beam to scan the path. Due to changes in the jitter parameters of the galvanometer, the scanning speed of the galvanometer will also change accordingly. The scanning speed of the galvanometer is adjusted according to the jitter parameters of the adjusted galvanometer so that the new flight cutting trajectory is compensated for the preset trajectory.

[0095] The present invention not only copes with the change of instantaneous speed by real-time adaptive adjustment of the jitter parameters, but is also suitable for high-speed motion conditions in metal foil cutting. Especially when the material strip runs at a high speed of more than 80m / min, the present invention can dynamically compensate for the relative speed change between the laser and the metal foil to ensure the accuracy and consistency of cutting, which greatly improves the application ability in efficient and high-speed production. The present invention has the function of adjusting the jitter parameters in segments to adapt to the cutting of metal foils of different shapes: the present invention targets different trajectory segments and adaptively adjusts the jitter parameters of the galvanometer in real time according to the different cutting characteristics of each segment, so that each segment can obtain the best laser cutting effect. However, the prior art fails to solve the problem of motion speed changes caused by the complexity of different shapes, which limits its application in nonlinear trajectories. The present invention uses a machine learning method to automatically establish a relationship model between the instantaneous speed of the galvanometer and the jitter parameters. The neural network does not need to mechanically find the relationship between the instantaneous speed and the jitter parameters. The neural network will adaptively learn the best jitter adjustment strategy through training, thereby optimizing the jitter parameters. It also has a dynamic adjustment function. Under different metal foil cutting conditions, the neural network can dynamically adjust the jitter parameters according to historical data and real-time input, so that the system has stronger adaptability to different working conditions. Flexible adaptation to a variety of cutting requirements: The present invention is not only suitable for the cutting of single-pole ears, but can also be extended to the laser cutting of double-pole ears, multi-pole ears and unconventional-shaped ears, and supports the cutting requirements of ears with positioning holes. Through the setting of jitter parameters and adaptive control based on real-time feedback, the system can adapt to different workpiece shapes and trajectory requirements, and has a wide range of application scenarios. The existing technology is limited to the single shape and trajectory of the welding workpiece and has poor adaptability.

[0096] Example 2

[0097] The present embodiment provides an ultrafast laser precision cutting system for parallel burr removal, comprising a material conveying module, a first galvanometer cutting module, a second galvanometer cleaning and cutting module, and a same-path parallel cutting module. The material conveying module is responsible for transporting the metal foil with a portion of coating to the processing area; the nanosecond laser in the first galvanometer cutting module cuts the coated metal foil portion 14 along a straight cutting path 20; the ultrafast laser in the second galvanometer cutting module cleans the area cut by the first galvanometer cutting module in the same path, and cuts the uncoated metal foil portion 17; the same-path parallel cutting module is responsible for realizing the path selection and repeatability accuracy control of laser cutting according to the tab specification parameters.

[0098] like Figure 7As shown, the material conveying module is mainly used for unwinding electrode materials with the bearing structure 15, and the electrode materials are conveyed to the processing area in a straightened and stable state. The plane where the bearing is located is the same vertical or horizontal plane after the rolled electrode materials are unwound and straightened. Further, optionally, the material conveying module can be static or move at a constant speed during the processing.

[0099] The first galvanometer cutting module is mainly composed of nanosecond laser 1, nanosecond laser transmission optical path 3, nanosecond laser adaptation two-dimensional galvanometer 5, nanosecond laser adaptation field mirror 6, nanosecond laser galvanometer signal input port 7, and nanosecond laser processing beam 8. When the tab specification parameters are set and the material is transported normally, the control signal controls the nanosecond laser adaptation two-dimensional galvanometer 5 through the nanosecond laser galvanometer signal input port 7, thereby outputting nanosecond lasers that meet the corresponding cutting power, and realizing the Figure 2 The straight cutting path 20 in FIG. 1 is used to pre-cut the coated metal foil portion 14 .

[0100] The second galvanometer cleaning and cutting module is mainly composed of an ultrafast laser 2, an ultrafast laser transmission optical path 11, an ultrafast laser adapted two-dimensional galvanometer 9, an ultrafast laser adapted field mirror 16, an ultrafast laser galvanometer signal input port 12, and an ultrafast laser processing beam 10. When the tab specification parameters are set and the material is transported normally, the control signal controls the ultrafast laser adapted two-dimensional galvanometer 9 through the ultrafast laser galvanometer signal input port 12. Figure 3 The total path 21 of the tab cutting in the electrode sheet performs secondary cutting on the metal foil, and cleans the path pre-cut by the nanosecond laser on the metal foil to reduce burrs and melt beads generated during the nanosecond laser cutting process, and completes the tab cutting work on the electrode sheet without coating. The pulse width of the ultrafast laser 2 can be selected according to the industrial design requirements, and an ultrafast picosecond laser or an ultrafast femtosecond laser can be selected.

[0101] The same-path parallel cutting module mainly includes a laser fixing frame 4, a nanosecond laser galvanometer signal input port 7, an ultrafast laser galvanometer signal input port 12, a laser fixing frame base 13, a CCD image sensor 18, a host computer 19 and an external control system. It mainly realizes the simultaneous control of nanosecond laser cutting and ultrafast laser cleaning cutting functions, ensures that the ultrafast laser cleaning path is consistent with the nanosecond laser pre-cutting path, and completes the overall cutting process.

[0102] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An ultrafast laser precision cutting method for parallel burr removal, characterized in that: The steps include: Conveying materials to cutting positions of the first galvanometer system and the second galvanometer system, the materials comprising metal foil and a coating applied on the metal foil; Determine the total path for cutting the tab and the straight-line cutting path in the total path for cutting the tab according to the specification parameters of the tab; First, the first galvanometer system pre-cuts the metal foil along a straight cutting path; The second galvanometer system then cuts the metal foil along the total path of the tab cutting and cleans the pre-cut path of the metal foil; During the process of cutting the metal foil material along the total path of the tab cutting by the second galvanometer system, the cutting path of the second galvanometer system is monitored and automatically controlled in real time by the CCD image sensor and the host computer, so that the path of the pre-cutting path of the metal foil material cleaned by the second galvanometer system coincides with the pre-cutting path of the first galvanometer system; The first galvanometer system uses a nanosecond laser as a laser light source, and the second galvanometer system uses an ultrafast laser as a laser light source; The straight cutting path is located at the coating portion of the metal foil, and the end of the straight cutting path forms a circular arc corner and extends through the circular arc corner to the uncoated coating portion of the metal foil.

2. The ultrafast laser precision cutting method for parallel burr removal according to claim 1, characterized in that: The power of the nanosecond laser is 10-200W, and the power of the ultrafast laser is 10-30W.

3. The ultrafast laser precision cutting method for parallel burr removal according to claim 1, characterized in that: The specific process of real-time monitoring and automatic control of the cutting path of the second galvanometer system through the CCD image sensor and the host computer is as follows: The instantaneous motion state of the galvanometer in the second galvanometer system relative to the metal foil is obtained in real time by using a CCD image sensor; The host computer processes the instantaneous motion state of the galvanometer in the second galvanometer system relative to the metal foil in real time to obtain the instantaneous speed of the galvanometer relative to the metal foil, and substitutes the instantaneous speed into the relationship model between the instantaneous speed and the jitter parameters of the galvanometer that is automatically established through a machine learning method. The jitter parameters of the galvanometer are dynamically adjusted in real time, and a new metal foil cutting path is synchronously generated according to the jitter parameters of the galvanometer adjusted in real time, and the scanning speed of the galvanometer is adjusted according to the adjusted jitter parameters. The galvanometer is adjusted with the new cutting path and scanning speed, and the output of the laser in the second galvanometer system is controlled to continue to run to complete the metal foil cutting operation.

4. The ultrafast laser precision cutting method for parallel burr removal according to claim 3, characterized in that: A multi-layer perceptron is used to automatically establish the relationship model between instantaneous speed and jitter parameters.

5. The ultrafast laser precision cutting method for parallel burr removal according to claim 4, characterized in that: The method of automatically establishing a relationship model between instantaneous speed and jitter parameters using a multi-layer perceptron includes: Acquire a data set of instantaneous speeds V and jitter parameters d of different galvanometers relative to the metal foil during the metal foil cutting process, and divide the data set into a training set and a validation set; Construct a multi-layer perceptron network structure and randomly initialize weights; The instantaneous velocity V of the galvanometer relative to the metal foil in the training set is input through each layer of the neural network to obtain the predicted jitter parameters , calculate the loss function, and use the Adam optimizer to update the model weights so that the loss function gradually decreases; After multiple rounds of training iterations, the training is terminated when the preset upper limit of the number of training rounds is reached or the loss function converges, and the model training is completed; The prediction ability of the model is evaluated on the validation set and the trained model is saved.

6. The ultrafast laser precision cutting method for concurrent burr removal according to claim 5, characterized in that: The multilayer perceptron has three hidden layers, the number of neurons in the first and second layers is 64, the number of neurons in the third layer is 32, and the hidden layers all use the ReLU activation function.

7. The ultrafast laser precision cutting method for concurrent burr removal according to claim 1, characterized in that: During the cutting process, the material is in a static or uniformly transported state.

8. The ultrafast laser precision cutting method for concurrent burr removal according to claim 1, characterized in that: The first galvanometer system and the second galvanometer system cut the material synchronously and in parallel, and the cut material forms a pole ear.

Images