A ceramic fiber composite gasket zonal cutting adaptive control system and method

Abstract

The embodiment of the application provides a kind of ceramic fiber composite gasket partition cutting adaptive control system and method, it is applied to composite material processing technical field, first obtains fiber trend, microscopic damage, cutting resistance, cutter pose and the data of the cutting topography when cutting, input stress distribution calculation program in combination with preset design data, stress distribution and deformation trend data are obtained by anisotropic solution and geometric nonlinear analysis;Again input partition stress evolution calculation program, virtually remove candidate partition stress redistribution, determine cutting order according to cumulative deformation after cutting in ascending order;Subsequently, extract stress concentration zone and deformation direction, and iteratively optimize adaptive cutting speed, energy and offset compensation path;If the deviation of the cut exceeds the limit after executing the cutting, recalculate the parameters, until all partition cutting is completed, which can effectively suppress stress release deformation and improve the consistency of the cut.

Description

Technical Field

[0001] This application relates to the field of composite material processing technology, and in particular to an adaptive control system and method for partitioned cutting of ceramic fiber composite liner. Background Technology

[0002] Ceramic fiber composite liners are characterized by anisotropy, brittleness, and low interlaminar strength, making them prone to defects such as edge delamination, fiber pull-out, and edge chipping during cutting. Traditional cutting processes typically employ fixed parameters and a single path, lacking effective process control methods. They are ill-suited to adapting to differences in local fiber orientation and stiffness distribution within the material, resulting in poor cut quality consistency and low yield.

[0003] Most existing zoning methods rely on static zoning based on geometric shapes and a fixed cutting sequence, failing to consider the overall deformation of the gasket caused by stress redistribution during the cutting process. As material is gradually removed, the release of residual stress can lead to warping or displacement of the processed areas, resulting in dimensional deviations and difficulties in subsequent assembly in severe cases.

[0004] Currently, there is a lack of adaptive control methods that can integrate material state information in real time, such as fiber orientation and microscopic damage, with processing information, such as cutting resistance and pose data. Stress distribution calculation and deformation trend prediction are not integrated online into the cutting decision-making process, making it impossible to dynamically optimize process parameters and paths, thus hindering the realization of high-precision cutting. Summary of the Invention

[0005] Embodiments of this application provide an adaptive control system and method for partitioned cutting of ceramic fiber composite liners. This system introduces refined process control logic, suppressing deformation and defects during the cutting of ceramic fiber composite liners, and significantly improving cutting accuracy and consistency. To achieve the above objectives, this application adopts the following technical solution: An adaptive control method for partitioned cutting of ceramic fiber composite gaskets, the method comprising: During the cutting process, acquire the material state information and processing information of the current cutting area of ​​the liner; The material state information, the processing information, and the preset design data of the liner are respectively input into a preset stress distribution calculation program, and the stress distribution calculation program is run to obtain stress distribution calculation data and deformation trend data. The stress distribution calculation data and the deformation trend data are input into a preset partitioned stress evolution calculation program, and the partitioned stress evolution calculation program is run to obtain stress evolution prediction information. Based on the stress evolution prediction information, the cutting order of multiple partitions is determined; Based on the stress distribution calculation data and the deformation trend data, adaptive process parameters and cutting path instructions are calculated. The partition cutting is performed according to the cutting sequence, the adaptive process parameters, and the cutting path instructions.

[0006] In some possible implementations, obtaining the material state information and processing information of the current cutting area of ​​the liner includes: Acquire a surface image of the current cutting area of ​​the pad, perform texture analysis on the surface image, and obtain fiber orientation feature data; Acoustic emission signals from the current cutting area of ​​the liner are collected, and feature extraction operations are performed on the acoustic emission signals to obtain microscopic damage feature data. The reaction force signal experienced by the cutting actuator is collected, and numerical conversion calculation is performed on the reaction force signal to obtain cutting resistance data; Collect cutting head pose data and cut area topography data; The fiber orientation characteristic data, the micro-damage characteristic data, the cutting resistance data, the cutting head pose data, and the morphology data of the cut area are used as the material state information and processing information.

[0007] In some possible implementations, the step of inputting the material state information, the processing information, and the preset design data of the gasket into a preset stress distribution calculation program, running the stress distribution calculation program, and obtaining stress distribution calculation data and deformation trend data includes: The fiber orientation characteristic data and the cutting resistance data are obtained from the material state information and processing process information; The acquired fiber orientation characteristic data and cutting resistance data, combined with preset padding material stiffness parameters and thermophysical parameters, are input into the stress distribution calculation program. Run the anisotropic material mechanics solution module in the stress distribution calculation program to calculate the stress distribution data inside the gasket under the current constraint conditions; Run the geometric nonlinear analysis module in the stress distribution calculation program to calculate the deformation trend data of the offset of each node position of the liner.

[0008] In some possible implementations, the step of calculating adaptive process parameters and cutting path instructions based on the stress distribution calculation data and the deformation trend data includes: Extract the coordinate information of the stress concentration area from the stress distribution calculation data, and extract the deformation direction vector from the deformation trend data; Using the coordinate information of the stress concentration area and the deformation direction vector as constraints, a preset process parameter adjustment calculation program is called to iteratively optimize the cutting speed and cutting energy values ​​until the preset convergence condition is met, and the adjusted cutting speed and cutting energy values ​​are output. Adaptive process parameters are generated based on the adjusted cutting speed and cutting energy values. The offset compensation amount of the cutting tool is calculated based on the deformation direction vector, and the offset compensation amount is superimposed with the preset path to obtain the cutting path instruction.

[0009] In some possible implementations, the step of inputting the stress distribution calculation data and the deformation trend data into a preset partitioned stress evolution calculation program, running the partitioned stress evolution calculation program, and obtaining stress evolution prediction information includes: Extract the actual contour data of the cut partitions from the processing information; The actual contour data is compared with the preset partitioning plan file to obtain a list of candidate partitions that have not yet been cut. Read the geometric feature data of each candidate partition one by one from the candidate partition list; By removing the mesh cells representing candidate partitions in the simulation environment, stress redistribution data for the remaining portion of the liner is calculated. The stress redistribution data corresponding to each candidate partition are summarized to form stress evolution prediction information.

[0010] In some possible implementations, determining the cutting order of multiple partitions based on the stress evolution prediction information includes: Read the stress redistribution data corresponding to each candidate partition from the stress evolution prediction information; For each candidate partition, the stress redistribution data corresponding to the candidate partition is input into a preset deformation calculation program, the deformation calculation program is run, and the cumulative deformation value of the entire pad structure after the candidate partition is removed is output. Perform a sorting and comparison operation on the cumulative deformation values ​​corresponding to each candidate partition, select the candidate partition based on the sorting and comparison results, and mark the selected candidate partition as the current cutting target; Write the identifier of the current cutting target into the cutting order queue.

[0011] In some possible implementations, the method further includes: Collect the processing information, adaptive process parameters, cutting path instructions, and cut quality data obtained after performing partitioned cutting generated during the cutting process, and combine them into a process data record; The multiple process data records accumulated from multiple partitioning and splitting tasks are combined to generate a historical dataset.

[0012] In some possible implementations, the method further includes: The process data records are subjected to de-identification processing to obtain de-identified process data; The desensitization process data is sent to a remote computing terminal, which performs material parameter identification calculations based on the desensitization process data and returns the corrected material stiffness parameters. Receive the corrected material stiffness parameters returned by the remote computing terminal; In subsequent partitioning and cutting tasks, the modified material stiffness parameters are called to run the stress distribution calculation program.

[0013] In some possible implementations, after performing the partitioning, the process further includes: Acquire images of the actual cut morphology of the cut area; Edge extraction operations are performed on the actual cut morphology image to obtain measured edge geometric parameters characterizing the cut quality; The measured edge geometric parameters are compared with the preset cut quality benchmark parameters to calculate the cut deviation value. When the notch deviation value is greater than the preset deviation threshold, the recalculation operation of the adaptive process parameters is triggered.

[0014] A ceramic fiber composite liner partition cutting adaptive control system, the system comprising: The information acquisition unit is used to acquire material status information and processing information of the current cutting area of ​​the liner during the cutting process; The stress distribution calculation unit is used to input the material state information, the processing information, and the preset design data of the liner into a preset stress distribution calculation program, run the stress distribution calculation program, and obtain stress distribution calculation data and deformation trend data. The partitioned stress evolution calculation unit is used to input the stress distribution calculation data and the deformation trend data into a preset partitioned stress evolution calculation program, run the partitioned stress evolution calculation program, and obtain stress evolution prediction information. The cutting sequence determination unit is used to determine the cutting sequence of multiple partitions based on the stress evolution prediction information. The process parameter and path calculation unit is used to calculate adaptive process parameters and cutting path instructions based on the stress distribution calculation data and the deformation trend data. The execution control unit is used to perform partitioned cutting according to the cutting sequence, the adaptive process parameters, and the cutting path instructions.

[0015] As can be seen from the above technical solution, this application has the following beneficial effects: 1. This application obtains material state information and processing information in real time, combines it with preset design data input into a stress distribution calculation program to obtain stress distribution calculation data and deformation trend data, and then obtains stress evolution prediction information through a partitioned stress evolution calculation program, thereby scientifically determining the cutting order of multiple partitions. This order is not based on a static sorting of geometric shapes, but rather on a dynamic optimization based on the cumulative deformation value of the remaining structure after each candidate partition is cut, prioritizing the cutting of the partition that causes the least deformation, thereby fundamentally suppressing the overall warping of the gasket and residual stress concentration. Compared with fixed-sequence cutting, this invention can reduce the final deformation caused by cutting, significantly improving the contour accuracy and dimensional stability of large-size ceramic fiber composite gaskets, and is especially suitable for applications with strict requirements for gasket shape tolerances in aerospace, high-temperature insulation and other fields.

[0016] 2. This application adaptively calculates cutting speed, cutting energy, and cutting path instructions with offset compensation based on stress distribution and deformation trend data, achieving dynamic matching between process parameters and real-time material conditions. For example, it automatically reduces cutting energy in stress concentration areas and adjusts tool offset in the deformation trend direction, effectively avoiding fiber fraying, edge chipping, or overheating caused by improper parameters. The system collects processing information, process parameters, and cut quality data during the cutting process, forming a historical dataset that can be uploaded to a remote computing terminal for material parameter identification and correction, continuously optimizing the stiffness parameter input for subsequent tasks. As cutting tasks accumulate, the geometric deviation of the cut edge can gradually converge to within a preset deviation threshold, improving the automation level and batch consistency of ceramic fiber composite gasket zonal cutting. Attached Figure Description

[0017] The invention will now be further described with reference to the accompanying drawings.

[0018] Figure 1 This is a flowchart of the adaptive control for partitioning and cutting provided in an embodiment of this application. Figure 2 A flowchart illustrating the calculation of stress distribution and deformation trend provided in this application embodiment; Figure 3 A flowchart of the partitioned stress evolution prediction and sequence decision-making sub-process provided in the embodiments of this application; Figure 4 This is a flowchart illustrating the adaptive process parameters and cutting path planning sub-process provided in the embodiments of this application. Figure 5 The flowchart for quality feedback and remote parameter identification provided in the embodiments of this application is as follows. Detailed Implementation

[0019] The terms "first," "second," and "third," etc., used in this application specification, claims, and drawings are used to distinguish different objects, not to limit a specific order.

[0020] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0021] Research has found that existing ceramic fiber composite gasket cutting methods mostly use fixed process parameters and static partitioning sequence based on geometry, which cannot adapt to the anisotropy of materials and stress redistribution during the cutting process, resulting in frequent defects such as delamination and edge chipping. At the same time, they lack the ability to perceive and adaptively control the material state and processing process in real time, making it difficult to suppress the overall deformation caused by residual stress release, resulting in poor cutting accuracy and consistency.

[0022] To address the aforementioned issues, this application provides a ceramic fiber composite gasket partitioning and cutting adaptive control system and method. This composite gasket product is applicable to Baosteel's ultra-high grade grain-oriented silicon steel green manufacturing project. Example 1: To solve the above problems, such as Figure 1 As shown, the following description, taking the ultrasonic cutting of irregularly shaped ceramic fiber composite liners for aerospace thermal insulation components as an example, will provide a detailed, complete, and clear description of the technical solution of this invention, combined with a specific processing scenario. In this embodiment, the control system runs on an industrial control computer of a CNC cutting machine tool equipped with a six-axis industrial robotic arm, a high-speed industrial camera, an acoustic emission sensor array, a six-dimensional force / torque sensor, and a line laser profilometer.

[0023] I. System Hardware Configuration and Data Definition.

[0024] (1) Fiber orientation characteristic data: denoted as The unit is degrees (°), and the range of values ​​is... to This represents the angle between the principal direction of the ceramic fiber bundles on the surface of the current cutting area and the X-axis of the machine tool coordinate system. This data is obtained by acquiring surface images through a high-speed industrial camera and then calculating using a gray-level co-occurrence matrix and a deep convolutional neural network.

[0025] (2) Microscopic damage characteristic data: denoted as , dimensionless. Among them Characterizing the degree of fiber breakage damage, Characterizes the degree of matrix cracking. The degree of interlayer debonding is characterized. This data is obtained by acquiring stress wave signals through a resonant acoustic emission sensor, followed by wavelet packet decomposition and principal component analysis.

[0026] (3) Cutting resistance data: denoted as The unit is Newton (N). The tangential resistance in the tool feed direction. It is a lateral force. This refers to the normal pressure perpendicular to the liner surface, i.e., the cutting depth direction. This data is acquired through a six-dimensional force / torque sensor integrated into the tool holder and obtained after calibration matrix transformation.

[0027] (4) Cutting head pose data: denoted as ,in The coordinates (in mm) of the center point of the tool tip in the machine tool coordinate system. This is the Euler angle (in degrees) along the tool axis. This data is output in real time by the robotic arm controller.

[0028] (5) Topographic data of the cut area: denoted as , is a set of three-dimensional point cloud coordinates This represents the surface geometry of the liner after it has been cut and stress-released. This data was obtained by scanning with a line laser profilometer mounted on the cutter.

[0029] (6) Pre-set design data: including the three-dimensional computer-aided design model of the padding. (Triangular mesh or NURBS surface form), nominal material stiffness parameters (Sixth-order tensor), nominal thermophysical parameters (thermal conductivity) Specific heat capacity Anisotropic thermal expansion coefficient ), and zoning planning documents pre-defined by process engineers. The zoning plan divides the entire padding into N geometric sub-regions, or zoning zones, each zone... From its boundary curve The identifier (ID) is uniquely identified.

[0030] (7) Material stiffness parameter (after real-time correction): denoted as It is a stiffness tensor that, based on nominal parameters, is obtained by identifying remote parameters using historical cutting data, and is closer to the true properties of the current batch of materials. Specifically, it includes: longitudinal Young's modulus. Transverse Young's modulus Normal Young's modulus (Unit: GPa), Shear Modulus and Poisson's ratio .

[0031] (8) Thermophysical parameters: Nominal values ​​are used in the initial stage and subsequently corrected through the same identification process.

[0032] (9) Stress distribution calculation data: denoted as The unit is megapascals (MPa), representing any point inside the gasket. The stress tensor components at each point. After finite element discretization, this data is represented as an array of stress values ​​at each mesh integration point.

[0033] (10) Deformation trend data: denoted as The unit is millimeters (mm), which is the three-dimensional displacement vector of any point inside the liner relative to its original undeformed position. This data is also stored in the form of finite element nodal displacements.

[0034] (11) Stress redistribution data: denoted as This indicates whether a virtual or physical partition is cut off. After that, the change in the stress field within the remaining structure.

[0035] (12) Cumulative deformation value: denoted as The unit is square millimeters (mm²). It is defined as the weighted square sum of the displacements of several predefined key points on the gasket, such as the center of the assembly positioning hole or the highest point of the outer contour, under a certain cutting sequence.

[0036] (13) Adaptive process parameters: determined by cutting speed (mm / s) and cutting energy (For ultrasonic cutting, it is an amplitude percentage, 0% to 100%) consisting of two scalars.

[0037] (14) Cutting path command: a set of offset-compensated spatial position sequences with tool orientation. Each point includes its coordinates and the direction of the tool axis.

[0038] (15) Material equivalent stress limit: denoted as The unit is megapascals (MPa). This parameter represents the equivalent stress failure threshold of the ceramic fiber composite gasket under uniaxial tensile conditions. In this embodiment, the smaller value between the tensile strength along the fiber direction and the tensile strength perpendicular to the fiber direction is taken as... A conservative estimate. If measured data is lacking, the factory nominal value can be used, such as 300 MPa. This parameter is used to determine the stress concentration threshold, i.e., the value in step 5.1.1. .

[0039] II. Detailed steps of the method.

[0040] Step 1: During the cutting process, obtain the material status information and processing information of the current cutting area of ​​the liner.

[0041] During the cutting process, the control system performs the following acquisition and calculation operations in parallel at a fixed sampling period, for example, once every 10 milliseconds.

[0042] 1.1 Calculation of fiber orientation characteristic data.

[0043] 1.1.1 Trigger a high-speed industrial camera to capture a high-resolution grayscale image of the current cutting area, approximately 10mm × 10mm in front of the tool, with an image size of 512 × 512 pixels.

[0044] 1.1.2 Image preprocessing: First, apply medium filtering with a kernel size of 3×3 to remove salt-and-pepper noise, and then use histogram equalization to enhance the contrast between the fiber bundles and the matrix.

[0045] 1.1.3 Calculate the gray-level co-occurrence matrix. Set the orientation parameters. Pixel distance Grayscale compression to Level. For each direction ,calculate co-occurrence matrix , of which elements Indicates grayscale With grayscale The normalized frequency of pixel pairs appearing in this direction.

[0046] 1.1.4 Extract three statistical features from the co-occurrence matrix in each direction: Contrast Ratio: ; Correlation: ,in These are the mean and standard deviation of the row and column, respectively; energy: .

[0047] This results in a 12-dimensional feature vector: .

[0048] 1.1.5 The feature vector is input into a pre-trained three-layer fully connected neural network. The network structure is as follows: 12 nodes in the input layer, 32 nodes in the hidden layer, using ReLU activation function; and 1 node in the output layer, using a linear activation function (directly outputting the angle value). The network training data comes from 5000 images of similar materials labeled with fiber orientation angles collected previously. The network output value is the fiber orientation feature data. .

[0049] 1.2 Calculation of microscopic damage characteristic data.

[0050] 1.2.1 Acquire acoustic emission sensor signals at a sampling rate of 2 MHz, and record time-series data with a length of 4096 points. .

[0051] 1.2.2 Selection of db8 wavelet basis pairs Perform 3-level wavelet packet decomposition to obtain 8 frequency band nodes. Each node corresponds to a set of wavelet packet coefficients. .

[0052] 1.2.3 Calculate the reconstructed signal energy for each node: Using discrete approximation ,in Microseconds are the sampling interval.

[0053] 1.2.4 Calculate the total energy Then the energy ratio vector is obtained. There are a total of 8 dimensions.

[0054] 1.2.5 (Regarding) Principal component analysis (PCA) was performed for dimensionality reduction. The PCA model was pre-built using historical data, retaining the first three principal components, with a cumulative variance contribution rate >85%. After projection, a three-dimensional vector was obtained. This refers to the microscopic damage characteristic data. It mainly reflects the proportion of high-frequency energy (>500kHz) and is related to fiber breakage; It reflects the proportion of mid-frequency energy (250-500kHz) and is related to matrix cracking; It reflects the proportion of low-frequency energy (<250kHz) and is related to delamination or debonding.

[0055] 1.3 Obtaining cutting resistance data.

[0056] 1.3.1 Reading the analog voltage signal vector output by the six-dimensional force / torque sensor (Unit: volts)

[0057] 1.3.2 Perform zero-point calibration: ,in This is the zero-point voltage vector calibrated when the tool is unloaded.

[0058] 1.3.3 Application of the calibration matrix (A 6×6 matrix, obtained from sensor factory calibration or field calibration) converts voltage into force and torque: This embodiment only considers the force component, therefore cutting the resistance data. .

[0059] 1.4 Acquisition of pose and shape data.

[0060] 1.4.1 The position and pose of the tool end effector in the machine tool coordinate system are read in real time by the robotic arm controller to obtain... .

[0061] 1.4.2 A trigger-line laser profilometer (mounted approximately 20mm behind the tool) scans the area that has just been cut. The laser line is perpendicular to the feed direction, and each scan obtains a 3D point cloud along the profilometer's contour line. The point cloud from the profilometer's coordinate system is transformed to the machine tool's coordinate system through coordinate transformation and then stitched with historical scan data to form the topographic data of the cut area. .

[0062] 1.5 Data summary.

[0063] The fiber orientation characteristic data obtained above Microscopic damage characteristic data Cutting resistance data Cutting head pose data and topographic data of the cut area The information is packaged into a single structure, called the current material state information and processing information, denoted as... This information will be used in subsequent calculation steps.

[0064] Step Two: As Figure 2 As shown, the material state information, the processing information, and the preset design data are input into the stress distribution calculation program to obtain stress distribution calculation data and deformation trend data.

[0065] The control system will obtain the results from step one. The pre-defined design data (geometric model, nominal material parameters, etc.) read from the design database is input into a pre-defined stress distribution calculation program. This program uses the finite element method and considers the anisotropy and geometric nonlinearity of the material.

[0066] 2.1 Finite element model initialization.

[0067] 2.1.1 Read the three-dimensional geometric model of the liner from the preset design data The mesh is generated using tetrahedral or hexahedral elements. To ensure computational accuracy, local refinement is applied near the expected cutting path, with the mesh size controlled within 0.5 mm, resulting in an overall element count of approximately 500,000.

[0068] 2.1.2 Based on fiber orientation characteristic data Assign orientation values ​​to the material properties of each element. For the first element... Each unit, with its centroid at... The value is used as the local material coordinate system orientation angle for this element. Specifically, the principal material direction of the element is defined as: the angle between direction 1 (fiber direction) and the global X-axis is... The angle between the two directions (the directions perpendicular to the fiber in the plane) and the global X-axis is... 3. The direction (thickness direction) is perpendicular to the liner plane.

[0069] 2.2 Constitutive model of anisotropic materials.

[0070] The construction matrix of orthogonal anisotropic linear elastic materials in the principal direction coordinate system (1-2-3). for:

[0071] in These are Young's moduli (in GPa) in three directions. Poisson's ratio, This is the shear modulus. The initial value is the nominal material stiffness parameter. If the corresponding value in the table has been corrected by remote parameter identification, then the corrected value will be used. .

[0072] Transform the constitutive matrix from the principal material directions to the global coordinate system: ,in The coordinate transformation matrix is ​​derived from the direction angle. And the ply angle is determined.

[0073] 2.3 Boundary conditions and load application.

[0074] 2.3.1 Displacement Boundary Conditions: The area where the bottom surface of the liner (the side away from the cut surface) contacts the vacuum adsorption fixture is set as a fixed constraint, meaning that the displacement of all nodes in this area is limited. For the topographic data of the cut area The boundary marked as the completed cut is subjected to a zero-pressure boundary condition.

[0075] 2.3.2 Force Load: Cutting resistance data As a concentrated force, it is applied to the finite element node at the current tool contact position. Simultaneously, considering the high-frequency vibrations of ultrasonic cutting, The stress distribution is decomposed into static components (average force) and dynamic components (amplitude). In this embodiment, only the static components are considered for a first approximation of the stress distribution.

[0076] 2.4 Solving module for anisotropic material mechanics (linear elasticity solution).

[0077] This module is used to calculate the stress distribution under the current load and boundary conditions.

[0078] Establish the global finite element equation: ,in The unknown vector is composed of the displacements of all nodes (dimension 3 × number of nodes). The external load vector of the node. This is the overall stiffness matrix. From the stiffness matrix of all elements Assembled , This is the strain-displacement matrix.

[0079] Solving this large sparse linear system of equations employs the preprocessed conjugate gradient (PCG) method, using incomplete Chollisky decomposition as the preprocessor. The iteration terminates when the Euclidean norm of the residual vector is less than 1. .

[0080] The displacement field is obtained by solving. Then, through strain-displacement relationship and constitutive equations Calculate the stress components at each integration point to obtain stress distribution calculation data. .

[0081] 2.5 Geometric Nonlinear Analysis Module (Large Deformation Prediction).

[0082] Because the ceramic fiber composite gasket may warp significantly during the cutting process (the displacement in the thickness direction can be comparable to the thickness of the sheet), geometric nonlinearity must be considered. This module uses an updated Lagrangian scheme for incremental solution.

[0083] 2.5.1 Decompose the total cutting force load into Each increment step, with the load increment at each increment step. .

[0084] 2.5.2 In the... At the start of the step, based on the current configuration, the node coordinates are... Calculate the tangent stiffness matrix The tangent stiffness matrix includes the elastic stiffness matrix and the geometric stiffness matrix, i.e., the initial stress stiffness.

[0085] 2.5.3 Solving for incremental displacement: Update total displacement. Update node coordinates .

[0086] 2.5.4 Calculate the internal forces under the current configuration With external load Unbalanced forces between .if If the increment step is successful, proceed to the next step; otherwise, execute the Newton-Raphson iteration to solve the problem. ,renew Recalculate This continues until the convergence condition is met.

[0087] 2.5.5 After all incremental steps are completed, output the final nodal displacement field. This refers to deformation trend data. The displacement vector at any point can be extracted from this displacement field, where the deformation direction vector is defined as... .

[0088] 2.6 Data output.

[0089] After step two is completed, two main data objects are output: stress distribution calculation data. (Stress tensor at each node or integration point) and deformation trend data (Displacement vectors at each node). These two data points will be used for subsequent prediction of zoned stress evolution and adaptive process parameter planning.

[0090] Step 3: As Figure 3 As shown, the stress distribution calculation data and the deformation trend data are input into the partitioned stress evolution calculation program to obtain stress evolution prediction information.

[0091] The goal of this step is to predict how the stress field of the remaining structure of the liner will change if one of the candidate sections is subsequently removed, based on the current stress distribution and deformation state, thus providing a basis for deciding the optimal cutting sequence. This process is automatically executed by a section stress evolution calculation program.

[0092] 3.1 Identify the cut partitions and candidate partitions.

[0093] 3.1.1 Topographic data of the cut area obtained from step one The point cloud containing the actual cutting boundaries is extracted. A region growing algorithm is used to segment the point cloud into different connected components. Each component corresponds to a partition that has been completely cut and removed, or a partition that has been cut but not removed, with its boundaries being completely free.

[0094] 3.1.2 Read the preset partitioning plan file This file records all the partitions created during the design phase. and its ideal boundary curve .

[0095] 3.1.3 Perform comparison operations: For each partition Calculate its ideal boundary The Hausdorff distance between the measured cut boundary and the actual cut boundary. If this distance is less than a threshold, such as 0.5 mm, then it is marked. Partitions that have been cut are marked as such; otherwise, they are marked as uncut. All uncut partitions constitute a candidate partition list. .

[0096] 3.2 Virtual removal and stress redistribution calculation for each candidate partition.

[0097] Next, the program will... Each candidate partition Perform the following virtual simulation operations. These operations are performed on a copy of the model in computer memory and do not change the actual processing state.

[0098] 3.2.1 Copy the finite element model at the current moment, including all material properties and stress distribution. Displacement ), to obtain a copy of the model .

[0099] 3.2.2 In Identify partitions in the middle The set of interface nodes between the padding and the remaining portion .

[0100] 3.2.3 Calculate the release force at the interface. For each interface node... Based on the current stress distribution The stress at the boundary of the integral element is obtained from the stress at this node. Unit nodal forces on one side In mathematics, ,in This is the shape function corresponding to the node. In practical finite element software, the nodal force vector can be directly extracted through the post-processing interface. Therefore, to achieve equilibrium in the remaining structure, equal and opposite external forces need to be applied to the interface nodes. .

[0101] 3.2.4 Remove all partitions from the model copy. The mesh elements and nodes are removed, along with any loads and constraints on these elements.

[0102] 3.2.5 In the new model, the release force will be calculated only on the remaining portion. Apply to the interface nodes. Keep other boundary conditions unchanged.

[0103] 3.2.6 Solving the Static Equilibrium Equations ,in This is the stiffness matrix of the remaining structure.

[0104] 3.2.7 Calculate stress redistribution data ,in Derived from the new displacement field. This data reflects the excised partitions. Afterwards, the change in the stress field in the remaining structure increases or decreases.

[0105] 3.2.8 Recording Partitions and its corresponding stress redistribution data Simultaneously, the predicted displacements of key points on the remaining structure after removal are recorded for use in the next step of cumulative deformation calculation.

[0106] Repeat sub-steps 3.2.1 to 3.2.8 for each partition in the candidate partition list to obtain a set of prediction results, called stress evolution prediction information, denoted as... .

[0107] Step 4: Determine the cutting order of multiple partitions based on the stress evolution prediction information.

[0108] This step is based on stress evolution prediction information, selects the next partition to be cut according to the criterion of minimizing cumulative deformation, and dynamically constructs the cutting sequence queue.

[0109] 4.1 Calculation of cumulative deformation values.

[0110] For each candidate partition Calculate the cumulative deformation of the entire pad structure if it is removed. .

[0111] 4.1.1 Predefine a set of key points, denoted as... These critical points are typically located where dimensional accuracy needs to be maintained during final assembly or inspection of the workpiece, such as the center of locating holes and extreme points of the outer contour. Each critical point... The coordinates are known in the undeformed state.

[0112] 4.1.2 Obtained from step three Extract each key point Predicted displacement vector .Notice It is the absolute displacement relative to the original undeformed state, while cumulative deformation focuses on the displacement increment newly generated due to the current cutting operation, therefore it should be calculated. ,in It is the current (before resection) displacement field.

[0113] 4.1.3 Define the weight of each key point (Value range 0-1). The weight is set according to the functional importance of the key point in the final product. For example, the weight of the center of the positioning hole is set to 1.0, and the weight of the non-key contour point is set to 0.2.

[0114] 4.1.4 Calculate the cumulative deformation value:

[0115] in This represents the square of the displacement vector magnitude. The unit of this value is square millimeters. The smaller the value, the smaller the impact on the overall shape of the workpiece after removing this section (i.e., the gentler the deformation trend).

[0116] 4.2 Ranking and Decision Making.

[0117] 4.2.1 For all candidate partitions ,according to The values ​​are sorted in ascending order. That is, the partition with the smallest cumulative deformation is ranked first.

[0118] 4.2.2 Select the candidate partition that ranks first in the sorting as the current cutting target, denoted as... .

[0119] 4.2.3 Will Identifiers, such as partition numbers, are added to a first-in-first-out queue. The tail.

[0120] It is important to note that the cutting order in this embodiment is dynamically determined, rather than planned all at once. That is, before each actual cutting of a partition, steps one through four are re-executed to obtain the latest state and make predictions, thereby determining the next partition to be cut. This allows for real-time adjustments to the material's state after each cutting step. (Queue) Only the partitioning order that has been decided but not yet executed is stored. In actual control, the queue often contains only one element before each cutting step, i.e. After the cutting is completed, the element is popped, and the decision-making process is re-executed to determine the next partition. For the first cut, when there are no cut partitions yet, the candidate partition list is all partitions, and the cumulative deformation calculation... Take it as zero field.

[0121] Step 5: As Figure 4 As shown, adaptive process parameters and cutting path instructions are calculated based on the stress distribution calculation data and the deformation trend data.

[0122] After determining the next partition to be cut... Next, the system needs to plan the specific machining parameters and toolpaths for this zone. This step is based on the current stress distribution data. and deformation trend data Perform the calculation.

[0123] 5.1 Adaptive process parameter optimization.

[0124] 5.1.1 Extract the coordinate information of the stress concentration area.

[0125] From stress distribution data In the middle, calculate the Mises equivalent stress at each point:

[0126] Will Exceeding the threshold A set of continuous elements is defined as a stress concentration region, and its coordinate range is denoted as... These areas are where the impact load should be minimized during cutting.

[0127] 5.1.2 Extract the deformation direction vector.

[0128] For any point in the path planning From deformation trend data The displacement vector of a point is obtained by interpolation using shape functions. The deformation direction vector is defined as follows: When the displacement is close to zero, the default direction is taken as the direction perpendicular to the liner plane.

[0129] 5.1.3 Process parameter optimization model.

[0130] Decision variable: Cutting speed (Unit: mm / s, range of values) Cutting energy (For ultrasonic cutting, amplitude percentage, value range) ).

[0131] Define the objective function This function takes into account processing efficiency, heat-affected zone width, and burr height.

[0132] in: This represents the predicted width of the heat-affected zone (in mm). This embodiment uses an empirical model. This model was developed by performing cutting experiments on the same type of material with different parameter combinations, measuring the depth of the heat-affected zone near the cut, and then using nonlinear least squares fitting. The experiments covered a range of speeds. Energy range A total of 25 experiments were conducted. The fitted parameters are as follows: Dimensionless exponent , .in The unit is mm, which makes The dimensions are correct. The maximum allowable width of the heat-affected zone.

[0133] The predicted burr height (in mm). An empirical model is used. This model was also obtained by fitting experimental data, using an optical microscope to measure the burr height at the cut edge. Fitting parameters: , , .Notice The dimensions are consistent with the unit of burr height. This refers to process tolerances.

[0134] This is a weighting coefficient used to balance the three objectives of thermal influence, burrs, and processing efficiency. This embodiment defaults to using... These values ​​are compromise solutions determined on the Pareto front based on orthogonal experiments. Users can adjust these values ​​via the human-machine interface according to specific product requirements, such as prioritizing edge quality or improving efficiency.

[0135] objective function The smaller the value, the better the overall effect.

[0136] 5.1.4 Constraint handling.

[0137] When the cutting path passes through a stress concentration area To prevent sudden material fracture caused by high energy input, rigid boundary constraints are applied to the process parameters in this region. Specifically, when the tool is located... When the time comes, the feasible region of the decision variables is temporarily narrowed to and Once the tool leaves the stress concentration region, it returns to the original feasible region. This constraint is implemented in the gradient descent iteration through a projection step: if the updated variable exceeds the currently allowed boundary, it is pruned to the nearest boundary value. This scheme does not add an additional penalty term to the objective function, thus avoiding logical conflicts caused by double processing.

[0138] 5.1.5 Iterative optimization calculation.

[0139] The optimal process parameters are solved using the gradient descent method. Initialization. . No. The update rules for the next iteration are as follows:

[0140] Learning rate , Partial derivatives are approximated using finite difference methods, for example... ,Pick After each update, it depends on whether the current point is located within... within, will Projecting onto the corresponding feasible region, i.e., clipping to the boundary. The iteration termination condition is: the absolute value of the change in the objective function between two consecutive iterations. Or reach the maximum number of iterations of 50.

[0141] After optimization and convergence, the optimal cutting speed is output. and optimal cutting energy Combined into adaptive process parameters .

[0142] 5.2 Generation of cutting path instructions.

[0143] 5.2.1 Obtain the preset path.

[0144] According to the zoning plan document, for the zoning Its ideal cutting path consists of a series of continuous trajectory points: Each point Including ideal blade tip position coordinates The theoretical height of the upper surface of the liner and the direction of the tool normal are given.

[0145] 5.2.2 Calculate the offset compensation amount.

[0146] Because the gasket has warped and deformed, the actual material surface position deviates from the ideal position. For each instruction point on the path... The offset compensation amount needs to be calculated based on the deformation trend data.

[0147] First, interpolation is used to analyze the deformation trend data. Obtain the displacement vector at that point Let the unit normal vector of the ideal cutting plane be... Assuming the upper surface of the pad is horizontal when it is initially undeformed, the normal offset caused by deformation is: To ensure that the tool tip always conforms to the actual deformed surface, the compensation vector is taken as... This means moving downwards or upwards by a distance equal to the warp, in the opposite direction to the warp. For in-plane offset... Generally, no compensation is needed because the cutting trajectory is defined by projection onto a plane.

[0148] More generally, if the actual deformation includes significant tilting, the compensation vector should include an in-plane component. The simplified method used in this embodiment is as follows: When more precise compensation is required, the normal component of the deformation direction vector can be projected onto the normal line.

[0149] 5.2.3 Path overlay.

[0150] Generate the compensated actual path points: .Will Arrange the cutter path instructions in the original order, adding the tool orientation to keep the tool axis perpendicular to the ideal plane, or adjust them according to the deformed local normals to form the final cutting path instructions. .

[0151] Step Six: As Figure 5 As shown, partition cutting is performed according to the cutting sequence, the adaptive process parameters, and the cutting path instructions.

[0152] 6.1 Perform the cut.

[0153] The control system will queue the cutting sequence determined in step four. The first partition in, namely Step 5 generates adaptive process parameters and cutting path instructions The command is sent to the servo drive and the cutting head power supply. The robotic arm executes the cutting according to the command. During the execution, the real-time sensing in step one continues to operate. Once a major anomaly in the material state is detected, such as a sudden increase in acoustic emission signal energy, an emergency stop or online replanning can be triggered.

[0154] 6.2 Cut quality inspection and trigger recalculation.

[0155] After a section is cut, the line laser profilometer performs a fine scan of the cut shortly after the tool is removed to obtain an image of the actual cut shape.

[0156] 6.2.1 Edge extraction.

[0157] Edge extraction is performed on the acquired 3D point cloud data. First, the curvature of each point is calculated to identify the fold positions of the cut edges. Then, the planes on both sides of the cut are fitted to calculate the deviation of the cut edge from the ideal contour. The measured edge geometric parameters include: Layering depth : Maximum depth of fibrous layer separation on the incision sidewall (mm).

[0158] Burr height : Maximum height of the protrusion at the cut edge (mm).

[0159] Straightness of the cut edge : The maximum distance (mm) from the measured edge point to the ideal straight line.

[0160] 6.2.2 Deviation Judgment.

[0161] Compare the measured parameters with the preset cut quality benchmark parameters. Perform a comparison. Calculate the incision deviation value:

[0162] like This indicates that the cutting quality of the current partition does not meet the requirements, and the system triggers a recalculation of the adaptive process parameters. At this point, the actual cutting parameters and quality data of the current partition are used as new samples, and the optimization process in step five is re-executed to update the process parameters used in subsequent partitions. Note that steps three and four are not re-executed (the cutting order remains unchanged), because the order decision has already been made based on historical states, and changing the order may introduce additional risks.

[0163] 6.2.3 Adaptive process parameter recalculation operation.

[0164] when At this time, the system triggers a recalculation operation of the adaptive process parameters. This recalculation operation, based on the actual feedback data from the current partition cutting, incrementally corrects the process parameter optimization model in step five, specifically including: (i) Construct new samples.

[0165] The actual cutting speed, actual cutting energy, average fiber orientation characteristics, and area ratio of stress concentration regions in the current partition are used as input features, and the measured layer depth, burr height, and edge straightness are used as output labels to form new samples and store them in the historical sample set.

[0166] (ii) Incremental update of model parameters for the heat-affected zone.

[0167] A recursive least squares method is used to incrementally update the undetermined coefficients in the empirical model of the heat-affected zone width. Specifically, the nonlinear relationship related to cutting speed and cutting energy in the model is transformed into a linear form through logarithmic transformation. The deviation between the model output value and the measured value is used as a feedback signal, and the model parameters are successively corrected using a recursive formula. A forgetting factor is introduced during the recursion process to control the influence of historical data on the current parameter estimation, enabling the model to track changes in processing conditions more quickly. This method does not require storing all historical data, requires only a small number of matrix operations in each iteration, has high computational efficiency, and is suitable for online applications.

[0168] (iii) Incremental update of burr height model parameters.

[0169] Using the same method as in step (ii), the undetermined coefficients in the empirical model of burr height are incrementally updated. The linear relationship between the model and the reciprocal of the cutting speed and the cutting energy is used as input, and the deviation between the measured burr height and the model prediction is used to drive parameter correction. The recursive method is the same as in step (ii).

[0170] (iv) Adaptive adjustment of the weight coefficients of the objective function.

[0171] If the cutting deviation values ​​of three consecutive partitions are all greater than 1.0, then the adaptive adjustment of the objective function weight coefficients is triggered. First, calculate the average values ​​of various quality indicators for several recently segmented partitions, and then calculate the deviation rate of each indicator's average value relative to its preset benchmark value.

[0172] Then, for each quality indicator, the relative severity of the indicator is determined using the average deviation rate of all indicators as a benchmark: if the deviation rate of an indicator is higher than the average deviation rate, its corresponding weight coefficient is increased; otherwise, it is decreased. Each weight adjustment must satisfy the constraint that its sum is 1, and each is limited to a preset reasonable range.

[0173] (v) Update parameters take effect and become fixed.

[0174] After recalculation, the system temporarily stores the updated model parameters and weight coefficients in the cache and tries them out in the next partition cutting and checks the cutting quality again: if the cutting deviation value is still greater than 1.0, then repeat steps (i) to (iv) and use the data as a new sample to continue iterative updates; if the cutting deviation value drops to below 1.0, then the updated parameters are fixed in the local configuration file for use in all subsequent processing tasks.

[0175] 6.3 Iterate to the next partition.

[0176] Pop queue Completed in Repeat steps one through six until all partitions have been cut.

[0177] III. Additional functions: data recording and remote parameter identification.

[0178] To continuously optimize the prediction accuracy of the system, this embodiment also includes a data accumulation and model update mechanism.

[0179] 7.1 Process data recording.

[0180] After each complete partition cutting task is completed, the control system collects the following data generated during the entire processing into a process data record: Time series: Fiber orientation characteristic data Microscopic damage characteristic data Cutting resistance data ; The adaptive process parameter sequence used in this mission ; Cutting path instruction sequence ; Final inspection results for incision quality data And actual geometric dimensions.

[0181] The record is stored in a local database in JSON format and uploaded to an enterprise-level data server when the network is idle.

[0182] 7.2 Remote parameter identification and correction.

[0183] 7.2.1 Desensitization treatment.

[0184] Before sending the data to a remote computing terminal, such as a cloud server, an anonymization process is performed: non-physical attribute fields such as workpiece batch number, operator information, and processing timestamp are deleted; random translation (translation range ±5mm) is applied to the cutting path coordinates to blur the specific shape, but the relative geometric relationship and stress-strain response characteristics are preserved. The anonymized process data is then obtained.

[0185] 7.2.2 Sending and Reverse Requesting.

[0186] The anonymized data is transmitted to a remote computing terminal via HTTPS (HTTPS). The remote computing terminal runs a material parameter identification program based on Bayesian inversion. This program takes the cutting resistance sequence from the anonymized data as input, and the measured deformation and notch mass as observations. It samples the posterior distribution using the Markov chain Monte Carlo method to inversely determine the material stiffness parameters that most likely explain the observed data. (Specifically including etc.). The inversion process uses finite element simulation as the forward model.

[0187] 7.2.3 Receiving and updating.

[0188] The control system receives the corrected material stiffness parameters returned by the remote computing terminal. Stored in a local configuration file, overwriting the original. In the subsequent partitioning task, the stress distribution calculation program in step two automatically loads new stiffness parameters, thereby obtaining more accurate simulation results.

[0189] The foregoing has shown and described the basic principles, main features, and advantages of this application. Those skilled in the art should understand that this application is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of this application. Various changes and modifications can be made to this application without departing from the spirit and scope thereof, and all such changes and modifications fall within the scope of this application as claimed. The scope of protection of this application is defined by the appended claims and their equivalents.

Claims

1. A method for adaptive control of zoned cutting of ceramic fiber composite gaskets, characterized in that, The method includes: During the cutting process, acquire the material state information and processing information of the current cutting area of ​​the liner; The material state information, the processing information, and the preset design data of the liner are respectively input into a preset stress distribution calculation program, and the stress distribution calculation program is run to obtain stress distribution calculation data and deformation trend data. The stress distribution calculation data and the deformation trend data are input into a preset partitioned stress evolution calculation program, and the partitioned stress evolution calculation program is run to obtain stress evolution prediction information. Based on the stress evolution prediction information, the cutting order of multiple partitions is determined; Based on the stress distribution calculation data and the deformation trend data, adaptive process parameters and cutting path instructions are calculated. The partition cutting is performed according to the cutting sequence, the adaptive process parameters, and the cutting path instructions.

2. The method according to claim 1, characterized in that, The acquisition of material state information and processing information of the current cutting area of ​​the liner includes: Acquire a surface image of the current cutting area of ​​the pad, perform texture analysis on the surface image, and obtain fiber orientation feature data; Acoustic emission signals from the current cutting area of ​​the liner are collected, and feature extraction operations are performed on the acoustic emission signals to obtain microscopic damage feature data. The reaction force signal experienced by the cutting actuator is collected, and numerical conversion calculation is performed on the reaction force signal to obtain cutting resistance data; Collect cutting head pose data and cut area topography data; The fiber orientation characteristic data, the micro-damage characteristic data, the cutting resistance data, the cutting head pose data, and the morphology data of the cut area are used as the material state information and processing information.

3. The method according to claim 2, characterized in that, The process involves inputting the material state information, the processing information, and the preset design data of the liner into a preset stress distribution calculation program, running the stress distribution calculation program, and obtaining stress distribution calculation data and deformation trend data, including: The fiber orientation characteristic data and the cutting resistance data are obtained from the material state information and processing process information; The acquired fiber orientation characteristic data and cutting resistance data, combined with preset padding material stiffness parameters and thermophysical parameters, are input into the stress distribution calculation program. Run the anisotropic material mechanics solution module in the stress distribution calculation program to calculate the stress distribution data inside the gasket under the current constraint conditions; Run the geometric nonlinear analysis module in the stress distribution calculation program to calculate the deformation trend data of the offset of each node position of the liner.

4. The method according to claim 3, characterized in that, The step of calculating adaptive process parameters and cutting path instructions based on the stress distribution calculation data and the deformation trend data includes: Extract the coordinate information of the stress concentration area from the stress distribution calculation data, and extract the deformation direction vector from the deformation trend data; Using the coordinate information of the stress concentration area and the deformation direction vector as constraints, a preset process parameter adjustment calculation program is called to iteratively optimize the cutting speed and cutting energy values ​​until the preset convergence condition is met, and the adjusted cutting speed and cutting energy values ​​are output. Adaptive process parameters are generated based on the adjusted cutting speed and cutting energy values. The offset compensation amount of the cutting tool is calculated based on the deformation direction vector, and the offset compensation amount is superimposed with the preset path to obtain the cutting path instruction.

5. The method according to claim 1, characterized in that, The step of inputting the stress distribution calculation data and the deformation trend data into a preset zoned stress evolution calculation program, running the zoned stress evolution calculation program, and obtaining stress evolution prediction information includes: Extract the actual contour data of the cut partitions from the processing information; The actual contour data is compared with the preset partitioning plan file to obtain a list of candidate partitions that have not yet been cut. Read the geometric feature data of each candidate partition one by one from the candidate partition list; By removing the mesh cells representing candidate partitions in the simulation environment, stress redistribution data for the remaining portion of the liner is calculated. The stress redistribution data corresponding to each candidate partition are summarized to form stress evolution prediction information.

6. The method according to claim 5, characterized in that, The step of determining the cutting order of multiple partitions based on the stress evolution prediction information includes: Read the stress redistribution data corresponding to each candidate partition from the stress evolution prediction information; For each candidate partition, the stress redistribution data corresponding to the candidate partition is input into a preset deformation calculation program, the deformation calculation program is run, and the cumulative deformation value of the entire pad structure after the candidate partition is removed is output. Perform a sorting and comparison operation on the cumulative deformation values ​​corresponding to each candidate partition, select the candidate partition based on the sorting and comparison results, and mark the selected candidate partition as the current cutting target; Write the identifier of the current cutting target into the cutting order queue.

7. The method according to claim 1, characterized in that, The method further includes: Collect the processing information, adaptive process parameters, cutting path instructions, and cut quality data obtained after performing partitioned cutting generated during the cutting process, and combine them into a process data record; The multiple process data records accumulated from multiple partitioning and splitting tasks are combined to generate a historical dataset.

8. The method according to claim 7, characterized in that, The method further includes: The process data records are subjected to de-identification processing to obtain de-identified process data; The desensitization process data is sent to a remote computing terminal, which performs material parameter identification calculations based on the desensitization process data and returns the corrected material stiffness parameters. Receive the corrected material stiffness parameters returned by the remote computing terminal; In subsequent partitioning and cutting tasks, the modified material stiffness parameters are called to run the stress distribution calculation program.

9. The method according to claim 1, characterized in that, After performing the partitioning, the process also includes: Acquire images of the actual cut morphology of the cut area; Edge extraction operations are performed on the actual cut morphology image to obtain measured edge geometric parameters characterizing the cut quality; The measured edge geometric parameters are compared with the preset cut quality benchmark parameters to calculate the cut deviation value. When the notch deviation value is greater than the preset deviation threshold, the recalculation operation of the adaptive process parameters is triggered.

10. A ceramic fiber composite liner partitioning and adaptive control system, characterized in that, The system includes: The information acquisition unit is used to acquire material status information and processing information of the current cutting area of ​​the liner during the cutting process; The stress distribution calculation unit is used to input the material state information, the processing information, and the preset design data of the liner into a preset stress distribution calculation program, run the stress distribution calculation program, and obtain stress distribution calculation data and deformation trend data. The partitioned stress evolution calculation unit is used to input the stress distribution calculation data and the deformation trend data into a preset partitioned stress evolution calculation program, run the partitioned stress evolution calculation program, and obtain stress evolution prediction information. The cutting sequence determination unit is used to determine the cutting sequence of multiple partitions based on the stress evolution prediction information. The process parameter and path calculation unit is used to calculate adaptive process parameters and cutting path instructions based on the stress distribution calculation data and the deformation trend data. The execution control unit is used to perform partitioned cutting according to the cutting sequence, the adaptive process parameters, and the cutting path instructions.

Images