Titanium alloy component machining deformation simulation compensation control method

Abstract

The application relates to the technical field of mechanical processing process control, and discloses a titanium alloy component processing deformation simulation compensation control method; the method is characterized in that: firstly, the titanium alloy component is clamped on a five-axis numerical control machining center, a corresponding relationship between a clamping unit, a supporting unit, a working condition sensing channel and a machining coordinate and a time reference is established; before cutting, an initial stress state description of a current in-process component is established according to clamping force disturbance and reaction force response; during the machining process, the stress state description is updated according to machining process nodes, and in the clamped cutting state and the loose clamped stable state, the tool deflection deformation and the stress release deformation are separated, and then the supporting amount, the tool position offset amount, the clamping force time sequence and the remaining machining allowance are adjusted; the application can improve the wall thickness consistency, the contour precision and the machining stability of titanium alloy thin-walled parts and frame overall structural parts.

Description

Technical Field

[0001] This invention relates to the field of machining process control technology, specifically a simulation compensation control method for deformation during the machining of titanium alloy components. Background Technology

[0002] Titanium alloys, due to their high specific strength, good heat resistance, and strong corrosion resistance, are often used in the manufacture of thin-walled components and frame-like integral structural parts for aerospace applications. These components typically have small wall thicknesses and weak overall rigidity. Titanium alloys themselves have low elastic modulus and poor thermal conductivity, making them susceptible to the combined effects of cutting forces and residual stresses within the material during CNC milling. On the one hand, cutting forces can cause recoverable elastic deformation of the tool in weak rigid areas; on the other hand, residual stresses within the blank are released and redistributed as material is removed layer by layer, leading to irreversible warping deformation, making it difficult to stably control the component's wall thickness and profile dimensions. To reduce machining deformation, existing technologies typically employ finite element simulation prediction and toolpath compensation before machining. For example, the published invention patent application CN109840337B discloses a milling method for controlling the wall thickness uniformity of thin-walled structural parts, which mainly determines the machining path that minimizes deformation through finite element simulation and toolpath experiments. Another example is the published invention patent application CN108182325B, which discloses a method for predicting and analyzing deformation during the cutting of thin-walled structural parts. This method mainly predicts the stress evolution and deformation of a blank containing initial residual stress during machining by establishing a finite element model. While these methods can predict deformation trends before machining or optimize the machining path based on the prediction results, they usually rely on a preset nominal initial residual stress field and an offline simulation model for one-time compensation. However, the residual stress distribution of titanium alloy blanks formed by actual forging or rolling may vary between different batches and within the same blank. Furthermore, the true residual stress field is difficult to obtain non-destructively across the entire field before machining, making it difficult for offline models to accurately reflect the actual stress state of the workpiece and its dynamic redistribution as material is removed. Simultaneously, the elastic tool deformation caused by cutting force and the irreversible deformation caused by residual stress release differ in their causes, reversibility, and time response characteristics. Existing offline compensation methods struggle to separate and accurately attribute these two types of deformation in real time during machining. This leads to a deviation between the compensation amount and the actual deformation of the workpiece, easily resulting in overcompensation or undercompensation, which in turn affects the wall thickness consistency, contour accuracy, and machining pass rate of thin-walled titanium alloy parts and frame-type integral structural parts. Therefore, it is necessary to propose a simulation compensation control method for deformation during the machining of titanium alloy components that can identify the true residual stress state online, distinguish different deformation sources, and dynamically correct the compensation toolpath during machining. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a simulation compensation and control method for deformation during the processing of titanium alloy components, which solves the problems mentioned in the background section.

[0004] To achieve the above objectives, the present invention provides the following technical solution: A method for simulation compensation and control of deformation during the processing of titanium alloy components includes: S1. Clamp the titanium alloy component in a five-axis CNC machining center and establish the correspondence between the flexible and controllable clamping unit, the conformal support unit, the working condition sensing channel and the machining coordinates and time reference. S2. Before formal cutting, apply controllable clamping force changes to each clamping point, collect the reaction force response of each clamping point, and establish an initial stress state description of the current component under manufacturing based on the reaction force response. S3. Perform material removal according to the preset processing steps. Collect clamping reaction force, support force and cutting condition data at each processing step and update the stress state description of the current step based on the collected data. S4. Collect the changes in component position and clamping reaction force under clamping and cutting conditions and loosening and stabilizing conditions, and separate the tool deflection deformation caused by cutting force and the deformation caused by residual stress release. S5. Adjust the conformal support top support amount, tool position offset amount, clamping force timing and remaining machining allowance according to the deformation source after separation, and use the adjusted machining action as the basis for the execution of the next machining process node.

[0005] Furthermore, S1 includes: Based on the nominal digital model of the component and the measured coordinates of the reference positioning point, establish the transformation relationship between design coordinates, workpiece coordinates and machine tool coordinates, and preset the actual geometric coordinate recording field; The clamping support points and top support points are arranged in the weak rigidity area of ​​the component and its adjacent support positions; The CNC system interpolation clock is used to correlate clamping reaction force, support force, cutting condition data, and tool interpolation position. After zero-point and static load verification of the force measuring element, a point mapping table and data record entries are generated.

[0006] Furthermore, S2 includes: The foundation clamping force is set according to the component wall thickness, local stiffness, and cutting load; After the clamping reaction force stabilizes, clamping force disturbances with limited amplitude are applied sequentially to each clamping support point and the reaction force difference is collected. Arrange the reaction force response vectors in the order of disturbance to form the reaction force response matrix; The stress zones are divided by combining the a priori interval of residual stress and the point mapping table, and the deviation is evaluated based on the back-calculated reaction force response matrix to generate an initial stress state description with offset direction, offset intensity and confidence level.

[0007] Furthermore, after the clamping reaction force stabilizes, clamping force disturbances with limited amplitude are sequentially applied to each clamping support point, and the reaction force difference is collected, including: After the basic clamping state satisfies the reaction force stability criterion, the disturbance amplitude is determined according to the ratio of the basic clamping force and the amplitude is limited. Apply clamping force disturbance point by point according to the circumferential sequence of the component or the priority sequence of the weak rigid area, while keeping the remaining clamping support points in the basic clamping state; After the disturbance, the reaction forces before and after the disturbance are collected according to the stability waiting condition and the difference is processed to generate the reaction force response vector. Clamping support points that do not meet the stability waiting conditions are marked as points to be reviewed.

[0008] Furthermore, S3 includes: The machining process nodes are divided according to the machining depth, cavity characteristics, and stress-sensitive areas; Set the clamping force and initial support force of the current node according to the stress state description of the previous node; Collect clamping reaction force, support force, spindle load, feed axis current and tool interpolation position, and write them into the node data table; Choose between simplified or full perturbation based on changes in the reaction force reference, changes in the offset direction, and changes in the low-confidence region. The stress state description is updated based on the nodal reaction force response matrix, and a nodal state vector is formed.

[0009] Furthermore, simplified or full perturbations are selected based on changes in the reaction force reference, changes in the offset direction, and changes in the low-confidence region, including: The first machining process node applies clamping force disturbance to all clamping support points; In subsequent processing steps, when the reaction force reference change between adjacent nodes is within a stable range and the offset direction has not changed, clamping force disturbance is only applied to the medium and high offset trend areas and their adjacent clamping support points. When there is a sudden change in the reaction force reference, a change in the offset direction, or an increase in the number of low-confidence stress sections, clamping force disturbance is applied to all clamping support points. Further, S4 includes: Synchronize the position fluctuations and clamping reaction force fluctuations with the cutting load in time and match their trends. After the clamping force is reduced to a low holding force and the clamping is in a stable state, the stable position deviation and reaction force reference drift are collected. Based on the degree of position recovery, the consistency voting results of multiple clamping support points, and the status of the top support point, the total deviation of the same processing area is decomposed, and the tool deformation component, stress release deformation component, separation confidence level, and verification mark are generated.

[0010] Furthermore, after the clamping force is reduced to a low holding force and the clamping reaches a stable state, the stable position deviation and reaction force reference drift are collected, including: After cutting stops, adjust the clamping force of each flexible and controllable clamping unit to a low holding force; After the component position and clamping reaction force benchmark meet the stability criteria, the loosening position deviation and clamping reaction force benchmark are collected; Restore the basic clamping force and perform a short-term retest; The continuous positional deviation and the drift of the reaction force in the same direction are used as the data for determining stress release deformation.

[0011] Furthermore, S5 includes: Based on the deflection direction, deflection amplitude, separation confidence level, and the corresponding position of the support point and the machining area, determine the support force increment and the normal tool position offset; Adjust the tool position offset ratio according to the increase in the support force; The clamping force timing and area margin table are modified based on the stress release offset direction; After the expected residual deviation convergence determination and low confidence anomaly handling, a machining action package is generated that records at least the toolpath correction segment, execution confidence, and anomaly flag.

[0012] Furthermore, the clamping force timing and regional margin table are modified based on the stress release offset direction, including: Identify excess material side offset or insufficient material side offset for stress relief deformation components; For areas with multiple material offsets, reduce the working clamping force of the next processing node and correct the clamping force timing to retain or increase the subsequent removal allowance. For areas with insufficient material, increase the working clamping force of the next processing node and correct the clamping force timing to reduce the subsequent removal amount or retain the finishing allowance. Generate a region balance table based on the balance adjustment amount, recording the region number, current remaining balance, and execution status.

[0013] Compared with existing technologies, this invention provides a simulation compensation control method for machining deformation of titanium alloy components, which has the following beneficial effects: 1. This invention establishes a correspondence between flexible and controllable clamping units, conformal support units, working condition sensing channels, and machining coordinates and time references. Before formal cutting, it uses clamping force disturbance and reaction force response to establish an initial stress state description of the component currently in production. This allows the compensation basis to be combined with the actual reaction force response of the current workpiece, reducing the deviation caused by relying solely on offline nominal residual stress field for one-time compensation. By continuously collecting clamping reaction force, top support force, and cutting condition data at each machining process node and updating the stress state description, the dynamic redistribution of residual stress caused by material removal is continuously characterized. By collecting position changes and reaction force changes under clamping cutting state and loosening stable state respectively, recoverable tool-relief deformation and unrecoverable stress release deformation are separated. The top support amount, tool position offset amount, clamping force timing, and remaining machining allowance are adjusted respectively, thereby reducing overcompensation and undercompensation caused by mismatch between compensation amount and actual deformation, and improving the wall thickness consistency, contour accuracy, and machining qualification rate of titanium alloy thin-walled parts and frame-type integral structural parts.

[0014] 2. This invention establishes node data tables, node state vectors, separation confidence levels, verification identifiers, and machining action packages at each machining process node. Clamping reaction force, support force, cutting load, tool position, area allowance, and abnormal states are recorded and transmitted according to a unified data structure. This allows the state judgment, abnormal marking, and compensation actions of the previous machining process node to be continuously transmitted to the control process of the next machining process node. Simultaneously, by setting conservative machining, encrypted sampling, degraded control, and manual verification conditions for low-confidence regions, abnormal support points, missing working condition data, and convergence failures, the direct interference of abnormal data on compensation actions is reduced, thereby improving the data traceability, compensation decision stability, and control reliability under abnormal working conditions. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the process of a simulation compensation control method for machining deformation of titanium alloy components according to the present invention; Figure 2 This is a schematic diagram of the clamping support and working condition sensing arrangement of the present invention; Figure 3 This is the logic diagram for identifying clamping force disturbance and initial stress state in this invention; Figure 4 This is a schematic diagram illustrating the dynamic update of stress state at the processing nodes of the present invention; Figure 5 This is a schematic diagram illustrating the separation of tool deformation and stress relief deformation in this invention; Figure 6 A closed-loop diagram is generated for the differentiated compensation action of this invention. Detailed Implementation

[0016] 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.

[0017] Example 1: Figure 1 - Figure 6 A simulation compensation and control method for machining deformation of titanium alloy components is presented, including: S1. Clamp the titanium alloy component in a five-axis CNC machining center and establish the correspondence between the flexible and controllable clamping unit, the conformal support unit, the working condition sensing channel and the machining coordinates and time reference. S2. Before formal cutting, apply controllable clamping force changes to each clamping point, collect the reaction force response of each clamping point, and establish an initial stress state description of the current component under manufacturing based on the reaction force response. S3. Perform material removal according to the preset processing steps. Collect clamping reaction force, support force and cutting condition data at each processing step and update the stress state description of the current step based on the collected data. S4. Collect the changes in component position and clamping reaction force under clamping and cutting conditions and loosening and stabilizing conditions, and separate the tool deflection deformation caused by cutting force and the deformation caused by residual stress release. S5. Adjust the conformal support top support amount, tool position offset amount, clamping force timing and remaining machining allowance according to the deformation source after separation, and use the adjusted machining action as the execution basis for the next machining process node. This method is applicable to aerospace titanium alloy thin-walled parts, frame-type integral structural parts, and integral milled components with weak rigidity machining areas; the machining object can be selected as TC4 titanium alloy components, the component wall thickness can be selected from 1.5mm to 3mm, the cavity depth can be selected from 20mm to 40mm, the blank can be obtained by forging, rolling or pre-stretching process, the final wall thickness tolerance can be selected from ±0.05mm to ±0.1mm, and the contour requirement is not greater than 0.1mm; this method relies on a five-axis CNC machining center to perform milling, and the flexible controllable clamping unit, conformal support unit, force measuring element, working condition sensing channel and CNC system clock together form the basis for data acquisition and closed-loop control of the machining process; the input information includes at least the nominal digital model of the titanium alloy component, the geometric shape of the blank, the material property range, the a priori range of initial residual stress, and the pre- Information such as machining allowance distribution, nominal cutting process conditions, clamping reaction force data, support force data, spindle load data, feed axis current data, and position offset data during component machining are included. This method establishes a stress state description on each in-work component and continuously updates the stress state along machining process nodes. It separates recoverable tool-relief deformation that occurs synchronously with the cutting load from irrecoverable stress-release deformation that persists after clamping. Based on the separation results, it generates the support amount, tool position offset, clamping force sequence, and remaining machining allowance adjustment scheme for the next machining process node. After processing, it outputs machining action commands corresponding to each machining process node. These commands include at least the fields of clamping force sequence, conformal support support amount, tool position offset, remaining machining allowance allocation, execution confidence, and anomaly flags. Specifically, for example... Figure 2 As shown: After the titanium alloy component completes blank inspection, datum surface cleaning, and initial positioning, it is fixed on the worktable of a five-axis CNC machining center. This process is executed after the titanium alloy component enters the machining station and the machining area, clamping datum, and blank outer dimensions are confirmed. The input information includes the component's nominal digital model, blank outer dimensions, reserved machining allowance, clamping datum coordinates, preset arrangement positions of clamping support points and top support points, and initial coordinate parameters of the CNC system. Based on the nominal numerical model of the component, the design coordinate system is determined. After the blank is actually clamped, the actual coordinates of no less than three non-collinear reference positioning points are obtained through probe measurement, in-machine probe measurement, or laser displacement measurement, preferably obtaining more than four reference positioning points. According to the coordinates of each reference positioning point in the design coordinate system and the corresponding measured coordinates, a rigid body coordinate transformation relationship from the design coordinate system to the workpiece coordinate system is established to obtain the translation amount and attitude angle correction amount of the workpiece coordinate system relative to the design coordinate system, and the workpiece coordinate system is aligned. Subsequently, the workpiece coordinate system is matched with the machine tool coordinate system of the five-axis CNC machining center so that the real-time interpolation coordinates of the CNC system can correspond to the machining position on the surface of the component. Among them, no less than three non-collinear reference positioning points are used to determine the workpiece coordinate attitude, and more than four reference positioning points are used to form redundant measurement and error verification, thereby reducing the impact of clamping and alignment errors on subsequent reaction force identification, node update, and tool position offset. Flexible and controllable clamping units are arranged around the component, below the web, at the intersection of frame beams, or near the weakly rigid area. Each flexible and controllable clamping unit corresponds to one clamping support point. Conformal support units are arranged on the back side of the weakly rigid area, with each conformal support unit corresponding to one top support point. The number of clamping support points can be selected from four to twelve, determined according to the component's external dimensions, wall thickness, cavity depth, and the distribution of the weakly rigid area. The rationale for this range is that for small, thin-walled components, using fewer than four clamping support points makes it difficult to simultaneously constrain the component's translation, rotation, out-of-plane bending, and local warping tendencies. For large-sized frame components, using more than twelve clamping support points increases the complexity of clamping calibration and the degree of multi-point constraint coupling, and may suppress the residual stress release tendency of the component due to excessively dense constraints, affecting the accuracy of subsequent reaction response identification. For components with large profile dimensions, thin webs, or deep open cavities, conformal support units are added on the back side of the weakly rigid area, so that the top support point corresponds to... Adjacent clamping support points form local support pairs; after establishing the coordinate correspondence, a unified time reference is configured for the flexible controllable clamping unit, conformal support unit, and working condition sensing channel; the time reference adopts the CNC system interpolation clock of the five-axis CNC machining center, and the force measuring element, conformal support force feedback channel, spindle load acquisition channel, and feed axis current acquisition channel all generate timestamps with this interpolation clock; the sampling frequency of the force measuring element is not less than 500Hz. The reason for this setting is that the change in clamping reaction force mainly reflects the change in clamping constraint, low-frequency structural response, and residual stress redistribution trend. The frequency of such changes is lower than the impact frequency of the cutting teeth. 500Hz sampling can provide millisecond-level reaction force change records and meet the subsequent time alignment requirements with the working condition sensing data; the top support force feedback frequency of the conformal support unit is not less than 500Hz; in areas where tool deformation is obvious, the top support force feedback frequency is increased to not less than 1000Hz in order to synchronize and compare with the changes in cutting conditions; The sampling frequency of the working condition sensing channel is no less than 1000Hz, preferably between 1000Hz and 5000Hz. This sampling frequency is determined based on the tooth passage frequency, which is determined by the spindle speed and the number of tool teeth. In milling thin-walled titanium alloy parts, two- to four-tooth tools and spindle speeds of several thousand to over ten thousand revolutions per minute are commonly used, causing the cutting load to exhibit periodic fluctuations on the order of several hundred hertz. Therefore, sampling at 1000Hz to 5000Hz can cover the rapid changes in spindle load and feed axis current, and reserve sampling margin for subsequent identification of reaction force fluctuations synchronized with changes in cutting force. During time alignment, each set of clamping reaction force data, support force data, spindle load data, and feed axis current data is associated with the corresponding tool interpolation position to form a data record entry. The data record entry includes at least the sampling time, tool position, clamping support point number, clamping reaction force, support point number, support force, spindle load, and feed axis current. Fields such as load, feed axis current, and current machining segment number are included. To enable subsequent reaction force response, tool position, and component deformation to be retrieved under the same spatial reference, a three-layer coordinate mapping relationship is established. The first layer is the design coordinates of the component's nominal digital model, used to represent the theoretical contour, theoretical wall thickness, cavity boundary, and reserved allowance. The second layer is the workpiece coordinates after the blank is actually clamped, used to represent the actual posture of the blank, the position of the clamping support point, and the position of the top support point. The third layer is the actual geometric coordinates of the component acquired by subsequent machining nodes, used to represent the deviation of the component from the nominal shape in the clamping cutting state or the loosening stable state. In this processing stage, the recording field and point number of the third layer of actual geometric coordinates are pre-established, and the actual geometric coordinates are written by the measurement data of subsequent machining nodes. For each clamping support point and top support point, its corresponding number in the design coordinates, workpiece coordinates, and actual geometric coordinates is recorded to form a point mapping table. The preferred reference point deviation threshold is 0.02 mm. If the reference point deviation measured after clamping exceeds this threshold, the workpiece coordinate system alignment is re-executed. The reason for setting this value is that the target profile of thin-walled titanium alloy parts and frame-type integral structural parts is usually required to be no greater than 0.1 mm. If the initial coordinate error exceeds one-fifth of the target profile, it will affect the subsequent stress identification and tool position offset accuracy. For components with a wall thickness tolerance of no more than ±0.05 mm, the reference point deviation threshold can be selected from 0.01 mm to 0.02 mm. The above thresholds are clamping and alignment control thresholds before stress identification and are not used as a final machining accuracy limit. If a force measuring element fails to return a stable zero point signal, zero point reset and static load verification are performed first. If the difference between two static load verification readings under the same standard load exceeds 1% of the full scale of the force measuring element, the corresponding clamping support point is marked as low confidence. The 1% threshold is used for data availability verification before on-site clamping. The rationale for this threshold is that when the difference in repeated readings under the same standard load exceeds 1% of the full scale, it usually indicates that zero-point drift of the force measuring element, changes in installation preload, or local contact anomalies are sufficient to affect the judgment of reaction force trends. Data from low-confidence clamping support points is not directly deleted; its weight is reduced to 30% to 60% of the weight of normal clamping support points, and it is used in conjunction with data from adjacent high-confidence clamping support points in subsequent reaction force identification, node updates, and deformation separation. The rationale for this weight range is that low-confidence data may still retain local reaction force change trends, and direct deletion would reduce spatial coverage; however, if given the same weight as normal data, abnormal readings might dominate stress state judgment. Therefore, a weight reduction range of 30% to 60% is used to balance trend preservation and anomaly suppression. After the above processing, a machining data acquisition link with a unified time reference and spatial mapping relationship is formed, as well as a point mapping table between clamping support points, top support points, tool positions and machining areas; this machining data acquisition link and point mapping table are used to identify the initial stress state of the currently manufactured titanium alloy component before formal cutting.

[0018] Specifically, such as Figure 3As shown: After the titanium alloy component is clamped, the workpiece coordinate system is aligned, the time reference is configured, and the data channel is verified, but before the component enters the cavity material removal stage, the initial reaction force of each clamping support point is identified. The input information includes the aforementioned point mapping table, the initial reaction force of each clamping support point, the adjustable range of the clamping force of the flexible controllable clamping unit, the component material properties, the a priori range of initial residual stress, the blank thickness distribution, and the reserved machining allowance. First, a basic clamping force is applied to each flexible controllable clamping unit to keep the titanium alloy component stably positioned. The basic clamping force can be selected from 300N to 3000N, and is determined according to the component's external dimensions, wall thickness, local stiffness, support span, and expected cutting load. For thin-walled areas with a wall thickness of 1.5mm to 2mm and a large support span, the basic clamping force can be selected from 300N to 100N. 0N; For wall thicknesses of 2mm to 3mm, frame beam areas, or areas with high local rigidity, the basic clamping force can be selected from 1000N to 3000N; the engineering basis is that when the clamping force is less than 300N, it is difficult to resist the transverse cutting force and machining vibration during the milling process of titanium alloy, which easily leads to unstable clamping state; when the clamping force is greater than 3000N, it is easy to forcibly flatten the weak rigid component, so that the clamping reaction force cannot truly reflect the offset trend caused by the residual stress inside the component; after the basic clamping force is applied, when the reaction force of each effective clamping support point fluctuates within 3 consecutive seconds and does not exceed 2% of the corresponding basic clamping force, the basic clamping state is determined to be stable; this 2% threshold is used to exclude noise from the force measuring element and micro-movement of clamping contact, and is lower than the subsequent 5% high confidence deviation threshold, so that the clamping state can be confirmed to be stable before disturbance identification; After the basic clamping state stabilizes, a small clamping force disturbance is sequentially applied to each clamping support point. The small clamping force disturbance is determined to be 5% to 15% of the basic clamping force, and the amplitude of each disturbance is limited to between 30N and 300N. When the proportionally obtained disturbance amplitude is lower than 30N, 30N is used; when the proportionally obtained disturbance amplitude is higher than 300N, 300N is used. The engineering basis for this amplitude limit rule is that when the disturbance amplitude is lower than 30N, the reaction force redistribution may be lower than the effective resolution of the force measuring element and the small fluctuation amplitude of the clamping contact, making it difficult to form a recognizable reaction force response. When the disturbance amplitude is higher than 300N, it may change the local contact state or introduce new elastic bending errors into thin-walled areas. The disturbance sequence can be executed according to the spatial order of the clamping support points in the circumferential direction of the component, or according to the priority order of weakly rigid areas. Each disturbance only... The force is applied to one clamping support point, while the remaining clamping support points remain in a basic clamped state. After each minor clamping force disturbance, a stabilization waiting time of 5 to 30 seconds is maintained, and reaction force data of each clamping support point is collected at the end of the waiting period. The reason for setting this waiting time is that after a change in local clamping constraint, the titanium alloy thin-walled component will undergo a short-term elastic recovery and redistribution of support reaction force. If the waiting time is less than 5 seconds, the reaction force reading may still be in the transition phase. If the waiting time is more than 30 seconds, the improvement in reaction force stability is limited, and it will increase auxiliary processing time. During the waiting period, the same reaction force stability criterion as the basic clamping state is used, or the reaction force is determined to have reached a stable state when the change in reaction force no longer shows a unidirectional increasing trend. If the stability criterion is not met after 30 seconds, the current reaction force data is recorded, and the corresponding clamping support point is marked as a point to be verified. After the reaction force data acquisition is completed, the reaction force data of each clamping support point after the disturbance is processed by the difference between the reaction force data of each clamping support point before the disturbance, to obtain the reaction force response vector corresponding to the current disturbance. Each reaction force response vector is composed of the reaction force change of each clamping support point under the same disturbance, and is used to characterize the reaction force transmission and redistribution on the component under the disturbance. After the disturbance of all clamping support points is completed, the reaction force response vectors are arranged in the order of disturbance to form a reaction force response matrix. This reaction force response matrix is ​​used to represent the overall reaction force transfer relationship of the titanium alloy component under different clamping force disturbances, and serves as the basic data for establishing the initial stress state description. After establishing the reaction force response matrix, based on the reaction force response matrix and the initial residual stress... The a priori range for residual stress is established to describe the initial stress state of the components currently under manufacturing. The a priori range for initial residual stress is determined based on the source of the blank process. For pre-stretched TC4 titanium alloy forgings, the absolute values ​​of surface compressive stress and core tensile stress can be selected as 100MPa to 300MPa as the initial constraint range. The engineering basis for this range is that the residual stresses commonly found in titanium alloy forgings, rolling, and pre-stretched blanks are usually within this order of magnitude, which can cover the differences between normal batches and exclude abnormal results that are obviously inconsistent with the material processing rules. For titanium alloy components that have not undergone pre-stretching treatment, have different heat treatment regimes, or have different blank sources, the a priori range can be corrected based on residual stress test samples from the same batch, blank process records, or reaction force responses of test cut pieces. When establishing the initial stress state description, the component is divided into multiple stress segments according to the point mapping table. Each stress segment corresponds to at least one clamping support point or top support point. For each stress segment, the direction of reaction force change at each clamping support point represents the offset direction of the corresponding stress segment, and the ratio of the reaction force change amplitude to the disturbance amplitude represents the offset trend intensity. When the reaction force change directions of adjacent clamping support points are consistent, the offset trend intensity of the stress segment is increased; when the reaction force change directions of adjacent clamping support points are opposite, the offset trend intensity of the stress segment is decreased, and the verification mark is retained. The offset trend mark can be selected as three categories: excess material side offset, insufficient material side offset, and stable. Excess material side offset indicates an offset on one side that may still retain excess material after processing; insufficient material side offset indicates an offset on one side that may be lower than the theoretical profile or cause insufficient wall thickness after processing; stable indicates that the segment has not formed under the current disturbance. A clear directional reaction force transfer is observed; the offset trend intensity can be selected as low, medium, or high, or a normalized numerical value between 0 and 1 can be used as the basis for process node division, node reaction force re-identification, and clamping force timing adjustment; to improve the reliability of the initial stress state description, the reaction force response matrix is ​​evaluated for deviation and assigned a confidence level; when evaluating the deviation, a back-calculated reaction force response matrix is ​​generated based on the initial stress state description; the back-calculated reaction force response matrix is ​​jointly determined by the offset trend intensity of each stress segment, the local stiffness coefficient of the corresponding clamping support point, and the clamping force disturbance amplitude. The local stiffness coefficient can be determined based on the nominal modulus of the component, wall thickness, support span, and the aforementioned clamping force disturbance response; the measured reaction force response matrix is ​​compared point by point with the back-calculated reaction force response matrix, and the mean absolute deviation, mean square deviation, or maximum deviation is used to characterize the back-calculated deviation of the reaction force response at each clamping support point; When the reaction response back-calculation deviation is less than 5% of the corresponding clamping force of the foundation at the clamping support point, the confidence level of that clamping support point is set to high; when the reaction response back-calculation deviation is between 5% and 15%, the confidence level of that clamping support point is set to medium; when the reaction response back-calculation deviation is greater than 15%, the confidence level of that clamping support point is set to low, and repeated disturbances of that clamping support point are triggered. The reason for setting the above 5% to 15% range is that mechanical clamping contact, elastic rebound of thin-walled components, and field acquisition errors of force measuring elements will form a certain comprehensive engineering deviation; when the deviation is less than 5%, the measured response and the back-calculated response can usually be considered to be basically consistent; when the deviation is greater than 15%, there may usually be poor support contact, chip clogging, etc. Zero-point drift of the force measuring element or local defects in the blank are sufficient to significantly affect the initial stress state judgment. For clamping support points that still have low confidence after repeated perturbation, their data are not directly deleted. When there are two or more high-confidence clamping support points nearby, the average reaction force response of the nearby high-confidence clamping support points is used for weighted correction. When there are fewer than two nearby high-confidence clamping support points, the clamping force perturbation in this area is re-executed before the formal cutting. If an effective response cannot be obtained after the re-perturbation, the area is marked as a conservative machining area or a manually verified area. This processing method can suppress the influence of abnormal reaction force data on the initial stress state judgment while preserving the local reaction force change trend. After the above processing, an initial stress state description is generated. The initial stress state description includes at least the fields of stress section number, offset direction, offset intensity, confidence level, verification identifier and conservative machining identifier, and serves as the initial control benchmark for process node division, node removal order, node reaction force re-identification, dynamic stress update, clamping force timing adjustment and remaining machining allowance allocation.

[0019] Specifically, such as Figure 4As shown: After the initial stress state description of the current component under construction has been formed, and the high-confidence and medium-confidence regions cover the main weak rigid areas to be processed, material removal is performed according to the preset processing steps. Clamping reaction force, support force, and cutting condition data are collected at each processing step. The input information includes the initial stress state description, nominal machining toolpath, reserved machining allowance, machining depth, cavity feature distribution, point mapping table of clamping support points and support points, and cutting process parameters. The material removal process is divided into processing steps based on the machining depth, cavity features, and stress-sensitive areas. The cumulative material removal thickness between adjacent processing steps can be selected from 2mm to 5mm. For thin-walled areas with a wall thickness close to 1.5mm, frame beam intersection areas, and areas with high offset trend intensity, the cumulative material removal thickness interval can be reduced to 1mm to 2mm. For areas with strong local rigidity, low offset trend intensity, and stable reaction force benchmarks at continuous nodes, the cumulative material removal thickness interval can be widened to 5mm to 8mm. However, this interval is not suitable for areas with wall thickness close to the final size or medium to high offset trend intensity. The engineering basis for the above numerical range is that during the milling process of thin-walled titanium alloy components, residual stress release and structural stiffness reduction usually accumulate as material is removed layer by layer. If the node interval is too large, reaction force drift and deformation deviation will accumulate between two updates. If the node interval is too small, it will increase the number of pauses, sampling, and reaction force identification, thereby reducing machining efficiency. Among them, an interval of 1mm to 2mm is suitable for intensive tracking in deformation-sensitive areas, and an interval of 5mm to 8mm is suitable for efficiency optimization in low-risk areas. The above cumulative material removal thickness is used to determine the node update interval and is not used as the single cutting depth. Before each machining operation node begins, the clamping force of each flexible and controllable clamping unit is set according to the stress state description and clamping force sequence output by the previous node, and the initial support force of the conformal support unit is set according to the conformal support scheme output by the previous node. When the current node performs material removal, the clamping reaction force, support force, spindle load, feed axis current, and tool interpolation position are collected in real time according to the timestamp of the aforementioned data recording entries and the tool interpolation position. The clamping reaction force is used to reflect the change of the support reaction force of the component under clamping constraints, the support force is used to reflect the local support state on the back side of the weak rigid area, the spindle load and feed axis current are used to characterize the change of cutting load, and the tool interpolation position is used to determine the machining space position corresponding to the reaction force change. The collected data is written into the node data table, which is used to save the original acquisition sequence of the current node, including at least The system includes fields such as node number, machining section number, material removal area, cumulative removal thickness, tool position sequence, clamping reaction force sequence, top support force sequence, spindle load sequence, and feed axis current sequence. To reduce the impact of occasional noise on node judgment, the clamping reaction force sequence and top support force sequence are smoothed using a sliding window. The window length can be selected from 0.1s to 0.5s. The engineering basis for this window length is that the sampling frequency of the force measuring element is not less than 500Hz, and the sampling frequency of the working condition sensing channel is not less than 1000Hz. Within a window of 0.1s to 0.5s, it can cover tens to hundreds of sampling points, which is sufficient to reduce high-frequency noise and occasional spikes from the sensor. If the window length is less than 0.1s, it is difficult to form effective smoothing. If the window length is greater than 0.5s, it may weaken the reaction force change characteristics that occur synchronously with the tool position, affecting the subsequent tool deformation recognition. After the material removal is completed at the current processing node, the feed is paused while maintaining the basic clamping state, and the stress state of the current node is updated. During the stress state update, the aforementioned micro-clamping force disturbance rule is followed, applying either simplified or full disturbance to the clamping support points. For the first processing node, full disturbance is used. For subsequent processing nodes, when the change in the reaction force benchmark between the previous node and the node before that is less than 3% of the basic reaction force, and the offset direction has not changed, the stress state change of the previous node is considered stable, and simplified disturbance is used, applying only to clamping supports with medium or high offset trend intensity. Perturbation is applied to the bearing point and adjacent clamping support points; when there is a sudden change in reaction force, a change in offset direction, or an increase in the number of low-confidence stress sections, a complete perturbation is used, applying the perturbation to all clamping support points sequentially; where a sudden change in reaction force refers to a change in the reaction force reference of any effective clamping support point exceeding 5% of the foundation reaction force; the perturbation amplitude is 30N to 300N, and the stabilization waiting time is 5s to 30s, so that the node update and initial stress identification use the same reaction force response scale; a 3% threshold is used to identify slight changes in the reaction force reference between nodes, and a 5% threshold is used in conjunction with the aforementioned high-confidence deviation threshold. The values ​​are linked to identify significant reaction force changes that may affect the stress state assessment. After re-obtaining the nodal reaction force response matrix, the direction and amplitude of the reaction force changes at each clamping support point in the current node are compared with the offset direction and offset intensity of the corresponding region of the previous node. The stress state description of the current node is updated by combining the actual material area removed, the cumulative removed thickness, and the remaining wall thickness. If the directions are consistent and the amplitude levels are the same or adjacent, the confidence level of the corresponding region is maintained or increased. If the directions are opposite or the intensity levels differ by more than two levels, the confidence level of the corresponding region is decreased, but a review is retained. Identification; When a processing area experiences a continuous shift in the reaction force reference value of adjacent clamping support points after material removal, and the shift direction is consistent with the predicted direction of the previous node, the shift trend intensity of that area is increased; When the shift direction of the reaction force reference value is opposite to the predicted direction of the previous node, that area is marked as an area requiring re-separation; When the change in the reaction force reference value is less than 3% of the basic reaction force, and two consecutive nodes meet this condition, the shift trend intensity of that area is reduced by one level; This 3% threshold follows the aforementioned node stability judgment boundary and is used to distinguish between normal clamping micro-motion and reaction force reference drift that has node update significance; For updating the stress state at nodes, a node state vector is established. The node state vector stores the state features extracted from the original acquisition sequence, including at least the region number, remaining wall thickness, cumulative removed thickness, clamping reaction force reference value, reaction force reference change, top support force reference value, cutting load characteristics, offset direction, offset intensity, and confidence level. The cutting load characteristics include at least one or more of the following: spindle load peak value, spindle load average value, feed axis current peak value, and feed axis current fluctuation amplitude. The clamping reaction force reference value and top support force reference value are determined by the smoothed sequence of the current node's stable phase. The updated stress state description and node state vector serve as inputs for subsequent deformation separation between the clamping cutting state and the loosening stable state, and provide a node-level data foundation for generating the next node's machining action. If data is missing at the current node, it is processed according to the duration and data type of the missing data. The number of consecutive missing values ​​for a single clamping support point does not exceed 1. When processing the reaction force data, interpolation is performed using effective sampling points before and after the clamping support point. The engineering basis for this 1s threshold is that, under sampling conditions above 500Hz, sufficient forward and backward trend information is still retained within 1s, which can be used for short-term missing data compensation and should not be directly judged as node data failure. If the clamping reaction force data is missing for more than 5s, the current node data of the clamping support point is marked as invalid, and extrapolation is performed based on the reaction force change trend of adjacent high-confidence clamping support points. This 5s threshold corresponds to the minimum stabilization waiting time after the aforementioned minor disturbance. Exceeding this time will affect the node reaction force benchmark judgment. If the spindle load data or feed axis current data is missing, the current node retains the stress state update, but cutting synchronization matching is not used in the subsequent deformation separation. Only the reaction force and position differences in the clamping cutting state and the loosening stable state are used for preliminary separation, and supplementary verification is performed after the working condition sensing data is recovered.

[0020] Specifically, such as Figure 5 As shown: After material removal and stress state update are completed at the current machining node, and if clamping reaction force, support force, cutting condition and component position data are available for comparison at the current node, component position changes and clamping reaction force changes are collected in both clamping cutting state and loosening stable state; the input information includes the current node state vector, the reaction force sequence and position sequence in the clamping cutting state, the reaction force reference and position deviation in the loosening stable state, the spindle load sequence, the feed axis current sequence and the tool interpolation position sequence; The clamping cutting state refers to the state where the component is under normal operating clamping force constraint, and the tool performs the current machining segment or maintains the same clamping force for node backtesting after machining is completed. The loosening and stabilizing state refers to the state where, after cutting stops, the clamping force of the flexible and controllable clamping unit is recovered to the low holding force, and the component position and reaction force reference are stabilized. The low holding force can be selected as 10% to 20% of the basic clamping force, but not less than 100N. When the component has a large self-weight or vertical or lateral clamping is used, the low holding force is not less than the minimum anti-detachment clamping force to keep the component from detaching from the support. The engineering basis for this is that when the low holding force is less than 100N, the component may loosen due to its own weight, inertia, or local springback. When the force exceeds 20% of the basic clamping force, external constraints may still significantly affect the free springback and stress release pattern after loosening, making it difficult to observe irreversible deformation. During clamping and cutting, the changes in component position and clamping reaction force are collected when the tool passes through a weakly rigid area. Component position changes can be obtained through laser displacement sensors, contact displacement gauges, machine tool built-in displacement monitoring data, or tool backtesting data. When it is impossible to safely arrange an external displacement measurement device during cutting, tool backtesting data or in-machine probe node backtesting data after cutting is paused are used as the position sequence. The position sequence and tool interpolation position are aligned with the same timestamp and written into a separate data table according to the machining area number. For each machining area, the clamping reaction force fluctuation, position fluctuation, spindle load fluctuation, and feed axis current fluctuation are time-aligned, and synchronous judgment is performed using time difference threshold and trend matching degree. When the clamping reaction force fluctuation, position fluctuation, and spindle load or feed axis current change at a certain position meet the synchronous judgment, and the change recovers to near the pre-cutting reference after the tool leaves, this part of the change is taken as a candidate component of tool deflection deformation. The component position recovers to near the pre-cutting reference, which means that the residual position deviation after the tool leaves does not exceed 20% of the peak displacement at that point, or does not exceed 0.01mm to 0.02mm. The engineering basis for this range is that recoverable tool deflection deformation will rebound significantly as the cutting force is removed. If the residual deviation still exceeds 20% of the peak displacement or exceeds 0.02mm, the deviation is more likely to include residual stress release or clamping state change components. 0.01mm to 0.02mm matches the aforementioned clamping alignment control accuracy and the contour accuracy requirements of titanium alloy thin-walled parts. The time difference threshold can be selected to be no greater than 0.2s, and the corresponding tool interpolation position deviation is within the allowable spatial matching range of the current machining area. The spatial matching range can be determined according to the length of the current machining area, feed rate, and sampling period. The engineering basis for this time difference threshold is that the elastic tool deflection response caused by cutting force is usually approximately synchronous with the change in cutting load. Under the conditions of kilohertz-level sampling and unified timestamp alignment, 0.2s can cover enough sampling points for trend judgment. Lag exceeding 0.2s is more likely to come from changes in clamping contact, measurement delay, or non-cutting synchronous disturbances. The trend matching degree can be represented by the normalized correlation coefficient between the clamping reaction force fluctuation sequence and the spindle load sequence, or between the clamping reaction force fluctuation sequence and the feed axis current sequence. A correlation coefficient of not less than 0.7 is considered high matching, between 0.4 and 0.7 is considered medium matching, and below 0.4 is considered low matching. 0.7 is used to confirm that the clamping reaction force fluctuation and the cutting load fluctuation have strong synchronicity. 0.4 is used to eliminate weak correlation fluctuations caused by random noise, loose supports, abnormal contact, or sensor drift. In the stable state after loosening the clamp, after cutting stops and the clamping force is recovered, a stable waiting time of 10s to 60s is maintained to collect the component position and clamping reaction force reference. The engineering basis for this waiting time is that the component undergoes a short-term elastic recovery and damping decay process after loosening. Less than 10s may not be stable, and more than 60s provides limited improvement in position stability and increases auxiliary time. The criteria for determining the stable state after loosening the clamp are that the component position change does not exceed 0.005mm to 0.01mm within 5 consecutive seconds, or the change in the clamping reaction force reference does not exceed 2% of the base reaction force. The 0.005mm to 0.01mm position change threshold is lower than the aforementioned 0.02mm clamping and alignment control threshold, making it suitable for judging short-term position stability after loosening the clamp. The 2% reaction force reference change threshold is consistent with the base clamping stability criterion and is used to eliminate normal contact micro-motion and sensor noise. If the component position deviates continuously from the nominal profile, and this deviation still manifests as a drift of the reaction force reference in the same direction during a short-term retest after the basic clamping force is restored, then this deviation is considered a candidate component of unrecoverable deformation caused by residual stress release. The candidate component of unrecoverable deformation is confirmed by consensus voting among multiple clamping support points. For the same processing area, if more than half of the effective clamping support points have the same drift direction of the reaction force reference and the corresponding position deviation direction is the same, then the area is determined to have formed effective stress release deformation. If the number of effective clamping support points is less than three, or if the trends among multiple effective clamping support points are dispersed, then the area is marked as a low-confidence separation area, and sampling is intensified at the next processing node. Using more than half as the voting criterion can suppress the influence of abnormal contact of a single clamping support point on deformation attribution. At least three effective clamping support points can provide basic spatial constraints for judging the local deformation direction of thin-walled components. During the separation process, the total deviation of the component is the offset of the same machining area relative to the nominal contour or nominal wall thickness under the current measurement state, and it is decomposed into a tool-deflection deformation component and a stress-relieving deformation component. The tool-deflection deformation component can be the difference between the peak displacement under clamping cutting conditions and the residual displacement after the tool leaves, or it can be the displacement fluctuation amplitude synchronized with the cutting load. The stress-relieving deformation component is the stable position deviation of the component relative to the nominal contour under the loosening stable state, and the same-direction reaction force reference drift after the basic clamping force is restored is used as the confirmation basis. For the total deviation of the same machining area, if the synchronous tool-deflection candidate component is consistent with the direction of the loosening stable deviation, it is allocated according to the time response characteristics. First, the displacement component that can be recovered after cutting stops is deducted as the tool-deflection deformation component, and the remaining stable deviation is used as the stress-relieving deformation component. If the two are in the same direction, the deviation is determined by the following criteria: Conversely, two components are retained to avoid merging deformations from different physical sources into a single compensation amount. If the lack of working condition sensing data prevents trend matching from being completed, a conservative separation strategy is adopted, confirming only the displacement that is fully recovered after cutting stops as the tool deformation component, and temporarily classifying the displacement that cannot be confirmed as recoverable as the component to be reviewed, and collecting it again at the next machining process node. If the top support force feedback of a certain conformal support unit deviates from the set value by more than 10% in the clamping cutting state and lasts for more than 3 seconds, it is determined that the corresponding top support point is abnormally fitted. The tool deformation component in this area is determined only based on the clamping reaction force and position fluctuation, and the feedback data of this top support point is not used. The 10% deviation threshold is used to identify poor fitting of conformal support units, execution lag, or obstructed action. The 3-second judgment condition is used to exclude short-term impact and transient adjustment error. The separation confidence level is jointly determined based on the trend matching degree level, the consensus voting results of multiple clamping support points, the position recovery criterion, and the status of the top support point. When the trend matching degree is high and the consensus vote passes, the separation confidence level is high; when the trend matching degree is medium or the number of effective clamping support points is close to three, the separation confidence level is medium; for machining areas corresponding to the aforementioned conservative separation, dispersed voting, or abnormal top support points, the separation confidence level is low, and a verification mark is retained. After the above processing, a separation result is generated for each machining area. The separation result includes at least the following fields: tool deformation component, stress relief deformation component, separation confidence level, matching degree level, and voting status. The above separation result serves as input for adjusting the conformal support top support amount, tool position offset amount, clamping force timing, and remaining machining allowance. Specifically, for example... Figure 6As shown: After deformation separation has been completed at the current machining node, and the tool deflection deformation component and stress relief deformation component of the main weak rigid area have obtained a medium confidence level or above, the compensation action for the next machining node is generated. The input information includes the tool deflection deformation component, stress relief deformation component, separation confidence level, matching degree level, voting status, current clamping force timing, current support force, remaining machining allowance, nominal toolpath, and the area to be removed in the next machining node for each machining area. The compensation action is divided into two categories according to the deformation source: for recoverable tool deflection deformation, the conformal support support amount and tool position offset amount are adjusted; for unrecoverable stress relief deformation, the clamping force timing and remaining machining allowance are adjusted. For the tool deflection component, the increment of the supporting force at the corresponding supporting point is determined based on the deflection direction, deflection width, local stiffness level, and distance from the supporting point to the machining area. For machining areas with a deflection width less than 0.01mm, the current supporting force is not changed, only the monitoring mark is retained. For machining areas with a deflection width between 0.01mm and 0.05mm, the supporting force is increased by 10N to 50N based on the current supporting force. For machining areas with a deflection width greater than 0.05mm, the supporting force is increased by 50N to 200N based on the current supporting force, and the tool position offset is set simultaneously. The engineering basis for the above numerical range is that 0.01... mm is below the clamping and alignment control threshold and the wall thickness tolerance control level, and can usually be treated as a monitoring quantity; 0.05 mm is close to the lower limit of the common wall thickness tolerance of titanium alloy thin-walled components, and needs to be corrected by both the top support force and the tool position; 10N to 50N is suitable for local support correction of slight tool deflection, and 50N to 200N is suitable for local support enhancement of more obvious tool deflection, and is still within the commonly used adjustment range of tens to hundreds of Newtons of conformal support unit; after the top support force increment is applied, the top support force feedback is kept stable and the abnormal contact judgment of the top support point is avoided to prevent the thin-walled area from being pushed out of the nominal shape or introducing new out-of-plane deformation; The tool position offset is applied along the normal to the component surface, with the offset direction opposite to the tool deflection direction, and is written into the toolpath correction segment of the corresponding machining area. The tool position offset can be selected as 60% to 90% of the tool deflection deformation component. Among them, for machining areas with high separation confidence, the offset ratio can be selected as 80% to 90%; for machining areas with medium separation confidence, the offset ratio can be selected as 60% to 70%; for machining areas with low separation confidence, a large tool position offset is not directly generated, but the cutting parameters are reduced and verified at the next machining operation node. The engineering basis for this ratio range is that part of the tool deflection deformation has been corrected by the conformal support top support amount, and the tool position offset should not be completely equal to the tool deflection deformation component. Using an offset ratio of 60% to 90% can retain a compensation margin on the basis of top support correction, and avoid the top support effect and tool position offset from exceeding the actual tool deflection amount after superposition. When the top support force increment is in the range of 50N to 200N, the tool position offset ratio is preferentially taken as the lower limit of the corresponding confidence range to reduce overcompensation. Risks: For stress-relieving deformation components, adjust the clamping force sequence and remaining machining allowance of the next machining step node according to the continuous deviation direction under the stable state of loose clamping; for machining areas with excess material deviation, when removing adjacent material in the next machining step node, reduce the current working clamping force by 5% to 20% to allow the area to have release space to return to the nominal contour, and retain or increase the subsequent removal allowance for the area; for machining areas with insufficient material deviation, increase the current working clamping force by 5% to 20% to limit the area from continuing to deviate towards the insufficient material side, and reduce the subsequent removal amount or retain the finishing allowance for the area; the engineering basis for the 5% to 20% clamping force adjustment range is that below 5%, the control effect on the subsequent deformation direction of weak rigid components is not obvious; above 20%, it is easy to disrupt the original clamping balance or introduce new local clamping deformation; this adjustment belongs to the working clamping force sequence correction of the next machining step node, and is not equivalent to the micro clamping force disturbance used for reaction force identification; The remaining machining allowance can be selected from 0.3mm to 0.8mm; for thin-walled areas with a wall thickness of 1.5mm to 2mm, the remaining machining allowance can be selected from 0.3mm to 0.5mm; for areas with a wall thickness of 2mm to 3mm or frame beam areas, the remaining machining allowance can be selected from 0.5mm to 0.8mm. The engineering basis for this range is that 0.3mm to 0.8mm is generally greater than the node-level residual deviation control threshold of 0.02mm to 0.03mm, which can cover the contour deviation caused by subsequent stress release, while not significantly increasing the final finishing removal load, avoiding excessive final removal amount from inducing significant stress release again. After the allowance is adjusted, a regional allowance table is generated. The regional allowance table should include at least the regional number, current remaining allowance, allowance increase or decrease, allowance adjustment reason, and next node execution status. The top support amount, tool position offset amount, clamping force sequence, and regional allowance are generated. After the table, convergence determination is performed. The input for convergence determination is the current node stress state description, the aforementioned deformation separation results, the adjusted clamping force timing, the toolpath correction segment, and the region allowance table. The node prediction rule uses the offset change of the same machining area in the historical nodes as the benchmark, converts the clamping force change, the top support force change, and the tool position offset into the corresponding direction correction amount, and superimposes it with the stress release trend caused by the expected removal amount to obtain the expected residual deviation after the remaining machining is completed. The expected residual deviation threshold can be selected from 0.02mm to 0.03mm, and 0.02mm is preferred for aerospace thin-walled titanium alloy components. The engineering basis for this threshold is that the final profile of titanium alloy thin-walled parts and frame-type integral structural parts is usually required to be no more than 0.1mm. The node-level control threshold should be significantly stricter than the final tolerance in order to reserve margin for subsequent cutting errors, measurement errors, and clamping recovery errors, and to be consistent with the 0.02mm clamping alignment control threshold. When the expected residual deviation does not exceed the corresponding threshold, the adjusted machining action is written into the execution queue of the next machining operation node. When the expected residual deviation exceeds the corresponding threshold, the clamping force timing, support amount, tool position offset, and area allowance table are readjusted, with the number of adjustments not exceeding three. The engineering basis for the upper limit of three adjustments is that three adjustments can cover the sequential correction of the three main adjustment quantities: support amount, clamping force timing, and allowance distribution. If the expected residual deviation still cannot be converged after more than three adjustments, it usually indicates that there is a measurement anomaly, contact anomaly, or local material stress anomaly at the current node. Continuing automatic iteration will have limited benefits and may amplify machining errors. If convergence is still not achieved after reaching the upper limit of the number of adjustments, a downgrade process is entered, reducing the cutting depth of the next machining operation node, retaining a thicker remaining machining allowance, reducing the tool position offset ratio, and marking the corresponding machining area as a manual review area. When the separation result has low confidence or equipment feedback anomalies, conservative machining is generated according to the anomaly type. Actions: If the tool separation matching degree is low, only a conservative support amount is used, which does not exceed 50% of the normal calculated value, and the tool position offset does not exceed 30% of the tool separation candidate amount; of which, 50% is used to retain part of the support suppression effect, and 30% is used to limit the tool position offset to a low proportion range that will not dominate the machining error; if the stress release voting state is dispersed, the reduction action of the allowance in this area is suspended, and only the allowance is allowed to be retained or increased to prevent undercutting due to misjudgment; if a certain conformal support unit feedback failure, the support compensation of the corresponding support point is stopped, and the compensation method is switched to a small proportion of tool position offset and reduction of single-layer cutting depth; if a certain clamping support point has low confidence and has not recovered for two consecutive machining process nodes, the clamping support point is downgraded from an active control point to a monitoring point, and the clamping force timing adjustment is undertaken by the adjacent high confidence clamping support point; two consecutive nodes as judgment conditions can eliminate the misjudgment of the control point state by single-node occasional noise or short-term contact anomalies; The execution confidence level is determined based on the separation confidence level, convergence determination result, anomaly flag, and number of adjustments. If the separation confidence level is high and convergence occurs on the first attempt, the execution confidence level is high. If the separation confidence level is medium or converges after two to three adjustments, the execution confidence level is medium. The processing regions corresponding to the aforementioned conservative processing actions and the processing regions that have not converged even after reaching the upper limit of the number of adjustments have a low execution confidence level, and a review flag is retained. After the above processing, a machining action package for the next machining process node is generated. The machining action package includes at least the following fields: node number, toolpath correction segment, tool position offset, conformal support top force curve, clamping force timing curve, region margin table, execution confidence level, and anomaly flag. The machining action package is used for execution of the next machining process node, enabling material removal, reaction force update, deformation separation, and differential compensation to form a continuous closed loop. Example 2: Based on Example 1, the specific application process of a simulation compensation control method for deformation during the machining of titanium alloy components is further explained: This study focuses on the cavity milling of thin-walled TC4 titanium alloy components for aerospace applications. Online identification and differentiated compensation are performed for recoverable tool deflection caused by cutting forces and unrecoverable deformation caused by residual stress release during machining. The titanium alloy component includes an outer frame, web, and several internal cavity reinforcing ribs. The blank is formed from a pre-stretched forging. The machining area includes two deep cavity regions, four frame beam intersection regions, and several web regions with relatively thin walls. The target wall thickness is 1.8mm to 2.5mm, the final profile requirement is no greater than 0.1mm, and a machining allowance of 0.4mm to 0.7mm is reserved before finishing. The machining equipment is a five-axis CNC machining center. The clamping system consists of eight flexible and controllable clamping units and four conformal support units. The flexible and controllable clamping units are distributed near the outer frame and frame beam intersection regions, while the conformal support units are arranged on the back side of the web and the weakest points of the open cavities. The rigid area is located on the back side. Force measuring elements, top support force feedback channels, spindle load acquisition channels, feed axis current acquisition channels, and position measurement channels are all connected to the same CNC system time reference. Before machining, a design coordinate system is established based on the nominal digital model of the component. The actual coordinates of four non-collinear reference positioning points are acquired through an in-machine probe, forming a transformation relationship between the design coordinate system, workpiece coordinate system, and machine tool coordinate system. When the measured deviation of any reference positioning point exceeds 0.02mm, the workpiece coordinate system is re-aligned. This control value is less than one-fifth of the final contour requirement, reducing the impact of initial alignment errors on subsequent stress identification and tool position offset. After coordinate alignment is completed, each clamping support point, top support point, and corresponding machining area are written into a point mapping table, ensuring that the subsequently acquired clamping reaction force, top support force, tool interpolation position, and component position offset can all be attributed to the defined machining area. Before the component enters the cavity material removal stage, the initial reaction force of eight clamping support points is identified. Based on the component wall thickness, local stiffness, and expected cutting load, the basic clamping force is set, with the basic clamping force near the web support points ranging from 600N to 900N, and the basic clamping force near the frame beam support points ranging from 1200N to 1800N. After the basic clamping force is applied, the initial clamping state is considered stable when the reaction force at each clamping support point fluctuates by no more than 2% of the corresponding basic clamping force within 3 consecutive seconds. Then, in the circumferential order of the component, a small clamping force disturbance was applied to each clamping support point sequentially. The disturbance amplitude was determined according to 5% to 15% of the corresponding foundation clamping force, and limited to between 30N and 300N. Only one clamping support point was disturbed at a time, while the remaining clamping support points remained in the foundation clamping state. After waiting for 5 to 30 seconds, the reaction force changes of all clamping support points were collected, thereby forming multiple reaction force response vectors, which were arranged in the order of disturbance to form a reaction force response matrix. Based on the reaction force response matrix, the component was divided into the outer frame area, web area, and frame beam area. Multiple stress zones, including the intersection zone and the edge zone of the deep cavity, are identified. When a stress zone exhibits a reaction force transfer in the same direction after disturbance at adjacent clamping support points, and the ratio of the reaction force change amplitude to the disturbance amplitude is high, this stress zone is marked as a high offset trend zone. When the reaction force change amplitude is small and the direction is unstable, this stress zone is marked as a low offset trend zone. In this machined object, the initial reaction force response of the first deep cavity edge region shows that after a slight disturbance, both adjacent clamping support points generate a reaction force transfer towards the inner side of the web. Therefore, [the following is a continuation of the previous sentence, which is missing from the original text]. The area marked as having a high offset trend on the less material side is shown; the reaction force transfer direction in the intersection area of ​​the second frame beam is consistent with the nominal margin side, therefore this area is marked as having a medium offset trend on the more material side; then the measured reaction force response matrix and the calculated reaction force response matrix are compared for deviation. When the calculated deviation is less than 5% of the foundation clamping force, the confidence level of the corresponding section is set to high; when the calculated deviation is between 5% and 15%, the confidence level of the corresponding section is set to medium; when the calculated deviation exceeds 15%, repeated disturbance is triggered or the corresponding area is marked as an area to be reviewed. After entering the material removal stage, instead of using a one-time fixed compensation toolpath, the machining process nodes are divided according to the cumulative material removal thickness. For general cavity areas, the cumulative material removal thickness between adjacent nodes is 3mm. For web areas with a wall thickness close to 1.5mm and deep cavity edge areas with a high initial offset trend, the node interval is 1mm to 2mm. For outer frame areas with high local stiffness and stable continuous node reaction force references, the node interval can be widened to 5mm. Before each node starts, the clamping force sequence and initial value of the top support force are set according to the stress state description of the previous node. During node machining, the clamping reaction force, top support force, spindle load, feed axis current, and tool interpolation position are collected in real time and written into the node data table according to the machining area. The node data table includes at least the following fields: node number, machining section number, material removal area, cumulative removal thickness, tool position sequence, clamping reaction force sequence, top support force sequence, spindle load sequence, and feed axis current sequence. The clamping reaction force and top support force... The sequence is smoothed using a sliding window of 0.1s to 0.5s to reduce the impact of high-frequency noise and occasional spikes on node judgment. After the first processing node removes approximately 2mm of material, the feed is paused while maintaining the basic clamping state, and the node reaction response is updated again. Since the change in the reaction reference in the edge region of the first deep cavity exceeds 5% of the basic reaction force, and the offset direction is further enhanced compared to the initial predicted direction, a complete perturbation is performed on all clamping support points. The re-obtained node reaction response matrix shows that the offset trend on the material-deficient side in the edge region of the deep cavity still exists, and the intensity has increased from medium to high. The change in the reaction reference in the frame beam intersection region is less than 3% of the basic reaction force, and two consecutive nodes remain stable. Therefore, the offset trend intensity in this region is reduced by one level. This generates the current node state vector, which includes at least the following fields: region number, cumulative removed thickness, remaining wall thickness, clamping reaction reference value, top support force reference value, cutting load characteristics, offset direction, offset intensity, and confidence level. After updating the stress state at the nodes, the deformation sources of the current nodes are separated. For the edge region of the first deep cavity, the peak value of the local position offset of the component under clamping cutting condition is 0.042 mm. The peak value of the spindle load and the fluctuation of the clamping reaction force appear synchronously within 0.2 s, and the normalized correlation coefficient between the clamping reaction force fluctuation sequence and the spindle load sequence is 0.76. After the tool leaves this region, the residual position deviation drops to 0.012 mm, which is below the control range of about 20% of the peak displacement. Therefore, the recoverable part is identified as a candidate component of tool deflection deformation. Subsequently, the clamping force is recovered to 1 / 3 of the basic clamping force. 5%, and not less than 100N, wait for the component to enter the loosened and stable state; when the position change is less than 0.01mm within 5 consecutive seconds, collect the loosened and stable position deviation; if there is still a stable position deviation of 0.018mm in this area, and the reaction force reference of adjacent clamping support points still drifts in the same direction after the basic clamping force is restored, then the stability deviation is judged as a candidate component of unrecoverable deformation caused by residual stress release; since the reaction force reference drift direction of more than half of the effective clamping support points in the same area is consistent, and the number of effective clamping support points is not less than three, the stress release deformation in this area is judged to be of medium to high confidence. When decomposing the total deviation in the same machining area, the difference between the peak displacement under clamping cutting conditions and the residual displacement after the tool leaves is taken as the tool deflection component, approximately 0.030 mm; the stable position deviation under loosening stable conditions, combined with the reaction force reference drift confirmation, is taken as the stress release deformation component, approximately 0.018 mm. Since both are in the same direction, the recoverable 0.030 mm is first taken as the tool deflection component and entered into the instant compensation channel, and then the stable 0.018 mm is taken as the stress release deformation component and entered into the node-level correction channel. When the deviation between the support force feedback and the set value of a certain support point exceeds 10% and lasts for more than 3 seconds during machining in this area, the feedback data of that support point is not used to participate in the tool deflection component confirmation, but instead, the clamping reaction force and position fluctuation are used for conservative separation. The compensation action for the next machining process node is generated based on the separation result. For the tool deflection component of approximately 0.030 mm... For areas with a diameter of mm, increase the current support force at the corresponding support point by 10N to 50N, and set the tool position offset along the normal direction of the component surface. Since the separation confidence level is high in this area, the tool position offset ratio is set to 80% to 90% of the tool deformation component. When the support force increment enters the range of 50N to 200N, the offset ratio is set to the lower limit of the corresponding confidence level range to avoid overcompensation caused by the superposition of support effect and tool position offset. For areas where the stress release deformation component is manifested as a material shortage side offset, increase the working clamping force of the corresponding clamping support point by 5% to 20% in the next machining process node to limit the area from continuing to offset towards the material shortage side and reduce the subsequent removal amount or retain the finishing allowance. For areas that are manifested as a material excess side offset, reduce the working clamping force of the corresponding clamping support point by 5% to 20%, and retain or increase the subsequent removal allowance to provide release space for regression to the nominal contour. Regarding the allocation of allowance, the remaining machining allowance in the edge area of ​​the first deep cavity is controlled within the range of 0.3mm to 0.5mm due to the wall thickness being close to 1.8mm and the tendency for material to shift to the side with less material. The area number, current remaining allowance, allowance reduction, reason for allowance adjustment, and execution status of the next node are recorded in the area allowance table. For the frame beam area, due to the large wall thickness and high local stiffness, the allowance can be set to 0.5mm to 0.8mm. The above allowance range is greater than the node-level residual deviation control threshold of 0.02mm to 0.03mm and will not significantly increase the final finishing removal load. After generating tables for top support, tool position offset, clamping force timing, and area allowance, convergence is determined. Using the offset changes in the same machining area from historical nodes as a benchmark, the clamping force changes, top support force changes, and tool position offset are converted into corresponding directional corrections, which are then superimposed with the stress release trend caused by the expected removal amount in the next node to obtain the expected residual deviation. When the expected residual deviation does not exceed 0.02mm to 0.03mm, the adjusted machining action is written into the execution queue of the next node. When the expected residual deviation exceeds this threshold, the clamping force timing, top support, tool position offset, and area allowance are readjusted. The number of adjustments does not exceed three. If convergence is still not achieved after three adjustments, the cutting depth of the next node is reduced, a thicker allowance is retained, the tool position offset ratio is reduced, and the corresponding area is marked as a manually reviewed area. When the tool separation matching degree in a certain area is low, only a conservative support amount is used, and the conservative support amount does not exceed 50% of the normal calculated value, and the tool position offset does not exceed 30% of the tool separation candidate amount; when the stress release voting state is dispersed, the margin reduction action in that area is suspended, and only the margin can be retained or increased; when a clamping support point is in a low confidence state for two consecutive machining process nodes, the clamping support point is downgraded from an active control point to a monitoring point, and the clamping force timing adjustment is undertaken by the adjacent high confidence clamping support point. After the above processing, a machining action package for the next machining process node is generated. The machining action package includes at least the following fields: node number, toolpath correction segment, tool position offset, conformal support top force curve, clamping force timing curve, region margin table, execution confidence level, and anomaly flag. The execution confidence level is determined based on the separation confidence level, convergence judgment result, anomaly flag, and number of adjustments. If the separation confidence level is high and convergence occurs on the first attempt, the execution confidence level is high. If the separation confidence level is medium or converges after two to three adjustments, the execution confidence level is medium. If there are conservative machining actions or convergence has not occurred even after reaching the maximum number of adjustments, the execution confidence level is low and a review flag is retained. This machining action package is used for execution of the next machining process node, ensuring that material removal, reaction force update, deformation separation, and differential compensation at each node form a continuous closed loop. It should be noted that this invention can be deployed on the device itself for embedded applications or run on a PC or other terminal with a user interface, thus meeting various hardware environments and usage requirements.

[0021] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented in software, the above embodiments can be implemented in whole or in part by a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions of the embodiments of this application are implemented in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted wirelessly or wiredly from one website, computer, server, or data center to another website, computer, server, or data center. Wired methods include optical fiber, twisted pair, coaxial cable, etc. Wireless methods include infrared, microwave, etc. Available media include any available media that can be accessed by a computer or data storage devices such as servers and data centers that contain one or more sets of available media. Available media can be magnetic media (floppy disks, hard disks, magnetic tapes), optical media (DVDs), or semiconductor media. Semiconductor media can be solid-state drives. The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0022] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for simulation compensation and control of deformation during the processing of titanium alloy components, characterized in that, include: S1. Clamp the titanium alloy component in a five-axis CNC machining center and establish the correspondence between the flexible and controllable clamping unit, the conformal support unit, the working condition sensing channel and the machining coordinates and time reference. S2. Before formal cutting, apply controllable clamping force changes to each clamping point, collect the reaction force response of each clamping point, and establish an initial stress state description of the current component under manufacturing based on the reaction force response. S3. Perform material removal according to the preset processing steps. Collect clamping reaction force, support force and cutting condition data at each processing step and update the stress state description of the current step based on the collected data. S4. Collect the changes in component position and clamping reaction force under clamping and cutting conditions and loosening and stabilizing conditions, and separate the tool deflection deformation caused by cutting force and the deformation caused by residual stress release. S5. Adjust the conformal support top support amount, tool position offset amount, clamping force timing and remaining machining allowance according to the deformation source after separation, and use the adjusted machining action as the basis for the execution of the next machining process node.

2. The method for simulation compensation and control of deformation during the processing of titanium alloy components according to claim 1, characterized in that, S1 includes: Based on the nominal digital model of the component and the measured coordinates of the reference positioning point, establish the transformation relationship between design coordinates, workpiece coordinates and machine tool coordinates, and preset the actual geometric coordinate recording field; The clamping support points and top support points are arranged in the weak rigidity area of ​​the component and its adjacent support positions; The CNC system interpolation clock is used to correlate clamping reaction force, support force, cutting condition data, and tool interpolation position. After zero-point and static load verification of the force measuring element, a point mapping table and data record entries are generated.

3. The method for simulation compensation and control of deformation during the processing of titanium alloy components according to claim 1, characterized in that, S2 includes: The foundation clamping force is set according to the component wall thickness, local stiffness, and cutting load; After the clamping reaction force stabilizes, clamping force disturbances with limited amplitude are applied sequentially to each clamping support point and the reaction force difference is collected. Arrange the reaction force response vectors in the order of disturbance to form the reaction force response matrix; The stress zones are divided by combining the a priori interval of residual stress and the point mapping table, and the deviation is evaluated based on the back-calculated reaction force response matrix to generate an initial stress state description with offset direction, offset intensity and confidence level.

4. The method for simulation compensation and control of deformation during the processing of titanium alloy components according to claim 3, characterized in that, After the clamping reaction force stabilizes, clamping force disturbances with limited amplitude are applied sequentially to each clamping support point, and the reaction force difference is collected, including: After the basic clamping state satisfies the reaction force stability criterion, the disturbance amplitude is determined according to the ratio of the basic clamping force and the amplitude is limited. Apply clamping force disturbance point by point according to the circumferential sequence of the component or the priority sequence of the weak rigid area, while keeping the remaining clamping support points in the basic clamping state; After the disturbance, the reaction forces before and after the disturbance are collected according to the stability waiting condition and the difference is processed to generate the reaction force response vector. Clamping support points that do not meet the stability waiting conditions are marked as points to be reviewed.

5. The method for simulation compensation and control of deformation during the processing of titanium alloy components according to claim 1, characterized in that, S3 includes: The machining process nodes are divided according to the machining depth, cavity characteristics, and stress-sensitive areas; Set the clamping force and initial support force of the current node according to the stress state description of the previous node; Collect clamping reaction force, support force, spindle load, feed axis current and tool interpolation position, and write them into the node data table; Choose between simplified or full perturbation based on changes in the reaction force reference, changes in the offset direction, and changes in the low-confidence region. The stress state description is updated based on the nodal reaction force response matrix, and a nodal state vector is formed.

6. The method for simulation compensation and control of deformation during the processing of titanium alloy components according to claim 5, characterized in that, Simplified or full perturbations are selected based on changes in the reaction force reference, the offset direction, and the low-confidence region, including: The first machining process node applies clamping force disturbance to all clamping support points; In subsequent processing steps, when the reaction force reference change between adjacent nodes is within a stable range and the offset direction has not changed, clamping force disturbance is only applied to the medium and high offset trend areas and their adjacent clamping support points. When there is a sudden change in the reaction force reference, a change in the offset direction, or an increase in the number of low-confidence stress sections, clamping force disturbance is applied to all clamping support points.

7. The method for simulation compensation and control of deformation during the processing of titanium alloy components according to claim 1, characterized in that, S4 includes: Synchronize the position fluctuations and clamping reaction force fluctuations with the cutting load in time and match their trends. After the clamping force is reduced to a low holding force and the clamping is in a stable state, the stable position deviation and reaction force reference drift are collected. Based on the degree of position recovery, the consistency voting results of multiple clamping support points, and the status of the top support point, the total deviation of the same processing area is decomposed, and the tool deformation component, stress release deformation component, separation confidence level, and verification mark are generated.

8. The method for simulation compensation and control of deformation during the processing of titanium alloy components according to claim 7, characterized in that, After the clamping force is reduced to a low holding force and the system enters a stable, unclamped state, the stable position deviation and reaction force reference drift are collected, including: After cutting stops, adjust the clamping force of each flexible and controllable clamping unit to a low holding force; After the component position and clamping reaction force benchmark meet the stability criteria, the loosening position deviation and clamping reaction force benchmark are collected; Restore the basic clamping force and perform a short-term retest; The continuous positional deviation and the drift of the reaction force in the same direction are used as the data for determining stress release deformation.

9. The method for simulation compensation and control of deformation during the processing of titanium alloy components according to claim 1, characterized in that, S5 includes: Based on the deflection direction, deflection amplitude, separation confidence level, and the corresponding position of the support point and the machining area, determine the support force increment and the normal tool position offset; Adjust the tool position offset ratio according to the increase in the support force; The clamping force timing and area margin table are modified based on the stress release offset direction; After the expected residual deviation convergence determination and low confidence anomaly handling, a machining action package is generated that records at least the toolpath correction segment, execution confidence, and anomaly flag.

10. The method for simulation compensation and control of deformation during the processing of titanium alloy components according to claim 9, characterized in that, The clamping force timing and area margin table are corrected based on the stress release offset direction, including: Identify excess material side offset or insufficient material side offset for stress relief deformation components; For areas with multiple material offsets, reduce the working clamping force of the next processing node and correct the clamping force timing to retain or increase the subsequent removal allowance. For areas with insufficient material, increase the working clamping force of the next processing node and correct the clamping force timing to reduce the subsequent removal amount or retain the finishing allowance. Generate a region balance table based on the balance adjustment amount, recording the region number, current remaining balance, and execution status.

Images