Optimization method of process parameters for hole making in laser ultrasonic vibration milling of composite materials

Abstract

The present invention provides a method for optimizing process parameters of composite material laser ultrasonic vibration milling hole making, which relates to the technical field of composite material hole making process. The present invention systematically optimizes the process parameters of composite material laser ultrasonic vibration milling hole making through multivariate regression analysis and genetic algorithm, thereby improving the prediction accuracy of tool wear rate and hole making defect rate. Combined with the real-time PID control algorithm, the dynamic adjustment of process parameters is further realized to ensure the stability and consistency of the processing process. Compared with the traditional method, this scheme greatly reduces the number of experiments and costs, improves the data quality and analysis depth, reduces tool wear and hole wall defects, and significantly improves the processing quality and production efficiency.

Description

Technical Field

[0001] The invention relates to the technical field of composite material hole making technology, in particular to a method for optimizing process parameters of composite material laser ultrasonic vibration milling hole making. Background Art

[0002] Composite materials are widely used in high-end manufacturing fields such as aerospace, automobiles, etc. due to their excellent mechanical properties and lightweight advantages. However, the high hardness and brittleness of composite materials make it easy for problems such as tool wear and hole wall cracks to occur during their processing, affecting the processing quality and efficiency. Especially in the laser ultrasonic vibration milling hole making process, the selection of process parameters is crucial to the processing results. Traditional process parameter optimization methods often rely on experience and simple experimental designs, and it is difficult to systematically consider the complex relationship between multiple process parameters, and there is a lack of scientific optimization methods. In addition, existing detection methods usually rely on manual or simple geometric measurements, which cannot accurately reflect the actual situation of tool wear and hole wall defects.

[0003] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0004] The purpose of the present invention is to provide a method for optimizing process parameters of laser ultrasonic vibration milling hole making in composite materials to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A method for optimizing process parameters of laser ultrasonic vibration milling hole making in composite materials, the specific steps include:

[0007] S1: Orthogonal experimental design is performed based on multiple sets of process parameters, and several hole-making tests are carried out under different process parameters;

[0008] S2: Before each hole-making test, the initial image of the tool is collected, and after each hole-making test, the final image of the tool and the hole wall image are collected;

[0009] S3: performing image analysis on the initial image and final image of the tool and the hole wall image of the hole, and respectively generating the tool wear rate and hole defect rate of the hole making test;

[0010] S4: Use multivariate regression analysis to establish a relationship model between tool wear rate, hole defect rate and process parameters, and use the minimum tool wear rate and hole defect rate as the optimization goal. Use genetic algorithm to solve the relationship model to obtain the optimal process parameters.

[0011] Preferably, the process parameters include laser power, ultrasonic frequency, ultrasonic amplitude, milling speed and feed speed.

[0012] Preferably, the generation logic of the tool wear rate is:

[0013] Perform image cropping and grayscale conversion on the initial image and the final image of the tool respectively, and obtain grayscale images after the conversion;

[0014] Use Gaussian filter to denoise the grayscale images corresponding to the initial image and the final image respectively, and then use Laplacian operator to perform edge detection on the two groups of grayscale images and generate corresponding edge images;

[0015] An initial fluctuation coefficient and a final fluctuation coefficient are generated based on edge images corresponding to the initial image and the final image, respectively, and a tool wear rate is generated according to the initial fluctuation coefficient and the final fluctuation coefficient.

[0016] Preferably, the calculation method for denoising a grayscale image using a Gaussian filter is:

[0017] ;

[0018] In the formula , Respectively represent the grayscale value before denoising and the grayscale value after denoising, Represents the horizontal and vertical coordinates of the pixel point. represents the Gaussian kernel, represents the blur radius of the Gaussian kernel, where the Gaussian kernel is defined as:

[0019] ;

[0020] In the formula represents the standard deviation of the Gaussian kernel;

[0021] The Laplacian operator is used to process the denoised grayscale image. The calculation method is:

[0022] ;

[0023] In the formula Represents the edge detection value;

[0024] The denoised grayscale image is converted into an 8-bit image according to the edge detection value. The calculation method is:

[0025] ;

[0026] In the formula Represents the 8-bit detection value under the edge image, Respectively represent the maximum and minimum absolute values ​​of the edge detection value in the grayscale image;

[0027] Taking each pixel as the center, calculate the detection mean of the 8-bit detection values ​​of all pixels in the surrounding 3*3 grid, and compare the 8-bit detection value of the pixel with the detection mean. If the 8-bit detection value is greater than the detection mean, the pixel is marked as an edge point. If the 8-bit detection value is less than or equal to the detection mean, the pixel is marked as a non-edge point, thereby generating the corresponding edge image.

[0028] Preferably, the generation logic of the initial volatility coefficient and the final volatility coefficient is:

[0029] Calculate the curvature of each edge point in the edge image except the first and last points:

[0030] ;

[0031] in Indicates the curvature of the edge point, subscript Indicates the number of the edge point, , Indicates the maximum number of edge points;

[0032] The curvature of each edge point except the first and last points in the edge images corresponding to the initial image and the final image is divided into sets and collection ,gather and collection Can be expressed as , and then calculate the set ,gather The mean and standard deviation are calculated as follows:

[0033] ;

[0034] ;

[0035] ;

[0036] ;

[0037] In the formula , They represent the curvature of the edge point in the edge image corresponding to the initial image and the curvature of the edge point in the edge image corresponding to the final image, respectively. , Respectively represent sets The mean and set of The mean of , Respectively represent sets The standard deviation and set The standard deviation of and Calibrated as initial fluctuation coefficient and final fluctuation coefficient.

[0038] Preferably, the tool wear rate is calculated as follows:

[0039] ;

[0040] In the formula Indicates tool wear rate.

[0041] Preferably, the generation logic of the hole making defect rate is:

[0042] The Yolov5 model was constructed, and several groups of historical images of the borehole wall with cracks were collected. The historical images were divided into the first training set and the first validation set in a ratio of 8:2, and the Yolov5 model was trained and optimized respectively.

[0043] The hole wall image is input into the Yolov5 model for image recognition, and all the crack regions in the hole wall image and the crack area of ​​each crack region are detected;

[0044] The total crack area is calculated based on the crack areas of all crack regions, and the hole making defect rate is generated according to the ratio between the total crack area and the total image area of ​​the hole wall image. The calculation formula is: , where represents the hole making defect rate, , represent the total crack area and the total image area respectively.

[0045] Preferably, based on the multivariate regression analysis method, the relationship models between the tool wear rate, the hole making defect rate and the process parameters are respectively constructed, and the regression model between the tool wear rate and the process parameters is expressed as:

[0046] ;

[0047] The regression model between hole making defect rate and process parameters is expressed as:

[0048] ;

[0049] In the formula , , , , Respectively represent the model regression coefficients of tool wear rate, represents the model error regarding tool wear rate, , , , , They represent the model regression coefficients for hole-making defect rate, represents the model error regarding the hole making defect rate, , , , , They represent laser power, ultrasonic frequency, ultrasonic amplitude, milling speed, and feed speed, respectively;

[0050] The tool wear rate, hole-making defect rate and process parameters collected from multiple hole-making tests were divided into a second training set and a second validation set in a ratio of 8:2, and the regression model of tool wear rate-process parameters and the regression model of hole-making defect rate-process parameters were trained and optimized respectively;

[0051] The objective function is constructed, and the genetic algorithm is used for multi-objective optimization to obtain the optimal process parameters. The tool wear rate and hole defect rate under the optimal process parameters are defined as the target wear rate and target defect rate, respectively. The objective function is defined as:

[0052] ;

[0053] The fitness function is defined as:

[0054] ;

[0055] In the formula , represent the objective function and fitness function respectively, , denote the first weight and the second weight respectively, and .

[0056] Preferably, after obtaining the optimal process parameters, the composite material is trial-processed with the optimal process parameters, the real-time process parameters of the composite material during the trial processing are collected, and the real-time process parameters are adjusted using a PID control algorithm. The calculation method of the PID control algorithm is:

[0057] ;

[0058] In the formula Indicated in The laser power during the first trial processing, , , Respectively represent the proportional coefficient, integral coefficient, and differential coefficient of laser power, Indicated in The total error during the first trial processing, The calculation method is:

[0059] ;

[0060] In the formula , Respectively Tool wear rate and target wear rate during the first trial machining, , Respectively Hole making defect rate and target defect rate during the first trial processing;

[0061] The control method of ultrasonic frequency, ultrasonic amplitude, milling speed and feed speed is similar to that of laser power. By adjusting the proportional coefficient, integral coefficient and differential coefficient corresponding to different process parameters, the real-time process parameters are made closer to the optimal process parameters. When the error between the real-time process parameters and the optimal process parameters is less than 5%, the PID control algorithm optimization is considered to be completed and the trial processing is ended.

[0062] Compared with the prior art, the present invention has the following beneficial effects:

[0063] The present invention systematically optimizes the process parameters of laser ultrasonic vibration milling of composite materials through multivariate regression analysis and genetic algorithm, and improves the prediction accuracy of tool wear rate and hole defect rate. Combined with the real-time PID control algorithm, the dynamic adjustment of process parameters is further realized to ensure the stability and consistency of the processing process. Compared with traditional methods, this solution greatly reduces the number and cost of experiments, improves data quality and analysis depth, reduces tool wear and hole wall defects, and significantly improves processing quality and production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is a schematic diagram of the overall method flow of the present invention. DETAILED DESCRIPTION

[0065] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0066] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0067] Example:

[0068] See also Figure 1 , the present invention provides a technical solution:

[0069] A method for optimizing process parameters of laser ultrasonic vibration milling hole making in composite materials, the specific steps include:

[0070] S1: Orthogonal experimental design is performed based on multiple groups of process parameters, and several hole-making tests are carried out under different process parameters. The process parameters include laser power, ultrasonic frequency, ultrasonic amplitude, milling speed, and feed speed. The L9 (3^4) orthogonal table can be used in orthogonal experimental design to determine the value range of each process parameter.

[0071] In this step, through orthogonal experimental design, the value ranges of multiple parameters such as laser power, ultrasonic frequency, ultrasonic amplitude, milling speed and feed speed can be systematically explored, providing rich data support for subsequent multivariate regression analysis and optimization. Moreover, compared with the traditional full-factorial experimental design, fewer experiments are performed, and representative experimental data can be obtained in a relatively small number of experiments, which helps to reduce the number and cost of experiments and improve experimental efficiency and data quality.

[0072] S2: Before each hole making test, the initial image of the tool is collected. After each hole making test, the final image of the tool and the hole wall image are collected.

[0073] It is understandable that in order to maintain image consistency, the position of the tool and the shooting angle of the image should be kept consistent when capturing the initial and final images of the tool, so as to improve the accuracy of subsequent analysis. Specifically, the position and shooting angle of the shooting device, such as a high-definition camera, can be fixed at the initial position of the tool, and the tool is returned to the initial position before and after each hole-making test before shooting.

[0074] S3: Perform image analysis on the initial image and final image of the tool and the hole wall image of the hole to generate the tool wear rate and hole defect rate of the hole making test respectively.

[0075] The generation logic of tool wear rate is:

[0076] Perform image cropping and grayscale conversion on the initial image and the final image of the tool respectively, and obtain grayscale images after the conversion;

[0077] Use Gaussian filter to denoise the grayscale images corresponding to the initial image and the final image respectively, and then use Laplacian operator to perform edge detection on the two groups of grayscale images and generate corresponding edge images;

[0078] An initial fluctuation coefficient and a final fluctuation coefficient are generated based on edge images corresponding to the initial image and the final image, respectively, and a tool wear rate is generated according to the initial fluctuation coefficient and the final fluctuation coefficient.

[0079] The calculation method for grayscale image denoising using Gaussian filter is:

[0080] ;

[0081] In the formula , Respectively represent the grayscale value before denoising and the grayscale value after denoising, Represents the horizontal and vertical coordinates of the pixel point. represents the Gaussian kernel, represents the blur radius of the Gaussian kernel, where the Gaussian kernel is defined as:

[0082] ;

[0083] In the formula represents the standard deviation of the Gaussian kernel;

[0084] The Laplacian operator is used to process the denoised grayscale image. The calculation method is:

[0085] ;

[0086] In the formula Represents the edge detection value;

[0087] The denoised grayscale image is converted into an 8-bit image according to the edge detection value. The calculation method is:

[0088] ;

[0089] In the formula Represents the 8-bit detection value under the edge image, Respectively represent the maximum and minimum absolute values ​​of the edge detection value in the grayscale image;

[0090] Taking each pixel as the center, calculate the detection mean of the 8-bit detection values ​​of all pixels in the surrounding 3*3 grid, and compare the 8-bit detection value of the pixel with the detection mean. If the 8-bit detection value is greater than the detection mean, the pixel is marked as an edge point. If the 8-bit detection value is less than or equal to the detection mean, the pixel is marked as a non-edge point, thereby generating the corresponding edge image.

[0091] By combining the Gaussian filter with the Laplacian operator, effective suppression of image noise and precise extraction of edge information are achieved, which improves the accuracy and stability of the image processing results. The same processing steps are performed on the initial image and the final image to ensure data consistency, making the initial fluctuation coefficient and the final fluctuation coefficient comparable, which provides a basis for the accurate calculation of tool wear rate. At the same time, the automated image processing process avoids subjective errors caused by manual operation and can process a large amount of experimental data, making it suitable for high-throughput experimental environments.

[0092] The generation logic of the initial volatility coefficient and the final volatility coefficient is:

[0093] Calculate the curvature of each edge point in the edge image except the first and last points:

[0094] ;

[0095] in Indicates the curvature of the edge point, subscript Indicates the number of the edge point, , Indicates the maximum number of edge points.

[0096] In this step, the curvature of the edge point reflects the degree of shape change of the point, that is, the greater the curvature of a certain edge point, the more severe the convexity or concave of the point. It is understandable that when the tool is not worn before use, the overall edge is relatively smooth. In other words, in the edge image corresponding to the initial image, the curvature of the edge point will not be too large, and the change range will not be too large. After the hole making test, when the tool is worn, the edge will become convex or concave to a certain extent, which is reflected in the edge image corresponding to the final image, and will also cause the curvature of some edge points to change greatly.

[0097] The curvature of each edge point except the first and last points in the edge images corresponding to the initial image and the final image is divided into sets and collection ,gather and collection Can be expressed as , and then calculate the set ,gather The mean and standard deviation are calculated as follows:

[0098] ;

[0099] ;

[0100] ;

[0101] ;

[0102] In the formula , They represent the curvature of the edge point in the edge image corresponding to the initial image and the curvature of the edge point in the edge image corresponding to the final image, respectively. , Respectively represent sets The mean and set of The mean of , Respectively represent sets The standard deviation and set The standard deviation of and Calibrated as initial fluctuation coefficient and final fluctuation coefficient.

[0103] The tool wear rate is calculated as:

[0104] ;

[0105] In the formula Indicates tool wear rate.

[0106] From the calculation formula and definition of the initial fluctuation coefficient and the final fluctuation coefficient, it can be seen that the two coefficients reflect the discreteness of the curvature of all edge points. It can be understood that the larger the value of the initial fluctuation coefficient or the final fluctuation coefficient, the more violent the curvature fluctuation of the edge point, that is, the more edge points with large curvature, in other words, the more depressions and protrusions appear on the edge of the tool, and the more serious the tool wear. The tool wear rate represents the rate of change of the final fluctuation coefficient relative to the initial fluctuation coefficient, so it can be used to reflect the wear of the tool. The larger the tool wear rate, the greater the change of the final fluctuation coefficient relative to the initial fluctuation coefficient, and the more serious the tool wear.

[0107] In this step, the tool wear rate is automatically generated through image processing and mathematical calculation methods, which not only reduces manual intervention but also improves detection efficiency. It is suitable for high-throughput, multi-batch experimental environments. Compared with the subjective judgment or simple geometric measurement methods in the existing technology, the curvature calculation, mean and standard deviation quantitative analysis can more accurately and objectively reflect the tool wear condition, help improve the quality of experimental data and the depth of analysis, and provide reliable data support for multiple regression analysis and process parameter optimization.

[0108] The generation logic of hole making defect rate is:

[0109] The Yolov5 model was constructed, and several groups of historical images of the borehole wall with cracks were collected. The historical images were divided into the first training set and the first validation set in a ratio of 8:2, and the Yolov5 model was trained and optimized respectively.

[0110] The hole wall image is input into the Yolov5 model for image recognition, and all the crack regions in the hole wall image and the crack area of ​​each crack region are detected;

[0111] The total crack area is calculated based on the crack areas of all crack regions, and the hole making defect rate is generated according to the ratio between the total crack area and the total image area of ​​the hole wall image. The calculation formula is: , where represents the hole making defect rate, , represent the total crack area and the total image area respectively.

[0112] Yolov5 is a real-time target detection network based on deep learning. It is efficient and accurate and can quickly identify crack areas in borehole wall images. Compared with traditional image processing methods, Yolov5 has significant advantages in processing complex backgrounds and small cracks. By training and verifying the Yolov5 model, automatic detection of crack areas in borehole wall images can be achieved, reducing manual intervention and errors, and improving detection efficiency and consistency.

[0113] In this step, by constructing and optimizing the Yolov5 model, the cracks in the hole wall image are automatically detected, and the hole making defect rate is generated based on the ratio of the total crack area to the total image area. This can quantify and reflect the defects of the hole wall from multiple angles, and provide reliable data support for multivariate regression analysis and process parameter optimization, thereby realizing the optimization of the process parameters for laser ultrasonic vibration milling hole making in composite materials.

[0114] S4: Use multivariate regression analysis to establish a relationship model between tool wear rate, hole defect rate and process parameters, and use the minimum tool wear rate and hole defect rate as the optimization goal. Use genetic algorithm to solve the relationship model to obtain the optimal process parameters.

[0115] Based on the multivariate regression analysis method, the relationship models between tool wear rate, hole making defect rate and process parameters were constructed respectively. The regression model between tool wear rate and process parameters is expressed as:

[0116] ;

[0117] The regression model between hole making defect rate and process parameters is expressed as:

[0118] ;

[0119] In the formula , , , , Respectively represent the model regression coefficients of tool wear rate, represents the model error regarding tool wear rate, , , , , They represent the model regression coefficients for hole-making defect rate, represents the model error regarding the hole making defect rate, , , , , They represent laser power, ultrasonic frequency, ultrasonic amplitude, milling speed, and feed speed, respectively;

[0120] The tool wear rate, hole-making defect rate and process parameters collected under multiple hole-making tests were divided into the second training set and the second validation set in a ratio of 8:2. The regression model of tool wear rate-process parameters and the regression model of hole-making defect rate-process parameters were trained and optimized respectively. Through multivariate regression analysis, the influence of multiple process parameters (laser power, ultrasonic frequency, ultrasonic amplitude, milling speed, feed speed) on tool wear rate and hole-making defect rate can be considered at the same time, so as to establish a more comprehensive and accurate relationship model. Moreover, regression modeling based on experimental data can truly reflect the tool wear rate and hole-making defect rate under actual process conditions, avoiding errors caused by subjective assumptions and empirical formulas.

[0121] The objective function is constructed, and the genetic algorithm is used for multi-objective optimization to obtain the optimal process parameters. The tool wear rate and hole defect rate under the optimal process parameters are defined as the target wear rate and target defect rate, respectively. The objective function is defined as:

[0122] ;

[0123] The fitness function is defined as:

[0124] ;

[0125] In the formula , represent the objective function and fitness function respectively, , denote the first weight and the second weight respectively, and The first weight and the second weight can be flexibly adjusted according to actual needs, so as to achieve customized target optimization in different application scenarios and meet different processing requirements. Specifically, if the user is not sensitive to tool cost and pays more attention to hole quality, It can be set to less than , that is, the optimization direction tends to reduce the hole defect rate. Conversely, if the user wants to save tool costs, then You can set it to be greater than , the optimization direction tends to reduce the tool wear rate, and when equal When , it is considered to be the default equilibrium state. Using genetic algorithms for multi-objective optimization can search for the optimal solution in a wide range of parameter spaces and find the best balance point between tool wear rate and hole-making defect rate, thereby achieving the process parameter combination with the best comprehensive performance.

[0126] In this step, the relationship models between tool wear rate, hole defect rate and process parameters were constructed respectively through multivariate regression analysis, and multi-objective optimization was performed using genetic algorithm. The influence of process parameters on machining quality was comprehensively analyzed and optimized from multiple angles, thereby effectively reducing tool wear rate and hole defect rate, and achieving comprehensive improvement in machining quality and efficiency.

[0127] After obtaining the optimal process parameters, the composite material is trial-processed with the optimal process parameters, the real-time process parameters of the composite material during the trial processing are collected, and the PID control algorithm is used to adjust the real-time process parameters. The calculation method of the PID control algorithm is:

[0128] ;

[0129] In the formula Indicated in The laser power during the first trial processing, , , Respectively represent the proportional coefficient, integral coefficient, and differential coefficient of laser power, Indicated in The total error during the first trial processing, The calculation method is:

[0130] ;

[0131] In the formula , Respectively Tool wear rate and target wear rate during the first trial machining, , Respectively Hole making defect rate and target defect rate during the first trial processing;

[0132] The control method of ultrasonic frequency, ultrasonic amplitude, milling speed and feed speed is similar to that of laser power. By adjusting the proportional coefficient, integral coefficient and differential coefficient corresponding to different process parameters, the real-time process parameters are made closer to the optimal process parameters. When the error between the real-time process parameters and the optimal process parameters is less than 5%, the PID control algorithm optimization is considered to be completed and the trial processing is ended.

[0133] In this step, by setting up the trial processing link, it is possible to detect the difference between the process parameters of the processing equipment in actual operation and the optimal process parameters found theoretically using the genetic algorithm. The real-time process parameters are adjusted through the PID control algorithm, realizing closed-loop control from parameter optimization to trial processing and then to parameter adjustment, thereby improving the automation and intelligence level of the overall solution, and further improving the effect of the process parameters after multivariate regression analysis and genetic algorithm optimization, making the processing results closer to the expected goals, improving the optimization effect of the overall solution, and further reducing the tool wear rate and hole defect rate.

[0134] In summary, through multivariate regression analysis and genetic algorithm, the process parameters of laser ultrasonic vibration milling of composite materials are systematically optimized, and the prediction accuracy of tool wear rate and hole defect rate is improved. Combined with the real-time PID control algorithm, the dynamic adjustment of process parameters is further realized to ensure the stability and consistency of the processing process. Compared with traditional methods, this scheme greatly reduces the number and cost of experiments, improves data quality and analysis depth, reduces tool wear and hole wall defects, and significantly improves processing quality and production efficiency.

[0135] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0136] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0137] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0138] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for optimizing process parameters of laser ultrasonic vibration milling hole making in composite materials, characterized in that: The specific steps include: S1: An orthogonal experimental design is performed based on multiple sets of process parameters, and several hole-making tests are performed under different process parameters, wherein the process parameters include laser power, ultrasonic frequency, ultrasonic amplitude, milling speed, and feed speed; S2: Before each hole-making test, the initial image of the tool is collected, and after each hole-making test, the final image of the tool and the hole wall image are collected; S3: performing image analysis on the initial image and final image of the tool and the hole wall image of the hole, and respectively generating the tool wear rate and hole defect rate of the hole making test; S4: Use multivariate regression analysis to establish a relationship model between tool wear rate, hole defect rate and process parameters, and use the minimum tool wear rate and hole defect rate as the optimization goal, and use genetic algorithm to solve the relationship model to obtain the optimal process parameters; The generation logic of the tool wear rate is: Perform image cropping and grayscale conversion on the initial image and the final image of the tool respectively, and obtain grayscale images after the conversion; Use Gaussian filter to denoise the grayscale images corresponding to the initial image and the final image respectively, and then use Laplacian operator to perform edge detection on the two groups of grayscale images and generate corresponding edge images; The calculation method for grayscale image denoising using Gaussian filter is: In the formula , Respectively represent the grayscale value before denoising and the grayscale value after denoising, Represents the horizontal and vertical coordinates of the pixel point. represents the Gaussian kernel, represents the blur radius of the Gaussian kernel, where the Gaussian kernel is defined as: In the formula represents the standard deviation of the Gaussian kernel; Use the Laplacian operator to process the denoised grayscale image. The calculation method is: In the formula Represents the edge detection value; The denoised grayscale image is converted into an 8-bit image according to the edge detection value. The calculation method is: In the formula Represents the 8-bit detection value under the edge image, Respectively represent the maximum and minimum absolute values ​​of the edge detection value in the grayscale image; Taking each pixel as the center, calculate the detection mean of the 8-bit detection values ​​of all pixels in the surrounding 3*3 grid, and compare the 8-bit detection value of the pixel with the detection mean. If the 8-bit detection value is greater than the detection mean, the pixel is marked as an edge point. If the 8-bit detection value is less than or equal to the detection mean, the pixel is marked as a non-edge point, thereby generating the corresponding edge image; Generate an initial fluctuation coefficient and a final fluctuation coefficient based on edge images corresponding to the initial image and the final image respectively, and generate a tool wear rate according to the initial fluctuation coefficient and the final fluctuation coefficient; The generation logic of the initial volatility coefficient and the final volatility coefficient is: Calculate the curvature of each edge point in the edge image except the first and last points: in Indicates the curvature of the edge point, subscript Indicates the number of the edge point, , Indicates the maximum number of edge points; The curvature of each edge point except the first and last points in the edge images corresponding to the initial image and the final image is divided into sets and collection ,gather and collection Can be expressed as , and then calculate the set ,gather The mean and standard deviation are calculated as follows: In the formula , They represent the curvature of the edge point in the edge image corresponding to the initial image and the curvature of the edge point in the edge image corresponding to the final image, respectively. , Respectively represent sets The mean and set of The mean of , Respectively represent sets The standard deviation and set The standard deviation of and Calibrated as initial fluctuation coefficient and final fluctuation coefficient; The tool wear rate is calculated as follows: In the formula Indicates tool wear rate; Based on the multivariate regression analysis method, the relationship models between tool wear rate, hole making defect rate and process parameters were constructed respectively. The regression model between tool wear rate and process parameters is expressed as: The regression model between hole making defect rate and process parameters is expressed as: In the formula , , , , Respectively represent the model regression coefficients of tool wear rate, represents the model error regarding tool wear rate, , , , , They represent the model regression coefficients for hole-making defect rate, represents the model error regarding the hole making defect rate, , , , , They represent laser power, ultrasonic frequency, ultrasonic amplitude, milling speed, and feed speed, respectively; The tool wear rate, hole-making defect rate and process parameters collected from multiple hole-making tests were divided into a second training set and a second validation set in a ratio of 8:2, and the regression model of tool wear rate-process parameters and the regression model of hole-making defect rate-process parameters were trained and optimized respectively; The objective function is constructed, and the genetic algorithm is used for multi-objective optimization to obtain the optimal process parameters. The tool wear rate and hole defect rate under the optimal process parameters are defined as the target wear rate and target defect rate, respectively. The objective function is defined as: The fitness function is defined as: In the formula , represent the objective function and fitness function respectively, , denote the first weight and the second weight respectively, and ; The generation logic of the hole making defect rate is: The Yolov5 model was constructed, and several groups of historical images of the borehole wall with cracks were collected. The historical images were divided into the first training set and the first validation set in a ratio of 8:2, and the Yolov5 model was trained and optimized respectively. The hole wall image is input into the Yolov5 model for image recognition, and all the crack regions in the hole wall image and the crack area of ​​each crack region are detected; The total crack area is calculated based on the crack areas of all crack regions, and the hole making defect rate is generated according to the ratio between the total crack area and the total image area of ​​the hole wall image. The calculation formula is: , where represents the hole making defect rate, , represent the total crack area and the total image area, respectively; After obtaining the optimal process parameters, the composite material is trial-processed with the optimal process parameters, the real-time process parameters of the composite material during the trial processing are collected, and the PID control algorithm is used to adjust the real-time process parameters. The calculation method of the PID control algorithm is: In the formula Indicated in The laser power during the first trial processing, , , Respectively represent the proportional coefficient, integral coefficient, and differential coefficient of laser power, Indicated in The total error during the first trial processing, The calculation method is: In the formula , Respectively Tool wear rate and target wear rate during the first trial machining, , Respectively Hole making defect rate and target defect rate during the first trial processing; The control method of ultrasonic frequency, ultrasonic amplitude, milling speed and feed speed is similar to that of laser power. By adjusting the proportional coefficient, integral coefficient and differential coefficient corresponding to different process parameters, the real-time process parameters are made closer to the optimal process parameters. When the error between the real-time process parameters and the optimal process parameters is less than 5%, the PID control algorithm optimization is considered to be completed and the trial processing is ended.

Images