T-slot vibration suppression and shape control method based on force-position sensing

Abstract

This invention discloses a method for vibration suppression and shape control of T-slots based on force-position sensing, relating to the field of workpiece machining technology. The vibration suppression and shape control method includes: acquiring actual cutting force and vibration displacement signals during T-slot machining using a sensor array; constructing a digital twin model and refining the local mesh in its key areas; dynamically reconstructing the real dynamic displacement of the tool and workpiece using a joint state observer; superimposing the tool and workpiece displacements to obtain a comprehensive vibration, generating a compensation control signal; simultaneously, predicting machining deformation errors based on the simulation twin model, generating a dynamic compensation path based on the comprehensive vibration, and decomposing it into path compensation components and vibration compensation components, which are then fused with the basic path and compensation control signal respectively to achieve synergistic optimization of vibration suppression and geometric error compensation. This invention achieves unified control of high-frequency vibration and geometric deformation during T-slot machining, improving machining accuracy and surface quality.

Description

Technical Field

[0001] This invention relates to the field of workpiece processing technology, specifically to a method for suppressing and controlling the vibration of a T-groove based on force and position sensing. Background Technology

[0002] As a critical positioning, connection, or assembly reference surface, T-slots have stringent requirements for machining accuracy (including slot width, slot depth, sidewall perpendicularity, and surface roughness) and geometric integrity. When milling T-slots with high aspect ratios and distinctive sidewall features, the machining equipment is prone to harmful vibrations due to the excitation of dynamic cutting forces. These vibrations can affect the internal structural state of the T-slots, thus impacting their usability. Therefore, it is necessary to reduce these harmful vibrations.

[0003] For example, patent publication number "CN111694320A", entitled "An Adaptive Control Method for Machine Tools Considering Chatter Suppression", monitors the spindle motor power signal and spindle vibration signal in real time during CNC machining, and adaptively adjusts the feed rate and spindle speed in real time based on the spindle power signal and vibration signal. The entire signal acquisition process does not affect normal machining. The adaptive control method for machine tools considering chatter suppression can effectively improve machining efficiency, monitor and suppress chatter generated during machining, maximize machining quality while improving machining efficiency, extend the service life of tools and machine tools, and reduce production costs.

[0004] The aforementioned patents rely on the analysis of spindle motor power and overall vibration signals, which are essentially macroscopic and indirect monitoring methods. This makes it difficult to directly and accurately obtain the dynamic cutting force at the tool-workpiece contact point, and also makes it impossible to accurately detect key state quantities such as the micro-deformation displacement of the workpiece. The above methods are not sensitive enough to the local vibration and deformation modes unique to T-slot machining, and the suppression measures taken are often not targeted. During T-slot machining, the cutting force can easily cause geometric errors such as sidewall deflection and right-angle rounding. Simply suppressing the overall vibration cannot guarantee the dimensional accuracy of the slot width and the perpendicularity of the sidewall. Therefore, a T-slot vibration suppression and shape control method based on force and position sensing was invented. Summary of the Invention

[0005] The purpose of this invention is to provide a method for suppressing and controlling the vibration of a T-groove based on force-position sensing, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a T-groove vibration suppression and shape control method based on force-position sensing, wherein the vibration suppression and shape control method includes:

[0007] S01, Information Acquisition: By installing a sensor group on the processing equipment to collect processing data in real time during the T-groove processing, the processing equipment acquires and performs pre-processing based on the processing parameters to obtain pre-processing data;

[0008] S02. Construct and run the digital twin model: Construct a digital twin model based on the three-dimensional geometric model of the T-slot workpiece and tool properties, and refine the local mesh of the T-slot area, side wall area and right angle area in the digital twin model;

[0009] S03, Ideal-Actual State Comparison and Interference Quantification: Input the processing parameters into the digital twin model, obtain the ideal processing data of the digital twin model, and calculate the difference data between the ideal processing data and the pre-processing data;

[0010] S04. Dynamic State Reconstruction: Establish and run a joint state observer. The joint state observer corrects the digital twin model through difference data to obtain a simulation twin model. Based on the pre-processing data, the workpiece-side state variables during the processing are calculated. The workpiece-side state variables include the actual dynamic displacement of the tool and the actual dynamic displacement of the workpiece.

[0011] S05. Vibration Suppression: The actual dynamic displacement of the workpiece and the actual dynamic displacement of the tool are superimposed in the cutting area to obtain the comprehensive vibration at the cutting area. A compensation control signal is generated by the control strategy to suppress the comprehensive vibration by acting on the actuator.

[0012] S06. Cooperative optimization of vibration suppression and shape control: Obtain the basic path, simulate the machining deformation error using a twin model, and establish a dynamic compensation path based on the machining deformation error and the overall vibration.

[0013] The dynamic compensation path is decomposed into path compensation component and vibration compensation component by decomposition method.

[0014] The path compensation component is integrated with the basic path to generate a dynamic application path;

[0015] The vibration compensation component is fused with the compensation control signal used to suppress vibration to generate a vibration suppression command.

[0016] Furthermore, the machining data includes the actual cutting force signal and the vibration displacement signal of the machining equipment;

[0017] The local mesh encryption includes a partition encryption strategy, which includes: establishing encryption squares with the inner right-angle vertex where the sidewall of the T-shaped groove intersects the bottom of the groove and the endpoints of the two sides of the opening of the T-shaped groove as base points; connecting two adjacent encryption square units through an adjacent encryption rectangle; performing mesh encryption on the interior of the encryption square and the adjacent encryption rectangle respectively; the mesh density inside the encryption square is greater than the mesh density inside the adjacent encryption rectangle; the side length of the encryption square is 0.5-1.2 times the depth of the bottom of the T-shaped groove; the chamfer inside the T-shaped groove is regarded as the inner right-angle vertex where the sidewall intersects the bottom of the groove; performing mesh encryption on the interior region of the T-shaped groove; and the mesh density inside the adjacent encryption rectangle is greater than the mesh density inside the T-shaped groove.

[0018] Furthermore, the joint state observer adopts a dual-time-scale estimation architecture, which includes a high-frequency estimation loop for estimating the actual dynamic cutting force and the actual dynamic displacement, and a low-frequency estimation loop for estimating the perturbation components of the machining process.

[0019] The high-frequency estimation loop and the low-frequency estimation loop are connected by a disturbance compensation channel. The machining disturbance component output by the low-frequency estimation loop is fed back to the high-frequency estimation loop through the disturbance compensation channel, and the estimated values ​​of the real dynamic cutting force and the real dynamic displacement are corrected.

[0020] Furthermore, the control strategy includes: establishing a workpiece-tool relative motion coordinate system, treating the workpiece as stationary, calculating in real time the vibration component that the tool needs to counteract the comprehensive vibration through comprehensive vibration, and calculating in real time the compensation control signal input to the driver and acting on the spindle through the vibration component and the simulation twin model.

[0021] Furthermore, methods for establishing dynamic compensation paths include:

[0022] Based on the simulation twin model, the geometric errors in the densified square and adjacent densified rectangle during the T-slot processing are predicted. The geometric errors include sidewall deformation, right-angle transition zone fillet error and groove bottom flatness deviation.

[0023] Geometric errors and combined vibrations are mapped into a compensation vector for the tool in three-dimensional space;

[0024] By combining a compensation strategy with a tool dynamics model, a material removal model, and a compensation vector, a dynamic compensation path that changes in real time with the machining position is generated.

[0025] The compensation strategy includes: calculating the compensation feed rate of the tool in real time based on the tool dynamics model, the material removal model and the compensation vector, and generating a dynamic compensation path that changes in real time with the machining position based on the compensation feed rate.

[0026] Furthermore, the decomposition method includes: the dynamic compensation path includes several control commands, which are divided by the update frequency of the control commands, a frequency threshold is established, and the control commands in the dynamic compensation path are divided into path compensation components and vibration compensation components according to the update frequency: control commands with an update frequency lower than or equal to the frequency threshold are divided into path compensation components for correcting the tool feed trajectory; control commands with an update frequency higher than the frequency threshold are divided into vibration compensation components for real-time compensation of vibration effects.

[0027] Furthermore, the dynamic compensation path includes an iterative optimization mechanism, which includes:

[0028] After a single processing is completed, the actual geometric dimensions of the T-slot are measured and compared with the simulated geometric dimensions generated by the simulation twin model to obtain shape deviation data;

[0029] The shape deviation data is fed back to the simulation twin model, and the parameters inside the simulation twin model are adaptively corrected to obtain a corrected simulation twin model that is applied to the next processing.

[0030] Compared with the prior art, the beneficial effects of the present invention are:

[0031] The force-position sensing-based T-groove vibration suppression and shape control method integrates vibration control and geometric error compensation into a unified framework. The simulation twin model can not only simulate the vibration state but also predict the geometric deformation errors caused by machining (such as sidewall bending and right-angle fillet errors) with high fidelity. The predicted deformation errors are fused with real-time comprehensive vibration information and mapped into a dynamic compensation path in three-dimensional space. Through a decomposition method, this path is intelligently decomposed into a low-frequency path compensation component (used to correct the tool feed trajectory to compensate for deformation) and a high-frequency vibration compensation component (used to cancel vibration in real time). The path compensation component is fused with the basic path to generate a dynamic application path, while the vibration compensation component is fused with the aforementioned compensation control signal. This achieves vibration suppression during tool movement while ensuring the geometric accuracy of the T-groove.

[0032] Meanwhile, the joint state observer dynamically reconstructs the actual dynamic displacement of the tool and the actual dynamic displacement of the workpiece by using the difference data between the actual force and position signals collected and the ideal output of the model. This allows for the accurate calculation of the comprehensive vibration at the cutting point. The control strategy generates a compensation control signal in real time based on this comprehensive vibration, driving the actuator to produce an anti-phase canceling motion. This achieves a leap from "macro-machine tool vibration monitoring" to "precise cancellation of micro-vibrations at the tool-workpiece contact point". In particular, it can effectively suppress regenerative chatter and forced vibration that are easily caused by T-slot sidewalls and right-angle structures, significantly reduce the surface roughness of the machined surface, and improve the stability of the machining process.

[0033] By installing a multimodal sensor array for information acquisition and combining it with pre-processing data reflecting the dynamics of the actual system, a foundation for accurate perception is laid. The joint state observer adopts a unique dual-timescale estimation architecture, with the high-frequency loop and the low-frequency loop connected in a closed loop through a disturbance compensation channel. This can quickly estimate dynamic cutting forces and displacements to suppress vibrations, while simultaneously identifying slow-changing disturbances such as tool wear and material inhomogeneity. This enables accurate observation of the machining state at multiple scales and in all directions. The system also incorporates an iterative optimization mechanism. After a single machining operation, the actual geometric dimensions are measured and compared with the simulation results to obtain shape deviation data and provide feedback to correct the parameters of the simulation twin model, thereby ensuring the practicality of the simulation twin model. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the simulated twin model of the present invention;

[0035] Figure 2 This is a schematic diagram of the compensation control signal of the present invention;

[0036] Figure 3 This is a schematic diagram illustrating the synergistic optimization of vibration suppression and shape control in this invention;

[0037] Figure 4 This is a schematic diagram of the vibration suppression and shape control method of the present invention;

[0038] Figure 5 This is a schematic diagram of the mesh refinement in the T-groove region of the present invention;

[0039] Figure 6 This is a schematic diagram illustrating the processing of the simulated twin model of the present invention;

[0040] Figure 7 This is a schematic diagram of the vibration suppression method of the present invention. Detailed Implementation

[0041] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0042] like Figure 1 - Figure 7 As shown, the present invention provides a technical solution: a T-groove vibration suppression and shape control method based on force-position sensing, the vibration suppression and shape control method comprising:

[0043] S01, Information Acquisition: By installing a sensor group on the processing equipment to collect processing data in real time during the T-groove processing, the processing equipment acquires and performs pre-processing based on the processing parameters to obtain pre-processing data;

[0044] S02. Construct and run the digital twin model: Construct a digital twin model based on the three-dimensional geometric model of the T-slot workpiece and tool properties, and refine the local mesh of the T-slot area, side wall area and right angle area in the digital twin model;

[0045] S03, Ideal-Actual State Comparison and Interference Quantification: Input the processing parameters into the digital twin model, obtain the ideal processing data of the digital twin model, and calculate the difference data between the ideal processing data and the pre-processing data;

[0046] S04. Dynamic State Reconstruction: Establish and run the joint state observer. The joint state observer corrects the digital twin model through the difference data to obtain the simulation twin model. Based on the pre-processing data, the workpiece-side state variables during the processing are calculated. The workpiece-side state variables include the actual dynamic displacement of the tool and the actual dynamic displacement of the workpiece.

[0047] S05. Vibration Suppression: The actual dynamic displacement of the workpiece and the actual dynamic displacement of the tool are superimposed in the cutting area to obtain the comprehensive vibration at the cutting area. A compensation control signal is generated by the control strategy to suppress the comprehensive vibration by acting on the actuator.

[0048] S06. Cooperative optimization of vibration suppression and shape control: Obtain the basic path, simulate the machining deformation error using a twin model, and establish a dynamic compensation path based on the machining deformation error and the overall vibration.

[0049] The dynamic compensation path is decomposed into path compensation component and vibration compensation component by decomposition method.

[0050] The path compensation component is integrated with the basic path to generate a dynamic application path;

[0051] The vibration compensation component is fused with the compensation control signal used to suppress vibration to generate a vibration suppression command.

[0052] The machining data includes the actual cutting force signal and the vibration displacement signal of the machining equipment;

[0053] Local mesh refinement includes a partitioned refinement strategy. The partitioned refinement strategy includes: establishing a refinement square with the inner right-angle vertex where the sidewall of the T-shaped groove intersects the bottom of the groove and the endpoints of the two sides of the opening of the T-shaped groove as base points. Two adjacent refinement squares are connected by an adjacent refinement rectangle. Mesh refinement is performed on the interior of the refinement square and the adjacent refinement rectangle respectively. The mesh density inside the refinement square is greater than the mesh density inside the adjacent refinement rectangle. The side length of the refinement square is 0.5-1.2 times the depth of the bottom of the T-shaped groove. The chamfer inside the T-shaped groove is regarded as the inner right-angle vertex where the sidewall intersects the bottom of the groove. Mesh refinement is performed on the interior region of the T-shaped groove. The mesh density inside the adjacent refinement rectangle is greater than the mesh density inside the T-shaped groove.

[0054] The joint state observer adopts a dual-time-scale estimation architecture, which includes a high-frequency estimation loop for estimating the actual dynamic cutting force and the actual dynamic displacement, and a low-frequency estimation loop for estimating the disturbance components of the machining process.

[0055] The high-frequency estimation loop and the low-frequency estimation loop are connected through a disturbance compensation channel. The machining disturbance component output by the low-frequency estimation loop is fed back to the high-frequency estimation loop through the disturbance compensation channel, and the estimated values ​​of the actual dynamic cutting force and the actual dynamic displacement are corrected.

[0056] The control strategy includes: establishing a workpiece-tool relative motion coordinate system, treating the workpiece as stationary, calculating in real time the vibration component that the tool needs to counteract the overall vibration through comprehensive vibration, and calculating in real time the compensation control signal input to the driver and acting on the spindle through the vibration component and the simulation twin model.

[0057] Methods for establishing dynamic compensation paths include:

[0058] Based on the simulation twin model, the geometric errors in the densified square and adjacent densified rectangle during the T-slot processing are predicted. The geometric errors include sidewall deformation, fillet error in the right-angle transition zone, and groove bottom flatness deviation.

[0059] Geometric errors and combined vibrations are mapped into a compensation vector for the tool in three-dimensional space;

[0060] By combining a compensation strategy with a tool dynamics model, a material removal model, and a compensation vector, a dynamic compensation path that changes in real time with the machining position is generated.

[0061] The compensation strategy includes: calculating the compensation feed rate of the tool in real time based on the tool dynamics model, material removal model and compensation vector, and generating a dynamic compensation path that changes in real time with the machining position based on the compensation feed rate.

[0062] The decomposition method includes: the dynamic compensation path consists of several control commands, which are divided according to the update frequency of the control commands. A frequency threshold is established, and the control commands in the dynamic compensation path are divided into path compensation components and vibration compensation components according to the update frequency: control commands with an update frequency lower than or equal to the frequency threshold are divided into path compensation components used to correct the tool feed trajectory; control commands with an update frequency higher than the frequency threshold are divided into vibration compensation components used to offset the vibration effect in real time.

[0063] The dynamic compensation path includes an iterative optimization mechanism, which includes:

[0064] After a single processing is completed, the actual geometric dimensions of the T-slot are measured and compared with the simulated geometric dimensions generated by the simulation twin model to obtain shape deviation data;

[0065] The shape deviation data is fed back to the simulation twin model, and the parameters inside the simulation twin model are adaptively corrected to obtain a corrected simulation twin model that is applied to the next processing.

[0066] Workflow: Pre-processing data is collected through a sensor array to construct a digital twin model, and the key areas are meshed. The actual processing data is compared with the ideal processing data to quantify the differences. The model is then corrected using a joint state observer to obtain a high-precision simulation twin model and dynamic displacement state. Based on this model, two core operations are executed in parallel: First, the dynamic displacements of the tool and the workpiece are superimposed to generate a real-time compensation control signal to suppress comprehensive vibration. Second, a dynamic compensation path is generated by combining the predicted processing deformation and vibration. The dynamic compensation path is decomposed into low-frequency shape compensation components and high-frequency vibration compensation components, which are fused with the basic path and vibration suppression signal to form a collaborative control command. After processing, the model parameters are iteratively optimized using measured data to form a closed loop.

[0067] Vibration, in its physical essence, can be viewed as a short-period, small-amplitude relative motion between the tool and the workpiece. Therefore, it can be decomposed into the instantaneous superposition of displacements in various directions within a spatial coordinate system. Based on this, displacement and vibration are conceptually similar and can both be regulated through precise control of motion commands. In the decomposition method, the classification is based on the update frequency of the control commands: for displacement adjustment, multiple high-frequency corrections within a very short timeframe essentially constitute active vibration suppression; these high-frequency commands are received and executed by the driver to achieve real-time vibration cancellation. The path compensation component, on the other hand, is mainly used to guide the spindle or machine tool axes in macroscopic, continuous trajectory adjustment. Its command update frequency is relatively low. Due to the different functions and dynamic response characteristics of different actuators (such as high-speed drivers and CNC interpolators), the command signals they receive and process are also differentiated accordingly, thereby achieving coordinated allocation of vibration suppression and shape control at the command level.

[0068] High-precision sensor arrays, including force sensors (for acquiring actual cutting force signals) and displacement sensors (for acquiring vibration displacement signals of the machining equipment), are installed at key locations of the machining equipment (such as the spindle, tool clamping system, and workpiece fixture). When machining a T-slot on a workpiece, the corresponding machining parameters are first determined and acquired. These parameters include the workpiece properties, tool properties, tool rotation speed, tool feed rate, and the relative movement path and corresponding movement speed between the tool and the workpiece. The workpiece properties include the material of the workpiece, and the tool properties include the cutting edge type and surface hardness.

[0069] Pre-machining is performed on the T-slot based on the machining parameters to obtain pre-machining data. The machining data includes the actual cutting force signal and the vibration displacement signal of the machining equipment. The vibration displacement signal of the machining equipment includes the vibration displacement signal of the workpiece and the vibration displacement signal of the tool spindle. A digital twin model is established based on the properties of the workpiece and the tool. In the digital twin model, local mesh refinement is performed on the area where the T-slot is located, the side wall area of ​​the T-slot, and the right-angle transition area. Local mesh refinement can improve the simulation accuracy of subsequent machining of the T-slot.

[0070] Using the inner right-angle vertex where the T-slot cross-section sidewall intersects the bottom and the endpoints of the two sides of the T-slot opening as base points, a mesh refinement square is established. Two adjacent refinement square elements are connected by an adjacent refinement rectangle. Mesh refinement is applied to the interiors of both the refinement square and the adjacent refinement rectangle, with the mesh density inside the refinement square being greater than that inside the adjacent refinement rectangle. The side length of the refinement square is 0.5 to 1.2 times the depth of the T-slot bottom. The chamfer inside the T-slot is considered as the inner right-angle vertex where the sidewall intersects the bottom. Mesh refinement is applied to the interior of the T-slot, with the mesh density inside the adjacent refinement rectangle being greater than that inside the T-slot. This local mesh refinement strategy effectively improves the simulation accuracy of key areas during T-slot machining. Especially for the minute deformation and vibration characteristics of the sidewall area and right-angle transition area, the mesh density setting is based on the geometric characteristics and machining requirements of the T-slot. A higher mesh density can capture more subtle deformation and vibration patterns, thereby improving the accuracy of the digital twin model. The operation of the digital twin model requires input machining parameters, including the material properties of the workpiece, the geometric parameters of the tool, the tool rotation speed, the feed rate, and the relative motion path between the tool and the workpiece. These parameters directly affect the simulation results of the digital twin model. Through the operation of the digital twin model, ideal machining data can be generated, that is, theoretical data of the T-slot machining process under ideal conditions. These data will be used to compare with the actual pre-machining data to quantify the deviations in the machining process.

[0071] Discrepancy data refers to the deviation between ideal machining data and actual pre-machining data. These deviations reflect various interference factors in the actual machining process, including tool wear, material inhomogeneity, chip accumulation, and coolant retention. Ideal machining data is the output of the digital twin model under ideal conditions. This discrepancy is the basis for subsequent state reconstruction and compensation control. Then, by using this discrepancy, the state parameters of the model are continuously corrected, thereby achieving a high-precision estimate of the internal state of the system and obtaining a simulation twin model.

[0072] like Figure 1 As shown, the joint state observer is a key technology for vibration suppression and shape control in T-slot machining. Its working principle is based on the core idea of ​​a state observer: a software model (digital twin model) simulates the internal operating state of the physical system in real time, and the model is continuously corrected by comparing the model output with the actual system output. This allows the model's internal state to infinitely approximate the true internal state of the physical system, eliminating the discrepancies between pre-machining data and ideal machining data. After eliminating these discrepancies, a simulation twin model is obtained. Under simulation, the obtained simulation machining data closely approximates the pre-machining data, while retaining the original mathematical twin model that simulated the ideal machining data. This is achieved through... A multimodal sensor array is installed for information acquisition, and combined with pre-processing data to obtain data reflecting the dynamics of the actual system, laying the foundation for accurate perception. The joint state observer adopts a unique dual-time-scale estimation architecture, with the high-frequency loop and the low-frequency loop connected in a closed loop through a disturbance compensation channel. This can quickly estimate dynamic cutting forces and displacements to suppress vibrations, and at the same time identify slow-changing disturbances such as tool wear and material inhomogeneity, achieving accurate observation of the machining state at multiple scales and in all directions. The system is also designed with an iterative optimization mechanism. After a single machining operation, the actual geometric dimensions are measured and compared with the simulation results to obtain shape deviation data and feed back to correct the parameters of the simulation twin model, thereby ensuring the practicality of the simulation twin model.

[0073] like Figure 2As shown, the joint state observer adopts a dual-time-scale estimation architecture: the high-frequency estimation loop (update frequency ≥ 500Hz) focuses on estimating the actual dynamic cutting force at the tool's cutting edge and the actual dynamic displacement of key parts, which are the rapidly changing vibration-related state quantities of the system; the low-frequency estimation loop (update frequency 50-500Hz) focuses on estimating the comprehensive disturbance force equivalent to interference factors, including disturbances caused by chip accumulation and cooling medium retention. The observer extracts high-frequency abrupt change features and low-frequency drift features from the actual signal through wavelet transform or empirical mode decomposition, maps these features to disturbance parameters, and then feeds the output of the low-frequency estimation loop back to the high-frequency estimation loop through the disturbance compensation channel, forming a closed-loop estimation mechanism. This design allows the observer to continuously correct the state estimation value, making the estimated value gradually approach the actual state, thereby achieving high-precision reverse calculation of the actual dynamic cutting force of the tool acting on the T-groove workpiece, the actual dynamic displacement of key parts of the T-groove workpiece, and the comprehensive disturbance force equivalent to interference factors. The introduction of the disturbance compensation channel makes the low-frequency estimation loop... The circuit can correct the value of the high-frequency estimation loop, forming a closed-loop estimation mechanism. The disturbance compensation channel enables the low-frequency estimation loop (responsible for handling slowly changing interference factors, such as tool wear and workpiece thermal deformation) to feed back the calculated disturbance components to the high-frequency estimation loop (responsible for handling rapidly changing vibration-related state quantities) through the disturbance compensation channel, thereby correcting the estimated value of the high-frequency estimation loop and forming a closed-loop estimation mechanism. This ensures that the system can accurately estimate the dynamic displacement and cutting force at the tool-workpiece contact point. The joint state observer uses the difference data between the actual force and position signals collected and the ideal output of the model to dynamically reconstruct the actual dynamic displacement of the tool and the actual dynamic displacement of the workpiece, thereby accurately calculating the comprehensive vibration at the cutting point. The control strategy generates a compensation control signal in real time based on this comprehensive vibration, driving the actuator to produce an anti-phase cancellation motion. This achieves a leap from "macro-machine tool vibration monitoring" to "precise cancellation of micro-vibrations at the tool-workpiece contact point". In particular, it can effectively suppress regenerative chatter and forced vibration easily caused by T-slot sidewalls and right-angle structures, significantly reduce the surface roughness of the machined surface, and improve the stability of the machining process.

[0074] Pre-machining data only includes some measurement values ​​collected by force and displacement sensors during actual machining (such as overall cutting force and macroscopic workpiece displacement). These data are affected by sensor accuracy, installation position, and machining environment interference (such as vibration and noise), and cannot directly reflect the local dynamic force at the tool cutting edge, the small displacement of the thin sidewall plate area, or the transient deformation of the right-angle transition zone. Therefore, it is necessary to simulate it through a simulation twin model. After machining a T-slot, the parameters in the simulation twin model are modified by detecting the internal dimensions of the T-slot.

[0075] Pre-machining data does not directly acquire the actual dynamic cutting force of the tool acting on the T-slot workpiece, the actual dynamic displacement of key parts of the T-slot workpiece, and other state quantities. The pre-machining data only includes some measurement values ​​collected by force sensors and displacement sensors during the actual machining process (such as the overall cutting force and the macroscopic displacement of the workpiece). However, these measurement values ​​are limited by the accuracy of the sensors, the installation position, and the interference of the machining environment (such as vibration and noise), and cannot accurately reflect the local dynamic force at the tool cutting edge, the small displacement of the thin plate area of ​​the side wall, or the transient deformation of the right-angle transition zone. The joint state observer realizes the reverse calculation in the following way: it compares the pre-machining data (as the actual measurement input) with the ideal machining data generated by the digital twin model in real time, calculates the error, and then uses the error feedback mechanism, based on the system dynamic equation (such as the rigid-flexible coupling dynamic model) and the dual time scale estimation architecture (the high-frequency loop handles rapid vibration and the low-frequency loop handles interference), to gradually correct the state estimation value, and finally outputs the actual dynamic cutting force at the tool cutting edge, the actual dynamic displacement of key parts, and the comprehensive interference force.

[0076] like Figure 5 As shown, the bolded line represents the machining area of ​​the T-groove. After machining, the T-groove will conform to this bolded line. The two ends at the top of the bolded line can be considered as the endpoints of the two sides of the opening of the T-groove. Since the top of the T-groove is the opening direction, there will be endpoints of the two sides of the opening of the T-groove. A dense square is established using the inner right-angle vertex where the sidewall of the T-groove cross-section intersects with the bottom of the groove and the endpoints of the two sides of the opening of the T-groove as base points. This is represented by a bolded dashed line. Two adjacent dense square units are connected by an adjacent dense rectangle, the interior of which includes the sidewall area of ​​the T-groove. If there is a gap between two dense squares... In overlapping areas, the two encrypted squares are treated as a single unit. Then, the area inside the T-groove is further meshed. Since the machining of the sidewalls of the T-groove and the inner right-angle vertices where the sidewalls intersect with the bottom of the groove directly affects the shape control of the T-groove, the mesh density of the sidewalls of the T-groove and the inner right-angle vertices where the sidewalls intersect with the bottom of the groove should be high. Furthermore, the mesh density inside the encrypted square is greater than the mesh density inside the adjacent encrypted rectangle, and the mesh density inside the adjacent encrypted rectangle is greater than the mesh density inside the T-groove. The specific mesh density is affected by the machining accuracy requirements and the specific dimensions of the T-groove, which will not be specifically explained here, thus ensuring the simulation of the twin model.

[0077] The preferred control strategy is to reverse the vibration of the tool to counteract the overall vibration. The compensation control signal is input to the driver, which controls the vibration of the spindle. The spindle connects with the tool. Existing technologies, such as intelligent spindles, are the "smart brain" of modern manufacturing. They can monitor parameters such as speed and vibration in real time, automatically optimize machining strategies, and greatly improve efficiency and accuracy. Intelligent spindles can control the spindle and thus suppress overall vibration.

[0078] The joint state observer in vibration suppression and shape control methods is key to achieving high-precision state estimation. The joint state observer employs a dual-time-scale estimation architecture, where the collaborative work of the high-frequency and low-frequency estimation loops can simultaneously handle high-frequency vibrations and low-frequency disturbances, improving the accuracy and robustness of state estimation. In T-slot machining, high-frequency vibrations mainly originate from instantaneous force changes during the cutting process, while low-frequency disturbances mainly arise from slowly changing factors such as tool wear and workpiece thermal deformation. The dual-time-scale estimation architecture accurately distinguishes and processes these vibrations and disturbances of different frequencies, thereby improving the accuracy of state estimation. The introduction of a disturbance compensation channel allows the low-frequency estimation loop to correct the values ​​of the high-frequency estimation loop, forming a closed-loop estimation mechanism.

[0079] The implementation of vibration suppression control strategies requires treating the workpiece as stationary and establishing a workpiece-tool relative motion coordinate system. This simplifies vibration suppression calculations. In T-slot machining, due to the complex geometry of the T-slot, vibration suppression needs to consider the vibration characteristics at different locations. Therefore, it is necessary to calculate the vibration components that the tool needs to offset in real time. By using a digital twin model and comprehensive vibration calculation in real time, the targetedness and effectiveness of vibration suppression can be ensured. A "compensation strategy" is employed, which comprehensively considers the "tool dynamics model" (ensuring the stability of the compensation motion itself), the "material removal model" (ensuring the compensated material removal result conforms to the target geometry), and the aforementioned total adjustment vector, to calculate a "dynamic compensation path" that continuously changes with the machining position in real time. This path contains rich instructions from low to high frequencies. To match the bandwidth characteristics of different execution units of the machine tool, a "decomposition method" is needed to decompose it in the frequency domain. Specifically, a "frequency threshold" is set, which is usually determined by comprehensively considering the effective update frequency of the machine tool CNC system interpolator and the effective control bandwidth of each axis servo drive.

[0080] like Figure 4As shown, establishing a dynamic compensation path is the core of the synergistic optimization of vibration suppression and shape control. The dynamic compensation path is based on the machining deformation error and comprehensive vibration predicted by the simulation twin model. By decomposing the dynamic compensation path into path compensation components and vibration compensation components, it can simultaneously handle low-frequency geometric deformation and high-frequency vibration interference. The path compensation component is used to correct the tool feed trajectory to compensate for geometric deformation; the vibration compensation component is used to offset the vibration influence in real time. This decomposition method can improve the targeting and effectiveness of compensation, enabling vibration suppression and shape control to work synergistically to achieve high-precision machining. In T-slot machining, due to the special structure of the T-slot, the prediction of geometric deformation needs to pay special attention to the right-angle region and sidewall region, as the deformation in these regions has the greatest impact on the overall machining accuracy. By mapping the geometric error and comprehensive vibration into a compensation vector of the tool in three-dimensional space, the compensation vector can be accurately calculated, thereby generating a dynamic compensation path. The iterative optimization mechanism of the dynamic compensation path compares the actual machining results with the model prediction, continuously correcting the model parameters to make the model more accurate, thereby improving the accuracy of compensation.

[0081] Suppressing vibration alone cannot compensate for the static or quasi-static elastic deformation of the workpiece caused by cutting forces (such as sidewall deflection), while compensating only for the deformation path cannot eliminate high-frequency vibration. Therefore, synergistic optimization is necessary. This involves fusing the vibration compensation component with the compensation control signal used to suppress vibration to generate a vibration suppression command. The basic path refers to the ideal spindle movement path. Ideally, the relative positions of the spindle and the tool are fixed; therefore, the basic path can also be considered the tool movement path. It is the initial planning path for T-slot machining, without considering any vibration or deformation effects. The simulation twin model can simulate the geometric deformation trend of the T-slot during machining, including sidewall bending, right-angle rounding, and unevenness at the bottom of the slot. These deformation errors are common geometric errors in T-slot machining, directly affecting the geometric accuracy of the T-slot. Machining deformation error is the geometric error predicted by the simulation twin model based on the actual machining state. It reflects the deformation of the T-slot caused by vibration, cutting forces, and other factors during machining. The combination of machining deformation error and comprehensive vibration... This provides key input for establishing a dynamic compensation path, which is generated based on machining deformation errors and overall vibration. It is not fixed but updates in real-time with changes in machining position, material condition, and tool wear, achieving intelligent shape control with "correction while machining." The establishment of the dynamic compensation path includes predicting geometric errors within the dense square and adjacent dense rectangles during the T-slot machining process based on a simulation twin model. These geometric errors include sidewall deformation, fillet error in the right-angle transition zone, and groove bottom flatness deviation. The geometric errors and overall vibration are mapped into a compensation vector for the tool in three-dimensional space. Through a compensation strategy combined with the tool dynamics model, material removal model, and compensation vector, a dynamic compensation path that changes in real-time with the machining position is generated. The compensation strategy includes calculating the tool's compensation feed in real-time based on the tool dynamics model, material removal model, and compensation vector. A dynamic compensation path that changes in real-time with the machining position is generated based on the compensation feed. The decomposition method decomposes the dynamic compensation path into path compensation components and vibration compensation components.

[0082] like Figure 6 and Figure 7As shown, a digital twin model is constructed based on the 3D geometric model of the T-slot workpiece and tool properties, and the mesh of key areas is refined. Machining parameters are input into the digital twin model to generate ideal machining data. Simultaneously, the pre-machining data collected by the sensor array and the ideal machining data are sent together to the joint state observer. The observer adopts a dual-time-scale estimation architecture, including a high-frequency estimation loop and a low-frequency estimation loop. The disturbance component output by the low-frequency loop is fed back to the high-frequency loop through the disturbance compensation channel to correct the estimated value. The joint state observer corrects the digital twin model based on the difference data, ultimately obtaining a high-fidelity simulation twin model. It outputs state variables such as the actual dynamic displacement of the tool, the actual dynamic displacement of the workpiece, and the combined disturbance force, establishing a workpiece-tool relative motion coordinate system, treating the workpiece as stationary. The actual dynamic displacements of the tool and workpiece obtained by the joint state observer are superimposed in the cutting area to obtain the combined vibration. The control strategy calculates the vibration components that the tool needs to cancel in real time based on the combined vibration, and combines this with the tool dynamics characteristics provided by the simulation twin model to generate a compensation control signal acting on the actuator. This signal is input to the intelligent spindle or high-speed actuator, driving the tool to generate a reverse micro-motion, actively canceling the combined vibration, and achieving real-time suppression of high-frequency vibration, such as... Figure 3 As shown, the dynamic compensation path includes several control commands. It is divided based on the update frequency of these commands, establishing a frequency threshold. Control commands with update frequencies below or equal to the threshold are classified as low-frequency path compensation components used to correct tool feed trajectories. Control commands with update frequencies above the threshold are classified as high-frequency vibration compensation components used to offset vibration effects in real time. The path compensation components correct tool feed trajectories to compensate for geometric deformation; the vibration compensation components offset vibration effects in real time. This decomposition method improves the targeting and effectiveness of compensation, enabling vibration suppression and shape control to work synergistically for high-precision machining. The path compensation components are fused with the basic path to generate a dynamic application path, and the vibration compensation components are fused with the compensation control signal to generate vibration suppression commands. The dynamic application path and vibration suppression command work together on the actuator to achieve precise control of the T-groove machining process. By incorporating vibration control and geometric error compensation into a unified framework, the simulation twin model can not only simulate the vibration state, but also predict the geometric deformation errors caused by machining (such as sidewall bending and right-angle fillet errors) with high fidelity. The predicted deformation errors are fused with real-time comprehensive vibration information and mapped into a dynamic compensation path in three-dimensional space. Through the frequency threshold decomposition method, the path is intelligently decomposed into a low-frequency path compensation component (used to correct the tool feed trajectory to compensate for deformation) and a high-frequency vibration compensation component (used to cancel vibration in real time). The path compensation component is fused with the basic path to generate a dynamic application path, while the vibration compensation component is fused with the aforementioned compensation control signal, thereby suppressing vibration during tool movement and ensuring the geometric accuracy of the T-groove.

[0083] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended embodiments and their equivalents.

Claims

1. A method for vibration suppression and shape control of T-slot based on force-position sensing, characterized in that, The vibration suppression and shape control method includes: S01, Information Acquisition: By installing a sensor group on the processing equipment to collect processing data in real time during the T-groove processing, the processing equipment acquires and performs pre-processing based on the processing parameters to obtain pre-processing data; S02. Construct and run the digital twin model: Construct a digital twin model based on the three-dimensional geometric model of the T-slot workpiece and tool properties, and refine the local mesh of the T-slot area, side wall area and right angle area in the digital twin model; S03, Ideal-Actual State Comparison and Interference Quantification: Input the processing parameters into the digital twin model, obtain the ideal processing data of the digital twin model, and calculate the difference data between the ideal processing data and the pre-processing data; S04. Dynamic State Reconstruction: Establish and run a joint state observer. The joint state observer corrects the digital twin model through difference data to obtain a simulation twin model. Based on the pre-processing data, the workpiece-side state variables during the processing are calculated. The workpiece-side state variables include the actual dynamic displacement of the tool and the actual dynamic displacement of the workpiece. S05. Vibration Suppression: The actual dynamic displacement of the workpiece and the actual dynamic displacement of the tool are superimposed in the cutting area to obtain the comprehensive vibration at the cutting area. A compensation control signal is generated by the control strategy to suppress the comprehensive vibration by acting on the actuator. S06. Cooperative optimization of vibration suppression and shape control: Obtain the basic path, simulate the machining deformation error using a twin model, and establish a dynamic compensation path based on the machining deformation error and the overall vibration. The dynamic compensation path is decomposed into path compensation component and vibration compensation component by decomposition method. The path compensation component is integrated with the basic path to generate a dynamic application path; The vibration compensation component is fused with the compensation control signal used to suppress vibration to generate a vibration suppression command; The machining data includes the actual cutting force signal and the vibration displacement signal of the machining equipment; The local mesh encryption includes a partition encryption strategy, which includes: establishing encryption squares based on the inner right-angle vertex where the sidewall of the T-shaped groove intersects the bottom of the groove and the endpoints of the two sides of the opening of the T-shaped groove; connecting two adjacent encryption square units through an adjacent encryption rectangle; performing mesh encryption on the interior of the encryption square and the adjacent encryption rectangle respectively; the mesh density inside the encryption square is greater than the mesh density inside the adjacent encryption rectangle; the side length of the encryption square is 0.5-1.2 times the depth of the bottom of the T-shaped groove; the chamfer inside the T-shaped groove is regarded as the inner right-angle vertex where the sidewall intersects the bottom of the groove; performing mesh encryption on the interior region of the T-shaped groove; the mesh density inside the adjacent encryption rectangle is greater than the mesh density inside the T-shaped groove. The joint state observer adopts a dual-time-scale estimation architecture, which includes a high-frequency estimation loop for estimating the actual dynamic cutting force and the actual dynamic displacement, and a low-frequency estimation loop for estimating the disturbance components of the machining process. The high-frequency estimation loop and the low-frequency estimation loop are connected by a disturbance compensation channel. The machining disturbance component output by the low-frequency estimation loop is fed back to the high-frequency estimation loop through the disturbance compensation channel, and the estimated values ​​of the actual dynamic cutting force and the actual dynamic displacement are corrected. The decomposition method includes: the dynamic compensation path includes several control commands, which are divided by the update frequency of the control commands. A frequency threshold is established, and the control commands in the dynamic compensation path are divided into path compensation components and vibration compensation components according to the update frequency: control commands with an update frequency lower than or equal to the frequency threshold are divided into path compensation components for correcting the tool feed trajectory, and control commands with an update frequency higher than the frequency threshold are divided into vibration compensation components for real-time compensation of vibration effects.

2. The method for suppressing and controlling T-groove vibration based on force-position sensing according to claim 1, characterized in that: The control strategy includes: establishing a workpiece-tool relative motion coordinate system, treating the workpiece as stationary, calculating in real time the vibration component that the tool needs to counteract the comprehensive vibration through comprehensive vibration, and calculating in real time the compensation control signal input to the driver and acting on the spindle through the vibration component and the simulation twin model.

3. The method for suppressing and controlling T-groove vibration based on force-position sensing according to claim 1, characterized in that: Methods for establishing dynamic compensation paths include: Based on the simulation twin model, the geometric errors in the densified square and adjacent densified rectangle during the T-slot processing are predicted. The geometric errors include sidewall deformation, right-angle transition zone fillet error and groove bottom flatness deviation. Geometric errors and combined vibrations are mapped into a compensation vector for the tool in three-dimensional space; By combining a compensation strategy with a tool dynamics model, a material removal model, and a compensation vector, a dynamic compensation path that changes in real time with the machining position is generated. The compensation strategy includes: calculating the compensation feed rate of the tool in real time based on the tool dynamics model, the material removal model and the compensation vector, and generating a dynamic compensation path that changes in real time with the machining position based on the compensation feed rate.

4. The method for suppressing and controlling T-groove vibration based on force-position sensing according to claim 1, characterized in that: The dynamic compensation path includes an iterative optimization mechanism, which includes: After a single processing is completed, the actual geometric dimensions of the T-slot are measured and compared with the simulated geometric dimensions generated by the simulation twin model to obtain shape deviation data; The shape deviation data is fed back to the simulation twin model, and the parameters inside the simulation twin model are adaptively corrected to obtain a corrected simulation twin model that is applied to the next processing.

Images