Fastener precision machining control system based on real-time working condition monitoring

Abstract

The application relates to the field of fastener machining control, and discloses a fastener precision machining control system based on real-time working condition monitoring, which comprises a macro control unit for establishing an absolute coordinate system to complete reference feed trajectory planning; when the macro control unit executes the reference feed trajectory, a micro control unit determines load mutation abnormalities according to multi-dimensional feedback data and generates intervention instructions; a radial driver responds to the instructions to execute reverse micro tool retraction, and a main shaft driver executes speed switching. After the tool retraction, the micro control unit controls the radial driver to perform unloading detection verification, generates handover state data set after confirming the end of the abnormalities. The macro control unit reconstructs the absolute coordinate system according to the data set, plans new forming amount to generate a transition section trajectory. The micro control unit controls the driver to remove the intervention state and restore the closed-loop machining according to the trajectory. The application realizes accurate intervention and smooth recovery when the load is mutated, and guarantees machining size precision.

Description

Technical Field

[0001] This invention relates to the field of fastener processing control, specifically a fastener precision processing control system based on real-time working condition monitoring. Background Technology

[0002] In the precision machining of fasteners, uneven workpiece material hardness or fluctuations in blank allowance often trigger transient load changes. Existing machining control systems typically respond to such anomalies by issuing a fixed amount of tool retraction command from the upper layer of the system. This approach results in a long control chain, leading to response delays, and the fixed retraction amount cannot match the actual overload conditions. This can easily result in excessive tool retraction damaging the machined surface, or insufficient tool retraction causing overload damage to mechanical transmission components.

[0003] Meanwhile, the occurrence of anomalies and stress release can cause transient deformation of the workpiece, resulting in a shift in its actual physical position. Existing systems typically blindly cut back to the original stop coordinates when resuming machining, ignoring the changes in the actual contact point, which can easily lead to secondary load anomalies.

[0004] Furthermore, existing systems often clear accumulated servo errors at the underlying level by resetting the internal coordinate system when handling abnormal interruptions and attempting to resume machining. This operation disrupts the global position reference of the machining process, making it difficult to guarantee the dimensional accuracy of the resumed machining. Moreover, the lack of a smooth motion fusion mechanism during the feed recovery process makes the actuator prone to mechanical rigidity impact when re-entering the original feed trajectory, affecting the final forming quality of the fastener. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a fastener precision machining control system based on real-time working condition monitoring. This system solves the problem that existing control systems can only trigger a shutdown alarm or perform a fixed amount of tool retraction when encountering sudden load changes, and cannot achieve refined active intervention and coordinate reference restoration under abnormal working conditions. This results in frequent machining interruptions, high scrap rates, and difficulty in ensuring dimensional accuracy during remachining.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solution: a fastener precision machining control system based on real-time working condition monitoring, including a macro control unit, a micro control unit, a radial driver, and a spindle driver; The macro control unit establishes an absolute coordinate system based on the externally imported comprehensive machining parameter set to complete the reference feed trajectory planning; During the execution of the baseline feed trajectory, the micro control unit determines the abnormal load change based on the acquired multidimensional feedback data and generates intervention control commands. While the radial drive responds to the intervention control command to perform reverse micro-retraction, the spindle drive performs spindle speed switching and enters the intervention state. After performing reverse micro-retraction, the micro control unit controls the radial drive to perform unloading detection verification, and generates a handover status dataset after confirming that the load change anomaly has ended based on the verification result of the unloading detection verification. The macro control unit reconstructs the absolute coordinate system based on the handover status dataset and plans the additional forming amount to generate the transition segment trajectory. The micro control unit controls the radial driver and spindle driver to de-interfere with each other based on the transition section trajectory and resumes the closed-loop precision machining process of the fastener.

[0007] This invention separates global trajectory planning and low-level working condition response through a macro-micro dual-layer control architecture. It utilizes the micro control unit to perform abnormal tool retraction and detection verification, and combines the coordinate system reconstruction and transition trajectory generation of the macro control unit to achieve machining intervention and smooth continuity under transient abnormal working conditions.

[0008] Furthermore, the control system also includes a dual-port memory, which provides a dual-buffered communication architecture between the macro control unit and the micro control unit for data interaction. The dual-buffered communication architecture includes two buffers and is configured with a submission sequence number and an active buffer flag to indicate the currently active buffer. The buffer not indicated by the active buffer flag is an inactive buffer, and the data to be interacted contains a check value. When either the macro-control unit or the micro-control unit acts as the data sender, it writes the data to be exchanged and the verification value into the inactive buffer and simultaneously updates the submission sequence number and the active buffer switching flag. When either the macro-control unit or the micro-control unit acts as the data receiver, it accepts the data to be exchanged after comparing the consistency between the active buffer flag and the submission sequence number before and after reading, and confirming that the verification is correct. This dual-buffered communication architecture avoids cross-unit read conflicts and ensures the timing consistency and integrity of data interaction between the macro-control unit and the micro-control unit.

[0009] Furthermore, the control system also includes a torque acquisition module, an angular position acquisition module, a position feedback module, and a process acquisition module for jointly acquiring multidimensional feedback data; the micro control unit also calls a fixed-coefficient finite impulse response slope estimator as a digital differential filter, performs a weighted multiplication and addition operation on multiple continuously acquired torque sample values ​​and the weighted coefficient sequence built into the fixed-coefficient finite impulse response slope estimator to output a real-time torque change rate with smooth characteristics, and takes the current cumulative angular position of the spindle acquired by the angular position acquisition module modulo one spindle rotation cycle to obtain the current spindle rotation phase, and uses the current spindle rotation phase to query the phase torque compensation table in the offline process database to obtain the phase torque compensation value; The microcontroller subtracts the real-time torque change rate from the phase torque compensation value; the difference is defined as the real-time torque change rate residual. The aforementioned digital differential filtering and phase compensation calculation can eliminate periodic torque fluctuations introduced by mechanical transmission and extract load mutation characteristics reflecting real abnormal operating conditions, thereby improving the accuracy of abnormal operating condition identification.

[0010] Furthermore, the macro control unit generates trajectory data containing verification values ​​and sends it to the trajectory FIFO buffer of the micro control unit; the micro control unit is equipped with a unique radial servo direction and pulse physical output interface and has a built-in trajectory FIFO buffer to output control commands; the position feedback module also synchronously outputs the actual position encoding value and encoder status word to the micro control unit to obtain the actual radial feed position, actual radial feed speed, actual radial feed acceleration, and radial feed axis following error; the spindle driver is equipped with a normal machining speed interface and an intervention speed selection interface, and receives commands from the micro control unit to execute speed switching; The control system also includes a safety control unit, which operates independently and performs a functional safety shutdown operation when pre-configured functional safety shutdown conditions are triggered. The safety control unit contains a functional safety interface, which is connected in series in the control system's underlying power failure link or the servo driver's safe torque shutdown link to respond to out-of-limit faults. By configuring independent underlying safety interfaces and intervention speed interfaces, control commands are directly output from the micro-level to the drive level, shortening the control link delay for abnormal responses.

[0011] Furthermore, the micro control unit is also used for: parallel monitoring of shutdown conditions, including process protection shutdown conditions and functional safety shutdown conditions; when the shutdown conditions are not triggered, feature comparison is performed using the real-time torque change rate residual, and load mutation anomalies are identified when all preset operating condition feature judgment conditions are met simultaneously. Set the abnormal entry hysteresis amount, and use the sum of the preset upper limit of the normal torque change rate residual and the abnormal entry hysteresis amount as the abnormal entry threshold. The criteria for determining various operating conditions include: the residual of the real-time torque change rate is greater than the abnormal entry threshold; the transient spindle torque extracted from multi-dimensional feedback data is higher than the preset effective contact lower limit; the spindle speed change rate extracted from multi-dimensional feedback data is lower than the preset stability judgment threshold; and the reference feed acceleration corresponding to the reference feed trajectory does not undergo a sudden change exceeding the preset tolerance value. These multi-dimensional operating condition criteria combine torque changes and kinematic parameters, filtering out normal acceleration / deceleration and regular fluctuations during the no-load phase, thus avoiding misjudgments of abnormal conditions.

[0012] Furthermore, after confirming the abnormal load change, the micro control unit is also used to: stop sending the reference pulse corresponding to the reference feed trajectory to the radial driver, and send the speed adjustment trigger level through the intervention speed selection interface of the spindle driver while freezing the trajectory transmission, so as to control the spindle speed of the spindle driver to drop to the preset intervention speed range; continuously extract the real-time torque change rate residual after the abnormality occurs, and calculate the torque change rate over-limit integral index based on the underlying homogeneous time base. A piecewise linear mapping is performed on the integral index of the torque change rate exceeding the limit to simultaneously generate the reverse micro-retraction distance and reverse retraction speed. Combined with the maximum allowable acceleration of the radial actuator, intervention control commands including a limited deceleration curve and a reverse retraction speed curve are planned and generated. The micro-control unit independently controls the radial actuator to perform physical avoidance actions based on the intervention control commands, driving the radial actuator to perform limited deceleration and reverse micro-retraction. The above control logic dynamically calculates the retraction parameters through the integral of the torque change rate exceeding the limit, avoiding damage to the machined surface caused by a fixed retraction amount or overload of mechanical components due to insufficient retraction.

[0013] Furthermore, the micro control unit is also used to: control the radial driver to maintain the current physical position after completing the reverse micro retraction and perform unloading hold; after completing unloading hold, maintain the spindle within the preset intervention speed range, and use a radial detection feed speed lower than the normal radial feed speed to control the radial driver to run in the forward direction, so that the thread rolling wheel of the controlled machining equipment approaches the workpiece again and performs unloading detection verification; The average of a preset number of filtered spindle torques is used to obtain the unloading reference torque. Real-time torque samples from the multi-dimensional feedback data are read according to the control sampling cycle of the microcontroller unit, and filtered and denoised to obtain the actual torque during the detection phase. The difference between the actual torque during the detection phase and the unloading reference torque is used as the actual torque recovery amplitude. When the actual torque recovery amplitude exceeds a preset contact recovery amplitude threshold, the physical contact state is re-established. Utilizing low-speed detection feed and torque recovery amplitude verification allows for repositioning of the physical contact point after anomaly resolution, avoiding contact position deviations caused by transient workpiece deformation.

[0014] Furthermore, the macro control unit is also used to: extract the handover state dataset, combine the product of the intervention pulse difference and the command pulse equivalent recorded in the handover state dataset to compensate and freeze the macro reference command coordinates of the internal interpolation control module built into the macro control unit at the moment of the handover state dataset, so as to obtain the corrected absolute reference coordinates; prohibit the clearing of the servo error accumulated at the system bottom layer by resetting the internal coordinate system, use the corrected absolute reference coordinates as the position reference value, use the radial command velocity and radial command acceleration recorded in the handover state dataset as the velocity reference value and acceleration reference value respectively, and calculate the handover following error with the corresponding actual radial feed position, actual radial feed velocity and actual radial feed acceleration. The handover following error includes the handover position following error, the handover velocity following error and the handover acceleration following error. When the handover position following error is less than a preset position allowable threshold, the handover speed following error is less than a preset speed allowable threshold, and the handover acceleration following error is less than a preset acceleration allowable threshold, subsequent automatic recovery planning is allowed. By extracting handover status data and compensating for frozen coordinates, system position reference alignment before and after abnormal interruption is achieved, providing accurate reference coordinates for subsequent automatic recovery processing.

[0015] Furthermore, the macro control unit is also used to: extract the handover status dataset to determine the lower bound and upper bound of the effective tooth profile forming amount estimation; and calculate the lower bound and upper bound of the planned new forming amount estimation based on the final target forming amount of the target fastener stored in the offline process database and the lower bound and upper bound of the effective tooth profile forming amount estimation. The upper bound of the estimated planned increase in forming quantity is extracted and compared with the maximum permissible increase in effective tooth profile forming quantity. When the estimated upper bound of the planned increase in forming quantity exceeds the maximum permissible increase in effective tooth profile forming quantity, it is determined that the workpiece to be processed has a risk of forming breakage, and the process protection shutdown condition is triggered. When the estimated upper bound of the planned increase in forming quantity does not exceed the maximum permissible increase in effective tooth profile forming quantity, the fault tolerance calculation is confirmed to have passed. By verifying the upper and lower bounds of the planned increase in forming quantity, potential workpiece forming breakage risks during subsequent processing can be identified in advance, avoiding workpiece damage caused by forcibly resuming processing.

[0016] Furthermore, the macro control unit is also used to: allocate additional forming quantity to the remaining unexecuted rolling passes based on the externally imported comprehensive processing parameter set, and perform iterative reallocation in conjunction with preset multi-dimensional physical and process constraints; based on the reallocated additional forming quantity, generate a transition trajectory that sequentially includes a hold-off segment, a low-speed approach segment, a forming entry segment, and a trajectory fusion segment; in the forming entry segment of the transition trajectory, perform a physical entry action based on the maximum feed acceleration allowed by preset multi-dimensional physical and process constraints, extract the position, velocity, and acceleration at the end of the forming entry segment as the starting boundary conditions of the interpolation algorithm, and simultaneously use the position, velocity, and acceleration at the corresponding moment of the reference feed trajectory as the ending boundary conditions, and call a high-order spline interpolation algorithm to calculate and generate a continuous interpolation curve, thereby completing the motion state transition in the trajectory fusion segment. The above-mentioned multi-segment transition trajectory planning and high-order spline interpolation enable the actuator to smoothly transition from the detection contact point to the reference feed trajectory, reducing the mechanical impact generated during re-entry.

[0017] This invention provides a fastener precision machining control system based on real-time working condition monitoring. It has the following beneficial effects: 1. This invention uses a micro-control unit to identify load abrupt changes through real-time torque change rate residuals, and dynamically plans limited deceleration and reverse micro-retraction curves based on the torque change rate over-limit integral index, allowing intervention commands to take effect independently at the underlying drive level. This control method achieves rapid response to abnormal operating conditions and dynamically outputs matching retraction parameters according to the overload state, avoiding excessive retraction damage to the machined surface caused by conventional fixed retraction amounts or insufficient retraction leading to overload damage to mechanical transmission components.

[0018] 2. This invention uses a micro-control unit to control the radial driver to perform low-speed probing feed after unloading and holding. It then compares the difference between the actual torque during the probing phase and the unloading reference torque to see if it exceeds the recovery threshold, thereby reconfirming the physical contact state. This probing and verification process compensates for the physical positional shift caused by transient workpiece deformation during anomalies and stress release, accurately pinpointing the true contact point for resuming machining and preventing secondary load anomalies caused by blind cutting.

[0019] 3. This invention extracts the handover state dataset from the macro control unit to compensate for the frozen macro reference command coordinates and completes the absolute coordinate system reconstruction. Based on boundary conditions, it calls a high-order spline interpolation algorithm to generate a multi-segment transition trajectory. This planning method achieves global position reference alignment before and after abnormal interruption without clearing underlying accumulated errors, ensuring machining dimensional accuracy. Simultaneously, it smoothly integrates the motion state into the original reference feed trajectory, eliminating the mechanical rigidity impact during recovery. Attached Figure Description

[0020] Figure 1This is a diagram of the overall architecture of the control system of the present invention; Figure 2 This is a flowchart of the control system of the present invention; Figure 3 This is a timing diagram for the real-time operating condition monitoring and anomaly detection of the present invention; Figure 4 This is a schematic diagram illustrating the principle of local abnormal freezing and microscopic intervention control of the present invention. Figure 5 This is a diagram showing the state transition between unloading detection verification and handover maintenance in this invention. Figure 6 This is a graph showing the transient torque response of the spindle in this invention. Figure 7 This is a radial actual feed position tracking curve diagram of the present invention; Figure 8 This is a graph showing the real-time torque change rate residual monitoring curve of the present invention. Detailed Implementation

[0021] The technical solutions in 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.

[0022] Reference Figure 1 The present invention provides a fastener precision machining control system based on real-time working condition monitoring, including a macro control unit, a micro control unit, a dual-port memory, a torque acquisition module, an angular position acquisition module, a position feedback module, a process acquisition module, a radial driver, a spindle driver, and a safety control unit.

[0023] The macro control unit is responsible for offline process data management, baseline feed trajectory planning, absolute coordinate system reconstruction, remaining forming amount calculation, and transition section trajectory generation. The macro control unit establishes an absolute coordinate system based on the comprehensive machining parameter set and relies on the micro control unit to proxy the underlying high-frequency pulse conversion. The macro control unit does not directly intervene in the underlying physical control actions of the radial driver. In specific implementations, the macro control unit can use a microcontroller or a programmable logic controller as its hardware execution platform.

[0024] The microcontroller unit is responsible for acquiring multi-dimensional feedback data, extracting torque change rate residuals, determining load abrupt changes, generating intervention control commands, latching the operating condition dataset (including abnormal and handover state datasets), triggering spindle intervention speed, calculating intervention pulse differences, and sending underlying drive control signals. The microcontroller unit has a unique radial servo direction and pulse physical output interface and a built-in trajectory FIFO buffer, coordinating all command positions and sending control signals to the radial driver. To meet the high-frequency parallel processing and transient response requirements of the underlying signals, the internal logic operations of the microcontroller unit can be implemented using field-programmable gate arrays (FPGAs) or field-programmable logic devices (FPGAs).

[0025] A dual-port memory is deployed between the macro-control unit and the micro-control unit to facilitate the exchange of status datasets, trajectory data, control commands, and acknowledgment information. The micro-control unit and macro-control unit utilize the dual-port memory to construct a dual-buffered communication architecture. This architecture includes two buffers, each configured with a submission sequence number and an active buffer flag indicating the currently active buffer. Buffers not indicated by the active buffer flag are inactive. The status datasets, trajectory data, control commands, and acknowledgment information used for exchange serve as the data to be exchanged, each containing a checksum for integrity verification. When the macro-control unit or micro-control unit acts as the data sender, it writes the data to be exchanged and the checksum to the inactive buffer. After synchronization, it updates the submission sequence number of the corresponding inactive buffer and atomically switches the active buffer flag. When the macro-control unit or micro-control unit acts as the data receiver, it first reads the active buffer flag and submission sequence number, then reads the corresponding active buffer according to the active buffer flag, and reads the active buffer flag and submission sequence number again after completion. Only when the results of the two reads are consistent and the verification is error-free is the receiver accepting the read data, preventing data tearing caused by asynchronous read / write operations.

[0026] The torque acquisition module, angular position acquisition module, position feedback module, and process acquisition module work together to acquire multi-dimensional industrial data, unifying the multi-dimensional feedback data to the microcontroller unit and establishing a unified underlying time base to meet the needs of high-frequency transient condition monitoring. The position feedback module synchronously outputs the actual position encoded value and encoder status word; the microcontroller unit, based on the unified underlying time base and combined with continuous actual position encoded values, obtains the actual radial feed position and performs differential and filtering processing to obtain the actual radial feed speed, actual radial feed acceleration, and radial feed axis following error. The spindle driver is equipped with a normal machining speed interface and an intervention speed selection interface, receiving commands from the microcontroller unit to execute speed switching. The safety control unit includes a functional safety interface, connected in series in the underlying power failure link of the control system or the servo driver safety torque shutdown link, responding to over-limit faults and executing functional safety shutdown operations.

[0027] Reference Figure 2 This invention provides a method for controlling the precision machining of fasteners based on real-time operating condition monitoring, comprising the following steps: S10. Establish an offline process database and perform machining trajectory initialization. The control system acquires the material constitutive relation of the target fastener to calculate the material flow stress, and converts the material flow stress into nominal spindle torque by combining the contact geometry parameters of the target fastener and the thread rolling wheel of the controlled machining equipment. Simultaneously, the control system establishes an offline process database based on the rolling calibration data obtained from the pre-executed normal rolling test. The offline process database includes a working condition reference mapping table, a phase torque compensation table, forming error limits, a nominal spindle torque mapping table, and an effective tooth profile forming amount mapping table. On this basis, the macro control unit establishes an absolute coordinate system based on the comprehensive machining parameter set, which includes the target size, the preset normal machining speed, the preset machining parameters, and the preset multi-dimensional physical and process constraints.

[0028] Subsequently, the macro control unit generates a reference feed trajectory based on the absolute coordinate system and discretizes it into trajectory data containing check values. The macro control unit uses a dual-port memory and a dual-buffered communication architecture to write trajectory data and check values ​​into the inactive buffer, update the submission sequence number and active buffer flag, and then send the data to the micro control unit's first-in-first-out trajectory buffer. The micro control unit uses an internal sampling counter as the underlying time base and allocates an absolute pulse counter that does not reset with each processing epoch and a relative pulse counter that evolves with each epoch.

[0029] S20. Synchronously monitor real-time operating conditions and perform graded anomaly judgment. During the execution of the reference feed trajectory, the micro control unit reads trajectory data frame by frame from the trajectory first-in-first-out buffer and converts it into pulse signals for release. At this time, the absolute pulse counter and the relative pulse counter accumulate and count synchronously with the pulse release. During this period, the micro control unit synchronously obtains multi-dimensional feedback data through the torque acquisition module, angular position acquisition module, position feedback module and process acquisition module, and obtains the operating status of the radial drive and spindle drive.

[0030] The micro control unit uses a digital differential filter combined with a phase torque compensation table to query the spindle phase determined by the angular position acquisition module. It calculates the continuous torque sampling values ​​in the multidimensional feedback data to extract the real-time torque change rate residual, and uses the real-time torque change rate residual for subsequent comparison of load mutation anomalies.

[0031] The microcontroller unit monitors preset shutdown conditions in parallel, including process protection shutdown conditions and functional safety shutdown conditions. If a process protection shutdown condition is triggered, the microcontroller unit stops executing the current trajectory and continues to output deceleration pulses according to the preset local controlled deceleration curve. Pulse output stops after the radial command speed drops to zero, and the automatic recovery prohibition flag is set. If a functional safety shutdown condition is triggered, the microcontroller unit or an independent safety detection link outputs a safety trigger signal to the safety control unit, which disconnects the underlying safety power-off link or triggers the servo driver's safety torque shutdown state. If the preset shutdown conditions are not triggered, the microcontroller unit compares and analyzes the acquired operating status with the real-time torque change rate residual.

[0032] The micro control unit determines whether there is a sudden load change based on the operating condition characteristics and sampling continuity. The operating condition characteristics include: the torque change rate residual exceeding the limit, the effective contact lower limit, the spindle speed stability, and the reference feed acceleration state, as determined by the operating condition reference mapping table.

[0033] S30. Latch the abnormal state dataset and perform micro-level local autonomous intervention. Once a load mutation anomaly is confirmed, the micro-control unit stops releasing the reference pulse corresponding to the reference feed trajectory at the end boundary of the current hardware pulse execution frame, latches the abnormal state dataset, and simultaneously triggers the spindle driver to reduce to the preset intervention speed range through the intervention speed selection interface. Subsequently, the micro-control unit synchronizes the abnormal state dataset to the macro-control unit through a dual-buffered communication architecture with dual-port memory to freeze the interpolation control module inside the macro-control unit.

[0034] Subsequently, the microcontroller detaches from the reference feed trajectory, calculates the torque mutation value based on the abnormal state dataset, and generates intervention control commands based on the mapping result of the torque mutation value. The microcontroller drives the radial driver to perform limited deceleration and reverse micro-retraction through the radial servo direction and pulse physical output interface, thereby generating an accumulated intervention pulse difference in the relative pulse counter.

[0035] S40. Perform unloading detection verification and establish handover holding state. After completing the micro-retraction action and entering the unloading holding state, the micro-control unit controls the thread rolling wheel of the controlled machining equipment to re-contact the workpiece at a radial detection feed speed lower than the normal radial feed speed. The micro-control unit calculates the actual torque recovery amplitude based on the real-time torque sampling value in the multi-dimensional feedback data to confirm the contact state between the thread rolling wheel and the workpiece. At the same time, the micro-control unit calculates the spindle through angle based on the data from the angular position acquisition module, and then confirms the end of the load sudden change anomaly after eliminating periodic repetitive anomalies based on the verification results of the unloading detection verification.

[0036] After confirmation, the microcontroller controls the radial driver to perform micro-retraction, ensuring the thread rolling wheel is in a flexible, light-contact position, thus reducing the system speed to within the preset handover speed range. Finally, the microcontroller latches the handover status dataset, which includes intervention pulse difference, coordinate offset, radial command and feedback motion status, and current operating condition partition identifier, and synchronizes the handover status dataset to the macrocontroller via a dual-port memory double-buffered communication architecture.

[0037] S50. Reconstruct the absolute coordinate system and plan the remaining forming amount boundary. The macro control unit compares the submission sequence number before and after the read operation via the dual-port memory. After verification and confirmation that the numbers match, the macro control unit obtains the handover state dataset and reconstructs the absolute coordinate system based on the intervention pulse difference. Simultaneously, the macro control unit performs position, velocity, and acceleration tracking error checks to prevent the actual position encoding value provided by the position feedback module from directly overwriting the underlying position command being executed by the micro control unit.

[0038] After reconstructing the absolute coordinate system, the macro control unit constructs upper and lower bounds for estimating the forming amount based on the forming error limit to solve for the planned additional forming amount. Finally, under the premise of satisfying the preset multidimensional physical and process constraints in the integrated processing parameter set, the macro control unit allocates the planned additional forming amount to the remaining rolling rings according to weights.

[0039] S60. Generate a recovery transition trajectory and close the loop for precision machining. Based on the reallocated planned additional forming amount, the macro control unit generates a transition trajectory that sequentially includes a hold-off segment, a low-speed approach segment, a forming cut-in segment, and a trajectory fusion segment. The macro control unit sets a new trajectory epoch and sends the transition trajectory and the new trajectory epoch to the micro control unit through a dual-port memory dual-buffered communication architecture.

[0040] After receiving the transition trajectory and the new trajectory epoch, the microcontroller unit performs a status verification, including the buffer capacity of the trajectory FIFO buffer. Upon successful verification, the microcontroller unit resets the relative pulse counter and the intervention pulse difference to zero, while maintaining continuous counting of the absolute pulse counter, and switches to the new epoch takeover state. After completing the epoch switch, the microcontroller unit sends a command through the normal machining speed interface to control the spindle driver to smoothly release the intervention speed. Subsequently, the control system completes the fastener rolling according to the transition trajectory, and at the end, saves the process traceability data and jointly determines that the machining closed loop is complete, thereby achieving precision machining of fasteners based on real-time working condition monitoring.

[0041] The following section will provide a detailed explanation of the operational details of each work stage, taking into account the control system architecture and overall workflow.

[0042] For step S10, which involves establishing an offline process database and performing machining trajectory initialization, the control system needs to complete the construction of the offline process database and the allocation of the underlying timing logic before actual operation. This includes a joint mapping operation from material physical mechanisms to underlying digital pulse counting, and includes the following sub-steps: S11. Execute the mapping between the material physics model and the nominal spindle torque. The control system acquires the Johnson-Cook constitutive parameters of the target fastener material. During conventional machining, the material flow stress is mainly determined by the yield strength, strain hardening effect, and strain rate strengthening effect. The control system, considering the target fastener material type, initial blank size, thread rolling wheel tooth profile and pitch of the controlled machining equipment, and the preset rolling process range, calculates the material flow stress under different equivalent plastic strain and equivalent strain rate conditions.

[0043] After acquiring the material flow stress, the control system converts the stress into a nominal spindle torque using plastic forming geometry. Based on the actual radial feed position, the control system determines the current indentation depth between the thread rolling wheel and the workpiece to calculate the instantaneous contact area. Then, combining the friction coefficient and the effective lever arm of the transmitted torque, the integral of the material flow stress over the instantaneous contact area is equivalent to radial and tangential resistance, thereby calculating the nominal spindle torque. The spindle speed and the actual radial feed rate are used together according to a preset calibration relationship to calculate the equivalent forming rate.

[0044] S12. Execution Cycle Disturbance Compensation and Joint Envelope Table Generation. The control system identifies mechanical dynamic parameters and measurement link dynamic parameters by combining normal rolling calibration test data. Multiple normal rolling tests are performed using standard material specimens, collecting spindle transient torque, spindle cumulative angular position, spindle speed, and actual radial feed position. In actual transmission systems, due to periodic disturbances such as thread rolling wheel runout, workpiece installation eccentricity, and gear meshing, the torque signal exhibits significant periodic physical fluctuations. To eliminate the interference of these periodic physical fluctuations on anomaly detection, the control system divides a complete spindle rotation cycle into multiple continuous phase intervals, extracts the average torque change rate of the same phase interval from multiple normal rolling tests, and establishes a phase torque compensation table.

[0045] The difference between the real-time torque change rate and the corresponding compensation item in the phase torque compensation table constitutes the torque change rate residual. The control system statistically analyzes the torque change rate residuals recorded in multiple normal rolling tests, extracts the mean and standard deviation of the residuals at each radial actual feed position and equivalent forming rate node, and determines the upper and lower limits of the normal torque change rate residual by adding or subtracting three times the standard deviation from the mean residual. This forms a working condition reference mapping table with the radial actual feed position and equivalent forming rate as addressing inputs and the upper and lower limits of the normal torque change rate residual as output boundaries. The working condition reference mapping table constructs a reference surface for the control system to determine whether the machining working condition is abnormal. The control system also uses the measured spindle torque obtained from the normal rolling calibration test to correct the nominal spindle torque calculated in step S11 for different working conditions, forming a nominal spindle torque mapping table with the radial actual feed position, equivalent forming rate, and machining temperature as inputs.

[0046] S13. Establishment of Forming Error Limits. The control system establishes forming error limits and writes the effective tooth profile forming amount mapping relationship and forming error limits together into the offline process database. The control system establishes the effective tooth profile forming amount mapping relationship based on the cross-sectional profile measurement results of the workpiece before and after rolling. The effective tooth profile forming amount is determined by the product of the target tooth profile filling area increased by the current thread cross-section of the workpiece relative to the initial blank cross-section and the axial measurement length. The control system extracts the forming amount calibration error limit based on the absolute value of the statistical maximum deviation between the measured values ​​and the nominal calculated values ​​of multiple calibration tests, and then determines the lower bound and upper bound of the effective tooth profile forming amount estimation. Specifically, the lower bound and upper bound of the effective tooth profile forming amount estimation are equal to the effective tooth profile forming amount estimation value minus and plus the forming amount calibration error limit corresponding to the current working condition zone, respectively. The control system thereby establishes the spatial boundary of fault tolerance assessment in the offline stage.

[0047] S14. Generate the reference trajectory and initialize the underlying counting architecture. After completing the offline process database construction, the macro control unit generates the reference feed trajectory based on the comprehensive machining parameter set and the established absolute coordinate system. The macro control unit discretizes the reference feed trajectory into multiple sets of trajectory data and sends these sets of trajectory data to the micro control unit. Each set of trajectory data contains underlying control elements, including trajectory epoch number, trajectory segment number, execution start condition, signed position increment, target command velocity, target command acceleration, and check value.

[0048] The microcontroller unit uses an internal hardware sampling counter as the underlying time base to establish a global counting architecture and a local epoch counting architecture, and allocates an absolute pulse counter that is not cleared with each machining epoch. The absolute pulse counter is used to continuously record all signed control pulses actually sent to the radial driver by the control system since the start of the current machining cycle. For the currently executing independent trajectory execution cycle, the microcontroller unit is configured with a relative pulse counter. The relative pulse counter is specifically divided into a reference received cumulative amount, a reference expiration cumulative amount, and a relative final output cumulative amount.

[0049] The baseline received cumulative amount records the cumulative value of signed baseline position increments that have passed integrity verification and been received by the microcontroller unit within the current trajectory epoch. The microcontroller unit also sets a first-in-first-out (FIFO) occupancy count and a continuous execution time value, representing the number of trajectory frames or bytes cached in the trajectory FIFO buffer, and the continuous execution duration that can be maintained at the current execution speed, respectively. Buffer level warnings are based on the FIFO occupancy count, remaining available capacity, and continuous execution time value. The baseline expiration cumulative amount accumulates the baseline position increments that have reached the execution opportunity at the current time according to the trajectory execution start conditions.

[0050] The microcontroller accumulates the signed control pulses that have actually been sent to the radial driver within the current independent trajectory execution cycle relative to the final output cumulative amount. The control system calibrates the command pulse equivalent based on the hardware physical resolution of the radial driver and servo motor encoder. The microcontroller calculates and generates the underlying radial command position in real time based on the product of the command pulse equivalent and the relative final output cumulative amount, superimposed with the reference coordinate origin of the current trajectory epoch.

[0051] Under normal processing conditions without load surges, the microcontroller outputs signals based on multiple sets of trajectory data from the macrocontroller, ensuring that the relative final output cumulative amount remains consistent with the baseline due cumulative amount. To enable independent low-level intervention calculations when facing load surges, the microcontroller defines an intervention pulse difference based on the baseline due cumulative amount and the relative final output cumulative amount. The value of the intervention pulse difference is equal to the difference between the relative final output cumulative amount and the baseline due cumulative amount within the current trajectory epoch. The microcontroller, relying on its internally configured intervention pulse difference mechanism, achieves real-time decoupling mapping between the macro-track reception progress and the underlying actual physical output progress, and converts the due baseline position increment into direction and pulse signals that meet the electrical parameter requirements of the radial driver for physical transmission.

[0052] Reference Figure 3Regarding step S20, which involves synchronously monitoring real-time operating conditions and performing graded anomaly judgment, during the execution of the baseline feed trajectory, the microcontroller unit needs to perform high-frequency acquisition and noise reduction of the underlying transient physical quantities, and distinguish between normal processing fluctuations and abnormal resistance increases caused by material hard points within a multi-dimensional constrained space. This includes the following sub-steps: S21. Perform multi-dimensional synchronous acquisition of operating conditions and digital filtering and noise reduction. The microcontroller reads trajectory data frame by frame from the trajectory first-in-first-out buffer and converts it into pulse signals for release. At this time, the absolute pulse counter and the relative pulse counter accumulate synchronously with the release of the pulses. During this period, the microcontroller simultaneously acquires multi-dimensional feedback data through the torque acquisition module, angular position acquisition module, position feedback module, and process acquisition module. The multi-dimensional feedback data specifically covers the spindle transient torque, spindle cumulative angular position, spindle speed, actual radial feed position, actual radial feed speed, radial feed axis following error, reference speed and reference acceleration issued by the macrocontroller, lubrication status of controlled components, and machining temperature.

[0053] Because the torque acquisition module generates electrical interference during operation, and the mechanical transmission system exhibits high-frequency micro-vibrations, directly performing single-point differential calculations using adjacent sampling points would amplify the noise amplitude, increasing the risk of high-frequency false alarms in the subsequent judgment system. The microcontroller unit utilizes a fixed-coefficient finite impulse response slope estimator as a digital differential filter. It performs a weighted multiplication-addition operation on multiple consecutively acquired torque samples with the weighted coefficient sequence built into the fixed-coefficient finite impulse response slope estimator to output a real-time torque change rate with smooth characteristics. The weighting coefficients in the digital differential filter are determined by the tap coefficients of a pre-designed offline finite impulse response band-limited differentiator. Based on the sampling frequency, the maximum operating speed of the spindle, the target abnormal characteristic frequency band, and the characteristic meshing frequency of the transmission chain, the control system determines the effective differential frequency band and the high-frequency cutoff frequency, ensuring that the band-limited differentiator approximates the differential response within the target abnormal characteristic frequency band and attenuating high-frequency components exceeding the upper limit of the physical transmission system response.

[0054] S22. Calculate the real-time torque change rate residual and extract the dynamic boundary. The microcontroller unit calculates the equivalent forming rate based on the current actual radial feed rate and the current spindle angular velocity, according to the offline calibrated forming rate conversion relationship. Simultaneously, it takes a modulus of the current spindle cumulative angular position over one spindle rotation cycle to obtain the current spindle rotation phase. Using the spindle rotation phase as an index, it queries the phase torque compensation table in the offline process database to obtain the phase torque compensation value corresponding to the current spindle rotation phase. The microcontroller unit subtracts the real-time torque change rate from the phase torque compensation value, and the difference between the two is defined as the real-time torque change rate residual.

[0055] The microcontroller uses the actual radial feed position and equivalent forming rate as coordinate variables to query the operating condition reference mapping table in the offline process database. It employs a bilinear interpolation algorithm, combined with data from four adjacent nodes surrounding the input coordinate point, to smoothly extract the upper and lower limits of the normal torque change rate residual corresponding to the current machining condition. When the queried actual radial feed position or equivalent forming rate exceeds the effective addressing range of the operating condition reference mapping table, the microcontroller prohibits extrapolation of data outside the table and outputs an over-limit alarm to the safety logic link. The extracted upper limit of the normal torque change rate residual provides a dynamically changing reference surface for subsequent identification of abnormal states.

[0056] S23. Perform safety and process-level protection shutdown monitoring. The microcontroller monitors preset shutdown conditions in parallel during each control cycle. The shutdown conditions are logically divided into process protection shutdown conditions and functional safety shutdown conditions, each corresponding to different physical disconnection links and recovery permissions.

[0057] Process protection shutdown conditions include: process condition parameters exceeding the effective mapping range of the operating condition reference mapping table; lubrication status of controlled components below the preset effective lubrication threshold; machining temperature exceeding the effective temperature range of the corresponding operating condition partition in the offline process database; synchronization time of the dual-buffered communication architecture between the macro control unit and the micro control unit exceeding the preset timeout threshold; cumulative micro-retraction distance or repeated detection count generated by the control system self-adjustment exceeding the process safety limit; discontinuous or failed verification of trajectory data sequence numbers issued by the macro control unit; and numerical overflow of the absolute pulse counter inside the micro control unit. The preset timeout threshold is set based on the maximum tolerable transmission delay period of the control system's underlying communication bus; the process safety limit is calibrated based on the maximum permissible idle wear of the thread rolling wheel and the mechanical interference limit of the machine tool structure; and the preset effective lubrication threshold is set based on the minimum rated flow rate of the equipment's lubrication system or the minimum effective oil film thickness standard provided by the supplier.

[0058] When any process protection shutdown condition is met, the micro control unit stops executing the original trajectory and instead generates and outputs a local controlled deceleration pulse, causing the radial drive to decelerate according to a deceleration curve that does not exceed the maximum allowable deceleration. After both the radial command speed and the actual feed speed drop to the preset stop threshold, the pulse output stops, and the automatic recovery prohibition flag is set in the global system status. The preset stop threshold is set based on the zero-speed detection dead zone range of the servo drive system or the minimum resolution speed equivalent of the encoder.

[0059] Functional safety shutdown conditions include: the transient torque of the spindle in the multi-dimensional feedback data exceeding the safe torque limit of the mechanical transmission chain; sensor saturation or disconnection in the torque acquisition module resulting in no feedback; encoder status word indication failure in the position feedback module; and overcurrent error or following error in the radial drive feedback exceeding the hardware allowable limit. The safe torque limit is determined by the physical yield strength of the mechanical transmission chain. When any functional safety shutdown condition is met, the microcontroller or macrocontroller outputs an emergency trigger request to the safety control unit. The safety control unit cuts off the drive power supply via a safety relay connected in series at the hardware level, or directly activates the safe torque shutdown interface between the radial drive and the spindle drive, executing a safety stop to prevent physical damage.

[0060] S24. Hysteresis determination of load mutation anomalies based on multidimensional physical constraints. Under the premise that all shutdown conditions are not triggered and the lubrication status and processing temperature are within the effective range of the corresponding operating condition reference mapping table, the microcontroller unit performs feature comparison on the real-time torque change rate residual. When all operating condition feature determination conditions are met simultaneously, a load mutation anomaly consistent with the characteristics of internal hard points or local structural abnormalities in the material is identified. The microcontroller unit sets the anomaly entry hysteresis amount and the anomaly release hysteresis amount respectively. The sum of the upper limit of the normal torque change rate residual output from the operating condition reference mapping table and the anomaly entry hysteresis amount is used as the anomaly entry threshold, and the difference between the upper limit of the normal torque change rate residual and the anomaly release hysteresis amount is used as the anomaly release threshold. The anomaly entry hysteresis amount and the anomaly release hysteresis amount are comprehensively calibrated based on the maximum amplitude of the white noise measured by the multidimensional sensor during no-load operation and the width of the natural drift band of the torque signal.

[0061] The specific conditions for determining operating conditions include: meeting the condition of torque change rate residual exceeding the limit as determined by the operating condition reference mapping table, i.e., the real-time torque change rate residual is greater than the abnormal entry threshold; the spindle transient torque is higher than the preset effective contact lower limit; the spindle speed change rate calculated based on the continuous differential of the spindle speed is lower than the preset stability judgment threshold to meet the spindle speed stability; and the reference feed acceleration issued by the macro control unit has not undergone a sudden change exceeding the preset tolerance value, i.e., it is in a normal reference feed acceleration state. Among them, the stability judgment threshold and the preset tolerance value are set according to the background fluctuation characteristics of the controlled machining equipment during no-load operation and the response bandwidth of the driver, to eliminate the coupling interference of the feed axis acceleration / deceleration or spindle speed adjustment process on the torque characteristics; the effective contact lower limit is determined by adding the measurement noise margin to the base torque during no-load operation.

[0062] To avoid high-frequency oscillations in the judgment results near the signal boundary, the microcontroller sets an anomaly entry threshold higher than the anomaly release threshold, forming a hysteresis interval. When all operating condition characteristic judgment conditions remain met within a consecutive preset sampling number, or within a consecutive preset rotation angle range of the continuously rotating spindle, the microcontroller finally confirms a load mutation anomaly. The preset sampling number is determined jointly by the sampling period of the microcontroller and the elapsed time corresponding to the minimum effective physical size of the hard point defect inside the target fastener; the consecutive preset rotation angle range is established based on the minimum physical circumferential span angle corresponding to a typical hard point defect inside the target fastener or the minimum meshing rotation angle of the mechanical transmission chain. After confirming the load mutation anomaly, the control flow transitions to a micro-level local autonomous intervention process.

[0063] Reference Figure 4 For the latched abnormal state dataset in step S30, and to perform micro-level local autonomous intervention, after confirming the abnormal load change, the control system enters the micro-level hardware-level control intervention process, relying on the underlying high-frequency response to isolate macro-level communication delays, including the following sub-steps: S31. Execution Trajectory Freeze and Abnormal Snapshot Latching. At the end boundary of the current hardware pulse execution frame, the microcontroller stops sending the reference pulse corresponding to the reference feed trajectory to the radial driver, and simultaneously extracts the current transient physical condition elements and packages them into an abnormal state dataset. The abnormal state dataset includes the absolute pulse counter value at the moment the abnormality occurred, the cumulative amount of reference reception, the cumulative amount of reference expiration, the cumulative amount of relative final output, the intervention pulse difference, the spindle transient torque, the spindle cumulative angular position, the actual radial feed position, and the actual radial feed speed.

[0064] Simultaneously with sending the freeze trajectory, the microcontroller sends a speed adjustment trigger level through the spindle driver's intervention speed selection interface, controlling the spindle speed to drop to within the preset intervention speed range. The microcontroller imposes constraints on the spindle speed based on safety physical boundaries, ensuring that the preset intervention speed is lower than the normal machining speed but not lower than the minimum operating limit. The minimum operating limit is obtained based on the spindle motor nameplate parameters and no-load friction torque calibration; the preset intervention speed is determined based on the low-speed limit in material machining mechanics to avoid physical impacts of hard points on the tool.

[0065] S32. Execution Event Double-Buffer Notification and Interpolation Control Freeze. The micro-control unit calls the double-buffered communication architecture of the dual-port memory to synchronize the abnormal state with the macro-control unit, writes the data block containing the abnormal state dataset and check value into the inactive buffer, and updates the commit sequence number and active buffer flag.

[0066] When the control cycle arrives, the macro control unit reads the corresponding active buffer according to the active buffer flag, and checks the active buffer flag and submission sequence number before and after reading. After verifying that the submission sequence number matches and the verification is correct, it extracts the abnormal state dataset, freezes the internal interpolation control module based on the actual radial feed position, suspends the issuance of new trajectory data, and returns a future trajectory failure command through a double-buffered communication architecture. After receiving the future trajectory failure command, the micro control unit clears the reference feed trajectory data that has not yet been executed in the internal trajectory first-in-first-out buffer.

[0067] S33. Calculate the torque mutation value and restricted intervention command. The microcontroller unit deviates from the macroscopic reference feed trajectory and continuously extracts the real-time torque change rate residual after the anomaly occurs. Based on the underlying homogeneous time base, it calculates the torque change rate over-limit integral index: ; in, The sampling period number that first meets the abnormal entry conditions; This refers to the sampling period number when generating intervention control commands; This is the integral index for the rate of change of torque exceeding the limit, and its dimension is torque; Representing the Real-time torque change rate residuals extracted from each sampling period; Representing the The abnormal entry threshold corresponding to each sampling period; The unit sampling period span representing the underlying homogeneous time base; This represents the function that takes the maximum value. This represents the out-of-limit difference after removing negative values.

[0068] Furthermore, the microcontroller unit measures the integral index of the torque change rate exceeding the limit. A piecewise linear mapping is performed to synchronously generate the reverse micro-retraction distance and reverse retraction speed. The specific mapping rule is as follows: when the torque change rate exceeding the limit integral index has not reached the saturation inflection point, the reverse micro-retraction distance and reverse retraction speed increase proportionally with the torque change rate exceeding the limit integral index, respectively, according to preset distance and speed proportionality coefficients; when the torque change rate exceeding the limit integral index exceeds the saturation inflection point, the micro-control unit clamps the reverse micro-retraction distance and reverse retraction speed to preset maximum allowable retraction distance and preset maximum allowable retraction speed, respectively.

[0069] The saturation inflection point is set based on the over-limit integral index value corresponding to the transient torque increment reaching the allowable torque margin of the mechanical transmission chain in normal and abnormal simulation tests. The preset maximum allowable retraction speed is calibrated by back-calculation based on the maximum disengagement torque of the radial driver. The preset maximum allowable retraction distance is determined based on the physical distance of the mechanical position corresponding to the critical point where the thread rolling wheel and the workpiece disengage from plastic contact. The micro-control unit combines the reverse retraction speed, the reverse micro-retraction distance, and the maximum allowable acceleration of the radial driver to plan and generate intervention control commands that include the limited deceleration curve and the reverse retraction speed curve.

[0070] S34. Perform local autonomous intervention and intervention pulse difference accumulation. The micro control unit converts the intervention control command into a low-level pulse signal, and independently controls the radial driver to perform physical avoidance actions through the radial servo direction and pulse physical output interface. The radial driver is driven to perform limited deceleration and reverse micro retraction. Specifically, limited deceleration is performed according to the limited deceleration curve to reduce the actual radial feed rate to 0; then, the radial driver is controlled to move in the reverse direction and reverse micro retraction distance according to the reverse retraction speed curve.

[0071] During micro-level autonomous intervention, the macro-control unit does not intervene with low-level commands. The cumulative relative final output in the relative pulse counter within the micro-control unit decreases and increases in real time with the reverse pulses generated by the reverse micro-retraction. Simultaneously, because the reference feed trajectory is frozen, the cumulative reference expiration in the relative pulse counter remains constant. The intervention pulse difference accumulates negatively with the physical retraction action, thus recording the actual physical distance deviation of the control system from the reference feed trajectory in the low-level logic, providing the underlying physical basis for subsequent handover status confirmation and absolute position reconstruction.

[0072] Reference Figure 5 Regarding step S40, which involves unloading detection verification and establishing a handover holding state, after completing the retraction action with microscopic local autonomous intervention, the control system confirms whether the abnormal resistance increase event of the hard point feature has physically ended and establishes a stable control handover boundary, including the following sub-steps: S41. Perform unloading hold and low-speed re-contact detection. After the micro control unit controls the radial driver to complete the reverse micro-retraction, it maintains the current physical position and performs unloading hold. The duration of unloading hold is set according to the stress release time constant of the mechanical transmission system. The stress release time constant is calibrated offline based on the comprehensive torsional stiffness and structural damping characteristics of the mechanical transmission system to fully dissipate the elastic strain energy accumulated in the mechanical structure due to abnormal load changes.

[0073] After unloading and holding, the microcontroller, while maintaining the spindle within a preset intervention speed range, controls the radial driver to run forward at a radial probe feed speed lower than the normal radial feed speed, causing the thread rolling wheel of the controlled machining equipment to reapproach the workpiece. Before the end of the unloading and holding phase, the microcontroller averages a preset number of filtered spindle torques to obtain the unloading reference torque. The preset number of consecutive torques is set based on the current spindle speed and the cutoff frequency of the digital differential filter to ensure that the averaging calculation window covers at least one complete mechanical vibration cycle. During the re-contact probe phase, the microcontroller synchronously reads the real-time torque sampling value from the multidimensional feedback data according to the microcontroller's control sampling cycle to perform filtering and noise reduction, obtaining the actual torque during the probe phase. The difference between the actual torque during the probe phase and the unloading reference torque is used as the actual torque recovery amplitude.

[0074] To prevent the thread rolling wheel from detaching from the workpiece and entering a dead loop during the detection process, the microcontroller unit synchronously monitors the cumulative detection distance during forward operation. If the cumulative detection distance exceeds a preset detection overtravel threshold, which is calibrated based on the maximum permissible mechanical deformation limit of the radial feed mechanism, and if no torque recovery is detected even after exceeding this limit, a detection failure alarm is triggered and the current machining cycle is terminated. Provided the cumulative detection distance does not exceed the limit, when the actual torque recovery exceeds a preset contact recovery threshold, the microcontroller unit confirms that the thread rolling wheel and the workpiece have re-established physical contact. The preset contact recovery threshold is calibrated based on the critical contact torque resistance generated when the thread rolling wheel just overcomes the static elastic deformation of the workpiece surface, thus eliminating background noise interference during no-load operation.

[0075] S42. Perform spindle passing angle calculation and periodic anti-deadlock verification. After confirming the re-establishment of physical contact, the microcontroller unit synchronously calculates the spindle passing angle based on the multi-dimensional feedback data provided by the angle position acquisition module, and retrieves the cumulative spindle angle position recorded in the abnormal state data to confirm the time when the abnormality occurred. Combined with the currently read cumulative spindle angle position, the spindle passing angle is calculated. The spindle passing angle is determined by the difference between the current cumulative spindle angle position and the cumulative spindle angle position at the time when the abnormality occurred.

[0076] The microcontroller compares the spindle clearance angle with a preset safety clearance angle margin. The safety clearance angle margin is determined based on the maximum possible distribution span of impurities and defects within the target fastener material and the tooth pitch characteristics of the thread rolling wheel. Simultaneously, the microcontroller extracts the abnormal occurrence phase intervals from the historical buffer and determines whether the phase corresponding to the current physical contact state is within the same periodic mapping interval as the historical abnormal phase, thus preventing misjudgments of the state during mechanical springback.

[0077] If the spindle clearance angle does not reach the preset safe clearance angle margin, the microcontroller maintains radial elastic light contact or preset low load contact state and controls the spindle to continue running at the intervention speed, while continuously monitoring the torque change rate residual and the abnormal occurrence phase; when the spindle clearance angle reaches the preset safe clearance angle margin, and the abnormal entry condition is not met again in the same or adjacent spindle phase interval, proceed to step S43; if the abnormal entry condition is met again before reaching the safe clearance angle margin, or if the spindle clearance angle, detection duration, or repeated detection number exceeds the corresponding preset upper limit and the release condition is still not met, then process protection shutdown is executed and the automatic recovery prohibition flag is set; only when the functional safety shutdown condition is met simultaneously is a safety trigger signal output to the safety control unit.

[0078] S43. Confirmation of Abnormal State Relief. After ruling out periodic recurring abnormalities and ensuring the spindle clearance angle meets the preset safety clearance angle margin, the microcontroller unit continues to monitor the real-time torque change rate residual. It is required that within consecutive preset anti-jitter sampling and detection cycles, the real-time torque change rate residual remains within the abnormal relief envelope range formed by the sum of the lower limit of the normal torque change rate residual output from the operating condition reference mapping table and the abnormal relief hysteresis, and the difference between the upper limit of the normal torque change rate residual and the abnormal relief hysteresis. The number of preset anti-jitter sampling and detection cycles is set based on the group delay characteristics of the torque filter and the state anti-jitter time requirements of the control system. When the condition of continuous no out-of-tolerance recording is met, and the actual torque recovery amplitude remains stable, the microcontroller unit confirms in the underlying control logic that the abnormal resistance increase region has ended in physical space.

[0079] S44. Establish and latch the non-plastic forming handover holding state dataset. After confirming the abnormal load change ends, the microcontroller controls the radial driver to perform micro-retraction, bringing the thread rolling wheel of the controlled processing equipment to an elastic light contact position. Simultaneously, the actual torque is controlled to decrease until it meets the following handover boundary condition: the actual torque is not greater than the difference between the lower limit of the plastic forming torque and the judgment hysteresis. The lower limit of the plastic forming torque is obtained based on the critical spindle torque calibration during the offline normal rolling calibration test when the material undergoes plastic yield deformation; the judgment hysteresis is set based on the background white noise amplitude of the torque acquisition module to prevent high-frequency fluctuations in the contact state judgment near the critical point.

[0080] While maintaining the thread rolling wheel in the elastic light contact position, the microcontroller performs dynamic synchronization to ensure that the absolute values ​​of both the radial command speed and the actual feed speed do not exceed the handover speed threshold; that the absolute values ​​of both the radial command acceleration and the actual feed acceleration do not exceed the handover acceleration threshold; and that the position tracking error does not exceed the absolute value of the difference between the radial command position and the actual feed position. The handover speed threshold and the handover acceleration threshold are set based on the allowable fluctuation range of extremely low-frequency mechanical oscillations when the controlled machining equipment is in an elastic light contact steady state; the handover position error threshold is calibrated based on the steady-state tracking dead zone error limit of the radial feed system servo position loop.

[0081] After all motion state parameters have entered the handover threshold range and remained stable, the microcontroller latches the handover state dataset. The handover state dataset includes the trajectory epoch number, current absolute pulse counter value, cumulative reference expiration amount, cumulative relative final output amount, intervention pulse difference, coordinate offset (i.e., the product of intervention pulse difference and command pulse equivalent), current radial command position, radial command velocity, radial command acceleration, actual radial feed position, actual radial feed velocity, actual radial feed acceleration, current spindle angle position, spindle speed, machining temperature, and current operating condition partition identifier. Upon receiving the handover state dataset, the macrocontroller queries the effective tooth profile forming amount mapping table to obtain the estimated effective tooth profile forming amount based on the radial command position, actual radial position, and current operating condition partition identifier. It then calculates the lower and upper bounds of the effective tooth profile forming amount estimate based on the forming error limits.

[0082] Subsequently, the micro control unit synchronizes the handover status dataset to the macro control unit through a dual-buffered communication architecture with a dual-port memory. Specifically, it writes the handover status dataset and the corresponding verification value into the inactive buffer, updates the submission sequence number and the active buffer flag, and waits for the macro control unit to take over and perform absolute coordinate system reconstruction.

[0083] For step S50, after reconstructing the absolute coordinate system and planning the remaining forming amount boundary, and establishing the non-plastic forming handover and holding state in the micro-control unit, the macro-control unit takes over control, corrects the macro-command reference based on the underlying physical deviation, and replans the subsequent processing path, including the following sub-steps: S51. Perform absolute coordinate anti-jump reconstruction and feedback following verification. The macro control unit compares the submission sequence number before and after the read operation via a dual-buffered communication architecture of a dual-port memory. After verifying that they match and are error-free, the handover status dataset is extracted. To prevent a position step jump in the servo control loop at the moment of takeover, the macro control unit must not directly overwrite the radial command position currently maintained by the micro control unit with the actual position encoding value provided by the position feedback module.

[0084] Because the microcontroller autonomously intervened to retract the tool at the lower level, the coordinate reference within the macrocontroller deviated from the actual physical position. The macrocontroller then executed an incremental compensation mechanism based on the intervention pulse difference centrally recorded in the handover state data. Specifically, it added the product of the intervention pulse difference and the command pulse equivalent to the macro reference command coordinates at the moment the internal interpolation control module was frozen, thus obtaining the corrected absolute reference coordinates. The command pulse equivalent was pre-calibrated based on the hardware configuration of the radial driver and the physical resolution of the encoder.

[0085] After calculating the corrected absolute reference coordinates, the macro control unit performs position, velocity, and acceleration tracking error verification.

[0086] The macro control unit prohibits clearing accumulated servo errors at the system's underlying level by resetting the internal coordinate system. Instead, it uses the corrected absolute reference coordinates as position reference values ​​and the radial command velocity and radial command acceleration recorded in the handover state data as velocity and acceleration reference values, respectively. These are used to calculate the handover tracking error against the corresponding actual radial feed position, actual radial feed velocity, and actual radial feed acceleration. Subsequent automatic recovery planning is only allowed when the handover position tracking error is less than a preset position allowable threshold, the handover velocity tracking error is less than a preset velocity allowable threshold, and the handover acceleration tracking error is less than a preset acceleration allowable threshold. The position allowable threshold, velocity allowable threshold, and acceleration allowable threshold are obtained through comprehensive calibration based on the maximum backlash of the mechanical transmission chain and the transient dynamic response limit of the radial driver, thus ensuring the underlying physical continuity during the coordinate system reconstruction process.

[0087] S52. Perform boundary calculation and fault tolerance verification for the remaining forming amount. The macro control unit determines the lower bound and upper bound of the effective tooth profile forming amount estimate based on the handover status dataset. The macro control unit constructs the upper and lower bounds of the forming amount estimate based on the forming error limit to solve for the planned new forming amount. Specifically, based on the final target forming amount and the lower and upper bounds of the effective tooth profile forming amount estimate stored in the offline process database, the lower and upper bounds of the planned new forming amount estimate are calculated according to the following formulas: ; ; in, To estimate the lower bound for the planned increase in forming volume; The upper bound for estimating the planned increase in forming volume; The final target forming quantity; To estimate the upper limit of the effective tooth profile shaping amount, To estimate the lower bound of the effective tooth profile shaping amount; This represents the function that takes the maximum value. If... If the amount of material produced exceeds the sum of the final target forming quantity and the allowable forming tolerance, an over-forming risk is identified, and a process protection shutdown is implemented. This constitutes the tolerance space for the planned additional forming quantity.

[0088] The macro control unit performs fault tolerance calculations within the tolerance space to assess the maximum theoretical deformation required to complete the remaining processing. Specifically, it extracts the estimated upper bound of the planned additional forming quantity as the evaluation benchmark. The maximum permissible increase in effective tooth profile forming amount corresponding to the current processing temperature, material constitutive relation, and remaining number of rolling passes. A comparison is then made. The maximum permissible increase in effective tooth profile forming amount is calculated jointly from material tensile test data, material constitutive model, and rolling contact geometry mapping relationship. A material forming limit mapping table is generated by correcting the limit deformation data from normal rolling tests. The macro-control unit queries this table using the current processing temperature and remaining rolling cycles as indexes to obtain the maximum permissible increase in effective tooth profile forming amount, ensuring it uses the same dimensions as the effective tooth profile forming amount. If the estimated upper limit of the planned increase in forming amount exceeds the maximum permissible increase in effective tooth profile forming amount, the current workpiece is deemed to have a forming fracture risk, thus triggering the process protection shutdown condition; conversely, if the limit is not exceeded, the fault tolerance calculation is confirmed to have passed.

[0089] S53. Perform remaining rolling ring weight reallocation and multi-dimensional constraint adjudication. After completing the fault tolerance calculation, the macro control unit adds forming amount to the remaining unexecuted rolling ring allocation plan based on the comprehensive machining parameter set. The macro control unit extracts the preset feed weight distribution of each rolling ring in the initial reference feed trajectory. The preset feed weight distribution is obtained offline based on the volume distribution rate of the target fastener tooth profile. Subsequently, iterative reallocation is performed in combination with preset multi-dimensional physical and process constraints. The preset multi-dimensional physical and process constraints cover the maximum allowable feed amount, maximum allowable feed speed, maximum allowable feed acceleration, spindle torque predicted according to the nominal spindle torque mapping table, torque change rate safety margin determined according to the working condition reference mapping table, target geometric tolerance of the target fastener, maximum allowable load allowed by the machine tool structure, remaining rolling ring number, and maximum penetration depth per ring.

[0090] In each iteration of the allocation calculation, the macro control unit checks whether the new forming amount allocated to an independent single ring triggers any of the preset multi-dimensional physical and process constraints. If the forming amount allocated to the current ring triggers the constraint boundary, the excess portion is distributed to subsequent rolling rings according to the preset weight distribution ratio of the remaining rolling rings.

[0091] If, after a complete traversal iteration, the forming amount allocated in the final ring still fails to meet the preset multidimensional physical and process constraints, or if the allocation operation encroaches on the polishing allowance of the final ring, it indicates that the current operating condition has exceeded the system's safe forming capacity. In the event of exceeding the system's safe forming capacity, the macro control unit outputs a defective product interception signal and safely terminates the current processing cycle to prevent mechanical overload of the equipment. Conversely, if the allocation results of all remaining rolling rings meet the preset multidimensional physical and process constraints, a trajectory update point set is generated, and the process proceeds to step S60 to execute the recovery transition trajectory generation, new epoch takeover, and processing loop closure.

[0092] For step S60, which involves generating and restoring the transition trajectory and completing the closed-loop precision machining process, the macro control unit and the micro control unit, after confirming that the coordinate reconstruction and forming plan are correct, collaboratively execute the smooth transfer of control and the restoration of the physical forming action. This includes the following sub-steps: S61. Generate a multi-stage recovery transition trajectory and issue a new epoch. The macro control unit generates a transition trajectory based on the reallocated planned new forming amount, which sequentially includes a hold release segment, a low-speed approach segment, a forming cut-in segment, and a trajectory fusion segment. Specifically, in the hold release segment, the macro control unit uses the current radial command position in the handover state dataset as the starting position of the recovery transition trajectory, maintains the continuity of position, velocity, and acceleration, and gradually releases the position hold state; since the intervention pulse difference has already been used for absolute coordinate reconstruction in step S51, displacement compensation equivalent to the intervention pulse difference is not performed in this stage; in the low-speed approach segment, the thread rolling wheel of the controlled processing equipment is guided from the current elastic light contact position to approach the plastic forming boundary again with a constrained radial feed speed, wherein the constrained radial feed speed is limited to a preset proportion of the normal feed rate based on the strain rate requirement of the material yield stage, and the preset proportion is obtained offline calibration based on the experimental data of the material's strain rate sensitivity coefficient, with a value range of 5% to 10%.

[0093] In the forming cut-in section, the macro control unit performs a physical cut-in action based on the maximum feed acceleration allowed by the preset multidimensional physical and process constraints, and assigns the planned additional forming amount to the current number of rolling revolutions. To ensure the smoothness of the trajectory connection, the position, velocity and acceleration at the end of the forming cut-in section are further extracted as the starting boundary conditions of the interpolation algorithm. At the same time, the position, velocity and acceleration at the corresponding moment of the reference feed trajectory are used as the ending boundary conditions. A high-order spline interpolation algorithm (specifically a fifth-order polynomial spline interpolation algorithm) is used to calculate and generate a continuous interpolation curve, thereby completing the motion state transition in the trajectory fusion section.

[0094] Subsequently, the macro control unit packages the calculated transition segment trajectory data blocks into a new trajectory epoch structure. This new trajectory epoch data structure contains eight elements: trajectory epoch number, segment number, execution start condition type, execution start condition value, position increment, command velocity, command acceleration, and underlying checksum. The execution start condition value can be the spindle phase, underlying sample count value, or the value corresponding to the takeover confirmation flag; its specific physical meaning is determined by the execution start condition type. After encapsulation, the data is sent to the micro control unit via a dual-buffered communication architecture using a dual-port memory. Trajectory data and checksum are written to the inactive buffer. After write synchronization, the commit sequence number is updated, and the active buffer flag is atomically switched to ensure the integrity of trajectory data transmission across processor platforms.

[0095] S62. Perform low-level buffer verification and new epoch physical takeover. After receiving the transition segment trajectory and the new trajectory epoch, the microcontroller first writes them into a communication temporary storage area independent of the trajectory FIFO buffer. After completing the data integrity verification, a multi-dimensional low-level state verification is performed before the takeover pulse output. The low-level state verification includes the following 8 items: The new trajectory epoch number is higher than the current failed trajectory epoch number; the trajectory segment sequence number is continuous and there are no duplicate segments; the trajectory data verification value is correct, and the submission sequence number and active buffer flag before and after reading are consistent; the execution start condition type and execution start condition value are valid and consistent with the underlying source time base; the available capacity and continuous execution time of the trajectory first-in-first-out buffer meet the initial pre-loading requirements; the handover position following error, handover speed following error, and handover acceleration following error are lower than the preset position allowable threshold, preset speed allowable threshold, and preset acceleration allowable threshold, respectively; the radial drive and spindle drive are in a ready and fault-free state, and the safety link of the safety control unit is in a permissible operating state; the spindle speed, contact torque, and handover holding state meet the new epoch takeover conditions.

[0096] When performing the verification of the remaining buffer space of the trajectory first-in-first-out buffer, the remaining buffer space of the trajectory first-in-first-out buffer is extracted to determine whether the remaining buffer space of the trajectory first-in-first-out buffer is sufficient to accommodate the initial data volume of the transition segment trajectory, so as to prevent the interruption of instruction issuance during pulse execution.

[0097] If the trajectory FIFO buffer's remaining capacity is insufficient to accommodate the initial data volume, the microcontroller maintains its current position hold state, does not submit new trajectory epochs to the trajectory FIFO buffer, and notifies the macrocontroller to resegment and send data in segments according to the allowed segment length or delay submission, while starting a preset waiting timer; it must not rely on the physical execution of expired old trajectories to release buffer space. The preset duration is calculated and determined based on the maximum tolerable transmission delay period of the communication bus and the servo pulse consumption rate. If the waiting timer expires before the trajectory FIFO buffer's remaining capacity is met, a buffer overflow exception is reported to the macrocontroller, and a process protection shutdown is executed.

[0098] After the status verification is passed, the microcontroller first copies the current trajectory epoch number, the cumulative amount of reference reception, the cumulative amount of reference expiration, the cumulative amount of relative final output, and the intervention pulse difference to the current anomaly event tracing snapshot. Then, when the initial data for the new trajectory has been pre-loaded, the driver is in a ready state, and the new epoch takeover conditions are met, the microcontroller sets the new trajectory epoch as the current execution epoch using an atomic state switch. At the atomic state switch boundary, it initializes the relative counter corresponding to the new trajectory epoch to zero and clears the online intervention pulse difference. Through this zeroing operation, the microcontroller resets the local coordinate deviation, realigning the relative counting reference within the microcontroller with the new trajectory epoch. Simultaneously, it maintains continuous counting of the absolute pulse counter to preserve the global position physical reference point since the start of this machining cycle. After configuring the internal counters, it switches to the new epoch takeover state, granting the driver the sole output authority to send position direction and pulse commands based on the transition segment trajectory takeover.

[0099] S63. Smoothly release the intervention speed and complete the precision machining closed loop. After completing the epoch switch, the microcontroller sends a command through the normal machining speed interface to control the spindle driver to smoothly release the intervention speed. The spindle driver controls the spindle speed to return to the normal machining speed corresponding to the comprehensive machining parameter set. To avoid torque shock to the transmission gear system caused by speed jumps, the spindle driver must perform speed recovery according to a preset angular acceleration. The preset angular acceleration should meet the following requirements: ; in, Represents the preset angular acceleration; The maximum angular acceleration allowed by the spindle drive; This represents the maximum permissible torque that the transmission gear system can withstand. The machining load torque is estimated based on the current filtered spindle torque. This provides a safety torque margin for gear transmissions. Represents the equivalent moment of inertia of the mechanical system, which is pre-calibrated offline. This represents the function that takes the maximum value. This represents the function that takes the minimum value.

[0100] Subsequently, the control system controls the radial drive to complete the forming operation of the remaining rolled ring of the fastener according to the transition section trajectory. When the final rolling operation of the fastener is completed and the tool is retracted to the safe origin, the microcontroller collects the state parameters of the entire machining process and saves process traceability data including torque mutation values, absolute pulse counter values, the intervention pulse difference before zeroing saved in the current abnormal event traceability snapshot, the extreme value of the spindle transient torque, the actual number of interventions triggering micro-retraction, and the estimated deviation of the final forming size calculated based on the final radial actual position, the effective tooth profile forming amount mapping table, and the forming error limit. The process traceability data is determined based on the underlying data element requirements for the target fastener lifecycle quality traceability.

[0101] The macro control unit and micro control unit jointly determine the completion of the machining closed loop when the following conditions are met simultaneously: all valid trajectory segments in the new trajectory epoch have been executed and there are no missing, repeated or failed trajectory segments; the radial drive has reached the safe origin according to the retraction trajectory and the spindle drive has entered the preset end state; the final forming dimension estimation deviation is within the target geometric tolerance range; there are no active process protection shutdown conditions or functional safety shutdown conditions; and the abnormal event traceability snapshot and process traceability data of this machining cycle have been completely saved.

[0102] Load mutation anomaly records that have been resolved and closed-loop recovery completed are not used as the sole basis for determining incomplete processing. After the processing loop is completed, the system automatically resets the global status flag and controls the multi-dimensional sensor components to enter standby monitoring mode, waiting to receive the processing start trigger signal for the next workpiece.

[0103] Reference Figures 6 to 8 This embodiment takes the precision rolling process of Ti-6Al-4V titanium alloy high-strength bolts (thread specification M10×1.25) commonly used in the aerospace field as an example. During the billet smelting and heat treatment process, titanium alloy materials are prone to localized high-hardness carbide segregation (i.e., physical hard spots). If forced feeding is applied during rolling, it can lead to tooth breakage of the thread rolling wheel or scrapping of the workpiece. The specific implementation process is as follows: The system undergoes offline preparation and underlying timing initialization. The control system acquires the Johnson-Cook constitutive parameters of the Ti-6Al-4V titanium alloy, with its yield strength benchmark set at 825 MPa. Based on the preset thread rolling wheel tooth profile parameters and the initial billet diameter of 8.95 mm, a transient contact area integral model for the machining process is established. Simultaneously, the system uses machining time (s) as a unified horizontal axis benchmark to standardize the data tone for subsequent monitoring, including the spindle transient torque (N·m), actual radial feed position (mm), and torque change rate residual (N·m / s). For the underlying space vector pulse width modulation control of the servo motor, those skilled in the art can configure general-purpose AC servo drive hardware for implementation. The control system performs 15 cycles of normal rolling without hard spots on a standard titanium alloy specimen, extracts the average value to generate a phase torque compensation table, and writes the calculated upper and lower bounds of the residuals at each process node into a dual-port memory. The microcontroller unit is internally configured with a 200 MHz hardware timer as the underlying homogeneous time base, with a sampling control period set to 100 microseconds.

[0104] High-frequency operating condition synchronous acquisition and dynamic boundary extraction. Around the 2.0-second mark of the rolling process, the microcontroller reads the reference feed trajectory frame by frame. Absolute and relative pulses accumulate synchronously with pulse release. The torque acquisition module acquires analog signals through a strain gauge sensor and performs filtering and noise reduction using a digital differential filter. The microcontroller queries the compensation table based on angular position feedback to obtain the actual residual value of the real-time residual condition. Based on previous data, it generates dynamic upper and lower limit boundaries for monitoring the state tolerance. The upper and lower limits together constitute a safety tolerance band, used to monitor positive abrupt changes caused by hard point obstruction and negative abrupt changes caused by material voids or tool chipping, respectively.

[0105] Local anomaly detection and microscopic intervention control. At 2.03 seconds, the thread rolling wheel physically contacts the hard spot inside the titanium alloy. Combined with... Figure 8 The real-time torque change rate residual monitoring curve showed that the actual residual value rose sharply within a short period of time and exceeded the dynamic upper and lower limits. The microcontroller simultaneously monitored other parameters such as spindle speed. After confirming the condition for three consecutive sampling cycles, it determined a load mutation anomaly at the 2.05-second abnormal intervention trigger point and stopped pulse release. The intervention command was extracted and sent to the driver to execute a 0.15mm reverse micro-retraction. Combined with... Figure 7 The actual radial feed position tracking curve shows that the actual position deviates from the linear upward trend of the original macroscopic reference trajectory after 2.05 seconds, and controlled deceleration and reversal are executed. The cumulative deviation of the intervention is stored in the register as the intervention pulse difference.

[0106] Stress unloading detection and absolute coordinate reconstruction. The microcontroller maintains the tool retraction position to unload strain energy, then controls the driver at 8% of the normal speed for re-contact detection. The spindle speed is synchronously reduced to the intervention speed. After confirming the reconstruction of physical contact and that the safe pass angle margin is met, elastic light contact is established. Once all handover speeds and error thresholds are met, the handover state dataset is latched. The macrocontroller reads the handover state, corrects the absolute reference coordinates using the intervention pulse difference, avoids command steps, and redistributes the remaining machining amount according to the principle of equal volume plastic forming.

[0107] Transition trajectory generation and processing closed loop. The macro control unit generates a fifth-order polynomial transition trajectory including hold-and-release segments, forming and cutting segments, etc. After the micro control unit verifies that the current handover allowable thresholds are met, it switches to the new epoch takeover state using atomic operations at the epoch takeover recovery time of 3.5 seconds. Figure 6 The spindle transient torque response curve shows that the actual torque of the system smoothly recovers after 3.5 seconds, and the radial feed seamlessly enters normal machining according to the replanned transition trajectory. After all remaining rolling passes are completed and the system returns to the origin, the closed-loop precision machining is confirmed to be complete.

[0108] Based on the design of this invention and the above-mentioned figures and data, the following conclusions are drawn: The control system enhances its ability to respond to local physical anomalies through intelligent decoupling. Combined with... Figure 8 Within an extremely short processing time (s), abnormal fluctuations exceeding the dynamic upper and lower limits are rapidly identified by the micro-control unit, which then directly stops releasing trajectory pulses at the moment of abnormal intervention triggering; traditional macroscopic system communication delays make it difficult to respond to such transient characteristics in a timely manner.

[0109] Combination Figure 7 It can be seen that after the intervention was triggered, the actual position did not blindly advance along the original macroscopic baseline trajectory, but rather severed the root cause of the mechanical displacement that led to the sharp increase in spindle torque, thus reflecting in Figure 6 In this process, the escalating actual torque is forcibly cut off and quickly drops back before reaching the dangerous threshold, successfully preventing physical damage to the machining tool.

[0110] At the epoch takeover and recovery moment at 3.5 seconds, the system uses register intervention pulse difference to perform compensation. After regaining macroscopic control, all motion dimensions maintained a high degree of continuity without any abrupt changes, and the system perfectly cut into the forming depth based on the principle of equal volume plasticity, verifying that the mechanism can smoothly escape from hard point interference and ensure the final forming quality.

[0111] While specific embodiments of the present invention have been described above, those skilled in the art should understand that these specific embodiments are merely illustrative. Those skilled in the art can omit, substitute, and modify the details of the above methods and systems in various ways without departing from the principles and essence of the present invention. For example, combining the above method steps to perform substantially the same function and achieve substantially the same result using substantially the same method falls within the scope of the present invention.

Claims

1. A fastener precision machining control system based on real-time working condition monitoring, characterized in that, include: Macro control unit, micro control unit, radial drive, and spindle drive; The macro control unit establishes an absolute coordinate system based on an externally imported set of comprehensive machining parameters to complete the reference feed trajectory planning; During the execution of the reference feed trajectory, the micro control unit determines the load mutation anomaly based on the acquired multidimensional feedback data and generates intervention control commands. While the radial driver responds to the intervention control command to perform reverse micro-retraction, the spindle driver performs spindle speed switching, causing the fastener precision machining control system to enter the intervention state. After performing the reverse micro-retraction, the micro control unit controls the radial driver to perform unloading detection verification, and generates a handover status dataset after confirming that the load mutation anomaly has ended based on the verification result of the unloading detection verification. The macro control unit reconstructs the absolute coordinate system based on the handover status dataset and plans the additional forming amount to generate the transition segment trajectory. The micro control unit controls the radial driver and the spindle driver to release the intervention state and resume the closed-loop precision machining process of the fastener based on the transition segment trajectory.

2. The fastener precision machining control system based on real-time working condition monitoring according to claim 1, characterized in that, It also includes a dual-port memory that provides a double-buffered communication architecture between the macro control unit and the micro control unit for data interaction; The dual-buffered communication architecture includes two buffers and is configured with a submission sequence number and an active buffer flag to indicate the current active buffer. Buffers not indicated by the active buffer flag are inactive buffers, and all data to be interacted with contains a check value. When the macro control unit or the micro control unit acts as the data sender, it writes the data to be interacted and the verification value into the inactive buffer and synchronously updates the submission sequence number and the flag for switching the active buffer. When the macro control unit or the micro control unit acts as the data receiver, it accepts the data to be interacted after comparing the consistency between the activity buffer flag before and after reading and the submission sequence number and verifying that there are no errors.

3. The fastener precision machining control system based on real-time working condition monitoring according to claim 2, characterized in that, It also includes a torque acquisition module, an angular position acquisition module, a position feedback module, and a process acquisition module for jointly acquiring the multidimensional feedback data; The micro control unit also calls a fixed-coefficient finite impulse response slope estimator as a digital differential filter to perform a weighted multiplication and addition operation on multiple continuously acquired torque sample values ​​and the weighted coefficient sequence built into the fixed-coefficient finite impulse response slope estimator, so as to output a real-time torque change rate with smooth characteristics. Simultaneously, the current cumulative angular position of the spindle acquired by the angular position acquisition module is moduloed by one spindle rotation cycle to obtain the current spindle rotation phase, and the phase torque compensation value is obtained by querying the phase torque compensation table in the offline process database using the current spindle rotation phase. The microcontroller subtracts the real-time torque change rate from the phase torque compensation value, and the difference between the two is defined as the real-time torque change rate residual.

4. The fastener precision machining control system based on real-time working condition monitoring according to claim 3, characterized in that, The macro control unit also generates trajectory data containing the verification value and sends it to the trajectory first-in-first-out buffer of the micro control unit; The micro control unit is internally configured with a radial servo direction and a pulse physical output interface, and has a built-in trajectory first-in-first-out buffer to output control commands. The position feedback module also synchronously outputs the actual position encoding value and encoder status word to the micro control unit to obtain the actual radial feed position, actual radial feed speed, actual radial feed acceleration and radial feed axis following error. The spindle driver is equipped with a normal machining speed interface and an intervention speed selection interface, and receives instructions from the micro control unit to execute the speed switching; It also includes a safety control unit that operates independently and performs a functional safety shutdown operation when a pre-configured functional safety shutdown condition is triggered. The safety control unit includes a functional safety interface that is connected in series in the underlying power-off link of the control system or the safe torque shutdown link of the servo driver to respond to out-of-limit faults.

5. The fastener precision machining control system based on real-time working condition monitoring according to claim 4, characterized in that, The microcontroller unit is also used for: Parallel monitoring of shutdown conditions, including process protection shutdown conditions and functional safety shutdown conditions; when the shutdown conditions are not triggered, feature comparison is performed using the real-time torque change rate residual, and the load mutation anomaly is identified when all preset operating condition feature judgment conditions are met simultaneously. Set the abnormal entry hysteresis amount, and use the sum of the preset upper limit of the normal torque change rate residual and the abnormal entry hysteresis amount as the abnormal entry threshold. The conditions for determining various operating conditions include: the residual of the real-time torque change rate is greater than the abnormal entry threshold; the transient torque of the spindle extracted from the multidimensional feedback data is higher than the preset effective contact lower limit; the spindle speed change rate extracted from the multidimensional feedback data is lower than the preset stability determination threshold; and the reference feed acceleration corresponding to the reference feed trajectory has not undergone a sudden change exceeding the preset tolerance value.

6. The fastener precision machining control system based on real-time working condition monitoring according to claim 5, characterized in that, The microcontroller unit, after confirming the abnormal load change, is also used for: Stop sending the reference pulse corresponding to the reference feed trajectory to the radial driver. While freezing the trajectory transmission, send the speed adjustment trigger level through the intervention speed selection interface of the spindle driver to control the spindle speed of the spindle driver to drop to the preset intervention speed range. The residual of the real-time torque change rate after the anomaly occurs is continuously extracted, and the torque change rate over-limit integral index is calculated based on the underlying homogeneous time base. A piecewise linear mapping is performed on the torque change rate exceeding the limit integral index to synchronously generate the reverse micro-retraction distance and reverse retraction speed. Combined with the maximum allowable acceleration of the radial driver, the intervention control command containing the limited deceleration curve and the reverse retraction speed curve is planned and generated. The micro control unit independently controls the radial driver to perform physical avoidance actions according to the intervention control command, and drives the radial driver to perform limited deceleration and reverse micro-retraction.

7. The fastener precision machining control system based on real-time working condition monitoring according to claim 6, characterized in that, The microcontroller unit is also used for: After the radial driver completes the reverse micro-retraction, it maintains the current physical position to perform unloading hold. After the unloading and holding are completed, the spindle is kept within the preset intervention speed range, and the radial drive is controlled to run in the forward direction using a radial detection feed speed lower than the normal radial feed speed, so that the thread rolling wheel of the controlled machining equipment approaches the workpiece again and performs the unloading detection verification. The average of the filtered spindle torques of a continuous preset number is taken to obtain the unloading reference torque. The real-time torque sampling value in the multi-dimensional feedback data is read according to the control sampling period of the micro control unit, and the filtering and noise reduction are performed to obtain the actual torque in the detection stage. The difference between the actual torque in the detection stage and the unloading reference torque is taken as the actual torque recovery amplitude. When the actual torque recovery exceeds the preset contact recovery threshold, the physical contact state is confirmed to be re-established.

8. The fastener precision machining control system based on real-time working condition monitoring according to claim 7, characterized in that, The macro control unit is also used for: Extract the handover state dataset, and combine the product of the intervention pulse difference and the command pulse equivalent recorded in the handover state dataset to compensate and freeze the macroscopic reference command coordinates of the internal interpolation control module built into the macroscopic control unit at the time of the correction to obtain the corrected absolute reference coordinates. It is prohibited to clear the servo error accumulated at the system's underlying layer by resetting the internal coordinate system. The corrected absolute reference coordinates are used as the position reference values. The radial command velocity and radial command acceleration recorded in the handover state data are used as the velocity reference values ​​and acceleration reference values, respectively. The handover following error is calculated with the corresponding actual radial feed position, actual radial feed velocity, and actual radial feed acceleration. The handover following error includes the handover position following error, the handover velocity following error, and the handover acceleration following error. When the handover position following error is less than a preset position allowable threshold, the handover speed following error is less than a preset speed allowable threshold, and the handover acceleration following error is less than a preset acceleration allowable threshold, subsequent automatic recovery planning is allowed.

9. The fastener precision machining control system based on real-time working condition monitoring according to claim 8, characterized in that, The macro control unit is also used for: Extract the handover status dataset to determine the lower bound and upper bound of the effective tooth profile forming amount estimation. Based on the final target forming amount of the target fastener stored in the offline process database and the lower bound and upper bound of the effective tooth profile forming amount estimation, calculate the lower bound and upper bound of the planned additional forming amount estimation. The upper bound of the estimated planned new forming amount is extracted as the evaluation benchmark and compared with the maximum allowable new effective tooth profile forming amount. When the upper bound of the estimated planned new forming amount exceeds the maximum allowable new effective tooth profile forming amount, it is determined that the current workpiece to be processed has a forming crack risk and the process protection shutdown condition is triggered. When the upper bound of the estimated planned new forming amount does not exceed the maximum allowable new effective tooth profile forming amount, the fault tolerance calculation is confirmed to have passed.

10. The fastener precision machining control system based on real-time working condition monitoring according to claim 9, characterized in that, The macro control unit is also used for: Based on the externally imported comprehensive processing parameter set, the planned new forming amount is allocated to the remaining unexecuted rolling rounds, and iterative reallocation is performed in combination with preset multi-dimensional physical and process constraints; based on the reallocated planned new forming amount, the transition segment trajectory is generated, which sequentially includes a hold release segment, a low-speed approach segment, a forming cut-in segment, and a trajectory fusion segment. In the forming cut-in segment of the transition trajectory, a physical cutting action is performed based on the maximum feed acceleration allowed by the preset multidimensional physical and process constraints. The position, velocity, and acceleration at the end of the forming cut-in segment are extracted as the starting boundary conditions of the interpolation algorithm. At the same time, the position, velocity, and acceleration at the corresponding moment of the reference feed trajectory are used as the ending boundary conditions. A high-order spline interpolation algorithm is called to calculate and generate a continuous interpolation curve, thereby completing the motion state transition in the trajectory fusion segment.

Images