Intelligent self-adaption system and method in blade machining process

Abstract

The invention belongs to the technical field of tool machining, and discloses an intelligent self-adaption system and method in the blade machining process. The method comprises the steps of collecting physical parameters of a blade machining material; processing parameters of the real-time blade are obtained; acquiring a processing material surface image in unit time in real time; processing the surface image of the processing material in unit time to obtain a crystal orientation matrix set; inputting the physical parameters and the crystal orientation matrix set into a crystal orientation prediction model to obtain a crystal orientation matrix in future unit time; based on the physical parameters, the machining parameters, the crystal orientation matrix set and the crystal orientation matrix of the future unit time, the machining parameters of the blade are intelligently and adaptively adjusted; according to the scheme, physical parameters, machining parameters and crystal orientation prediction are combined, tool machining can be adjusted in a self-adaptive mode according to the actual state of a machined material, the machining parameters are dynamically optimized, and the machining quality stability and the tool durability are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of tool processing, and more specifically, to an intelligent adaptive system and method in the process of blade processing. Background Art

[0002] In the processing of superhard materials, the prior art usually takes the macroscopic properties of the materials as the main consideration factors, while ignoring the microscopic orientation differences inside the crystals. The complex crystal structure properties lead to significant anisotropy in the processing process, resulting in the difficulty of reaching the optimal level in the processing precision and efficiency of superhard materials. Although the prior art attempts to adjust the cutting parameters to adapt to the properties of different materials, these methods are usually based on the material type and hardness of the materials, etc., without fully considering the orientation changes inside the crystals.

[0003] When a blade processes superhard materials, even when the material type and hardness are known before processing, the prior art still fails to consider that there are significant differences in the cutting force, blade wear mechanism and friction characteristics corresponding to different crystal orientations, and fails to adaptively adjust the processing parameters for different crystal orientations, resulting in uneven loads on the blade, aggravating the local wear and micro-chipping phenomena of the blade, and affecting the processing precision of high-hardness materials. At the same time, the high-cost attribute of the blade determines that the service life and stability of the blade are crucial for the overall processing efficiency. Failing to optimize considering the crystal orientation will lead to a shortened blade life and an increased replacement frequency, thereby reducing the production beat and processing stability, and also making it difficult to reach the optimal level in processing quality and efficiency.

[0004] Therefore, there is an urgent need for an effective intelligent system that can dynamically identify the crystal orientation and dynamically optimize the cutting strategy during the processing, thereby optimizing the blade wear, improving the blade processing precision and the overall processing efficiency. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art and to achieve the above object, the present invention provides the following technical solutions: an intelligent adaptive method in the process of blade processing, including:

[0006] Collecting physical parameters of the blade processing material;

[0007] Obtaining the processing parameters of the real-time blade;

[0008] Collecting the surface image of the processing material per unit time in real time;

[0009] Processing the surface image of the processing material per unit time to obtain a crystal orientation matrix set;

[0010] Inputting the physical parameters and the crystal orientation matrix set into a crystal orientation prediction model to obtain the crystal orientation matrix of the future unit time;

[0011] Intelligently and adaptively adjust the machining parameters of the blade based on physical parameters, machining parameters, the set of crystal orientation matrices, and the crystal orientation matrix in the future unit time.

[0012] Furthermore, the method for intelligently and adaptively adjusting the machining parameters of the blade includes:

[0013] Statistically analyze the occurrence frequency of each crystal orientation matrix in the set of crystal orientation matrices, and select the crystal orientation matrix with the highest occurrence frequency, denoted as the main crystal orientation matrix;

[0014] Calculate the crystal orientation angle difference based on the main crystal orientation matrix and the crystal orientation matrix in the future unit time;

[0015] If the crystal orientation angle difference is greater than or equal to the preset crystal orientation angle difference threshold, it is regarded that the crystal orientation has changed and intelligent adaptive adjustment is required;

[0016] Obtain the crystal type from the physical parameters;

[0017] If the crystal type is single crystal material, perform intelligent adaptive adjustment of machining parameters for the single crystal material;

[0018] If the crystal type is polycrystalline material, perform intelligent adaptive adjustment of machining parameters for the polycrystalline material.

[0019] Furthermore, the method for intelligently and adaptively adjusting the machining parameters for the single crystal material includes:

[0020] Step 1: Set a first single crystal orientation angle difference threshold and a second single crystal orientation angle difference threshold for the single crystal material; the first single crystal orientation angle difference threshold is less than the second single crystal orientation angle difference threshold;

[0021] Step 2: If the crystal orientation angle difference is less than the first single crystal orientation angle difference threshold, there is no need to adjust the machining parameters of the single crystal material;

[0022] If the crystal orientation angle difference is greater than or equal to the first single crystal orientation angle difference threshold and less than the second single crystal orientation angle difference threshold, input the crystal type, crystal orientation angle difference, and machining parameters into the blade machining parameter setting model to obtain the target machining parameters, where the target machining parameters include target cutting force, target cutting speed, target cutting depth, target tool temperature, and target vibration amplitude; adjust the machining parameters of the blade to the target machining parameters;

[0023] If the crystal orientation angle difference is greater than or equal to the second single crystal orientation angle difference threshold, the crystal plane of the currently machined single crystal material is not suitable for machining, perform a crystal plane switching operation, and adjust to the adjacent crystal plane of the currently machined single crystal material crystal plane;

[0024] Step 3: Repeat Step 2 until the processing of the single crystal material is completely finished.

[0025] Further, the method for intelligent adaptive adjustment of processing parameters for polycrystalline materials includes:

[0026] First step: Set a threshold value for the difference in polycrystalline orientation angles for the polycrystalline material.

[0027] Second step: If the difference in crystal orientation angles is less than the threshold value for the difference in polycrystalline orientation angles, there is no need to adjust the processing parameters of the polycrystalline material.

[0028] If the difference in crystal orientation angles is greater than or equal to the threshold value for the difference in polycrystalline orientation angles, input the crystal type, the difference in crystal orientation angles, and the processing parameters into the blade processing parameter setting model to obtain the target processing parameters.

[0029] Third step: Adjust the processing parameters of the blade to the target processing parameters.

[0030] Fourth step: Repeat the second step to the third step until the processing of the polycrystalline material is completely finished.

[0031] Further, the method for obtaining the crystal orientation matrix set includes:

[0032] S100: Divide the unit time into N time points, let the initial value of n be 1, and the value range of n is from 1 to N.

[0033] S101: Obtain the surface image of the processed material at the nth time point; divide the surface image of the processed material into M sub-regions, denoted as the surface sub-images of the processed material.

[0034] S102: Perform diffraction band detection processing on each of the M surface sub-images of the processed material to obtain the Kikuchi diffraction bands corresponding to the M surface sub-images of the processed material.

[0035] S103: Obtain the crystal orientation matrices corresponding to the M surface sub-images of the processed material based on the M Kikuchi diffraction bands.

[0036] S104: Perform matrix numerical comparison on each of the M crystal orientation matrices to obtain the crystal orientation matrix of the surface image of the processed material at the nth time point.

[0037] S105: Let n = n + 1. If n is less than or equal to N, continue to execute S101 to S104. If n is greater than N, obtain the crystal orientation matrices of the surface images of the processed material corresponding to N time points, and execute S106.

[0038] S106: Construct the crystal orientation matrices of the surface images of the processed material corresponding to N time points into a crystal orientation matrix set.

[0039] Further, the method for obtaining the Kikuchi diffraction bands corresponding to the M sub-images of the machined material surface includes:

[0040] S200: Let the initial value of m be 1, and the value range of m is from 1 to M;

[0041] S201: Obtain the m-th sub-image of the machined material surface; update the m-th sub-image of the machined material surface to a machined material surface sub-image with two-dimensional coordinates through an image processing library;

[0042] S202: Denote the number of pixel points in the m-th sub-image of the machined material surface as R; calculate the corresponding diffraction band gradient matrices for each of the R pixel points;

[0043] S203: Input the physical parameters of the blade machined material and the R diffraction band gradient matrices into a diffraction band gradient threshold setting model to obtain a diffraction band gradient threshold;

[0044] S204: Construct the Kikuchi diffraction bands of the m-th sub-image of the machined material surface based on the R diffraction band gradient matrices and the diffraction band gradient threshold;

[0045] S205: Let m = m + 1. If m is less than or equal to M, continue to execute S201 to S204; if m is greater than M, obtain the Kikuchi diffraction bands corresponding to the M sub-images of the machined material surface, and end the current process.

[0046] Further, the method for constructing the Kikuchi diffraction bands of the m-th sub-image of the machined material surface based on the R diffraction band gradient matrices and the diffraction band gradient threshold includes:

[0047] S300: Let the initial value of r be 1, and the value range of r is from 1 to R;

[0048] S301: Obtain the r-th diffraction band gradient matrix, and perform diffraction band gradient threshold determination for each element value in the matrix. Denote the elements with element values greater than or equal to the diffraction band gradient threshold as diffraction band elements, and count the number of diffraction band elements, denoted as the number of diffraction band elements; denote the elements with element values less than the diffraction band gradient threshold as non-diffraction band elements, and count the number of non-diffraction band elements, denoted as the number of non-diffraction band elements;

[0049] S302: Sum the number of diffraction band elements and the number of non-diffraction band elements to obtain the total number of elements; divide the number of diffraction band elements by the total number of elements to obtain the diffraction band element ratio;

[0050] S303: If the proportion of diffraction band elements is greater than or equal to the preset diffraction band element proportion threshold, then mark the pixel points corresponding to the r-th diffraction band gradient matrix as diffraction band pixel points; if the proportion of diffraction band elements is less than the preset diffraction band element proportion threshold, then mark the pixel points corresponding to the r-th diffraction band gradient matrix as non-diffraction band pixel points;

[0051] S304: Let r = r + 1. If r is less than or equal to R, then continue to execute S301 to S303. If r is greater than R, then execute S305;

[0052] S305: Based on all diffraction band pixel points, adopt a pixel connection algorithm to connect the discrete diffraction band pixel points into complete diffraction band lines, and obtain the Kikuchi diffraction band of the m-th processed material surface sub-image.

[0053] Further, the method for obtaining the crystal orientation matrix corresponding to the M processed material surface sub-images includes:

[0054] S400: Let the initial value of m be 1, and the value range of m is from 1 to M;

[0055] S401: Obtain the Kikuchi diffraction band of the m-th processed material surface sub-image; select three mutually independent diffraction band normal vectors from the Kikuchi diffraction band, and denote them as diffraction band normal vector one, diffraction band normal vector two, and diffraction band normal vector three respectively;

[0056] S402: Construct a diffraction band normal vector matrix with diffraction band normal vector one, diffraction band normal vector two, and diffraction band normal vector three;

[0057] S403: Calculate diffraction direction angle one, diffraction direction angle two, and diffraction direction angle three based on the diffraction band normal vector matrix;

[0058] S404: Construct a diffraction direction cosine matrix one with diffraction direction angle one, construct a diffraction direction cosine matrix two with diffraction direction angle two, and construct a diffraction direction cosine matrix three with diffraction direction angle three;

[0059] S405: Multiply the diffraction direction cosine matrix one, the diffraction direction cosine matrix two, and the diffraction direction cosine matrix three matrix-wise to obtain the crystal orientation matrix corresponding to the m-th processed material surface sub-image;

[0060] S406: Let m = m + 1. If m is less than or equal to M, then continue to execute S401 to S405; if m is greater than M, then obtain the crystal orientation matrices corresponding to the M processed material surface sub-images and end the current process.

[0061] Further, the method for obtaining the crystal orientation matrix of the processed material surface image at the n-th time point includes:

[0062] S500: Let the initial value of m be 1, and the value range of m is from 1 to M; pre-construct a set of crystal orientation matrix key-value pairs; the set of crystal orientation matrix key-value pairs includes crystal orientation matrix key-value pairs; the crystal orientation matrix key-value pair includes a key and a value, the key is the crystal orientation matrix, the key is initially empty, and the value is the quantity corresponding to the crystal orientation matrix, and the initial value of the value is 0;

[0063] S501: Obtain the m-th crystal orientation matrix; if the m-th crystal orientation matrix is the same as the key of the crystal orientation matrix key-value pair in the set of crystal orientation matrix key-value pairs, then increment the value corresponding to the crystal orientation matrix key-value pair by 1; if the m-th crystal orientation matrix is different from the key of the crystal orientation matrix key-value pair in the set of crystal orientation matrix key-value pairs, then create a new crystal orientation matrix key-value pair, use the m-th crystal orientation matrix as the key of the new crystal orientation matrix key-value pair, and increment the value of the new crystal orientation matrix key-value pair by 1; add the new crystal orientation matrix key-value pair to the set of crystal orientation matrix key-value pairs;

[0064] S502: Let m = m + 1. If m is less than or equal to M, then continue to execute S501; if m is greater than M, then execute S503;

[0065] S503: Retrieve the crystal orientation matrix key-value pair with the maximum value from the set of crystal orientation matrix key-value pairs, and use the key of the crystal orientation matrix key-value pair with the maximum value as the crystal orientation matrix of the processed material surface image at the n-th time point, and end the current process.

[0066] The intelligent adaptive system during the blade processing implements the intelligent adaptive method during the blade processing, including:

[0067] The first acquisition module acquires the physical parameters of the blade processing material;

[0068] The second acquisition module is used to obtain the processing parameters of the real-time blade;

[0069] The third acquisition module is used to acquire the processed material surface image within a unit time in real time;

[0070] The first processing module processes the processed material surface image within a unit time to obtain a set of crystal orientation matrices;

[0071] The orientation prediction module is used to input the physical parameters and the set of crystal orientation matrices into the crystal orientation prediction model to obtain the crystal orientation matrix of the future unit time;

[0072] The intelligent optimization module makes intelligent adaptive adjustments to the processing parameters of the blade based on the physical parameters, processing parameters, set of crystal orientation matrices, and crystal orientation matrix of the future unit time.

[0073] Compared with the prior art, the technical effects and advantages of the intelligent adaptive system and method in the blade processing process of the present invention are as follows:

[0074] The present invention provides an intelligent adaptive system and method in the blade processing process, which can monitor the change of crystal orientation of the processed material in real time, and dynamically optimize the processing parameters in combination with the predicted crystal orientation matrix in the future unit time.

[0075] For single crystal materials, this solution sets a threshold value for the difference in single crystal orientation angles. By monitoring the change of the difference in crystal orientation angles, it can intelligently judge whether it is necessary to adjust the processing parameters or perform a face-changing operation. For the case where the difference in crystal orientation angles is small, the original processing parameters are maintained to avoid unnecessary adjustments and improve the processing continuity and efficiency; when the difference in crystal orientation angles enters the set range, the target processing parameters are obtained based on the blade processing parameter setting model and adaptive adjustments are made to ensure that the processing process is always on the optimal crystal plane, improving the tool life and processing accuracy. If the difference in crystal orientation angles exceeds the set threshold, it indicates that the current crystal plane is no longer suitable for continued processing, and the system automatically performs a face-changing operation to switch to an adjacent crystal plane to ensure processing stability and consistency.

[0076] For polycrystalline materials, due to the randomness of the grain orientations inside them, it is difficult to accurately judge the face-changing requirement solely relying on the difference in orientation angles. Therefore, this solution adopts the method of dynamically adjusting the processing parameters. For regions with small orientation variations, the original processing parameters are maintained to reduce unnecessary adjustments and improve the processing efficiency; when the orientation variation enters the set range, the system optimizes the target processing parameters based on the blade processing parameter setting model to adapt to complex orientation changes, reduce tool wear, improve the processing quality, and ensure processing stability.

[0077] The solution of the present invention combines physical parameters, processing parameters, and crystal orientation prediction, and realizes intelligent optimization through machine learning models and real-time monitoring data, enabling the tool processing to adaptively adjust according to the actual state of the processed material. Compared with the prior art, it significantly improves the processing efficiency, quality stability, and tool durability, and has important application value in the field of superhard blade processing. Brief Description of the Drawings

[0078] Figure 1 It is a schematic diagram of the intelligent adaptive system in the blade processing process of Embodiment 1 of the present invention;

[0079] Figure 2 It is a flowchart of the intelligent adaptive method in the blade processing process of Embodiment 2 of the present invention;

[0080] Figure 3 It is a flowchart of the method for intelligently and adaptively adjusting the processing parameters of the blade;

[0081] Figure 4Schematic diagram of Kikuchi diffraction bands for single-crystal materials;

[0082] Figure 5 Schematic diagram of Kikuchi diffraction bands for polycrystalline materials;

[0083] Figure 6 Schematic diagram of the curve of the difference in crystal orientation angle changing with time. Specific implementation manners

[0084] The following will combine the drawings in the embodiments of the present invention to describe the technical solutions in the embodiments of the present invention in detail, clearly and completely. It should be particularly noted that the following specific embodiments are only used to better illustrate and explain the technical solutions of the present invention, aiming to enable those skilled in the art to better understand and implement the present invention, and should not be construed as a limitation on the protection scope of the present invention. Without departing from the spirit and essence of the present invention, those skilled in the art can modify, adjust or equivalently replace it according to the content disclosed in the present invention, and these should all be regarded as the protection scope of the present invention.

[0085] Embodiment 1

[0086] Please refer to Figure 1 As shown, the intelligent adaptive system in the blade processing process of this embodiment includes a first acquisition module, a second acquisition module, a third acquisition module, a first processing module, an orientation prediction module, and an intelligent optimization module. Each module is connected by wire and / or wirelessly to achieve data transmission.

[0087] The first acquisition module acquires the physical parameters of the blade processing material; the physical parameters include crystal type, material hardness, elastic modulus, fracture toughness, thermal conductivity, and specific heat capacity, and the crystal type includes single-crystal materials and polycrystalline materials.

[0088] It should be noted that these physical parameters such as crystal type, material hardness, elastic modulus, fracture toughness, thermal conductivity, and specific heat capacity can be obtained from the technical manuals of manufacturers or suppliers, or can be obtained from material databases.

[0089] During the tool machining process, physical parameters such as the crystal type, material hardness, elastic modulus, fracture toughness, thermal conductivity, and specific heat capacity of the machining material have a direct impact on the machining performance. Reasonably collecting and utilizing these physical parameters is crucial for optimizing the machining process, improving the machining quality and efficiency. The crystal type determines the anisotropic characteristics of the material, affecting the directionality of the cutting force and the tool wear mode during the cutting process; the material hardness is directly related to the cutting difficulty, affecting the tool wear rate and the machining surface quality; the elastic modulus affects the deformation degree of the material under the action of the cutting force, determining the tool load response and machining accuracy; the fracture toughness determines the anti-cracking ability of the material, affecting the chip breaking mode of the material and the risk of tool chipping during the cutting process; the thermal conductivity and specific heat capacity determine the heat conduction and accumulation during the cutting process, directly affecting the tool temperature, thermal expansion effect, and machining stability.

[0090] Therefore, collecting the above physical parameters and making intelligent adaptive adjustments in combination with the real-time machining state can effectively optimize the cutting parameters, improve the tool life, reduce machining defects, and ensure the optimal level of machining efficiency and accuracy.

[0091] The blade machining material refers to the entire machining object, while the crystal is the microscopic structural characteristic of the machining material. For example, metal materials are composed of metal crystals, and metal crystals include single-crystal metals (such as single-crystal copper, single-crystal copper, etc.) and polycrystalline metals (such as steel, aluminum alloy, etc.). "Crystal orientation" refers to the arrangement direction of the crystal structure inside the material. Crystals are formed by the specific periodic arrangement of atoms, ions, or molecules. The atomic arrangement density and bonding mode in different directions are different, resulting in significant differences in the physical and mechanical properties (such as hardness, elastic modulus, thermal conductivity, etc.) of the material in different directions. This phenomenon is called anisotropy.

[0092] The second acquisition module is used to obtain the machining parameters of the real-time blade, and the machining parameters include cutting force, cutting speed, cutting depth, tool temperature, and vibration amplitude.

[0093] It should be noted that during the tool machining process, machining parameters such as cutting force, cutting speed, cutting depth, tool temperature, and vibration amplitude have an important impact on the machining quality and tool life. Reasonably collecting and utilizing these parameters is crucial for optimizing the machining process and improving the machining accuracy.

[0094] The cutting force can be collected in real time by a dynamic cutting force sensor, which reflects the mechanical interaction between the tool and the workpiece and directly affects the tool wear and the surface quality of the workpiece. The cutting speed is obtained by calculating the spindle speed and feed rate of the machine tool, which determines the material removal rate and tool heat load during the cutting process. The cutting depth can be measured by a displacement sensor, which affects the magnitude of the cutting force and the machining accuracy of the workpiece. The tool temperature can be collected by a thermocouple or an infrared thermometer, which reflects the heat accumulation during the cutting process and affects the thermal expansion and wear rate of the tool. The vibration amplitude can be monitored by an acceleration sensor, which reflects the stability of the system during the cutting process. Excessive vibration may lead to tool chipping, increased surface roughness of the machined surface, and increased equipment wear.

[0095] Therefore, collecting the above machining parameters and making intelligent adaptive adjustments in combination with the real-time machining state can effectively optimize the cutting path and tool load, improve tool life, reduce machining defects, and ensure the stability and efficiency of the machining process.

[0096] A third acquisition module is used to collect the surface image of the machining material per unit time in real time. The surface image of the machining material is obtained by a high-resolution scanning electron microscope.

[0097] A first processing module processes the surface image of the machining material per unit time to obtain a set of crystal orientation matrices.

[0098] The method for obtaining the set of crystal orientation matrices includes:

[0099] S100: Divide the unit time into N time points, let the initial value of n be 1, and the value range of n is from 1 to N;

[0100] S101: Obtain the surface image of the machining material at the nth time point; divide the surface image of the machining material into M sub-regions, denoted as the surface sub-images of the machining material;

[0101] S102: Perform diffraction band detection processing on the M surface sub-images of the machining material respectively to obtain the Kikuchi diffraction bands corresponding to the M surface sub-images of the machining material;

[0102] S103: Obtain the crystal orientation matrices corresponding to the M surface sub-images of the machining material based on the M Kikuchi diffraction bands;

[0103] S104: Perform matrix numerical comparison on the M crystal orientation matrices respectively to obtain the crystal orientation matrix of the surface image of the machining material at the nth time point;

[0104] S105: Let n = n + 1. If n is less than or equal to N, continue to execute S101 to S104. If n is greater than N, obtain the crystal orientation matrices of the surface images of the machining material corresponding to the N time points, and execute S106;

[0105] S106: Construct the crystal orientation matrices corresponding to N time points into a crystal orientation matrix set.

[0106] The method for obtaining the Kikuchi diffraction bands corresponding to M sub-images of the processed material surface includes:

[0107] S200: Let the initial value of m be 1, and the value range of m is from 1 to M;

[0108] S201: Obtain the m-th sub-image of the processed material surface; update the m-th sub-image of the processed material surface to a sub-image of the processed material surface with two-dimensional coordinates through an image processing library;

[0109] S202: Denote the number of pixel points in the m-th sub-image of the processed material surface as R; calculate the corresponding diffraction band gradient matrices for the R pixel points respectively;

[0110] S203: Input the physical parameters of the blade processed material and the R diffraction band gradient matrices into the diffraction band gradient threshold setting model to obtain the diffraction band gradient threshold;

[0111] S204: Based on the R diffraction band gradient matrices and the diffraction band gradient threshold, construct the Kikuchi diffraction band of the m-th sub-image of the processed material surface;

[0112] S205: Let m = m + 1. If m is less than or equal to M, continue to execute S201 to S204; if m is greater than M, obtain the Kikuchi diffraction bands corresponding to M sub-images of the processed material surface and end the current process.

[0113] The calculation method of the diffraction band gradient matrix includes:

[0114]

[0115] Among them, TDJZ max , , r , , r，x , ,

[0113] ,

[0114] , min ,

[0117] , r，y , ,

[0115] , r ,

[0116] , , is the diffraction band gradient matrix corresponding to the r-th pixel point, TD r，x is the gradient matrix of the r-th pixel point in the horizontal direction, TD r，y is the gradient matrix of the r-th pixel point in the vertical direction, XSZ r is the pixel value of the r-th pixel point, XSZ min is the minimum pixel value of the sub-image of the processed material surface where the r-th pixel point is located, XSZ max is the maximum pixel value of the sub-image of the processed material surface where the r-th pixel point is located, r ≤ R.

[0116] The training method of the diffraction band gradient threshold setting model includes:

[0117] Pre-collect a diffraction band gradient threshold setting dataset, where the diffraction band gradient threshold setting dataset includes K sets of diffraction band gradient threshold setting data and the corresponding diffraction band gradients of the K sets of diffraction band gradient threshold setting data. K is a positive integer greater than 0. The diffraction band gradient threshold setting data includes crystal type, material hardness, elastic modulus, fracture toughness, thermal conductivity, specific heat capacity, and R diffraction band gradient matrices; divide the diffraction band gradient threshold setting dataset into a training set and a validation set, where the training set is used to train the diffraction band gradient threshold setting model, and the validation set is used to evaluate the generalization performance of the diffraction band gradient threshold setting model;

[0118] During the training process of the diffraction band gradient threshold setting model, minimize the cross-entropy loss function as the optimization objective, use the early stopping strategy to monitor the performance of the validation set, and optimize the model performance by continuously adjusting the network parameters; when the prediction accuracy on the validation set reaches the expected accuracy, it is considered that the diffraction band gradient threshold setting model has converged and stop training; the diffraction band gradient threshold setting model is trained using a deep neural network based on a multi-layer perceptron;

[0119] Convert the diffraction band gradient threshold setting data into feature vectors; the input layer of the diffraction band gradient threshold setting model receives the feature vectors, extracts the non-linear relationships in the data through the hidden layer, and finally the output layer of the diffraction band gradient threshold setting model calculates the probability distribution of the diffraction band gradient threshold through the softmax activation function, and outputs the diffraction band gradient threshold corresponding to the maximum probability as the final prediction result.

[0120] The softmax activation function is:

[0121]

[0122] where, Softmax(XL num ) is the output probability corresponding to the num-th feature vector, XL num is the num-th feature vector, NUM is the total number of feature vectors, and e is a constant.

[0123] It should be noted that the Kikuchi diffraction bands of single crystal materials are as Figure 4 shown, and the Kikuchi diffraction bands of polycrystalline materials are as Figure 5As shown. The crystal type determines the internal structural characteristics of the material, affects the distribution law of Kikuchi diffraction bands. The material hardness and elastic modulus jointly determine the degree of deformation of the material during processing, thus affecting the change of the surface microstructure. The fracture toughness determines whether microcracks are likely to occur under local stress conditions of the material, thus affecting the clarity of the diffraction bands. The thermal conductivity and specific heat capacity determine the change of the temperature gradient during processing, and further affect the surface stress distribution and the range of the heat affected zone of the processed material, causing the contrast of the diffraction bands to change accordingly. Combining with the diffraction band gradient matrix, the change characteristics of the Kikuchi diffraction bands in the processed material surface image can be reflected.

[0124] Therefore, based on the crystal type, material hardness, elastic modulus, fracture toughness, thermal conductivity, specific heat capacity, and diffraction band gradient matrix, the present invention constructs a data-driven machine learning model to realize the dynamic setting of the diffraction band gradient threshold during the Kikuchi diffraction band detection process, so as to obtain the optimal diffraction band gradient threshold of the Kikuchi diffraction bands under different materials, realize the adaptive adjustment of the diffraction band gradient threshold, improve the accuracy and robustness of the Kikuchi diffraction band detection, and thus optimize the subsequent crystal orientation analysis and material processing quality evaluation.

[0125] The method for constructing the Kikuchi diffraction bands of the m-th processed material surface sub-image based on the R diffraction band gradient matrices and the diffraction band gradient threshold includes:

[0126] S300: Let the initial value of r be 1, and the value range of r is from 1 to R;

[0127] S301: Obtain the r-th diffraction band gradient matrix, and perform diffraction band gradient threshold determination for each element value in the matrix. Mark the elements with element values greater than or equal to the diffraction band gradient threshold as diffraction band elements, and count the number of diffraction band elements, denoted as the diffraction band element number; mark the elements with element values less than the diffraction band gradient threshold as non-diffraction band elements, and count the number of non-diffraction band elements, denoted as the non-diffraction band element number;

[0128] S302: Sum the diffraction band element number and the non-diffraction band element number to obtain the total element amount; divide the diffraction band element number by the total element amount to obtain the diffraction band element ratio;

[0129] S303: If the diffraction band element ratio is greater than or equal to the preset diffraction band element ratio threshold, mark the pixel points corresponding to the r-th diffraction band gradient matrix as diffraction band pixel points; if the diffraction band element ratio is less than the preset diffraction band element ratio threshold, mark the pixel points corresponding to the r-th diffraction band gradient matrix as non-diffraction band pixel points;

[0130] For example, the diffraction band element ratio threshold can be set to Or

[0131] S304: Let r = r + 1. If r is less than or equal to R, continue to execute S301 to S303. If r is greater than R, execute S305;

[0132] S305: Based on all the pixel points of the diffraction bands, use the pixel connection algorithm to connect the discrete pixel points of the diffraction bands into complete diffraction band lines, and obtain the Kikuchi diffraction bands of the m-th sub-image of the processed material surface.

[0133] It should be noted that the identification of Kikuchi diffraction bands is an important part of intelligent tool processing. Kikuchi diffraction bands refer to the high-intensity diffraction regions formed by Bragg scattering inside the crystal in electron backscatter diffraction or transmission electron microscope analysis. The distribution of Kikuchi diffraction bands is closely related to the crystal orientation and is a direct manifestation of the crystal orientation information of the material, that is, the distribution of Kikuchi diffraction bands strictly corresponds to the crystal orientation. Therefore, the crystal orientation of the material can be determined by analyzing the Kikuchi diffraction bands.

[0134] The method for obtaining the crystal orientation matrix corresponding to the M sub-images of the processed material surface includes:

[0135] S400: Let the initial value of m be 1, and the value range of m is from 1 to M;

[0136] S401: Obtain the Kikuchi diffraction bands of the m-th sub-image of the processed material surface; select three mutually independent diffraction band normal vectors from the Kikuchi diffraction bands, and denote them as diffraction band normal vector one, diffraction band normal vector two, and diffraction band normal vector three respectively;

[0137] S402: Construct a diffraction band normal vector matrix with diffraction band normal vector one, diffraction band normal vector two, and diffraction band normal vector three;

[0138] The diffraction band normal vector matrix is:

[0139]

[0140] where FXLJZ is the diffraction band normal vector matrix, (fxl 11 , fxl 21 , fxl 31 ) is the diffraction band normal vector one, (fxl 12 , fxl 22 , fxl 32 ) is the diffraction band normal vector two, (fxl 13 , fxl 23 , fxl 33 ) is the diffraction band normal vector three.

[0141] S403: Calculate diffraction direction angle one, diffraction direction angle two, and diffraction direction angle three based on the diffraction band normal vector matrix;

[0142] The calculation methods for the diffraction direction angle one, diffraction direction angle two, and diffraction direction angle three include:

[0143] θ1 = cos -1 (fxl 33 );

[0144]

[0145] where θ1 is the diffraction direction angle one, θ2 is the diffraction direction angle two, θ3 is the diffraction direction angle three, cos -1 () is the inverse cosine function, and tan -1 () is the arctangent function.

[0146] S404: Construct the first diffraction direction cosine matrix with the diffraction direction angle one, construct the second diffraction direction cosine matrix with the diffraction direction angle two, and construct the third diffraction direction cosine matrix with the diffraction direction angle three;

[0147] The acquisition methods for the first diffraction direction cosine matrix, the second diffraction direction cosine matrix, and the third diffraction direction cosine matrix include:

[0148]

[0149]

[0150] where RJZ z (θ1) is the first diffraction direction cosine matrix, RJZ x (θ2) is the second diffraction direction cosine matrix, and RJZ y (θ3) is the third diffraction direction cosine matrix.

[0151] S405: Multiply the first diffraction direction cosine matrix, the second diffraction direction cosine matrix, and the third diffraction direction cosine matrix to obtain the crystal orientation matrix corresponding to the m-th sub-image of the processed material surface;

[0152] The acquisition method for the crystal orientation matrix corresponding to the m-th sub-image of the processed material surface includes:

[0153] JTQX m =RJZ z (θ1)×RJZ x (θ2)×RJZ y (θ3);

[0154] where JTXQ mis the crystal orientation matrix corresponding to the sub-image of the surface of the m-th processed material.

[0155] S406: Let m = m + 1. If m is less than or equal to M, then continue to execute S401 to S405; if m is greater than M, then obtain the crystal orientation matrices corresponding to the M sub-images of the surface of the processed material, and end the current process.

[0156] The method for obtaining the crystal orientation matrix of the surface image of the processed material at the n-th time point includes:

[0157] S500: Let the initial value of m be 1, and the value range of m is from 1 to M; pre-construct a set of crystal orientation matrix key-value pairs; the set of crystal orientation matrix key-value pairs includes crystal orientation matrix key-value pairs; the crystal orientation matrix key-value pair includes a key and a value, the key is the crystal orientation matrix, the key is initially empty, and the value is the quantity corresponding to the crystal orientation matrix, and the initial value of the value is 0;

[0158] S501: Obtain the m-th crystal orientation matrix; if the m-th crystal orientation matrix is the same as the key of the crystal orientation matrix key-value pair in the set of crystal orientation matrix key-value pairs, then increment the value corresponding to the crystal orientation matrix key-value pair by 1; if the m-th crystal orientation matrix is different from the key of the crystal orientation matrix key-value pair in the set of crystal orientation matrix key-value pairs, then create a new crystal orientation matrix key-value pair, use the m-th crystal orientation matrix as the key of the new crystal orientation matrix key-value pair, and increment the value of the new crystal orientation matrix key-value pair by 1; add the new crystal orientation matrix key-value pair to the set of crystal orientation matrix key-value pairs;

[0159] S502: Let m = m + 1. If m is less than or equal to M, then continue to execute S501; if m is greater than M, then execute S503;

[0160] S503: Retrieve the crystal orientation matrix key-value pair with the maximum value from the set of crystal orientation matrix key-value pairs, and use the key of the crystal orientation matrix key-value pair with the maximum value as the crystal orientation matrix of the surface image of the processed material at the n-th time point, and end the current process.

[0161] The orientation prediction module is used to input physical parameters and a set of crystal orientation matrices into a crystal orientation prediction model to obtain the crystal orientation matrix for the next unit of time.

[0162] The training method of the crystal orientation prediction model includes:

[0163] Pre-collect a crystal orientation prediction dataset, where the crystal orientation prediction dataset includes F groups of crystal orientation prediction data and the corresponding crystal orientation matrices for the F groups of crystal orientation prediction data. F is a positive integer greater than 0. The crystal orientation prediction data includes physical parameters and a set of crystal orientation matrices; divide the crystal orientation prediction dataset into a training set and a validation set, where the training set is used to train the crystal orientation prediction model, and the validation set is used to evaluate the generalization performance of the crystal orientation prediction model;

[0164] During the training process of the crystal orientation prediction model, minimize the cross-entropy loss function as the optimization objective, use the early stopping strategy to monitor the performance of the validation set, and optimize the model performance by continuously adjusting the network parameters; when the prediction accuracy on the validation set reaches the expected accuracy, it is considered that the crystal orientation prediction model has converged, and stop the training; the crystal orientation prediction model is trained using a deep neural network based on a multi-layer perceptron;

[0165] Convert the crystal orientation prediction data into feature vectors; the input layer of the crystal orientation prediction model receives the feature vectors, extracts the non-linear relationships in the data through the hidden layer, and finally the output layer of the crystal orientation prediction model calculates the probability distribution of the crystal orientation matrix through the softmax activation function, and outputs the crystal orientation matrix corresponding to the maximum probability as the final prediction result.

[0166] It should be noted that the change of crystal orientation is usually a process that evolves over time. As the processing process progresses, new crystal planes are continuously exposed. Through time series modeling, the change pattern of the historical crystal orientation matrix can be captured, and then the future crystal orientation can be predicted. Specifically, the evolution of crystal orientation has time dependence, that is, the crystal orientation matrix at the previous time point determines the crystal orientation state at the subsequent time point.

[0167] In addition, physical parameters determine the evolution mechanism of crystal orientation. For example, material hardness and elastic modulus affect the deformation behavior and determine the lattice rotation in the stress concentration region; fracture toughness determines the microcrack propagation path, which in turn affects the local crystal orientation; thermal conductivity and specific heat capacity affect the temperature distribution and regulate the influence of thermal stress on the lattice. These physical parameters are used as inputs to the machine learning model and can be used to establish physical constraints on the change of crystal orientation, thereby improving the prediction accuracy.

[0168] Comprehensively consider the time series characteristics of the crystal orientation matrix and the material property constraints of physical parameters, establish a mapping relationship of the crystal orientation changing with time, and then predict the crystal orientation matrix in the future unit time on the premise of the given current state and material characteristics, providing a theoretical basis and optimization strategy for precisely controlling the microstructure evolution of the processed material.

[0169] The intelligent optimization module intelligently and adaptively adjusts the machining parameters of the blade based on physical parameters, machining parameters, the crystal orientation matrix set, and the crystal orientation matrix in the future unit time.

[0170] As Figure 3 shown, the method for intelligently and adaptively adjusting the machining parameters of the blade includes:

[0171] Statistically analyze the occurrence frequency of each crystal orientation matrix in the crystal orientation matrix set, and select the crystal orientation matrix with the highest occurrence frequency, denoted as the main crystal orientation matrix;

[0172] Calculate the crystal orientation angle difference based on the main crystal orientation matrix and the crystal orientation matrix in the future unit time;

[0173] The method for obtaining the crystal orientation angle difference includes:

[0174]

[0175] Among them, JCZ is the crystal orientation angle difference, cos -1 () is the inverse cosine function, ZJZ is the main crystal orientation matrix, WLJZ is the crystal orientation matrix in the future unit time, WLJZ T is the transpose matrix of the crystal orientation matrix in the future unit time, ZJZ×WLJZ T represents the multiplication of the main crystal orientation matrix and the transpose matrix of the crystal orientation matrix in the future unit time to obtain a new matrix, denoted as the change matrix, Tr(ZJZ×WLJZ T ) represents the trace of the change matrix, and the trace of the change matrix is the sum of the main diagonal elements of the change matrix. For example, Figure 6 is a schematic diagram of the curve of the crystal orientation angle difference changing with time, where the horizontal axis is time and the vertical axis is the crystal orientation angle difference.

[0176] If the crystal orientation angle difference is greater than or equal to the preset crystal orientation angle difference threshold, it is considered that the crystal orientation has changed and intelligent adaptive adjustment is required; for example, the crystal orientation angle difference threshold can be set to 5°.

[0177] Obtain the crystal type from the physical parameters;

[0178] If the crystal type is single crystal material, perform intelligent adaptive adjustment of machining parameters for single crystal material;

[0179] If the crystal type is polycrystalline material, perform intelligent adaptive adjustment of machining parameters for polycrystalline material.

[0180] The method for intelligently and adaptively adjusting the machining parameters for single crystal material includes:

[0181] Step 1: Set a first single-crystal orientation angle difference threshold and a second single-crystal orientation angle difference threshold for the single-crystal material; the first single-crystal orientation angle difference threshold is less than the second single-crystal orientation angle difference threshold.

[0182] Step 2: If the crystal orientation angle difference is less than the first single-crystal orientation angle difference threshold, there is no need to adjust the processing parameters of the single-crystal material.

[0183] If the crystal orientation angle difference is greater than or equal to the first single-crystal orientation angle difference threshold and less than the second single-crystal orientation angle difference threshold, input the crystal type, crystal orientation angle difference, and processing parameters into the blade processing parameter setting model to obtain target processing parameters, where the target processing parameters include target cutting force, target cutting speed, target cutting depth, target tool temperature, and target vibration amplitude; adjust the processing parameters of the blade to the target processing parameters.

[0184] If the crystal orientation angle difference is greater than or equal to the second single-crystal orientation angle difference threshold, the crystal plane of the currently processed single-crystal material is not suitable for processing, perform a crystal plane switching operation, and adjust to the adjacent crystal plane of the currently processed single-crystal material.

[0185] Step 3: Repeat Step 2 until the processing of the single-crystal material is completed.

[0186] It should be noted that the crystal orientation of the single-crystal material is continuous and stable, that is, the orientation of the same crystal plane is uniform throughout the material. Since different crystal planes of the single-crystal material have different physical properties (hardness, elastic modulus, friction coefficient, etc.), different thresholds can be used to evaluate whether to continue processing or change the crystal plane.

[0187] The method for intelligent adaptive adjustment of processing parameters for polycrystalline materials includes:

[0188] First step: Set a polycrystalline orientation angle difference threshold for the polycrystalline material.

[0189] Second step: If the crystal orientation angle difference is less than the polycrystalline orientation angle difference threshold, there is no need to adjust the processing parameters of the polycrystalline material.

[0190] If the crystal orientation angle difference is greater than or equal to the polycrystalline orientation angle difference threshold, input the crystal type, crystal orientation angle difference, and processing parameters into the blade processing parameter setting model to obtain target processing parameters.

[0191] Third step: Adjust the processing parameters of the blade to the target processing parameters.

[0192] Fourth step: Repeat the second step to the third step until the processing of the polycrystalline material is completed.

[0193] It should be noted that polycrystalline materials are composed of a large number of grains with randomly distributed orientations, and the internal crystal orientations of polycrystalline materials exhibit a high degree of discrete randomness. During the machining process, the cutting tool will continuously cut multiple grains with different orientations, resulting in a rapid change in crystal orientation in the local area. Therefore, compared with single-crystal materials, polycrystalline materials are not suitable for relying on the difference in crystal orientation angles to determine whether to change the cutting surface. Instead, it is more suitable to adopt the method of dynamically adjusting machining parameters to adapt to the influence of different grain orientations on cutting force, cutting temperature, and tool wear, so as to optimize machining stability and improve machining quality.

[0194] The training method of the cutting tool machining parameter setting model includes:

[0195] Pre-construct a cutting tool machining parameter setting data set, which includes B groups of cutting tool machining parameter setting data and the target machining parameters corresponding to the B groups of cutting tool machining parameter setting data. B is a positive integer greater than 0. The cutting tool machining parameter setting data includes equipment operation characteristic data; divide the cutting tool machining parameter setting data set into a cutting tool machining parameter setting data training set and a cutting tool machining parameter setting data verification set. Among them, the cutting tool machining parameter setting data training set is used for parameter learning of the cutting tool machining parameter setting model, and the cutting tool machining parameter setting data verification set is used for real-time evaluation of the generalization ability of the cutting tool machining parameter setting model.

[0196] During the training process of the cutting tool machining parameter setting model, a deep neural network structure based on a multi-layer perceptron is adopted. The cutting tool machining parameter setting data is converted into a feature vector as the input, and the non-linear features in the data are extracted through the hidden layer. Finally, the probability distribution of the target machining parameters is generated by using the softmax activation function in the output layer, and the target machining parameter corresponding to the maximum probability is output as the final prediction result; the training process aims to minimize the cross-entropy loss function, and at the same time introduce an early stopping strategy to monitor the performance of the cutting tool machining parameter setting data verification set. When the prediction accuracy on the cutting tool machining parameter setting data verification set reaches the preset accuracy, it is considered that the cutting tool machining parameter setting model has converged, and the training stops immediately.

[0197] An example of machining parameter adjustment for polycrystalline materials is as follows:

[0198] When machining WC-Co cemented carbide (polycrystalline material), due to the randomness of different grain orientations inside the polycrystalline material, the cutting force in the local area suddenly increases to 150 N, resulting in the cutting temperature rising to 500 °C, causing overheating wear of the cutting tool, and at the same time the vibration amplitude increases to 20 μm, affecting machining stability.

[0199] The machining parameters of the current cutting tool are shown in Table 1;

[0200] Table 1 Machining parameter table of the current cutting tool

[0201] Cutting force 120N Cutting speed 80 m / min Cutting depth 0.2 mm Tool temperature 450℃ Vibration amplitude 12 μm

[0202] The processing parameters of the adjusted blade are shown in Table 2;

[0203] Table 2 Processing parameter table of the adjusted blade

[0204] Cutting force 120N Cutting speed 70 m / min (Reduce heat accumulation) Cutting depth 0.15 mm (Reduce cutting load) Tool temperature 430 °C (Reduce tool wear) Vibration amplitude 10 μm (Improve machining stability)

[0205] Example 2

[0206] Please refer to Figure 2 As shown, this embodiment provides an intelligent adaptive method in the blade processing process, including:

[0207] Collect the physical parameters of the blade processing material;

[0208] Obtain the processing parameters of the real-time blade;

[0209] Collect the surface image of the processing material per unit time in real time;

[0210] Process the surface image of the processing material per unit time to obtain a crystal orientation matrix set;

[0211] Input the physical parameters and the crystal orientation matrix set into the crystal orientation prediction model to obtain the crystal orientation matrix of the future unit time;

[0212] Based on the physical parameters, processing parameters, crystal orientation matrix set and the crystal orientation matrix of the future unit time, perform intelligent adaptive adjustment on the processing parameters of the blade.

[0213] As mentioned above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

[0214] Finally: The above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included within the protection scope of the present invention.

Claims

1. An intelligent adaptive method during the blade processing, characterized in that, Including: Collecting the physical parameters of the blade processing material; Obtaining the processing parameters of the real-time blade; Collecting the surface image of the processing material per unit time in real time; Processing the surface image of the processing material per unit time to obtain a set of crystal orientation matrices; Inputting the physical parameters and the set of crystal orientation matrices into a crystal orientation prediction model to obtain the crystal orientation matrix for the next unit time; Based on the physical parameters, processing parameters, set of crystal orientation matrices, and crystal orientation matrix for the next unit time, intelligently and adaptively adjusting the processing parameters of the blade.

2. The intelligent adaptive method in the blade processing process according to claim 1, characterized in that, The method for intelligently and adaptively adjusting the processing parameters of the blade includes: Statistical frequency of each crystal orientation matrix in the set of crystal orientation matrices, and selecting the crystal orientation matrix with the highest frequency, denoted as the main crystal orientation matrix; Calculating the crystal orientation angle difference based on the main crystal orientation matrix and the crystal orientation matrix for the next unit time; If the crystal orientation angle difference is greater than or equal to a preset crystal orientation angle difference threshold, it is considered that the crystal orientation has changed and intelligent adaptive adjustment is required; Obtaining the crystal type from the physical parameters; If the crystal type is single crystal material, perform intelligent adaptive adjustment of processing parameters for single crystal material; If the crystal type is polycrystalline material, perform intelligent adaptive adjustment of processing parameters for polycrystalline material.

3. The intelligent adaptive method during the blade processing according to claim 2, wherein, The method for intelligently and adaptively adjusting the processing parameters for single crystal material includes: Step 1: Set a first single crystal orientation angle difference threshold and a second single crystal orientation angle difference threshold for the single crystal material; the first single crystal orientation angle difference threshold is less than the second single crystal orientation angle difference threshold; Step 2: If the crystal orientation angle difference is less than the first single crystal orientation angle difference threshold, there is no need to adjust the processing parameters of the single crystal material; If the crystal orientation angle difference is greater than or equal to the first single crystal orientation angle difference threshold and less than the second single crystal orientation angle difference threshold, input the crystal type, crystal orientation angle difference, and processing parameters into a blade processing parameter setting model to obtain target processing parameters, where the target processing parameters include target cutting force, target cutting speed, target cutting depth, target tool temperature, and target vibration amplitude; adjust the processing parameters of the blade to the target processing parameters; If the crystal orientation angle difference is greater than or equal to the second single crystal orientation angle difference threshold, the crystal plane of the currently processed single crystal material is not suitable for processing, perform a crystal plane switching operation, and adjust to the adjacent crystal plane of the currently processed single crystal material; Step 3: Repeat Step 2 until the processing of the single crystal material is completed.

4. The intelligent adaptive method in the blade processing process according to claim 3, wherein The method for intelligently and adaptively adjusting the processing parameters for polycrystalline material includes: The first step: Set a polycrystalline orientation angle difference threshold for the polycrystalline material; The second step: If the crystal orientation angle difference is less than the polycrystalline orientation angle difference threshold, there is no need to adjust the processing parameters of the polycrystalline material; If the crystal orientation angle difference is greater than or equal to the polycrystalline orientation angle difference threshold, input the crystal type, crystal orientation angle difference, and processing parameters into a blade processing parameter setting model to obtain target processing parameters; The third step: Adjust the processing parameters of the blade to the target processing parameters; The fourth step: Repeat the second step to the third step until the processing of the polycrystalline material is completed.

5. The intelligent adaptive method in the blade processing process according to claim 1, wherein, The method for obtaining the set of crystal orientation matrices includes: S100: Divide the unit time into N time points, let the initial value of n be 1, and the value range of n is from 1 to N; S101: Obtain the surface image of the processed material at the nth time point; divide the surface image of the processed material into M sub-regions, denoted as the surface sub-images of the processed material; S102: Perform diffraction band detection processing on the M surface sub-images of the processed material respectively to obtain the Kikuchi diffraction bands corresponding to the M surface sub-images of the processed material; S103: Obtain the crystal orientation matrices corresponding to the M surface sub-images of the processed material based on the M Kikuchi diffraction bands; S104: Perform matrix numerical comparison on the M crystal orientation matrices respectively to obtain the crystal orientation matrix of the surface image of the processed material at the nth time point; S105: Let n = n + 1. If n is less than or equal to N, then continue to execute S101 to S104. If n is greater than N, then obtain the crystal orientation matrices of the surface images of the processed material corresponding to N time points and execute S106; S106: Construct the crystal orientation matrices of the surface images of the processed material corresponding to N time points into a set of crystal orientation matrices.

6. The intelligent adaptive method during the blade processing according to claim 5, wherein The method for obtaining the Kikuchi diffraction bands corresponding to the M surface sub-images of the processed material includes: S200: Let the initial value of m be 1, and the value range of m is from 1 to M; S201: Obtain the mth surface sub-image of the processed material; update the mth surface sub-image of the processed material through an image processing library to a surface sub-image of the processed material with two-dimensional coordinates; S202: Denote the number of pixel points in the mth surface sub-image of the processed material as R; calculate the corresponding diffraction band gradient matrices for the R pixel points respectively; S203: Input the physical parameters of the blade processed material and the R diffraction band gradient matrices into a diffraction band gradient threshold setting model to obtain a diffraction band gradient threshold; S204: Construct the Kikuchi diffraction band of the mth surface sub-image of the processed material based on the R diffraction band gradient matrices and the diffraction band gradient threshold; S205: Let m = m + 1. If m is less than or equal to M, then continue to execute S201 to S204; if m is greater than M, then obtain the Kikuchi diffraction bands corresponding to the M surface sub-images of the processed material and end the current process.

7. The intelligent adaptive method during the blade processing according to claim 6, characterized in that, The method for constructing the Kikuchi diffraction band of the mth surface sub-image of the processed material based on the R diffraction band gradient matrices and the diffraction band gradient threshold includes: S300: Let the initial value of r be 1, and the value range of r is from 1 to R; S301: Obtain the rth diffraction band gradient matrix, and perform diffraction band gradient threshold determination for each element value in the matrix. Denote the elements with element values greater than or equal to the diffraction band gradient threshold as diffraction band elements, and count the number of diffraction band elements, denoted as the diffraction band element number; denote the elements with element values less than the diffraction band gradient threshold as non-diffraction band elements, and count the number of non-diffraction band elements, denoted as the non-diffraction band element number; S302: Sum the diffraction band element number and the non-diffraction band element number to obtain the total number of elements; divide the diffraction band element number by the total number of elements to obtain the diffraction band element ratio; S303: If the proportion of diffraction band elements is greater than or equal to the preset diffraction band element proportion threshold, then mark the pixel points corresponding to the r-th diffraction band gradient matrix as diffraction band pixel points; if the proportion of diffraction band elements is less than the preset diffraction band element proportion threshold, then mark the pixel points corresponding to the r-th diffraction band gradient matrix as non-diffraction band pixel points; S304: Let r = r + 1. If r is less than or equal to R, then continue to execute S301 to S303. If r is greater than R, then execute S305; S305: Based on all the diffraction band pixel points, use the pixel connection algorithm to connect the discrete diffraction band pixel points into complete diffraction band lines, and obtain the Kikuchi diffraction bands of the m-th processed material surface sub-image.

8. The intelligent adaptive method in the blade processing process according to claim 5, wherein The method for obtaining the crystal orientation matrices corresponding to M processed material surface sub-images includes: S400: Let the initial value of m be 1, and the value range of m is from 1 to M; S401: Obtain the Kikuchi diffraction bands of the m-th processed material surface sub-image; select three mutually independent diffraction band normal vectors from the Kikuchi diffraction bands, and denote them as diffraction band normal vector one, diffraction band normal vector two, and diffraction band normal vector three respectively; S402: Construct a diffraction band normal vector matrix from diffraction band normal vector one, diffraction band normal vector two, and diffraction band normal vector three; S403: Calculate diffraction direction angle one, diffraction direction angle two, and diffraction direction angle three based on the diffraction band normal vector matrix; S404: Construct a diffraction direction cosine matrix one with diffraction direction angle one, construct a diffraction direction cosine matrix two with diffraction direction angle two, and construct a diffraction direction cosine matrix three with diffraction direction angle three; S405: Multiply the diffraction direction cosine matrix one, the diffraction direction cosine matrix two, and the diffraction direction cosine matrix three matrix-wise to obtain the crystal orientation matrix corresponding to the m-th processed material surface sub-image; S406: Let m = m + 1. If m is less than or equal to M, then continue to execute S401 to S405; if m is greater than M, then obtain the crystal orientation matrices corresponding to M processed material surface sub-images, and end the current process.

9. The intelligent adaptive method during the blade machining process according to claim 5, wherein The method for obtaining the crystal orientation matrix of the processed material surface image at the n-th time point includes: S500: Let the initial value of m be 1, and the value range of m is from 1 to M; pre-construct a crystal orientation matrix key-value pair set; the crystal orientation matrix key-value pair set includes crystal orientation matrix key-value pairs; the crystal orientation matrix key-value pair includes a key and a value, the key is the crystal orientation matrix, the key is initially empty, and the value is the quantity corresponding to the crystal orientation matrix, and the initial value of the value is 0; S501: Obtain the m-th crystal orientation matrix; if the m-th crystal orientation matrix is the same as the key of the crystal orientation matrix key-value pair in the crystal orientation matrix key-value pair set, increment the value corresponding to the crystal orientation matrix key-value pair by one; if the m-th crystal orientation matrix is different from the key of the crystal orientation matrix key-value pair in the crystal orientation matrix key-value pair set, create a new crystal orientation matrix key-value pair, use the m-th crystal orientation matrix as the key of the new crystal orientation matrix key-value pair, and increment the value of the new crystal orientation matrix key-value pair by one; add the new crystal orientation matrix key-value pair to the crystal orientation matrix key-value pair set; S502: Let m = m + 1. If m is less than or equal to M, continue to execute S501; if m is greater than M, execute S503; S503: Retrieve the crystal orientation matrix key-value pair with the maximum value from the crystal orientation matrix key-value pair set, and use the key of the crystal orientation matrix key-value pair with the maximum value as the crystal orientation matrix of the surface image of the processed material at the n-th time point, and end the current process.

10. An intelligent adaptive system in the blade machining process, which implements the intelligent adaptive method in the blade machining process according to any one of claims 1-9, characterized in that, Including: The first acquisition module, which acquires the physical parameters of the blade processed material; The second acquisition module, which is used to obtain the processing parameters of the real-time blade; The third acquisition module, which is used to acquire the surface image of the processed material per unit time in real time; The first processing module, which processes the surface image of the processed material per unit time to obtain a crystal orientation matrix set; The orientation prediction module, which is used to input the physical parameters and the crystal orientation matrix set into the crystal orientation prediction model to obtain the crystal orientation matrix of the next unit time; The intelligent optimization module, which intelligently and adaptively adjusts the processing parameters of the blade based on the physical parameters, processing parameters, crystal orientation matrix set, and the crystal orientation matrix of the next unit time.

Images