Method for reducing damage to walls based on a penetration model for regional light control

Abstract

The present application belongs to the field of laser processing, and relates to a method for reducing damage to walls based on a penetration model and area light control, comprising the following steps: 1) constructing a penetration model; 2) judging the penetration model constructed in step 1) based on camera visual interpretation technology to obtain a judgment result; 3) capturing images of an actual production process in real time, comparing the images of the actual production process with the judgment result obtained in step 2), and sending a control signal to a light control system according to a comparison result; and 4) adjusting the distribution of light sources in time according to the control signal in step 3) to complete an actual hole forming process. The present application provides a method for reducing damage to walls based on a penetration model and area light control, which can make judgments in time and reduce strict requirements on real-time performance of a system.

Description

Technical Field

[0001] This invention belongs to the field of laser processing and relates to a method for monitoring and controlling the laser processing process, and more particularly to a method for mitigating wall damage by regional light control based on a penetration model. Background Technology

[0002] Ultrafast lasers, due to their high precision, high efficiency, low heat-affected zone, and wide applicability to difficult-to-machine materials, have been widely applied in recent years in various precision machining fields such as aerospace, electronic circuits, and shipbuilding. Laser processing is a non-contact process; after penetrating the machined surface, it continues to act on the back side of the material. However, ultrafast laser processing inevitably involves internal cavities in parts, and the distance between these cavities is generally small. During laser processing, the energy at the center of the laser beam spot is high, leading to rapid heat accumulation. This results in sufficient laser energy remaining within a certain range along the processing direction. If the high-energy laser beam spot penetrates the cavity and directly acts on the surface of the cavity wall, it is highly likely to remove the wall material, causing wall damage, and in severe cases, rendering the entire part unusable. To avoid wall damage, traditional protection techniques mainly involve filling the cavity with material to block the laser from acting on the wall material. However, this method is difficult to protect small cavities, and the removal of the filling material after processing affects the overall processing efficiency of the part. Therefore, in recent years, methods have emerged that monitor the processing to achieve wall protection. These methods detect penetration and immediately stop the laser, thus preventing wall damage. However, this method requires extremely high real-time performance, and the ultrafast laser's action time is extremely short, making it easy for wall damage to occur during the process of controlling the reduction of laser power.

[0003] Mei Xuesong et al. proposed a combined protection method and system for laser processing backwall based on interlayer differences. This method involves processing a target material filled with protective material, collecting monitoring signals during the processing, extracting monitoring signal features from these signals, and inputting these features into an interlayer state decision model to output the processing state. If the state is that the laser is on the protective material, the laser processing parameters are adjusted to modify the target material, and the laser is shut off before penetrating the protective material. This method establishes a feature-input interlayer state decision model, which is built using machine learning methods such as support vector machines, random forests, and convolutional neural networks. This method relies on real-time monitoring signal processing and complex machine learning models to determine the processing state, which requires the system to have powerful data processing capabilities and a high-performance computing platform, increasing the system's complexity and cost. Furthermore, model training requires a large amount of sample data and a meticulous parameter tuning process to ensure that the model can accurately identify different processing states. Data may vary under different materials and processing conditions; if the model's generalization ability is poor, it may lead to misjudgments or inaccurate control in practical applications. Summary of the Invention

[0004] In order to solve the above-mentioned technical problems in the background art, the present invention provides a method for mitigating wall damage by regional light control based on a penetration model, which can make timely judgments and reduce the strict requirements on system real-time performance.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A method for mitigating wall damage through regional light control based on a penetration model, characterized in that: the method includes the following steps:

[0007] 1) Construct a penetration model;

[0008] 2) Based on camera vision interpretation technology, interpret the penetration model constructed in step 1) to obtain the interpretation result;

[0009] 3) Capture images of the actual production process in real time, compare the images of the actual production process with the interpretation results obtained in step 2), and send control signals to the light control system based on the comparison results;

[0010] 4) The light control system adjusts the distribution of the light source in a timely manner according to the control signal in step 3) to complete the actual hole-making process.

[0011] The specific implementation method of step 1) above is as follows:

[0012] 1.1) Collect data from the previous penetration process;

[0013] 1.2) Establish a mathematical model between the penetration process and the processing process based on the data collected in step 1.1). The mathematical model between the penetration process and the processing process is a penetration model. The penetration model is a penetration process under the combined action of multiple variables. The variables include the number of layers, the workpiece tilt angle, the hole size, and specific process parameters.

[0014] The specific method for establishing the mathematical model in step 1.2) above is as follows:

[0015] 1.2.1) Establish the relationship between the penetration process and the number of processing layers based on the data collected in step 1.1);

[0016] 1.2.2) Based on step 1.2.1), establish the relationship between the penetration process and the workpiece tilt angle;

[0017] 1.2.3) Based on step 1.2.2), establish the relationship between the penetration process and the aperture;

[0018] 1.2.4) Based on step 1.2.3), establish the relationship between the penetration process and the process parameters, and finally obtain the mathematical model between the penetration process and the processing process.

[0019] The specific implementation method of step 2) above is as follows:

[0020] 2.1) Use a camera to capture images of the actual penetration process in real time; the images of the actual penetration process include images at the start of the actual penetration process and images at the point of complete penetration;

[0021] 2.2) The images captured in the actual penetration process in step 2.1) are processed and analyzed by a deep learning training model to obtain the interpretation results, which are the initial penetration layer and the complete penetration layer of the penetration model.

[0022] In step 2.2) above, the deep learning model is an image recognition model built by combining support vector regression, decision tree regression, long short-term memory neural network or Transformer;

[0023] The image processing method in step 2.2) is edge detection, pattern recognition, or dynamic analysis;

[0024] In step 2.2), the initial penetration layer is determined based on the micro-cracks, color changes, and / or brightness changes on the material surface.

[0025] The specific implementation method of step 3) above is as follows:

[0026] 3.1) Capture images of the actual production process in real time;

[0027] 3.2) Compare the images of the actual production process with the interpretation results obtained in step 2) to obtain the comparison results;

[0028] 3.3) If the comparison result is the initial penetration layer of the penetration model, then send the initial penetration signal to the light control system; if the comparison result is the complete penetration layer of the penetration model, then send the complete penetration signal to the light control system.

[0029] The aforementioned light control system includes a controller and a laser connected to the controller.

[0030] The specific implementation method of step 4) above is as follows:

[0031] 4.1) The controller receives the control signal sent in step 3);

[0032] 4.2) The control signal is converted into an electrical signal suitable for the laser through the drive circuit;

[0033] 4.3) Adjust the output power and wavelength of the laser according to the electrical signal to control the laser's output or off state. At the same time, adjust the light source distribution according to the penetration state to achieve the final actual hole-making process.

[0034] In step 4.3) above, the penetration state is the initial penetration state and the complete penetration state.

[0035] The specific implementation method for adjusting the light source distribution according to the penetration state in step 4.3) above is as follows:

[0036] When the controller receives the initial penetration signal, it determines that the laser is in the initial penetration state. The controller then controls the laser to continue working until the controller receives the complete penetration signal, at which point all light sources are turned off.

[0037] When the controller receives a complete penetration signal, it determines that the laser is in a complete penetration state and then controls the laser to stop working and turn off all light sources.

[0038] The advantages of this invention are:

[0039] This invention provides a method for mitigating wall damage through regional light control based on a penetration model. Compared to existing technologies, this invention adopts a more direct and efficient strategy. It no longer relies on deep learning models for continuous state recognition throughout the entire processing process, but instead focuses on two key penetration moments: initial penetration and complete penetration. The camera can capture image changes during processing in real time and uses image processing techniques to accurately determine these two penetration signals. Once the camera determines that initial or complete penetration has occurred, it immediately sends a signal to the light control system. Because this signal is sent only after penetration is confirmed, it eliminates the need for continuous state recognition and feedback like deep learning models, thus significantly reducing the system's real-time performance pressure. In summary, this invention simplifies the feedback mechanism, shifting the focus from continuous monitoring of the entire processing process to accurate judgment of key penetration events. This effectively reduces the system's direct dependence on real-time performance while still enabling regional light control based on the processing state to mitigate wall damage. Detailed Implementation

[0040] This invention provides a method for mitigating wall damage through regional light control based on a penetration model, comprising the following steps:

[0041] 1) The penetration process is qualitatively described from the actual penetration situation, leading to the derivation of a penetration model. This model involves collecting data on the penetration process and establishing a mathematical model relating the penetration to the machining process. The penetration model describes the penetration process under the combined influence of multiple variables, including the number of layers, workpiece tilt angle, aperture size, and specific process parameters. This model systematically describes how the penetration trend affects the specific location and size of the penetration area.

[0042] Specifically, the method used in this invention to establish the penetration model is as follows:

[0043] In studying the penetration process and its regional changes, the crucial step of data acquisition is paramount. To ensure data accuracy and reliability, a systematic iterative approach is employed, based on the relationship between penetration and the number of layers, to progressively collect data. Specifically, with fixed equipment parameters, process parameters, and workpiece angles, holes with identical parameters are repeatedly processed, with the number of layers gradually increasing each time until complete penetration is achieved. After each processing step, the distribution of the penetration area is precisely measured and the data is recorded. The mathematical model employed in this invention is established as follows: First, based on the data collected in step 1.1), the relationship between the penetration process and the number of processing layers is established; second, based on the aforementioned, the relationship between the penetration process and the workpiece tilt angle is established; subsequently, the relationship between the penetration process and the hole diameter is established; and finally, the relationship between the penetration process and the process parameters is established, ultimately yielding a mathematical model between the penetration process and the processing process. To eliminate the influence of potential random factors during processing, multiple rounds of testing are conducted, increasing the sample size to improve data representativeness and statistical significance. Subsequently, statistical analysis methods are used to process the collected data in depth, calculating the probability distribution of each penetration layer, thereby revealing the intrinsic connection between the penetration process and the number of layers.

[0044] After collecting all the data, the data first needs to be grouped according to the number of processing layers to ensure that each group corresponds to the penetration situation of the same layer. Then, the mean and variance of the penetration area for each layer are calculated to obtain the basic distribution of the penetration area. Finally, a probability distribution fitting is performed. For example, the distribution of the penetration area can be described by a normal distribution. Assume that the size of the penetration area of ​​the i-th layer is X. i Follows a normal distribution Where μ i It is the mean. It is the variance. For each layer i, its μ needs to be estimated. i and variance

[0045] The mean is calculated using the following formula:

[0046]

[0047] in:

[0048] n i It is the sample size of the i-th layer;

[0049] x ij It is the size of the penetration region of the j-th sample in the i-th layer.

[0050] The variance is calculated as follows:

[0051]

[0052] For a normal distribution, the probability density function is:

[0053]

[0054] The probability distribution of each penetrating region is calculated to describe the membership domain and confidence level of the penetrating region in subsequent modeling, while revealing the intrinsic relationship between the penetration process and the number of layers.

[0055] Because the penetration process is related not only to the number of processing layers but also to the workpiece angle, it is necessary to further test the relationship between penetration and workpiece angle, based on understanding the relationship between penetration and the number of layers, to make the established mathematical model more universal. According to the actual blade processing requirements, the workpiece tilt angle is adjusted, and the changes in the penetration area of ​​each layer under different tilt angles are recorded in detail. Through comparative analysis, the influence of the workpiece angle on the size of the penetration area is obtained, thus ensuring that the subsequently established mathematical model can adapt to different workpiece angles. Next, following the same logic, the relationship between penetration and aperture and process parameters is tested sequentially. In each test, other variables need to be strictly controlled to ensure that the influence of target factors on the penetration process can be accurately captured. Through these tests, the size of the penetration area of ​​different layers at different angles is obtained. These penetration area values ​​are used to directly construct a functional relationship related to the processing progress. When choosing an appropriate functional form, it is necessary to consider the characteristics of the change in the size of the penetration area with the processing progress, whether it is a linear or non-linear change. Finally, a mathematical model that can comprehensively describe the relationship between the size of the penetration area and the processing progress at different processing stages is constructed. This model not only considers multiple influencing factors such as the number of layers, workpiece angle, hole diameter, and process parameters, but also reveals the interaction mechanism between them.

[0056] In materials penetration technology, especially when multilayer materials or complex structures are involved, the penetration process often exhibits significant uncertainty and randomness. This uncertainty is mainly reflected in the unpredictable number of initial penetration layers; that is, penetration may begin from any layer of the material, depending on various factors such as the material's physical properties, the laser's power, and operating conditions.

[0057] The penetration model is established by describing the penetration process by obtaining the size of the hole penetration area after each layer of processing from the start of penetration to complete penetration (which can be obtained by two-dimensional detection or similar detection methods). However, in reality, the parameters affecting the penetration model are not only the number of layers required for penetration, but also the hole diameter, angle, and process parameters. Therefore, it is necessary to obtain the distribution of the penetration area under the number of layers, hole diameter, angle, and process parameters, and establish the penetration model based on the distribution of the penetration area and the processing progress.

[0058] 2) Introduce camera vision interpretation, use the camera to interpret the initial penetration layer and the fully penetration layer, and provide penetration signals for controlling light in the open area;

[0059] To address the uncertainties in the hole-making process and improve the accuracy and control precision of the penetration model, camera vision interpretation technology was introduced. The initial penetration layer and the fully penetrated layer are visually interpreted to provide signals for controlling the light in the opening area.

[0060] In step 2, visual interpretation utilizes a camera to capture images of the penetration process in real time. A deep learning model is used to train and analyze these images, enabling automatic interpretation of the processing. The final interpretation result is then output to the light control system. The deep learning model combines support vector regression, decision tree regression, long short-term memory neural networks, or Transformers to establish an image recognition model. This technology uses a camera to capture images of the processing process in real time, and then processes and analyzes these visualized images using image processing algorithms. Image processing algorithms are the core of this technology, utilizing image analysis techniques such as edge detection, pattern recognition, and dynamic analysis to identify and locate key events in the penetration process. Specifically, the image algorithm first detects and identifies image features of the initial penetration layer, such as micro-cracks on the material surface, color or brightness changes, etc. These features mark the start of the penetration process and are sent as the initial penetration signal to the regional light control system. Upon receiving the initial penetration signal, the regional light control system automatically adjusts the distribution and on / off state of the light source according to a preset penetration model. This process is dynamic, aiming to optimize the light source distribution based on real-time changes in the penetration process. As the penetration process continues, the camera continues to capture images and monitors the penetration status continuously through image processing algorithms. These algorithms employ deep learning, first collecting and labeling a large amount of image data, then constructing a suitable convolutional neural network model and defining its structure and parameters. Next, the model is trained using preprocessed image data, calculating predictions through forward propagation and optimizing model parameters using backpropagation to reduce prediction errors. After training, the model's performance is evaluated on independent test sets, and adjustments and optimizations are made as needed. Finally, the trained model is used to recognize new images. When complete penetration is detected, a complete penetration signal is emitted. This signal is then sent to the area control system, triggering it to shut down all light sources, marking the end of the aperture-making process.

[0061] The penetration signal in step 2 is a software signal and is sent to the control system.

[0062] 3) After receiving the penetration signal, the light control system is triggered to adjust the distribution of the light source according to the penetration state;

[0063] The light control system in step 3 consists of a controller and a laser. This system can receive and analyze commands from external or internal sources. Based on the analysis results, the controller generates control signals, which are then converted into electrical signals suitable for the laser via a drive circuit. These electrical signals directly act on the laser to adjust key parameters such as output power and wavelength, thereby controlling the laser's light output or off state and adjusting the light source distribution according to the transmission status.

[0064] 4) Once the complete penetration signal is obtained, turn off all the light to complete the hole making process.

[0065] Step 4 involves two penetration states: initial penetration and complete penetration. The light source distribution during the penetration process is adjusted according to the penetration model. Light is turned off in penetrated areas, and only in unpenetrated areas is light turned on, until a complete penetration signal is received, at which point the aperture making process ends. The complete penetration signal is determined by the visual algorithm; once complete penetration is confirmed, the signal is sent to the light control system, which turns off all light sources, ending the aperture making process.

Claims

1. A method for mitigating wall damage by regional light control based on a penetration model, characterized in that: The method for mitigating wall damage by regional light control based on a penetration model includes the following steps: 1) Construct a penetration model, specifically: 1.1) Collect data from the previous penetration process; 1.2) Based on the data collected in step 1.1), establish a mathematical model between the penetration process and the processing process. The mathematical model between the penetration process and the processing process is a penetration model. The penetration model is a penetration process under the combined action of multiple variables. The variables include the number of layers, the workpiece tilt angle, the hole size, and specific process parameters. 2) Based on camera vision interpretation technology, interpret the penetration model constructed in step 1) to obtain the interpretation result, wherein the interpretation result is the initial penetration layer and the complete penetration layer of the penetration model; 3) Capture images of the actual production process in real time, compare the images of the actual production process with the interpretation results obtained in step 2), and send control signals to the light control system based on the comparison results; 4) The light control system adjusts the distribution of the light source in a timely manner according to the control signal in step 3) to complete the actual hole-making process; the light control system includes a controller and a laser connected to the controller; step 4) specifically involves: 4.1) The controller receives the control signal sent in step 3); 4.2) The control signal is converted into an electrical signal suitable for the laser through the drive circuit; 4.3) Based on the electrical signal, adjust the output power and wavelength of the laser to control its output or off state. Simultaneously, adjust the light source distribution according to the penetration state, thus completing the actual hole-making process. The penetration state refers to the initial penetration state and the complete penetration state. The specific implementation method for adjusting the light source distribution according to the penetration state is as follows: When the controller receives the initial penetration signal, it determines that the laser is in the initial penetration state. The controller then controls the laser to continue working until the controller receives the complete penetration signal, at which point all light sources are turned off. When the controller receives a complete penetration signal, it determines that the laser is in a complete penetration state and then controls the laser to stop working and turn off all light sources.

2. The method for mitigating wall damage based on a penetration model according to claim 1, characterized in that: The specific method for establishing the mathematical model in step 1.2) is as follows: 1.2.1) Establish the relationship between the penetration process and the number of processing layers based on the data collected in step 1.1); 1.2.2) Based on step 1.2.1), establish the relationship between the penetration process and the workpiece tilt angle; 1.2.3) Based on step 1.2.2), establish the relationship between the penetration process and the aperture; 1.2.4) Based on step 1.2.3), establish the relationship between the penetration process and the process parameters, and finally obtain the mathematical model between the penetration process and the processing process.

3. The method for mitigating wall damage based on a penetration model according to claim 2, characterized in that: The specific implementation method of step 2) is as follows: 2.1) Use a camera to capture images of the actual penetration process in real time; the images of the actual penetration process include images at the start of the actual penetration process and images at the point of complete penetration; 2.2) The images captured in the actual penetration process in step 2.1) are processed and analyzed by a deep learning training model to obtain the interpretation results.

4. The method for mitigating wall damage based on a penetration model according to claim 3, characterized in that: In step 2.2), the deep learning model is an image recognition model built by combining support vector regression, decision tree regression, long short-term memory neural network or Transformer. The image processing method in step 2.2) is edge detection, pattern recognition, or dynamic analysis; In step 2.2), the initial penetration layer is determined based on the micro-cracks, color changes, and / or brightness changes on the material surface.

5. The method for mitigating wall damage based on a penetration model according to claim 4, characterized in that: The specific implementation method of step 3) is as follows: 3.1) Capture images of the actual production process in real time; 3.2) Compare the images of the actual production process with the interpretation results obtained in step 2) to obtain the comparison results; 3.3) If the comparison result is the initial penetration layer of the penetration model, then send the initial penetration signal to the light control system; if the comparison result is the complete penetration layer of the penetration model, then send the complete penetration signal to the light control system.

Images