Intelligent cutting and mold forming integrated processing system

Abstract

The application provides a kind of intelligent cutting and mould forming integrated processing system, it solves the problem such as poor machining precision of mould curved surface, its system calibration;Global multi-frequency orthogonal sinusoidal phase shift stripe projection, solve absolute phase and reconstruct dense three-dimensional point cloud, compared with theoretical model to generate correction initial path;Follow-up projection high-frequency phase shift stripe is carried out during processing local topography reconstruction, and cutting contact force is detected simultaneously, and geometric error and force deviation are weighted and fused into hybrid error, the tool center position is corrected in real time through PD control law, realize geometric-force dual closed-loop control.The application has the advantages of good self-adaptive stripe adjustment and good vibration active suppression capability, high machining precision and the like.

Description

Technical Field

[0001] This invention belongs to the field of mold processing technology, specifically relating to an intelligent integrated processing system for cutting and mold forming. Background Technology

[0002] Traditional precision machining of mold surfaces mainly relies on CNC milling followed by manual grinding and polishing, or offline coordinate measuring machines for inspection and rework. However, repeated clamping introduces positioning errors, and multiple removals of the workpiece significantly lengthen the machining cycle, making it difficult to achieve efficient online closed-loop control. Some machining systems have begun to integrate machine vision or laser displacement sensors to acquire the three-dimensional shape of the surface using structured light projection. However, existing structured light measurements mostly use single-frequency grid or parallel line projections, calculating the surface height by extracting the displacement of grid corner points or line centers. This type of method based on sparse feature points has inherent defects such as insufficient point cloud density and limited spatial resolution. It has limited ability to distinguish fine features such as tiny ripples, steep curvature abrupt changes, and chamfered edges on the mold surface. Furthermore, the fixed projection mode is difficult to adapt to different curvatures and surface reflection characteristics, easily causing local occlusion and extraction errors. As a result, the surface reconstruction accuracy is usually only at the tens of micrometer level, which cannot meet the requirements of high-precision mold machining.

[0003] Furthermore, existing machining systems generally lack the ability to directly sense cutting contact forces. Systems that rely solely on visual measurements cannot obtain the true contact state between the tool and the workpiece. They cannot quickly detect and compensate for dynamic disturbances such as elastic tool deflection, tool micro-deformation, and cutting chatter caused by uneven material hardness. Especially in optical measurement blind spots such as curved surface edges, deep grooves, and narrow slits, when visual signals are unreliable or completely ineffective, machining control loses its basis, easily leading to overcutting and surface chatter marks.

[0004] To address the aforementioned issues, this application proposes an intelligent integrated processing system for cutting and mold forming that enables in-situ synchronous online high-density full-field reconstruction of curved surface morphology and real-time sensing of cutting contact force. Summary of the Invention

[0005] The purpose of this invention is to address the above-mentioned problems by providing a rationally designed intelligent integrated processing system for cutting and mold forming that can achieve comprehensive surface perception.

[0006] Another objective of this invention is to address the aforementioned problems by providing an intelligent integrated processing control method for cutting and mold forming with geometric force perception dual closed-loop adaptive correction.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: an intelligent integrated cutting and mold forming processing system, comprising a processing platform, on which a cutting component is mounted via a drive assembly, forming the mechanical execution body of the system, responsible for workpiece fixation and multi-axis linkage cutting. Laser emitters and image acquisition modules are respectively installed around the processing platform and on the cutting component, forming a global multi-view structured light measurement subsystem. Multiple fixed-position lasers and cameras cover the inner curved surface of the mold from different angles, providing the hardware foundation for full-field 3D reconstruction. The laser emitters and image acquisition modules on the cutting component form a follow-up local fine measurement subsystem. They follow the movement of the tool, monitoring the surface morphology of a small area near the tool tip in real time. The laser emitters and image acquisition modules are connected to the data processing module to perform high-intensity algorithmic tasks such as image processing, phase calculation, 3D reconstruction, and error calculation. The data processing module is connected to the drive assembly and the cutting assembly via a main control module. This main control module, as the control core of the system, receives the analysis results from the data processing module, generates control commands, and controls the actions of the drive assembly and the cutting assembly, realizing a closed loop for path correction. The system achieves in-situ integrated integration of the measurement device and the processing device.

[0008] In the aforementioned integrated intelligent cutting and mold forming system, the processing platform includes a main table. A pneumatic clamp is mounted on the main table, providing a stable foundation for workpiece clamping. The pneumatic clamp offers fast response and uniform, controllable clamping force, effectively suppressing micro-displacement of the workpiece during processing and providing a unified benchmark for multiple in-situ measurements. The drive assembly includes first guide rails on both sides of the processing platform. The first guide rails are connected to slides via electric lead screws. A crossbeam connects the slides, and the crossbeam is connected to a translation seat via a second guide rail and an electric lead screw. A liftable cutting table is slidably mounted on the translation seat via limit guide rails and an electric lead screw. A ball end mill is connected to the lower end of the cutting table via a drive motor. The first guide rails and slides enable smooth movement of the crossbeam along the long axis (Y-axis) of the processing platform. The second guide rail on the crossbeam and the translation seat enable horizontal movement of the cutting assembly along the short axis (X-axis) of the platform. The limiting guide rail and the height-adjustable cutting table enable the vertical feed of the tool along the Z-axis. This gantry structure, combined with the electric lead screw drive, balances the large stroke and high rigidity required for machining internal curved surfaces. Precision transmission ensures the controllability and repeatability of the tool's center point spatial position.

[0009] In the aforementioned integrated intelligent cutting and mold forming system, the ball end mill includes a spherical cutting head with a cutting edge and intersecting movable grooves on its side. The cutting head is spherically shaped, and the cutting edge is responsible for material removal. The circumferential movable grooves, intersecting the cutting edge, structurally separate the sensing ring from the main body of the tool body, but do not affect the integrity of the cutting edge or chip removal. The movable grooves surround the circumference of the cutting head, and the sensing ring is movably mounted inside them. A pressure sensing element is placed between the sensing ring and the movable groove. The outer surface of the sensing ring maintains contact / near contact with the machined or unmachined surface near the tool tip, but does not participate in the main cutting. When the contact state between the tool and the workpiece changes, such as changes in allowance or sudden changes in material hardness, the normal / tangential contact force on the sensing ring will change slightly. The pressure sensing element can be a piezoelectric thin film, strain gauge, or MEMS force-sensitive element, which converts the micro-contact force transmitted from the sensing ring into a high-bandwidth electrical signal.

[0010] This ball end mill can sense local force changes at the tool-workpiece contact point, with a response speed sufficient to capture precursor signals of cutting chatter. A sensing ring surrounds the tool, giving it an approximately isotropic force response, allowing it to capture contact forces regardless of the tool's feed direction.

[0011] A method for integrated intelligent cutting and mold forming processing control includes the following steps: S1: System calibration provides accurate mathematical models and parameters for all subsequent online calculations. First, the parameters of the image acquisition module, laser emitter, ball end mill, and pressure sensing element installed on it are calibrated to obtain the camera's intrinsic and extrinsic parameters, the mapping relationship between the stripe pattern projected by the laser emitter and the tool coordinate system, and the force-voltage characteristics and normal contact stiffness coefficient of the pressure sensing element. The calibration objects include the optical sensing system and the force sensing system, as well as the spatial relationship between the two established through the tool coordinate system.

[0012] S2: Global Multi-Frequency Phase Shift Measurement and Initial Machining Path Correction, performed once before formal machining, replaces traditional trial cutting or tool setting, directly reconstructing the overall 3D surface shape of the blank or rough-machined surface. Using multi-frequency phase shift technology, a global dense point cloud is acquired to form an error map of the machining allowance. Based on this, an initial machining path that closely approximates the final shape is calculated, achieving measurement before cutting. Specifically, laser emitters around the machining platform project sinusoidal phase-shift fringes of multiple frequencies and orthogonal directions onto the curved surface inside the mold. The image acquisition module simultaneously acquires a sequence of images and calculates the absolute phase, reconstructing the global 3D point cloud. This is compared with the theoretical model to obtain the error distribution, and a corrected initial machining path for the ball end mill is generated accordingly. S3: Real-time local phase shift measurement and online dual closed-loop correction. During the cutting process, a laser emitter on the cutting assembly projects high-frequency phase shift fringes onto the cutting area, reconstructing the local three-dimensional topography of the cutting area in real time. Simultaneously, the cutting contact force is detected by the pressure sensing element on the ball end mill. Combined with geometric errors and force deviations, a hybrid error is generated to correct the machining position of the ball end mill in real time, achieving geometric-force dual closed-loop control. By fusing high-resolution geometric information with highly dynamic force information, it balances shape accuracy and physical process stability, achieving results far superior to feedback control from a single information source.

[0013] In the above-mentioned integrated intelligent cutting and mold forming processing control method, step S1 includes: S11: Calibrate the internal parameters of all image acquisition modules, including focal length, principal point, distortion coefficients, and coordinates relative to the world coordinate system of the processing platform. External parameters are used to establish a unified coordinate system for multiple cameras, enabling stereo matching of features and calculation of three-dimensional coordinates among multiple images; S12: For each laser emitter around the processing platform, calibrate its equivalent projector model, that is, treat it as an equivalent camera with a reverse optical path, and calibrate its focal length, principal point, etc. Alternatively, calibrate the phase-height mapping parameters relative to the reference plane. , The relative geometric positions of the projector and camera are determined; for the laser emitter on the cutting assembly, the hand-eye homogeneous transformation matrix between its coordinate system and the center point TCP of the ball end mill is calibrated. And the hand-eye transformation matrix between the image acquisition module on the cutting assembly and the TCP. They define the spatial pose relationship between the servo laser emitter and the camera relative to the center point of the ball end mill, respectively, and seamlessly transfer the measurement data of the servo system to the tool coordinate system, thereby directly guiding the tool movement.

[0014] S13: Calibrate the radius of the ball end mill using a reference block. And the precise pose of the TCP, the actual radius of the ball end mill will change due to wear, and needs to be accurately measured; S14: Calibrate the pressure sensing element installed on the ball end mill and determine its output voltage. and the normal force on the induction ring linear relationship between ,in For force sensitivity coefficient, The voltage is zero; and the normal contact stiffness coefficient of the tool-workpiece contact is calibrated using a standard force sensor. , The term refers to the force required to generate a normal indentation depth of 1 μm using a cutting tool, expressed in N / μm. Hand-eye calibration and force sensor stiffness calibration are the mathematical prerequisites for achieving local real-time measurement and force-position hybrid control.

[0015] In the above-mentioned integrated intelligent cutting and mold forming processing control method, step S2 includes: S21: The main control module controls the surrounding laser emitters to sequentially project at least two sets of sinusoidal stripe patterns with different frequencies and orthogonal directions onto the curved surface inside the mold. The frequencies are respectively... , , ..., The multi-frequency projection aims to resolve the ambiguity of phase wrapping in a single high-frequency fringe, simultaneously acquiring these sequential images from multiple cameras to obtain the raw data input of full-field 3D information. Each group of fringes is then processed... Phase shift, phase shift amount , And simultaneously trigger all image acquisition modules to acquire the corresponding images. Frame phase-shifted image sequence; S22: The data processing module performs a phase-shifting algorithm pixel-by-pixel on the image sequence acquired by each camera to calculate the wrap-around phase. ; In the formula: For the first The light intensity of the image acquired by phase shifting For the first Phase shift of the step, This represents the total number of phase shift steps. S23: Use multi-frequency wrapping phase to perform time phase expansion to obtain absolute phase. For frequency and The wrapping phase , The synthesized equivalent phase is: ; Based on this equivalent phase, the fringe order is determined step by step to obtain the unambiguous absolute phase at the highest frequency. The entire process is executed independently pixel by pixel, which can perfectly measure discontinuous steep surfaces; S24: Based on absolute phase With the calibrated projector model, in the world coordinate system The following uses the principle of triangulation to reconstruct the global dense 3D point cloud of the mold surface. Alternatively, the height can be obtained from the phase-height mapping relationship: ; in The phase difference relative to the reference plane, The highest frequency, , For calibration parameters; S25: Convert the global point cloud Mold Theory CAD Digital Model The best-fit alignment is performed using an iterative nearest-point algorithm, and the full-surface normal machining error distribution is calculated. For any alignment point, the error is defined as: ; In the formula: For the three-dimensional coordinates of the measurement point, Let be the coordinates of the corresponding point on the theoretical surface. For theoretical surfaces in The unit normal vector at the point is used to calculate the signed distance from the point to the surface. S26: The main control module generates a corrected initial ball end mill machining path based on the error distribution. The corrected nominal tool center position is: ; In the formula: The radius of the ball end mill. This step compensates for the deviation caused by uneven blank allowance by pre-compensating for the normal machining error at this point. This process replaces numerous trial cuts and offline inspections with a single rapid measurement, significantly reducing the interference of roughing errors on finishing and substantially improving the machining quality of the first pass.

[0016] In the above-mentioned integrated intelligent cutting and mold forming processing control method, step S3 includes: S31: When the ball end mill cuts along the initial path, the main control module controls the laser emitter on the cutting assembly to project a high-frequency sinusoidal fringe pattern onto the local area before and after the cutting point in a periodic phase-shifting manner after each predetermined distance the tool moves. The frequency is... It performs M-step rapid phase shifts, where M is 4 to 8, and synchronously triggers the image acquisition module on the cutting assembly to acquire M frames of local phase-shifted images; it only needs to measure a small area around the tool tip, achieving an extremely fast measurement cycle; S32: The data processing module calculates the high-frequency wrapping phase of the local image using the phase shift formula, performs spatial phase unfolding based on the constraints of known current tool position and continuous local surface, and obtains the local absolute phase; through the hand-eye relationship matrix... and Local 3D point clouds near the cutting edge are reconstructed in real time using triangulation in the tool coordinate system. And use this to calculate the geometric normal error: ; In the formula: To control the cycle number, For the corresponding points on the theoretical surface, Its unit normal vector; S33: The data processing module receives the voltage signal output by the pressure sensing element on the ball end mill, and calculates the current normal contact force based on the calibration relationship. And compared with the preset optimal cutting normal force target value By comparison, the force deviation is obtained: ; in, It is the optimal cutting force preset by the process, that is, the ideal value that ensures continuous cutting without excessive tool deformation or chatter. The positive or negative sign reflects whether the actual contact state is too strong or too weak.

[0017] S34: Integrate geometric errors and force deviations into a mixed error. : ; In the formula: The calibrated normal contact stiffness coefficient is expressed in N / μm. The range of values ​​for the fusion weighting coefficients is as follows: When optical conditions are good, When the value approaches 1, the system is guided by precise geometry control; when optical conditions deteriorate or force signals become abnormal, Decrease, force deviation term The weight is increased. Divided by the stiffness coefficient. Converting the force unit N to the length unit μm allows geometric quantities and forces with different physical meanings to be added or subtracted in the same dimension, at which point the system is guided by compliant force control.

[0018] S35: The main control module calculates the tool position correction amount in real time based on the mixed error. And superimposed on the current nominal knife core position. The corrected position of the cutting core is obtained. The control law used is: ; ; In the formula: This is the proportional gain coefficient. Here, is the differential gain coefficient, and is the control cycle duration. It is the unit normal vector; S36: Repeat steps S31 to S35 until the machining of the entire mold surface is completed.

[0019] In the aforementioned integrated intelligent cutting and mold forming processing control method, a weighting coefficient is incorporated. Dynamically adjust according to processing status: When the surface curvature is gentle and the optical measurement signal-to-noise ratio is higher than the preset value, take... Primarily relying on geometric feedback, the system trusts optical data and focuses on achieving micron-level contour accuracy. However, when in areas with steep curvature, optical obstruction, or where pressure signal fluctuations exceed the chatter threshold, the system reduces... The cutting depth is 0.2~0.4. It mainly relies on force feedback and aims to achieve constant small force contact to perform flexible advance and retreat, actively preventing overcutting and tool damage.

[0020] In the above-mentioned intelligent cutting and mold forming integrated processing control method, adaptive parameter adjustment is performed during step S31. During real-time local phase shift measurement, the data processing module monitors the curvature change of the local reconstructed point cloud or the modulation degree of the calculated phase in real time. When it is determined that the local curvature exceeds the preset threshold, the frequency of the local projected stripes is automatically reduced and the number of phase shift steps is increased. The wavelength of the low-frequency stripes is longer and the phase unfolding is more robust. Increasing the number of steps can suppress noise and improve the measurement success rate of the morphology change region. When the curvature returns to a smooth state, the stripe frequency is restored to the maximum value to ensure the highest measurement resolution and the densest point cloud.

[0021] In the above-mentioned intelligent cutting and mold forming integrated processing control method, the sinusoidal stripe pattern is generated by a laser emitter through MEMS micromirror scanning or digital micromirror device, and the phase shift is achieved by precisely deflecting the beam or moving the projection pattern. The fully digital structured light generation scheme is adopted, with no mechanical moving parts, so the phase shifting accuracy is high, the speed is fast, and the size is compact.

[0022] Compared with existing technologies, the advantages of this invention are: It integrates global and local optical measurement and force sensing in situ onto the processing platform, forming a closed-loop process of global initial measurement correction and local real-time online fine-tuning, eliminating offline handling and repeated clamping errors, and significantly shortening the processing cycle; it can automatically adjust the frequency and phase shift step of the projected fringes according to changes in surface curvature and signal quality, reducing the frequency in steep areas to ensure robustness and restoring high frequencies in smooth areas to ensure ultimate resolution, adapting to various complex surface characteristics; and it uses multi-frequency sinusoidal phase-shifting fringes instead of traditional square grids to achieve… Pixel-level phase calculation and full-field dense point cloud reconstruction can accurately restore fine features such as minute ripples and chamfered edges; the ball end mill structure with integrated pressure sensing element realizes high-bandwidth direct measurement of the cutting tip contact force; a weighted hybrid algorithm of geometric error and force deviation is proposed, combined with adaptive weight adjustment, to balance contour accuracy and cutting process stability; the pressure sensing element provides continuous force feedback in optically obstructed areas such as deep grooves and narrow slits, and actively compensates for elastic tool deflection caused by uneven material hardness or tool deformation through force-position conversion, thus making up for the inherent defects of pure optical measurement. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the structure of the present invention; Figure 2 This is a structural schematic diagram from another perspective of the present invention; Figure 3 This is a schematic diagram of the structure of the ball end mill of the present invention; Figure 4 This is a control structure diagram of the present invention; Figure 5 This is a flowchart of the processing control method of the present invention; Figure 6 This is a schematic diagram of the global multi-frequency phase shift measurement and initial processing path correction of the present invention; Figure 7 This is a schematic diagram of the dual closed-loop real-time correction control principle of the present invention; In the figure, the machining platform 1, main table 11, pneumatic fixture 12, drive assembly 2, first guide rail 21, slide table 22, crossbeam 23, second guide rail 24, translation seat 25, limit guide rail 26, cutting table 27, ball end mill 28, cutting assembly 3, cutting head 31, cutting edge 32, movable groove 33, sensing ring 34, laser emitter 4, image acquisition module 5, data processing module 6, and main control module 7 are shown. Detailed Implementation

[0024] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. Example 1

[0025] like Figure 1-7 As shown, the processing platform 1 is powered on, and the data processing module 6 and the main control module 7 complete self-tests. The pneumatic clamp 12 fixes the mold to be processed onto the main table 11. The system retrieves pre-calibrated parameters: the intrinsic parameters and extrinsic parameters relative to the world coordinate system of each image acquisition module 5, and the equivalent projector model parameters of the surrounding laser emitters 4. , The laser emitter 4 on the cutting assembly 3 and the hand-eye connection between the center point TCP of the ball end mill 28 tool and the cutting assembly 3. The image acquisition module 5 on the cutting assembly 3 and the hand-eye matrix of TCP The radius R of the ball end mill and the force sensitivity coefficient of the pressure sensing element. Contact stiffness coefficient with normal direction .

[0026] The main control module 7 instructs the laser emitters 4 around the processing platform 1 to sequentially project two sets of frequencies onto the curved surface inside the mold. =1, =8) and a sinusoidal stripe pattern with orthogonal directions (horizontal and vertical), each group of stripes undergoes 4-step phase shift (N=4, phase shift amount =0, π / 2, π, 3π / 2). All image acquisition modules 5 are synchronously triggered, acquiring a total of 2×2×4=16 frames of phase-shifted image sequences, and the image data is transmitted to the data processing module 6 in real time.

[0027] Data processing module 6 performs a phase-shifting algorithm on each pixel of each camera: ; After obtaining the wrapping phase at each frequency, the wrapping is unwrapped using multi-frequency heterodyne: ; By determining the fringe order step by step, the unambiguous absolute phase of the entire surface can be obtained. Using the phase-height mapping formula: ; Convert the absolute phase of each pixel into a 3D point cloud. A global dense surface model is formed by fusing point clouds from multiple cameras.

[0028] Data processing module 6 will With mold CAD theoretical digital model Alignment is achieved through an iterative nearest-point algorithm, and the normal machining error is calculated point by point. ; The main control module 7 generates a corrected initial machining path based on the error distribution, and corrects the tool center position as follows: ; This path has been pre-compensated for blank allowance deviation and stored as CNC machining code in main control module 7. Example 2

[0029] During the cutting process in this embodiment, the follow-up laser emitter 4 performs local fine measurements and adaptively adjusts the stripe parameters according to the curvature change.

[0030] The main control module 7 controls the operation of the drive component 2 according to the initial path generated in Embodiment 1: the first guide rail 21 drives the slide table 22 via an electric lead screw, which in turn moves the crossbeam 23 along the Y-axis; the second guide rail 24 on the crossbeam 23 drives the translation seat 25 along the X-axis via an electric lead screw; the limiting guide rail 26 on the translation seat 25 drives the cutting table 27 to rise and fall along the Z-axis via an electric lead screw. The drive motor at the lower end of the cutting table 27 drives the ball end mill 28 to rotate, and the cutting edge 32 on the cutting head 31 removes material from the mold surface. A local measurement is triggered every time the tool moves 0.5mm.

[0031] The laser emitter 4 on the cutting assembly 3 projects high frequency signals into local areas before and after the cutting point. =64) Sine fringes, performing 4-step rapid phase shifting. The image acquisition module 5 on the cutting assembly 3 synchronously acquires 4 frames of local phase-shifted images.

[0032] Data processing module 6 calculates the curvature of the locally reconstructed point cloud in real time. When the system detects that the curvature of the current processing area exceeds a preset threshold (abrupt surface transition), it automatically reduces the projection frequency of laser emitter 4 from 64 to 64. =16, and the phase shifting steps were increased from 4 to 8 to ensure the reliability of phase unpacking; when the tool enters the region of gentle curvature, the system restores the highest frequency. =64 and 4-step phase shifting to maintain ultimate resolution. This adaptive adjustment is achieved by the data processing module 6 through the main control module 7 to configure the laser emitter 4 in real time.

[0033] After calculating the high-frequency wrapping phase, data processing module 6 uses the constraint of the current tool position to complete the spatial phase unfolding. This is achieved using a calibrated hand-eye matrix. , Reconstructing local point clouds in tool coordinate system And calculate the geometric normal error: . Example 3

[0034] In this embodiment, the pressure sensing element inside the ball end mill 28 works in conjunction with optical measurement to achieve mixed error calculation and dual closed-loop position correction, and actively suppress cutting chatter.

[0035] When the cutting head 31 of the ball end mill 28 performs cutting, the sensing ring 34, which is movably mounted in the movable groove 33, remains in contact with the surface being machined. Changes in contact force caused by variations in material allowance or uneven hardness are transmitted through the sensing ring 34 to the pressure sensing element (in this embodiment, a miniature piezoelectric film) between the movable groove 33 and the sensing ring 34. The pressure sensing element outputs a voltage signal. The data is then transmitted to data processing module 6.

[0036] Data processing module 6 uses calibration parameters to convert the voltage signal into a normal force: ; Compared with the preset optimal cutting normal force target value By comparison, the force deviation is obtained: ; when >0 indicates excessive contact force, which may cause overcutting or tool deformation; when <0 indicates insufficient contact force, which may indicate elasticity causing tool deflection.

[0037] Data processing module 6 will process the geometric error obtained in Example 2. Combined with force deviation, it forms a mixed error: ; Fusion weights Dynamic adjustment based on processing status: When the current processing area is a gently curved surface and the optical signal-to-noise ratio is high, take... =0.85, with geometric feedback as the primary mechanism; when the system detects that the tool has entered a deep groove region and the optical signal is blocked and unreliable, it automatically... When the force feedback drops to 0.3, it becomes dominant; when the pressure sensing element detects that the high-frequency component of the force signal exceeds the flutter threshold, i.e., |dF n When / dt| is too large, When the value drops to 0.2, the main control module 7 prioritizes the response to force feedback.

[0038] The main control module 7 calculates the tool position correction amount based on the mixed error: ; The revised tool center position command is as follows: ; Among them, the proportion term The term responds immediately to the current error; the differential term The term provides active damping. When flutter precursors are detected, the differential term provides inverse suppression of rapid changes in the force signal, combined with weights. Tilting in the direction of force control causes the ball end cutter 28 to retract slightly to attenuate vibration and achieve active suppression of chatter.

[0039] The main control module 7 will send the corrected position command. The servo motors of each axis sent to the drive component 2 achieve real-time fine-tuning of the ball end mill 28 position through the linkage of the first guide rail 21, the second guide rail 24, and the limit guide rail 26. The local measurement of Embodiment 2 and the force detection, mixed error calculation, and position correction of this embodiment are repeated until the entire mold surface is machined. After machining, the global measurement steps of Embodiment 1 can be called again for final accuracy verification.

[0040] The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific embodiments or use similar methods to substitute them, without departing from the spirit of the invention or exceeding the scope defined by the appended claims.

[0041] Although this document frequently uses terms such as machining platform 1, main table 11, pneumatic fixture 12, drive assembly 2, first guide rail 21, slide table 22, crossbeam 23, second guide rail 24, translation seat 25, limit guide rail 26, cutting table 27, ball end mill 28, cutting assembly 3, cutting head 31, cutting edge 32, movable groove 33, sensing ring 34, laser emitter 4, image acquisition module 5, data processing module 6, and main control module 7, the possibility of using other terms is not excluded. The use of these terms is merely for the convenience of describing and explaining the essence of this invention; interpreting them as any additional limitation would contradict the spirit of this invention.

Claims

1. An intelligent integrated cutting and mold forming processing system, comprising a processing platform (1), wherein a cutting component (3) is mounted on the processing platform (1) via a drive component (2), characterized in that, A laser emitter (4) and an image acquisition module (5) are respectively installed around the processing platform (1) and the cutting component (3). The laser emitter (4) and the image acquisition module (5) are connected to the data processing module (6). The data processing module (6) is connected to the drive component (2) and the cutting component (3) through the main control module (7).

2. The intelligent cutting and mold forming integrated processing system according to claim 1, characterized in that, The processing platform (1) includes a main table (11), on which a pneumatic clamp (12) is installed; the drive assembly (2) includes a first guide rail (21) on both sides of the processing platform (1), the first guide rail (21) is connected to a slide table (22) by an electric screw drive, a crossbeam (23) is connected between the slide tables (22), the crossbeam (23) is connected to a translation seat (25) by a second guide rail (24) and an electric screw drive, the translation seat (25) is slidably mounted with a liftable cutting table (27) by a limiting guide rail (26) and an electric screw drive, and a ball end cutter (28) is connected to the lower end of the cutting table (27) by a drive motor drive.

3. The intelligent cutting and mold forming integrated processing system according to claim 2, characterized in that, The ball end mill (28) includes a spherical cutting head (31), the cutting head (31) has a cutting edge (32) on its side and a movable groove (33) that intersects with the cutting edge (32). The movable groove (33) surrounds the cutting head (31) circumferentially and a sensing ring (34) is movably installed inside it. A pressure sensing element is provided between the sensing ring (34) and the movable groove (33).

4. A method for integrated intelligent cutting and mold forming processing control, used in the integrated intelligent cutting and mold forming processing system described in any one of claims 1-3, characterized in that, Includes the following steps: S1: System calibration, parameter calibration of image acquisition module (5), laser emitter (4), ball end mill (28) and pressure sensing element set on it, to obtain camera internal and external parameters, mapping relationship between stripe pattern projected by laser emitter (4) and tool coordinate system, and force-voltage characteristics and normal contact stiffness coefficient of pressure sensing element; S2: Global multi-frequency phase shift measurement and initial machining path correction. Using the laser emitters (4) around the machining platform (1), sinusoidal phase shift stripes of multiple frequencies and orthogonal directions are projected onto the inner curved surface of the mold. The image acquisition module (5) synchronously acquires the sequence images and calculates the absolute phase, reconstructs the global three-dimensional point cloud, compares it with the theoretical model to obtain the error distribution, and generates the corrected initial machining path of the ball end mill (28). S3: Real-time local phase shift measurement and online dual closed-loop correction. During the cutting process, the laser emitter (4) on the cutting assembly (3) projects high-frequency phase shift stripes onto the cutting area to reconstruct the local three-dimensional shape of the cutting area in real time. At the same time, the cutting contact force is detected by the pressure sensing element on the ball end mill (28). Combined with geometric error and force deviation, a mixed error is generated to correct the machining position of the ball end mill (28) in real time, thereby realizing geometric-force dual closed-loop control.

5. The intelligent cutting and mold forming integrated processing control method according to claim 4, characterized in that, Step S1 includes: S11: Calibrate the internal parameters of all image acquisition modules (5), including focal length, principal point, distortion coefficient, and world coordinate system relative to the processing platform (1). External parameters are used to establish a unified coordinate system for multiple cameras; S12: For each laser emitter (4) around the processing platform (1), calibrate its equivalent projector model, or calibrate the phase-height mapping parameters relative to the reference plane. , For the laser emitter (4) on the cutting assembly (3), calibrate the hand-eye homogeneous transformation matrix between its coordinate system and the center point TCP of the ball end mill (28). and the hand-eye transformation matrix between the image acquisition module (5) on the cutting assembly (3) and the TCP. ; S13: Calibrate the radius of the ball end mill (28) using a reference block. and the precise pose of TCP; S14: Calibrate the pressure sensing element installed on the ball end mill (28) and determine its output voltage. The normal force on the induction ring (34) linear relationship between ,in For force sensitivity coefficient, The voltage is zero; and the normal contact stiffness coefficient of the tool-workpiece contact is calibrated using a standard force sensor. The unit is N / μm.

6. The integrated intelligent cutting and mold forming processing control method according to claim 4, characterized in that, Step S2 includes: S21: The main control module (7) controls the surrounding laser emitters (4) to sequentially project at least two sets of sinusoidal stripe patterns with different frequencies and orthogonal directions onto the inner curved surface of the mold. The frequencies are respectively , , ..., Perform each set of stripes Phase shift, phase shift amount , And simultaneously trigger all image acquisition modules (5) to acquire the corresponding images. Frame phase-shifted image sequence; S22: The data processing module (6) performs a phase-shifting algorithm pixel-by-pixel on each image sequence acquired by the camera to calculate the wrap phase. : ； In the formula: For the first The light intensity of the image acquired by phase shifting For the first Phase shift of the step, This represents the total number of phase shift steps. S23: Use multi-frequency wrapping phase to perform time phase expansion to obtain absolute phase. For frequency and The wrapping phase , The synthesized equivalent phase is: ； Based on this equivalent phase, the fringe order is determined step by step to obtain the unambiguous absolute phase at the highest frequency. ; S24: Based on absolute phase With the calibrated projector model, in the world coordinate system The following uses the principle of triangulation to reconstruct the global dense 3D point cloud of the mold surface. Alternatively, the height can be obtained from the phase-height mapping relationship: ； in The phase difference relative to the reference plane, The highest frequency, , These are calibration parameters; S25: Convert the global point cloud Mold Theory CAD Digital Model The best-fit alignment is performed using an iterative nearest-point algorithm, and the full-surface normal machining error distribution is calculated. For any alignment point, the error is defined as: ； In the formula: For the three-dimensional coordinates of the measurement point, Let be the coordinates of the corresponding point on the theoretical surface. For theoretical surfaces in The unit normal vector at that location; S26: The main control module (7) generates the corrected initial ball end mill (28) machining path based on the error distribution. The corrected nominal tool center position is: ； In the formula: The radius of the ball-end cutter (28) is... This represents the machining error in the normal direction at that point.

7. The intelligent cutting and mold forming integrated processing control method according to claim 4, characterized in that, Step S3 includes: S31: When the ball end mill (28) cuts along the initial path, the main control module (7) controls the laser emitter (4) on the cutting assembly (3) to project a high-frequency sinusoidal stripe pattern onto the local area before and after the cutting point in a periodic phase-shifting manner after each predetermined distance the tool moves. The frequency is... And perform M-step rapid phase shift, where M is 4 to 8, and synchronously trigger the image acquisition module (5) on the cutting component (3) to acquire M frames of local phase shift images; S32: The data processing module (6) calculates the high-frequency wrapping phase of the local image using the phase shift formula, performs spatial phase expansion based on the constraints of known current tool position and continuous local surface, and obtains the local absolute phase; through the hand-eye relationship matrix and In the tool coordinate system, the local three-dimensional point cloud near the cutting edge (32) is reconstructed in real time by triangulation. And use this to calculate the geometric normal error: ； In the formula: To control the cycle number, For the corresponding points on the theoretical surface, Its unit normal vector; S33: The data processing module (6) receives the voltage signal output by the pressure sensing element on the ball end mill (28) and calculates the current normal contact force based on the calibration relationship. And compared with the preset optimal cutting normal force target value By comparison, the force deviation is obtained: ； S34: Integrate geometric errors and force deviations into a mixed error. : ； In the formula: The calibrated normal contact stiffness coefficient is expressed in N / μm. The range of values ​​for the fusion weighting coefficients is as follows: ; S35: The main control module (7) calculates the tool position correction amount in real time based on the mixed error. And superimposed on the current nominal knife core position. The corrected position of the cutting core is obtained. The control law used is: ； ； In the formula: This is the proportional gain coefficient. The differential gain coefficient, To control the cycle duration, It is the unit normal vector; S36: Repeat steps S31 to S35 until the machining of the entire mold surface is completed.

8. The intelligent cutting and mold forming integrated processing control method according to claim 7, characterized in that, The fusion weight coefficient Dynamically adjust according to processing status: When the surface curvature is gentle and the optical measurement signal-to-noise ratio is higher than the preset value, take... It primarily uses geometric feedback; when in areas with steep curvature, optical obstruction, or when pressure signal fluctuations exceed the flutter threshold, it reduces... The range is from 0.2 to 0.4, with force feedback being the primary method.

9. The intelligent cutting and mold forming integrated processing control method according to claim 7, characterized in that, During the execution of step S31, adaptive parameter adjustment is performed. During the real-time local phase shift measurement, the data processing module (6) monitors the curvature change of the local reconstructed point cloud or the modulation degree of the calculated phase in real time. When it is determined that the local curvature exceeds the preset threshold, the frequency of the local projected stripes is automatically reduced and the number of phase shift steps is increased. When the curvature returns to a smooth state, the stripe frequency is restored to the maximum value.

10. The intelligent cutting and mold forming integrated processing control method according to claim 4, characterized in that, The sinusoidal stripe pattern is generated by the laser emitter (4) through MEMS micromirror scanning or digital micromirror devices, and the phase shift is achieved by precisely deflecting the beam or moving the projected pattern.

Images