An adaptive machining control system for elevator machine permanent magnets

Abstract

The application relates to the technical field of permanent magnet machining control, and discloses a self-adaptive machining control system for elevator traction machine permanent magnets, which comprises a multi-temperature-zone vacuum sintering furnace, a high-temperature-resistant radio frequency identification carrier, a numerical control machining center and a control system. The control system acquires temperature data in the sintering furnace to calculate the hardness deviation coefficient of the permanent magnet blank, and binds the radio frequency identification carrier. During numerical control machining, the control system maps the hardness deviation to the machining coordinate system in combination with the physical datum, analyzes the tool path to perform spatial interference comparison; for the abnormal hardness interference interval, the local feed rate and the spindle speed are reconstructed, and adaptive machining codes are generated to suppress the load impact caused by the hardness mutation. Meanwhile, the control system synchronously collects the actual transient cutting current of the spindle motor, calculates the dynamic load deviation rate, and reversely calibrates the hardness estimation model when the error exceeds the threshold value. The application realizes the closed-loop control of cross-process thermodynamic feedforward perception and entity mechanical load feedback iteration.

Description

Technical Field

[0001] This invention relates to the field of permanent magnet machining control technology, specifically an adaptive machining control system for permanent magnets in elevator traction machines. Background Technology

[0002] Elevator traction machines typically use high-performance permanent magnets as their core components. Their manufacturing process generally includes powder molding, vacuum sintering heat treatment, and subsequent CNC cutting or grinding. Because permanent magnet materials are inherently brittle, the machining process is highly sensitive to fluctuations in cutting forces.

[0003] In conventional production systems, CNC machining centers typically operate according to pre-programmed standard process codes, employing constant cutting feed rates and spindle speeds. However, during batch sintering and cooling in multi-zone vacuum sintering furnaces, the internal temperature gradient varies due to the water-cooling structure of the furnace wall and the stacking position of the material trays. This temperature gradient leads to inconsistent cooling rates within the same batch or even the same blank, resulting in discrete hardness distribution at the material's microscopic level, forming localized high-hardness regions. When the cutting tool cuts into these unknown localized high-hardness regions with constant parameters, the cutting resistance changes abruptly. This sharp increase in resistance causes mechanical vibration of the machine tool spindle, accelerating tool wear and easily causing chipping defects at the edges of brittle permanent magnets.

[0004] Currently, the sintering and machining processes in existing manufacturing systems are typically isolated from each other. The CNC system cannot obtain the spatial thermal history data generated in the upstream sintering process in advance, and cannot proactively adjust cutting parameters before the tool encounters a sudden change in hardness. Simultaneously, abnormal cutting load data exhibited in the machining process cannot be transmitted back to the sintering control unit, preventing the thermodynamic empirical model used to guide production at the sintering end from iteratively calibrating based on real physical feedback from the back end. This data fragmentation across processes makes it impossible for the system to adaptively intervene in material hardness fluctuations, making it difficult to guarantee the consistency of batch processing quality. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an adaptive machining control system for permanent magnets in elevator traction machines. This system solves the problems of uneven material hardness distribution caused by upstream sintering processes during the constant parameter machining of permanent magnets, which easily leads to sudden changes in cutting resistance and brittle chipping. Furthermore, it addresses the lack of data feedforward sensing and reverse calibration closed-loop mechanisms between processes.

[0006] To achieve the above objectives, the present invention provides the following technical solution: This invention provides an adaptive machining control system for permanent magnets in elevator traction machines, comprising a multi-zone vacuum sintering furnace, a high-temperature RFID carrier, a CNC machining center, and a control system. The multi-zone vacuum sintering furnace contains a multi-point thermocouple array for acquiring temperature data within the furnace chamber. The high-temperature RFID carrier is attached to a tray containing the permanent magnet blank. The CNC machining center includes a motion controller, a servo driver, and a spindle motor for machining the permanent magnet blank. The control system is communicatively connected to the multi-point thermocouple array, the high-temperature RFID carrier, and the CNC machining center. The control system acquires temperature data to calculate a hardness deviation coefficient and binds the hardness deviation coefficient to the high-temperature RFID carrier. During machining at the CNC machining center, the control system also analyzes the machining trajectory and reconstructs cutting parameters based on the hardness deviation coefficient, and calibrates the calculation model of the hardness deviation coefficient based on feedback from the actual transient cutting current of the spindle motor.

[0007] The control system acquires time and temperature sequence data from a multi-point thermocouple array and reconstructs the transient temperature response curves of each grid node within the material tray using an inverse distance weighted spatial interpolation algorithm. The control system then calculates the time derivative and time integral of the transient temperature response curves within a set cooling temperature range to obtain the local cooling rate integral for each grid node. The control system inputs the local cooling rate integral and the set ideal material cooling rate into a built-in hardness deviation coefficient calculation model to calculate the hardness deviation coefficient of the permanent magnet blank corresponding to each grid node. This invention utilizes a spatial interpolation algorithm to convert discrete temperature sensing signals into a continuous two-dimensional temperature field distribution, quantifying the non-uniform cooling gradient in the sintering process into specific discrete material hardness parameters, achieving a pre-mapping from thermodynamic processes to mechanical processing physical properties.

[0008] The control system generates a set of hardness deviation coefficients containing all grid nodes, obtains the unique identifier of the high-temperature RFID carrier, and establishes a mapping storage relationship between the unique identifier and the set of hardness deviation coefficients in the manufacturing execution system database. When the control system reads the unique identifier on the CNC machining center side, it issues the corresponding set of hardness deviation coefficients and initializes a two-dimensional traversal pointer in the internal memory. The two-dimensional traversal pointer is used to indicate the original storage coordinates of the permanent magnet blank to be processed in the two-dimensional array grid of the material tray.

[0009] The unmachined surface of the permanent magnet blank is formed with asymmetrical physical reference markings. The machine tool fixture of the CNC machining center is equipped with a mechanical limiting contour that matches the physical reference markings to constrain the clamping posture. The control system establishes a local workpiece coordinate system with the center of the mechanical limiting contour as the spatial reference origin, and maps the gradual change direction of the hardness deviation coefficient to the three-dimensional machining coordinate system of the CNC machining center to generate a hardness topology mapping table. This invention, through the physical reference markings formed by the forming mold and the mechanical limiting of the machine tool fixture, establishes a strict correspondence between the spatial posture of the blank on the machining platform and the posture within the sintering tray at the physical level, ensuring the consistency of the spatial reference for cross-process data mapping.

[0010] Before the spindle motor starts cutting, the control system parses the tool trajectory interpolation instructions in the original CNC machining program code to extract the tool target spatial coordinates. The control system performs spatial interference comparison between the tool target spatial coordinates and the hardness topology mapping table, extracting spatial coordinate regions where the hardness deviation coefficient exceeds the set normal physical threshold as abnormal hardness interference intervals. For abnormal hardness interference intervals, the control system calls the dynamic cutting parameter reconstruction algorithm to calculate the local reconstruction feed rate and local reconstruction spindle speed, generates local adaptive CNC machining program code, and sends it to the motion controller for execution. When calculating the local reconstruction feed rate, the dynamic cutting parameter reconstruction algorithm uses the process-preset standard feed rate minus the feed attenuation compensation value calculated based on the hardness deviation coefficient and the system-preset feed scaling sensitivity factor; when calculating the local reconstruction spindle speed, it uses the process-preset standard spindle speed plus the speed increase compensation value calculated based on the hardness deviation coefficient and the system-preset speed compensation factor.

[0011] The control system inserts a parameter smoothing transition code segment at the boundary between the retained conventional machining trajectory segment and the abnormal hardness interference range. Within this code segment, the control system employs a parametric linear interpolation algorithm to smoothly transition the actual feed rate and spindle speed to the locally reconstructed feed rate and spindle speed according to a set acceleration slope. Pre-execution of code interference comparison and generation of adaptive acceleration / deceleration codes avoids abrupt impacts when the cutting edge enters the high-hardness region, reducing the risk of brittle chipping.

[0012] During cutting operations in a CNC machining center, the control system continuously reads the actual transient cutting current of the spindle motor via a servo driver according to a set high-frequency sampling period. The control system inputs the currently executed cutting parameters into the built-in machine tool cutting mechanics model to calculate the theoretically expected cutting load. The control system compares the actual transient cutting current with the theoretically expected cutting load to determine the dynamic load deviation rate. The control system extracts the dynamic load deviation rate sequence of the tool across the entire abnormal hardness interference range and performs time integration on the sequence to obtain the cumulative error integral.

[0013] When the absolute value of the cumulative error integral exceeds the set error dead zone threshold, the control system determines that the calculation model of the hardness deviation coefficient has a systematic empirical bias and generates a reverse calibration network message to be transmitted back to the multi-temperature zone vacuum sintering furnace side. The control system transmits the cumulative error integral to the sintering control unit through the reverse calibration network message. The sintering control unit uses the product of the cross-process model learning rate and the cumulative error integral as a compensation amount to iteratively add the current initial empirical weight coefficients of the hardness deviation coefficient calculation model to obtain the updated model weight coefficients, and applies the updated model weight coefficients to the data analysis calculation of the next batch of permanent magnet blanks. This invention uses the physical and electrical parameters of machine tool processing to reverse verify the effectiveness of the upstream basic thermodynamic model, constructs a cross-process closed-loop iterative mechanism based on real mechanical load, and realizes the self-optimization of the overall manufacturing system parameters.

[0014] This invention provides an adaptive machining control system for permanent magnets in elevator traction machines. It offers the following advantages: 1. This invention collects temperature data inside the sintering furnace, uses a spatial interpolation algorithm to calculate the local cooling rate integral of each grid node of the permanent magnet blank, generates a hardness deviation coefficient, and combines the physical constraints of the RFID carrier and the machine tool fixture to map the hardness deviation to the three-dimensional machining coordinate system. This quantifies the temperature gradient difference generated by the upstream sintering process into spatially distributed hardness data, and ensures the consistency between the spatial orientation of the blank and the data matching when it is transferred to the machining center. This solves the problem of the unknown hardness distribution state inside the workpiece in conventional machining processes, and provides accurate basic data for subsequent adaptive machining.

[0015] 2. Before starting cutting, this invention extracts the target spatial coordinates of the tool and compares them with the hardness topology mapping table to locate the abnormal hardness interference range. Then, it calculates the local reconstruction feed rate and the local reconstruction spindle speed, and generates parameter smoothing transition code at the intersection of the normal trajectory and the interference range. This allows the system to identify local high hardness areas of the material in advance, actively reduce the feed rate and match the spindle speed before the tool enters the cutting area, avoids the surge in mechanical cutting force caused by the sudden change in material hardness under constant machining parameters, reduces the oscillation of the machine tool spindle, and reduces the probability of chipping defects in brittle permanent magnet materials.

[0016] 3. This invention synchronously collects the actual transient cutting current of the spindle motor during machining, calculates the dynamic load deviation rate between the actual value and the theoretically expected cutting load, calculates the cumulative error integral of the deviation rate, and sends back a reverse calibration network message when the integral value exceeds the error dead zone threshold. Iteratively updates the weight coefficient of the calculation model of the hardness deviation coefficient, transforming the CNC machining center into a physical verification node for detecting the accuracy of the upstream thermodynamic prediction model. Based on the actual machining load, the prediction parameters are objectively corrected, thus opening up a data feedback loop across processes and maintaining the accuracy of the calculation model in long-term, multi-batch production. Attached Figure Description

[0017] Figure 1 This is a diagram of the overall system architecture of the present invention; Figure 2 This is a flowchart illustrating the overall process of the method of the present invention. Figure 3 This is a two-dimensional mesh topology mapping diagram of the hardness deviation coefficient of the present invention; Figure 4 This is a comparison diagram of the actual transient cutting current of the spindle in conventional machining and adaptive machining according to the present invention. Detailed Implementation

[0018] 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.

[0019] See attached document Figure 1 The present invention provides an adaptive machining control system for permanent magnets of elevator traction machines, comprising: a multi-temperature zone vacuum sintering furnace, a high-temperature resistant radio frequency identification carrier, a CNC machining center, and a control system.

[0020] The multi-zone vacuum sintering furnace is equipped with a multi-point thermocouple array to acquire furnace chamber temperature data. The tray holding the permanent magnet blanks is attached with a high-temperature RFID carrier. The CNC machining center includes a motion controller, servo driver, and spindle motor. The control system is equipped with communication middleware, connecting the multi-zone vacuum sintering furnace's data interface to the CNC machining center's underlying communication interface via a manufacturing execution system network. The control system integrates the data flow between the heat treatment and machining processes.

[0021] See attached document Figure 2 The adaptive machining control system relies on the aforementioned hardware architecture to perform cross-process logical linkage, and the overall control process includes the following execution stages.

[0022] The system first performs sintering thermal history feature extraction and mapping. After powder pressing, the control system divides the material tray carrying the blank into a two-dimensional array grid, with each blank corresponding to a grid node. During the cooling stage of the multi-temperature zone vacuum sintering furnace, the control system collects time and temperature sequence data of the thermocouple array. The transient temperature response curve of each grid node is calculated through spatial interpolation. The control system calculates the cooling rate integral over a preset temperature range and outputs the hardness deviation coefficient of the grid node based on a hardness evaluation algorithm. The system establishes a database binding relationship between the set of hardness deviation coefficients of all grid nodes in the material tray and the unique identifier of the high-temperature RFID carrier.

[0023] The system then performs cross-process feature topology alignment and synchronization indexing. When the material tray flows to the CNC machining center station, the card reader reads the unique identifier of the high-temperature RFID carrier, the control system extracts the corresponding set of hardness deviation coefficients from the database, and initializes the mesh traversal pointer. The physical reference of the machine tool fixture with the permanent magnet blank defines the clamping posture. The control system establishes a local reference system based on this physical reference, mapping the hardness deviation coefficients of the mesh nodes to the three-dimensional machining coordinate system of the CNC machining center. The mesh traversal pointer steps according to the machine tool's cycle reset command to obtain the next blank parameters.

[0024] During the instruction execution phase, the system completes feedforward sensing and adaptive reconstruction of the machining code. Before the spindle motor starts cutting, the control system pre-reads and parses the interpolation instructions in the original CNC code to extract the tool target coordinates. The control system determines the hardness deviation coefficient corresponding to the target coordinates. When it is determined that the tool has entered a hardness deviation region, the system recalculates the cutting feed rate and spindle speed of this local cutting trajectory using the hardness deviation coefficient. The control system generates local adaptive CNC code containing reconstruction parameters and sends it to the motion controller for execution.

[0025] Finally, the system executes a reverse iterative calibration closed loop based on spindle load feedback. During the execution of the adaptive CNC code, the control system synchronously acquires the instantaneous cutting current of the spindle motor via the servo driver. The control system calculates the dynamic load deviation rate between the theoretically expected cutting load and the actual cutting current. The control system performs time integration on the dynamic load deviation rate, and when the integral value exceeds the set dead zone threshold, the system triggers a reverse calibration message. This calibration message is transmitted back via the network and used to adjust the initial weighting coefficients of the hardness assessment algorithm.

[0026] In the magnetic field press molding process, the inner wall of the forming mold has an asymmetrical geometric protrusion structure. After the powder material is pressed, a corresponding physical reference mark is formed on the non-machined surface of the permanent magnet blank. This physical reference mark is manifested as a single-sided chamfer or a positioning blind hole. A matrix of receiving slots is machined on the loading plane of the high-temperature resistant material tray. The control system divides the receiving slots into a two-dimensional array grid, defining each receiving slot as a grid node. The operating mechanism sequentially places the formed permanent magnet blanks into the receiving slots, ensuring that the physical reference marks of all permanent magnet blanks face the same reference direction of the material tray.

[0027] A tray loaded with permanent magnet blanks enters a multi-zone vacuum sintering furnace for sintering. The furnace chamber's inner wall is uniformly arranged with thermocouple arrays according to a spatial coordinate system. During the cooling phase of the sintering process, a spatial thermal history analysis module synchronously reads the temperature feedback signals from the thermocouple arrays via a programmable logic controller. The spatial thermal history analysis module acquires the time and temperature sequence data of each thermocouple node according to a set sampling period.

[0028] The spatial thermal history analysis module uses an inverse distance weighted spatial interpolation algorithm to process time and temperature sequence data. The control system acquires the three-dimensional spatial coordinates of the thermocouples and the two-dimensional planar coordinates of the material tray's location. The spatial thermal history analysis module calculates the Euclidean distance from the thermocouple temperature measurement point to each grid node. Using the reciprocal of the Euclidean distance as a weighting factor, the module maps the discrete thermocouple temperature data onto the two-dimensional array grid surface, reconstructing the transient temperature response curve corresponding to each grid node.

[0029] The space thermal history analysis module extracts the start and end times of the cooling interval according to the vacuum sintering process specifications. The module then calculates the absolute value of the time derivative of the transient temperature response curve within this time interval and performs a time integration operation to obtain the local cooling rate integral for that grid node. The expression for the local cooling rate integral is as follows: ; In the formula, Indicates grid coordinates as Integral of the local cooling rate at the location; Indicates the start time of the cooling temperature range; Indicates the end time of the cooling temperature range; Indicates grid coordinates as The transient temperature response curve at the location.

[0030] The system control module has a built-in calculation model for the hardness deviation coefficient. The system control module retrieves the standard ideal cooling rate of the material and inputs the integral of the local cooling rate into the calculation model for the hardness deviation coefficient to calculate the hardness deviation coefficient of the blank corresponding to that mesh node. The expression for calculating the hardness deviation coefficient is: ; In the formula, Indicates grid coordinates as Hardness deviation coefficient at the location; The standard ideal cooling rate for permanent magnet materials; This represents the initial empirical weighting coefficient of the calculation model for the hardness deviation coefficient.

[0031] The system control module traverses the two-dimensional array mesh to generate a set of hardness deviation coefficients containing all mesh nodes. A high-temperature resistant RFID carrier is fixedly installed on the side of the material tray. The system control module obtains the unique identifier of the high-temperature resistant RFID carrier through an RFID reader / writer. The system control module establishes a data table in the manufacturing execution system database, using the unique identifier as the primary key and mapping and storing the set of hardness deviation coefficients and the spatial vector data of the reference direction as related fields.

[0032] See attached document Figure 1 -Appendix Figure 2 The material trays that have completed the vacuum sintering process are transferred to the CNC machining center station via a conveyor mechanism. The CNC machining center is equipped with an RFID reader / writer terminal. The RFID reader / writer terminal scans the high-temperature resistant RFID carrier on the side of the material tray to obtain the corresponding unique identifier. The system control module sends a data request command containing this unique identifier to the database server through the manufacturing execution system network. The system control module receives the set of hardness deviation coefficients and the spatial vector data of the reference direction from the database server. The system control module allocates an independent index cache in its internal memory and stores the set of hardness deviation coefficients in the form of a two-dimensional array in the index cache.

[0033] The system control module establishes a two-dimensional traversal pointer within the index cache. This pointer indicates the original storage coordinates of the permanent magnet blank to be processed within the two-dimensional array grid of the material tray. Upon receiving the material tray positioning signal, the system control module initializes the two-dimensional traversal pointer, setting it to point to the starting coordinate node of the two-dimensional array. The starting coordinate node corresponds to the location of the first physical material within the material tray grid.

[0034] A machine tool fixture is mounted on the worktable of the CNC machining center. The positioning surface of the machine tool fixture is machined with a mechanical limiting profile that matches the asymmetric physical reference markings of the permanent magnet blank. The operating mechanism grasps the permanent magnet blank and moves it above the machine tool fixture to perform the clamping operation. The mechanical limiting profile of the machine tool fixture physically constrains the loading posture of the permanent magnet blank. Relying on this physical constraint, the permanent magnet blank can only fall into the machine tool fixture in a single preset posture. This mechanical constraint method ensures that the actual placement direction of the permanent magnet blank on the CNC machining center worktable is completely consistent with the initial reference direction set during the forming stage.

[0035] The system control module extracts the hardness deviation coefficient pointed to by the current two-dimensional traversal pointer. The system control module calls the workpiece coordinate system setting parameters from the CNC system. The system control module establishes a local workpiece coordinate system with the center of the mechanical limit contour on the machine tool fixture as the spatial reference origin. The system control module constructs a coordinate transformation matrix using the spatial vector data of the reference direction. Through the coordinate transformation matrix, the system control module maps the gradual hardness change direction inside the permanent magnet blank caused by the sintering temperature difference to the local workpiece coordinate system, generating a hardness topology mapping table aligned with the three-dimensional machining space.

[0036] The CNC machining center executes the cutting program for the permanent magnet blank according to the CNC code. After the cutting process is completed, the programmable logic control unit of the CNC system outputs auxiliary function code to control the release of the machine tool fixture. The auxiliary function code is represented by a preset machine tool action command signal. The system control module monitors the status variable of this auxiliary function code in real time. When the auxiliary function code is detected to be in the triggered state, the system control module determines that the current machining and unloading cycle of the permanent magnet blank has been completed.

[0037] The system control module drives a two-dimensional traversal pointer to perform step operations in the two-dimensional array according to a preset path sequence. The pointer steps to the next grid coordinate node and loads the hardness deviation coefficient corresponding to the next permanent magnet blank into the system control module. The system control module maintains synchronization between data parsing flow and the actual machining cycle by monitoring the state transitions of the underlying auxiliary function code.

[0038] Before the spindle motor starts executing the cutting action, the system control module performs pre-text parsing on the original CNC machining program code. The system control module scans and identifies tool trajectory interpolation instructions in the original CNC machining program code. These instructions include linear interpolation instructions for linear feed and circular interpolation instructions for surface machining. The system control module parses these tool trajectory interpolation instructions and extracts the associated tool target space coordinates. Combining the machine tool kinematics model and tool geometry parameters, the system control module constructs a theoretical tool motion envelope based on the tool target space coordinates.

[0039] See attached document Figure 2 The system control module performs a spatial interference comparison between the theoretical tool motion envelope and the hardness topology mapping table established in the preceding steps. The system control module traverses the hardness topology mapping table, extracts spatial coordinate regions where the hardness deviation coefficient exceeds a set normal physical threshold, and marks them as abnormal hardness interference intervals. When the spatial boundary of the theoretical tool motion envelope is determined to intersect this abnormal hardness interference interval, the system control module locks and extracts the preset CNC machining program code segment that triggers the spatial interference. For regular machining trajectory segments that do not overlap with abnormal hardness interference intervals, the system control module retains their original CNC machining program code content unchanged.

[0040] For the locked preset CNC machining program code segment, the system control module calls its internal dynamic cutting parameter reconstruction algorithm. The system control module extracts the local hardness deviation coefficient corresponding to the current interference position. The system control module obtains the preset standard feed rate and standard spindle speed for this type of material from the process database. Based on the local hardness deviation coefficient, the system control module performs correction calculations on the original machining parameters, calculating the local reconstruction feed rate and local reconstruction spindle speed. The calculation expressions for the local reconstruction feed rate and local reconstruction spindle speed are as follows: ; ; In the formula, Indicates the feed rate for local reconstruction; Indicates the spindle speed during partial reconfiguration; Indicates the standard feed rate preset in the process; Indicates the standard spindle speed preset for the process; Indicates the feed scaling sensitivity factor; Indicates the speed compensation factor; This represents the local hardness deviation coefficient corresponding to the tool trajectory in the current three-dimensional machining coordinate system.

[0041] The feed scaling sensitivity factor limits the control weight that limits the decrease in feed rate as material hardness increases. In machining engineering applications, reducing the single-tooth cutting thickness of the cutting edge by locally refactoring the feed rate is used to reduce the radial mechanical impact when the cutting edge enters the brittle, high-hardness material range. The spindle speed compensation factor is used to increase the spindle speed when the feed rate decreases. A moderate increase in spindle speed is used to maintain the cutting thermal balance in the cutting zone.

[0042] The system control module uses the calculated locally reconstructed feed rate and locally reconstructed spindle speed to replace the feed command characters and speed command characters in the original preset CNC machining program code segment. To avoid physical oscillations in the machine tool spindle mechanical structure caused by transient changes in cutting parameters, the system control module inserts a parameter smoothing transition code segment at the boundary between the normal machining trajectory segment and the abnormal hardness interference range. In the transition code segment, the system control module uses parameter linear interpolation to control the feed rate and spindle speed to smoothly transition to the locally reconstructed values ​​according to the set acceleration slope. The system control module compiles the modified and spliced ​​CNC machining program code into locally adaptive CNC machining program code and distributes it to the motion controller of the CNC machining center through a distributed CNC network to execute the physical machining.

[0043] During the execution of locally adaptive CNC machining program code in the CNC machining center, the system control module uses the machine tool's machining actuator as a physical probe node to verify the upstream thermodynamic prediction model. The system control module establishes a high-frequency data transmission link with the servo drive of the CNC machining center via an industrial communication bus. The servo drive monitors the electrical operating status of the spindle motor in a preset local area in real time. The system control module continuously reads the actual transient cutting current of the spindle motor during solid cutting from the servo drive according to a set high-frequency sampling period. The system control module simultaneously extracts the locally reconstructed feed rate, locally reconstructed spindle speed, and set constant depth of cut parameters for the currently executing program segment.

[0044] The system control module integrates a fundamental model of machine tool cutting mechanics. The module inputs the currently executed local reconfiguration feed rate and local reconfiguration spindle speed into this model to calculate the theoretically expected cutting load. The module then compares the actual transient cutting current with the theoretically expected cutting load to determine the dynamic load deviation rate during machining. The expression for the dynamic load deviation rate is: ; In the formula, The cutting time is indicated as Dynamic load deviation rate at that time; This represents the actual transient cutting current of the spindle motor; This represents the theoretically expected cutting load calculated based on the current cutting parameters.

[0045] In conventional machining environments, the current feedback of the spindle motor is mixed with high-frequency random mechanical noise caused by material microstructure inhomogeneities or chip friction. To eliminate the interference of such random noise on the system calibration logic, the system control module is configured with time integration processing logic and a dead zone determination mechanism. The system control module extracts the dynamic load deviation rate sequence of the tool throughout the entire abnormal hardness interference range and performs time integration on this sequence to obtain the cumulative error integral. The system control module retrieves the preset error dead zone threshold from the memory. The system control module calculates the absolute value of the cumulative error integral and compares this absolute value with the error dead zone threshold for determination.

[0046] When the absolute value of the cumulative error integral is less than or equal to the error dead zone threshold, the system control module determines that the current cutting resistance fluctuation is within a reasonable mechanical tolerance range and does not perform model intervention. When the absolute value of the cumulative error integral is greater than the error dead zone threshold, i.e., it satisfies... At that time, the system control module determined that the calculation model for the hardness deviation coefficient in the aforementioned sintering process stage had a systematic empirical bias.

[0047] In the formula, This represents the system-set error dead zone threshold. After confirming the existence of a systematic empirical deviation, the system control module initiates the reverse calibration procedure. The system control module encapsulates the accumulated error integral data into a reverse calibration network message according to a preset communication protocol. The system control module then sends the reverse calibration network message to the sintering control unit on the multi-temperature zone vacuum sintering furnace side via the manufacturing execution system network. The sintering control unit parses the message data and performs a reverse iterative calculation on the initial empirical weight coefficients in the hardness deviation coefficient calculation model based on the actual physical deviation. The iterative update expression for the weight coefficients is: ; In the formula, This represents the updated model weight coefficients; This represents the initial empirical weighting coefficients of the current hardness prediction model; This represents the cross-process model learning rate configured in the system.

[0048] The sintering control unit replaces the original initial empirical weighting coefficients with updated model weighting coefficients and applies the new parameter standards to the analytical calculation of the sintering thermal history of the next batch of permanent magnet blanks. This mechanism transforms the physical load of the CNC machining process into data correction and compensation for the upstream heat treatment process, realizing spontaneous optimization and closed-loop control of parameters throughout the entire manufacturing process.

[0049] Specific application examples: To further aid in understanding the technical solution of this invention, the following detailed description of the operation process and technical effects of this invention in a real industrial environment is provided, using a specific example of batch processing of permanent magnets for elevator traction machines and related comparative verification tests.

[0050] During the manufacturing process of a batch of permanent magnets for elevator traction machines, operators load arrayed, pressed blanks with physical reference markings into a high-temperature resistant tray. After processing and cooling in a multi-temperature zone vacuum sintering furnace, the control system calculates, based on data from a multi-point thermocouple array, that the local cooling rate integral of the grid coordinate region at the tray's edge is significantly higher than that in the center region. This leads to the calculation that the hardness deviation coefficient of the blanks in the edge region is positively high. This set of hardness deviation coefficients is bound to the high-temperature RFID carrier of the tray. When the tray reaches the CNC machining center station, the control system reads the unique identifier and downloads the corresponding data. Operators use the mechanical limit contour of the machine tool fixture and the physical reference markings of the blanks to complete the clamping. The control system establishes a local workpiece coordinate system and generates a hardness topology mapping table.

[0051] Before cutting begins, the control system parses the original CNC machining program code and determines that the tool cutting trajectory will cross the abnormal hardness interference range of the workpiece between the 4th and 6th second. (See attached...) Figure 3The control system inserts a parameter smoothing transition code segment at the boundary between the normal machining trajectory segment and the abnormal hardness interference range. Within the transition range of 4.0 seconds to 4.5 seconds, the actual feed rate smoothly decreases from the standard preset value to the locally reconstructed feed rate according to the set acceleration slope, while the actual spindle speed smoothly increases to the locally reconstructed spindle speed. This reconstruction parameter remains stable within the interference range, and after the tool exits this range at 6.0 seconds, the parameter smoothly returns to the standard state.

[0052] To verify the actual mechanical effect of the above adaptive control method, a comparative verification experiment was conducted. The experiment selected blanks from the same sintering batch located in the high-hardness region at the edge of the sintering tray as test samples, dividing them into a control group and an experimental group. The control group underwent conventional processing with constant standard parameters throughout the process, while the experimental group underwent adaptive processing using the parameter reconstruction method provided by this invention. (See attached diagram.) Figure 4 During the test, the control system continuously reads the actual transient cutting current of the spindle motor through the servo driver according to the set high-frequency sampling period.

[0053] Test data showed that when the control group entered the abnormal hardness interference range, the actual transient cutting current exhibited severe high-frequency pulse fluctuations, indicating that the machine tool spindle encountered a sudden impact of cutting resistance. Post-machining statistics revealed several defective parts with chipped edges in the control group. In contrast, when the experimental group passed through the same abnormal hardness interference range, due to the system's prior intervention of feed rate attenuation and speed increase, the actual transient cutting current showed a smooth load transition without destructive impact peaks. The machined parts had intact surfaces without chipping. Simultaneously, the control system compared the actual transient cutting current of the experimental group with the theoretically expected cutting load output by the machine tool cutting mechanics fundamental model to calculate the dynamic load deviation rate. The control system performed time integration calculations on the sequence within the interference range to obtain the cumulative error integral. Verification showed that the absolute value of the cumulative error integral for this machining did not exceed the set error dead zone threshold, indicating that the current hardness deviation coefficient calculation model matches the actual material properties well. The system maintained its current state and did not send a reverse calibration network message to the sintering end.

Claims

1. An adaptive machining control system for permanent magnets in elevator traction machines, characterized in that, include: A multi-temperature zone vacuum sintering furnace, wherein a multi-point thermocouple array is provided inside the multi-temperature zone vacuum sintering furnace, and the multi-point thermocouple array is used to acquire temperature data inside the furnace. A high-temperature resistant radio frequency identification carrier, wherein the high-temperature resistant radio frequency identification carrier is attached to a tray containing a permanent magnet blank; A CNC machining center, comprising a motion controller, a servo driver, and a spindle motor, is used to perform mechanical processing on the permanent magnet blank; The control system is communicatively connected to the multi-point thermocouple array, the high-temperature resistant radio frequency identification carrier, and the CNC machining center. The control system is used to acquire the temperature data to calculate the hardness deviation coefficient, and bind the hardness deviation coefficient to the high-temperature resistant radio frequency identification carrier. The control system is also used to analyze the machining trajectory and reconstruct the cutting parameters according to the hardness deviation coefficient during machining in the CNC machining center, and to calibrate the calculation model of the hardness deviation coefficient according to the actual transient cutting current feedback of the spindle motor.

2. The adaptive machining control system for permanent magnets of elevator traction machines according to claim 1, characterized in that, The control system acquires the time and temperature sequence data of the multi-point thermocouple array and reconstructs the transient temperature response curve of each grid node in the pan using the inverse distance weighted spatial interpolation algorithm. The control system calculates the time derivative and time integral of the transient temperature response curve within the set cooling temperature range to obtain the local cooling rate integral of each grid node. The control system integrates the local cooling rate and the set material standard ideal cooling rate into the calculation model of the hardness deviation coefficient, and calculates the hardness deviation coefficient of the permanent magnet blank corresponding to each of the grid nodes.

3. The adaptive machining control system for permanent magnets of elevator traction machines according to claim 2, characterized in that, The control system generates a set of hardness deviation coefficients containing all the grid nodes, obtains the unique identifier of the high-temperature resistant RFID carrier, and establishes a mapping storage relationship between the unique identifier and the set of hardness deviation coefficients in the manufacturing execution system database. When the control system reads the unique identifier on the CNC machining center side, it issues the corresponding set of hardness deviation coefficients and initializes a two-dimensional traversal pointer in the internal memory. The two-dimensional traversal pointer is used to indicate the original storage coordinates of the permanent magnet blank to be processed in the two-dimensional array grid of the material tray.

4. The adaptive machining control system for permanent magnets of elevator traction machines according to claim 1, characterized in that, The unmachined surface of the permanent magnet blank is formed with asymmetrical physical reference marks, and the machine tool fixture of the CNC machining center is provided with a mechanical limiting profile that matches the physical reference marks to constrain the clamping posture. The control system establishes a local workpiece coordinate system with the center of the mechanical limit contour as the spatial reference origin, and maps the gradual change direction of the hardness deviation coefficient to the three-dimensional machining coordinate system of the CNC machining center to generate a hardness topology mapping table.

5. The adaptive machining control system for permanent magnets of an elevator traction machine according to claim 4, characterized in that, Before the spindle motor starts cutting, the control system parses the tool trajectory interpolation instructions in the original CNC machining program code to extract the target space coordinates of the tool. The control system performs spatial interference comparison between the spatial coordinates of the tool target and the hardness topology mapping table, and extracts the spatial coordinate region where the hardness deviation coefficient exceeds the set normal physical threshold as the abnormal hardness interference interval. For the abnormal hardness interference range, the control system calls the dynamic cutting parameter reconstruction algorithm to calculate the local reconstruction feed and local reconstruction spindle speed, generates local adaptive CNC machining program code, and sends it to the motion controller for execution.

6. The adaptive machining control system for permanent magnets of an elevator traction machine according to claim 5, characterized in that, When calculating the local reconstruction feed amount, the dynamic cutting parameter reconstruction algorithm uses the standard feed amount preset by the process to subtract the feed attenuation compensation value calculated based on the hardness deviation coefficient and the feed scaling sensitivity factor preset by the system. When calculating the spindle speed of the local reconstruction, the speed increase compensation value is calculated by adding the standard spindle speed preset by the process to the speed compensation factor based on the hardness deviation coefficient and the speed compensation factor preset by the system.

7. The adaptive machining control system for permanent magnets of an elevator traction machine according to claim 5, characterized in that, The control system inserts a parameter smoothing transition code segment at the boundary between the retained conventional processing trajectory segment and the abnormal hardness interference range; The control system employs a parametric linear interpolation algorithm in the parameter smooth transition code segment to smoothly transition the actual feed rate and actual spindle speed to the locally reconstructed feed amount and the locally reconstructed spindle speed according to a set acceleration slope.

8. The adaptive machining control system for permanent magnets of elevator traction machines according to claim 1, characterized in that, During the cutting process performed by the CNC machining center, the control system continuously reads the actual transient cutting current of the spindle motor through the servo driver according to a set high-frequency sampling period; The control system inputs the currently executed cutting parameters into the built-in machine tool cutting mechanics basic model to calculate the theoretical expected cutting load; The control system compares the actual transient cutting current with the theoretical expected cutting load to obtain the dynamic load deviation rate.

9. The adaptive machining control system for permanent magnets of elevator traction machines according to claim 2, characterized in that, The control system extracts the dynamic load deviation rate sequence of the tool within the entire abnormal hardness interference range, and performs time integration on the sequence to obtain the cumulative error integral. When the absolute value of the cumulative error integral is greater than the set error dead zone threshold, the control system determines that the calculation model of the hardness deviation coefficient has a systematic empirical deviation, and generates a reverse calibration network message to be sent back to the multi-temperature zone vacuum sintering furnace side.

10. The adaptive machining control system for permanent magnets of an elevator traction machine according to claim 9, characterized in that, The control system transmits the cumulative error integral to the sintering control unit via the reverse calibration network message; The sintering control unit uses the product of the cross-process model learning rate and the cumulative error integral as a compensation amount to iteratively add the current initial empirical weight coefficients of the calculation model for the hardness deviation coefficient, thereby obtaining the updated model weight coefficients, and then applies the updated model weight coefficients to the data parsing calculation of the next batch of permanent magnet blanks.

Images