A self-adaptive gradual online compensation control method for tool wear of a numerical control machine tool

Abstract

The application provides a kind of numerical control machine tool tool wear adaptive progressive online compensation control method, related to precision machining adaptive control technical field, through the periodic collection distance data of infrared sensor in continuous machining process, tool instantaneous wear rate is calculated according to the difference between before and after collection;Real-time instantaneous wear rate is input into built-in prediction model to complete self-correction, and the predicted average wear rate of the next machining path segment is output.Combined with the scheduled execution time of the machining path segment, the estimated total wear is calculated and planned as a micro-compensation sequence synchronized with the machining feed process.During the machining process, the piezoelectric ceramic step-by-step axial deformation of each group is synchronously controlled to drive the clamping mechanism to complete the progressive step-by-step compensation movement.Tool wear online adaptive prediction and synchronous progressive compensation can be realized, effectively compensating for tool machining wear error, improving the precision of continuous machining of numerical control machine tool and the quality of workpiece machining, adapting to long-time continuous machining scene of machine tool.

Description

Technical Field

[0001] This invention belongs to the field of precision machining adaptive control technology, specifically relating to an adaptive progressive online compensation control method for tool wear in CNC machine tools. Background Technology

[0002] During continuous machining on CNC machine tools, tool wear is inevitable, causing the actual cutting position to deviate from the preset machining trajectory and affecting the workpiece machining accuracy. To compensate for tool wear, a tool wear self-compensating clamping device has been proposed in the prior art (such as patent publication number CN223848625U). This device detects the amount of tool wear using an infrared sensor and uses piezoelectric ceramic to drive the clamping mechanism to extend axially to compensate for the tool length loss caused by wear.

[0003] However, existing devices suffer from the following technical problems in practical applications: their operating mode involves post-processing inspection and one-time compensation after stopping or interrupting processing, meaning each compensation action occurs only after the current processing cycle is completed. During a single continuous processing cycle, tool wear accumulates gradually over time, and the aforementioned devices cannot compensate for tool wear in real time during processing. This results in significant tool length loss due to continuous wear in the latter half of the processing path, leading to a gradual decrease in processing accuracy. Furthermore, the large, one-time compensation action leaves tool marks on the workpiece surface, affecting the surface quality. Therefore, how to achieve online, gradual compensation for tool wear during continuous processing is a pressing technical problem that needs to be solved in this field. Summary of the Invention

[0004] In view of the above-mentioned defects or deficiencies in the prior art, an adaptive progressive online compensation control method for CNC machine tool tool wear is provided, which is applied to a tool wear self-compensating clamping device, the device comprising: The housing has a connecting part for connecting the machine tool spindle; A clamping mechanism is slidably mounted inside the housing along the axial direction of the housing and is used to clamp the cutting tool; An infrared sensor, fixedly installed, is used to detect the actual distance value representing the distance between the front end of the clamping mechanism and the machine tool workpiece clamping tool; Multiple sets of piezoelectric ceramics are fixedly installed and configured to drive the clamping mechanism to generate compensated movement along the axial direction by deformation upon power-on. The controller is electrically connected to both the infrared sensor and the piezoelectric ceramic sensor. The control method is executed by the controller and includes the following steps: During continuous processing, the actual distance value is periodically acquired from the infrared sensor, and the instantaneous wear rate is calculated based on the difference between two consecutive actual distance values. Multiple instantaneous wear rates obtained through continuous calculation are input into the prediction model built into the controller in real time. The prediction model performs self-correction based on the newly input instantaneous wear rate and outputs the predicted average wear rate corresponding to a future processing path segment. Obtain the predetermined execution time of the future processing path segment, calculate the estimated total wear based on the predetermined execution time and the predicted average wear rate, and plan the estimated total wear as a set of micro-compensation sequences synchronized with the processing feed process; During the execution of the future machining path segment by the CNC system, according to the micro-compensation sequence, a driving voltage corresponding to each micro-compensation amount is synchronously applied to each group of piezoelectric ceramics, causing each group of piezoelectric ceramics to produce a series of step-by-step axial elongation deformations, thereby driving the clamping mechanism to perform multiple step-by-step compensation movements along the axial direction.

[0005] According to the technical solution provided in this application, the following steps are also included: Each time the driving voltage is applied, the driving voltage is corrected according to the hysteresis model of the piezoelectric ceramic; Determine whether the corrected driving voltage exceeds the stroke safety threshold of the piezoelectric ceramic; If the limit is exceeded, an alarm will be triggered and compensation will be suspended. If the error does not exceed the limit, the clamping mechanism is driven to make one step compensation movement along the axis direction according to the corrected driving voltage.

[0006] According to the technical solution provided in this application, the prediction model includes a Kalman filter and a time-varying wear rate predictor connected in series with the Kalman filter; The prediction model performs self-correction based on the new input instantaneous wear rate, including the following steps: The Kalman filter receives the instantaneous wear rate calculated in the current sampling period as the observation value, and at the same time receives the predicted wear rate output by the time-varying wear rate predictor in the previous period as the state prediction value. The observation value and the state prediction value are fused by the Kalman gain to output the corrected wear rate for the current period. The time-varying wear rate predictor takes the corrected wear rate as input, fits an exponential time-varying curve characterizing the wear rate as a function of time, and extrapolates the time curve to the prediction time window corresponding to the next processing path segment on the time axis to obtain the predicted average wear rate. Wherein, the length of the prediction time window is the predetermined execution duration of the future machining path segment, the extrapolation step size of the time-varying wear rate predictor on the time axis is consistent with the sampling period, and the Kalman gain is dynamically adjusted according to the current machining state parameters, which include the spindle speed and feed rate.

[0007] According to the technical solution provided in this application, the time-varying wear rate predictor further performs the following steps: The instantaneous wear rate calculated in the current sampling period is compared with a preset sudden wear threshold. When the instantaneous wear rate exceeds the sudden wear threshold, a sudden wear event is determined to have occurred. The time-varying wear rate predictor discards the corrected wear rate of the current sampling period, uses the instantaneous wear rate as the current effective wear rate, and refits the time-varying curve to update the predicted average wear rate for the future processing path segment.

[0008] According to the technical solution provided in this application, the step of planning the estimated total wear amount into a set of micro-compensation sequences synchronized with the machining feed process includes the following steps: Obtain the curvature distribution information of the machining trajectory corresponding to the future machining path segment and the feed speed information of each sub-segment; Based on the curvature distribution information, the future processing path segment is divided into at least one high curvature sub-segment and at least one low curvature sub-segment; According to the preset allocation weight, the estimated total wear amount is non-equally allocated to each sub-segment, wherein the unit length compensation amount allocated to the high curvature sub-segment is greater than the unit length compensation amount allocated to the low curvature sub-segment. Within each sub-segment, based on the feed rate variation curve of that sub-segment, the total compensation amount allocated to that sub-segment is weighted and distributed to each micro-compensation amount according to the feed rate ratio, so that the higher the feed rate, the larger the compensation amount per unit time is allocated, and finally the micro-compensation sequence synchronized with the machining feed process is generated.

[0009] According to the technical solution provided in this application, after obtaining the curvature distribution information of the machining trajectory corresponding to the future machining path segment and the feed speed information of each sub-segment, the method further includes the following steps: Based on the curvature distribution information, calculate the total curvature change of the future processing path segment; The total curvature change is compared with a preset curvature threshold. The step of dividing the future processing path segment into at least one high-curvature sub-segment and at least one low-curvature sub-segment based on the curvature distribution information includes the following steps: If the total curvature change is greater than or equal to the preset curvature threshold, then based on the curvature distribution information, the future processing path segment is divided into at least one high curvature sub-segment and at least one low curvature sub-segment.

[0010] According to the technical solution provided in this application, after comparing the total curvature change with a preset curvature threshold, the method further includes the following steps: If the total curvature change is less than the preset curvature threshold, it is determined to be an approximately straight-line cutting scenario. The estimated total wear is then weighted and distributed according to the feed rate ratio throughout the entire future machining path segment to generate the micro-compensation sequence.

[0011] According to the technical solution provided in this application, driving the clamping mechanism to perform a step-compensation movement along the axial direction according to the corrected driving voltage includes the following steps: Within the time window of one step compensation movement, an initial driving voltage is applied to each group of piezoelectric ceramics. The initial driving voltage is calculated based on the micro-compensation amount and the nominal voltage-displacement curve of the piezoelectric ceramics. After the initial driving voltage is applied, the actual displacement of the front end of the clamping mechanism is obtained by the infrared sensor; The actual displacement is compared with the target micro-compensation amount of this step compensation to calculate the displacement deviation; The displacement deviation is input to a proportional-integral controller, and the proportional-integral controller outputs a compensation voltage correction amount; The compensation voltage correction is superimposed on the initial driving voltage and applied again to each group of piezoelectric ceramics as the corrected driving voltage to form a closed-loop displacement correction for a single step compensation movement.

[0012] According to the technical solution provided in this application, the sudden wear threshold is determined according to the following steps: Before the continuous machining process begins, the clamping device is controlled to perform a pre-run during an empty stroke without contacting the workpiece, and during the pre-run, multiple reference distance values ​​are collected by the infrared sensor; The standard deviation of the measurement noise of the infrared sensor under no-travel conditions is calculated based on the fluctuation range of multiple reference distance values. The sudden wear threshold is set to N times the standard deviation of the measurement noise, where N is an integer between 3 and 6, so that the sudden wear threshold can distinguish between normal measurement noise and real sudden wear signals.

[0013] According to the technical solution provided in this application, obtaining the actual displacement of the cutting edge using the infrared sensor includes the following steps: Within the time window of the single step compensation movement, the working mode of the infrared sensor is switched from wear detection mode to displacement feedback mode; In the displacement feedback mode, the sampling period of the infrared sensor is shortened, and the actual displacement of the front end of the clamping mechanism is measured with the infrared sensor reading at the moment the initial driving voltage is applied as the zero reference. After a step-compensation movement is completed, the infrared sensor is restored to the wear detection mode, and the actual distance value is acquired again at the originally set sampling period.

[0014] Compared with the prior art, the beneficial effects of this application are as follows: I. Online progressive compensation during machining is achieved. By periodically detecting actual distance values ​​and calculating instantaneous wear rates, a self-correcting predictive model is used to predict the average wear rate of a future machining path segment. The estimated total wear is pre-planned as a micro-compensation sequence synchronized with the machining feed process. A series of step-by-step compensation movements are completed simultaneously with the CNC system executing the machining path segment. This method shifts the compensation action from reactive to reactive, releasing the compensation amount synchronously with the machining process. This eliminates the machining accuracy decay caused by progressive tool wear, while avoiding tool contact marks caused by abrupt compensation.

[0015] Second, it improves the accuracy and adaptability of compensation. The prediction model continuously self-corrects based on the real-time input instantaneous wear rate, and can dynamically track changes in tool wear trends, making the predicted average wear rate more accurately reflect the current actual wear state of the tool, thereby improving the accuracy of the estimated total wear and micro-compensation sequence. Attached Figure Description

[0016] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 A flowchart illustrating the steps of the adaptive progressive online compensation control method for CNC machine tool tool wear provided in this application; Figure 2 A front view of the CNC machine tool tool wear self-compensation clamping device provided in this application; Figure 3 for Figure 2 A cross-sectional view along the AA direction of the self-compensating clamping device for tool wear on the CNC machine tool shown; Figure 4 for Figure 2 A cross-sectional view along the BB direction of the self-compensating clamping device for tool wear on the CNC machine tool shown; Figure 5 This is a top view of the CNC machine tool tool wear self-compensating clamping device provided in this application; Figure 6 for Figure 3 A cross-sectional view along the CC direction of the self-compensating clamping device for tool wear on the CNC machine tool shown. The text labels in the image represent: 1. Outer shell; 2. Clamping part; 3. Connecting part; 4. Tool clamping block; 5. Clamping space; 6. Clamping device; 7. Telescopic rod; 8. Push plate; 9. Guide column; 10. Lifting component; 11. First wedge block; 12. Second wedge block; 13. Piezoelectric ceramic; 14. Guide part; 15. Infrared sensor; 16. Controller; 17. Return spring; 18. Sealing part; 19. Slide rail. Detailed Implementation

[0017] The present application will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0018] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0019] Example 1 As mentioned in the background section, this embodiment proposes a self-compensating clamping device for CNC machine tool tool wear, such as... Figure 2-6 As shown, it includes: The housing 1 includes a clamping part 2 and a connecting part 3 located at the tail end of the clamping part 2. The clamping part 2 has an open front end and an internal installation space. The connecting part 3 is used to connect to the machine tool spindle. A clamping mechanism is slidably mounted in the clamping space, with the sliding direction along the axial direction of the outer shell 1. The clamping structure is used to clamp the cutting tool. The detection mechanism is installed on the outer side wall of the clamping part 2 and is used to detect the wear amount of the cutting tool. A compensation mechanism is provided within the installation space and connected to one end of the clamping mechanism near the connecting part. The compensation mechanism is electrically connected to the detection mechanism. The compensation mechanism is used to push out the clamping mechanism according to the amount of wear, so that the clamping mechanism extends out of the opening.

[0020] Specifically, the clamping device provided in this application is used to clamp and install a cutting tool on the spindle of a machine tool. The clamping device includes at least a housing 1, a clamping mechanism, a detection mechanism, and a compensation mechanism. The housing 1 includes a clamping part 2 and a connecting part 3. The connecting part 3 is integrally formed with the clamping part 2 and is used to install the housing 1 on the spindle of the machine tool. The clamping part 2 is equipped with a clamping mechanism and a compensation mechanism. The clamping mechanism is used to clamp and fix the cutting tool. The clamping mechanism is slidably connected to the inner wall of the clamping part 2 so that the clamping mechanism can slide along the axial direction of the housing 1 to compensate for tool wear. The compensation mechanism is installed at the end of the clamping mechanism away from the opening of the housing 1. The compensation mechanism serves as the power source for tool compensation and can drive the clamping mechanism along the axial direction of the housing 1, making it convenient to push the clamping mechanism out a certain distance from the opening of the clamping part 2 according to the amount of tool wear to compensate for the tool wear. A detection mechanism is installed on the outer wall of the clamping part 2. The detection mechanism is used to detect the amount of tool wear by measuring the length of the cutting tool and then sending the amount of wear to the compensation mechanism. By cooperating with the clamping mechanism, the detection mechanism, and the compensation mechanism, the clamping device provided in this application can accurately compensate for the tool wear based on the tool wear, thereby improving the machining accuracy and efficiency of CNC machine tools.

[0021] Furthermore, the clamping mechanism includes: Multiple tool clamping blocks 4 are arranged around the axis of the outer shell 1, and a tool clamping space 5 is formed between the multiple tool clamping blocks 4; Multiple clamping devices 6 are provided, each corresponding to a tool clamping block 4. The clamping device 6 is located between the clamping part 2 and the tool clamping block 4 and is slidably connected to the clamping part 2. The sliding direction is along the axial direction of the outer shell 1. Each clamping device 6 includes a pair of telescopic rods 7, the free end of which is connected to the tool clamping block 4 via a push plate 8.

[0022] Specifically, such as Figure 4 and Figure 6 As shown, the clamping mechanism includes at least a plurality of tool clamping blocks 4 and a plurality of clamping devices 6. The clamping device 6 and the tool clamping block 4 are arranged in a one-to-one correspondence. In this embodiment, three tool clamping blocks 4 and three clamping devices 6 are provided. The three tool clamping blocks 4 are arranged circumferentially around the axis of the outer shell 1. The outer sides of the three tool clamping blocks 4 are arc-shaped. The three tool clamping blocks 4 can slide along the radial direction of the outer shell 1 so that the clamping space 5 between the three tool clamping blocks 4 can be adjusted. The tool can be fixed by installing the tool in the clamping space 5 and tightening the three tool clamping blocks 4. Three clamping devices 6 are also provided corresponding to the tool clamping block 4. Each clamping device 6 is located between the corresponding tool clamping block 4 and the inner wall of the clamping part 2. One side of the clamping device 6 is installed on the inner wall of the clamping part 2 via a slide rail 19. The clamping device 6 can slide along the slide rail 19 along the axial direction of the outer shell 1. The other side of the clamping device 6 is provided with two telescopic rods 7. The telescopic rods 7 can extend and retract in the radial direction of the outer shell 1. A push plate 8 is fixedly installed on the free end of the telescopic rod 7. One side of the push plate 8 is in the arc shape that matches the tool clamping block 4. The push plate 8 is fixedly connected to the outer side of the tool clamping block 4. The clamping device 6 can drive the tool clamping block 4 to move through the telescopic rods 7, thereby adjusting the size of the clamping space. Optionally, the clamping device 6 is a cylinder or a hydraulic cylinder, or the telescopic rods 7 can be electrically controlled. The tool clamping block 4 is also provided with a blocking part 18 at its front end. The blocking part 18 extends out of the opening on the clamping part 2 and extends in the radial direction of the outer shell 1. The blocking part 18 can block the opening of the clamping part 2 to prevent metal chips generated during the processing from entering the clamping device and affecting the operation of the clamping device.

[0023] Furthermore, the compensation mechanism includes: Guide post 9, which is disposed in the installation space and abuts against the end of the tool clamping block 4 near the connecting part 3; Lifting member 10, which is sleeved on the guide post 9 and can slide along the axial direction of the outer shell 1, and the lifting member 10 is provided with a plurality of first wedge blocks 11 on its periphery; Multiple second wedge blocks 12, each corresponding to a first wedge block 11 and abutting against the first wedge block 11; Multiple sets of piezoelectric ceramics 13 are disposed on the inner wall of the clamping part 2. The piezoelectric ceramics 13 are fixed to the second wedge block 12 and electrically connected to the detection mechanism. The piezoelectric ceramics 13 are electrically deformed according to the wear amount to drive the second wedge block 12 to push the first wedge block 11 to slide.

[0024] Specifically, such as Figure 5 and Figure 6As shown, the guide post 9 is located inside the clamping part 2 and is integrally formed with the clamping part 2. The guide post 9 is cylindrical and coaxially arranged with the outer shell 1. The lifting member 10 is sleeved on the end of the guide post 9. The lifting member 10 and the guide post 9 can slide relative to each other. The lifting member 10 is supported by the guide post 9. The front end of the lifting member 10 is slidably connected to the push plate 8 and the tool clamping block 4 to facilitate the movement of the push plate 8 and the tool clamping block 4 along the radial direction of the outer shell 1. Multiple first wedge blocks 11 are arranged circumferentially on the outer side wall of the lifting member 10, corresponding to each The first wedge block 11 is further provided with a second wedge block 12, which is located on the outer periphery of the first wedge block 11. The second wedge block 12 and the first wedge block 11 are in contact with each other, and the second wedge block 12 supports the first wedge block 11. The second wedge block 12 is installed on the inner wall of the clamping part 2 by a piezoelectric ceramic 13. The number of piezoelectric ceramic 13, the second wedge block 12 and the first wedge block 11 are the same. In this embodiment, six piezoelectric ceramic 13, six second wedge blocks 12 and six first wedge blocks 11 are provided. Six piezoelectric ceramics 13 are evenly arranged circumferentially along the inner wall of the outer shell 1. Each piezoelectric ceramic 13 is fixed with a corresponding second wedge block 12. Each piezoelectric ceramic 13 is electrically connected to the detection mechanism. When the detection mechanism detects the wear of the tool, it energizes each piezoelectric ceramic 13 to make the pressure point ceramic deform and elongate precisely, thereby driving the second wedge block 12 to approach the axis of the outer shell 1 and press the first wedge block 11. Through the cooperation of the second wedge block 12 and the first wedge block 11, the radial extrusion of the outer shell 1 is transformed into axial extrusion of the outer shell 1, thereby pushing the clamping mechanism to move axially along the outer shell 1.

[0025] Furthermore, the second wedge block 12 is disposed on the side of the first wedge block 11 away from the tool clamping block 4. The top of the second wedge block 12 is provided with a first inclined surface, and the bottom of the first wedge block 11 is provided with a second inclined surface corresponding to the first inclined surface. The angle between the first inclined surface and the second inclined surface is 5°-15°.

[0026] Specifically, a first inclined surface is formed on the bottom surface of the first wedge 11 near the side of the second wedge 12, and a second inclined surface is formed on the top surface of the second wedge 12 near the side of the first wedge 11. The second wedge 12 and the first wedge 11 abut against each other through the first and second inclined surfaces. The angle between the first and second inclined surfaces and the radial direction of the outer shell 1 is between 5° and 15°. This allows the second wedge 12 to support the first wedge 11 while increasing the displacement of the first wedge 11 along the axial direction of the outer shell 1, even with a small displacement of the second wedge 12 along the radial direction of the outer shell 1, thereby increasing the supply amount. In this embodiment, the angle between the first and second inclined surfaces is 15°.

[0027] Furthermore, each group of piezoelectric ceramics 13 comprises a plurality of piezoelectric ceramic sheets arranged in a stacked manner.

[0028] Specifically, by arranging each group of piezoelectric ceramics 13 in a stacked manner using multiple piezoelectric ceramic sheets, the deformation and stretching of the piezoelectric ceramics 13 when energized can be increased, thereby enabling the compensation mechanism to have a larger compensation amount.

[0029] Furthermore, the clamping part 2 is provided with a plurality of guide parts 14 inside, and the guide parts 14 are located between two adjacent first wedge blocks 11.

[0030] Specifically, a plurality of guide portions 14 are formed on the inner sidewall of the clamping portion 2. The guide portions 14 are integrally formed with the clamping portion 2, extend along the axial direction of the outer shell 1, and are adapted to the shape of the first wedge block 11. By providing the guide portions 14, the movement direction of the first wedge block 11 can be limited, ensuring that it can move along the axial direction of the outer shell 1 under the push of the second wedge block 12. In addition, the guide portions 14 can also ensure the relative fixation between the compensation mechanism and the outer shell 1 when the machine tool spindle drives the clamping device to rotate, thereby reducing tool wobble.

[0031] Furthermore, the testing institution includes: Infrared sensor 15, the infrared sensor 15 is used to detect the actual distance between the front end of the housing 1 and the workpiece clamping tool on the machine tool; The controller 16 is electrically connected to the infrared sensor 15 and the piezoelectric ceramic 13 respectively. The controller 16 is used to calculate the wear amount based on the difference between the actual distance and the set distance, and to control the piezoelectric ceramic 13 to be powered on.

[0032] Specifically, the infrared sensor 15 is installed on the outer wall of the clamping part 2 near the opening on the upper part of the clamping part 2. When a tool is installed on the clamping mechanism, the tool is placed against the workpiece fixture of the machine tool. The infrared sensor 15 can measure the actual distance between the sealing part 18 and the workpiece fixture on the machine tool. Then, the controller 16 calculates the wear amount based on the difference between the actual distance and the set distance. The set distance is the distance between the sealing part 18 and the tool clamping tool on the machine tool measured by the infrared sensor 15 when the complete tool is installed on the clamping mechanism.

[0033] Furthermore, a reset spring 17 is fitted on the guide post 9. One end of the reset spring 17 is fixedly connected to the first wedge block 11, and the other end is fixedly connected to the bottom of the installation space.

[0034] Specifically, the return spring 17 is used to reset the lifting member 10. When the piezoelectric ceramic 13 is energized and deforms, causing the second wedge block 12 to push the first wedge block 11 to move, the return spring 17 is stretched and stores force as the lifting member 10 moves. When the piezoelectric ceramic 13 is de-energized, the second wedge block 12 resets as the piezoelectric ceramic 13 contracts. At this time, the first wedge block 11 loses the support of the second wedge block 12 and then retracts under the action of the return spring 17 until it abuts against the second wedge block 12, thereby resetting the clamping mechanism. By setting the return spring 17, the clamping mechanism can be easily reset when tool compensation is not required.

[0035] Working process: The clamping device is installed on the spindle of the machine tool. Multiple clamping devices 6 control multiple tool clamping blocks 4 to move away from each other to open the clamping space 5. Then, the tool is placed in the clamping space 5 and the multiple tool clamping blocks 4 are controlled to move closer to each other to clamp and fix the tool. Then, the spindle of the machine tool is operated to make the tool abut against the workpiece fixture on the machine tool. The infrared sensor 15 detects the actual distance between the sealing part 18 and the workpiece fixture, and the controller 16 calculates the difference between the actual distance and the set distance to obtain the wear amount. The controller 16 controls the piezoelectric ceramic 13 to be energized according to the wear amount. After the piezoelectric ceramic 13 is energized, it deforms and stretches to drive the second wedge block 12 to move in the radial direction of the outer shell 1. The movement of the second wedge block 12 pushes the first wedge block 11 to move in the axial direction of the outer shell 1, thereby causing the lifting member 10 to push the clamping mechanism out of the opening on the clamping part 2 to achieve tool compensation.

[0036] Example 2 Based on Example 1, this example proposes an adaptive progressive online compensation control method for CNC machine tool tool wear. The control method is executed by the controller, such as... Figure 1 As shown, it includes the following steps: S1. During continuous processing, the actual distance value is periodically obtained from the infrared sensor, and the instantaneous wear rate is calculated based on the difference between the two actual distance values ​​obtained. S2. The multiple instantaneous wear rates obtained by continuous calculation are input into the prediction model built into the controller in real time. The prediction model performs self-correction based on the newly input instantaneous wear rate and outputs the predicted average wear rate corresponding to a future processing path segment. S3. Obtain the predetermined execution time of the future processing path segment, calculate the estimated total wear based on the predetermined execution time and the predicted average wear rate, and plan the estimated total wear as a set of micro-compensation sequences synchronized with the processing feed process. S4. During the execution of the future machining path segment by the CNC system, according to the micro-compensation sequence, a driving voltage corresponding to each micro-compensation amount is synchronously applied to each group of piezoelectric ceramics, so that each group of piezoelectric ceramics produces a series of step-by-step axial elongation deformations, thereby driving the clamping mechanism to perform multiple step-by-step compensation movements along the axial direction.

[0037] Specifically, step S1: Periodically acquire the actual distance value and calculate the instantaneous wear rate. The instantaneous wear rate refers to the reduction in distance between the front end of the clamping mechanism and the workpiece clamping tool due to tool wear per unit time, and the unit is micrometers per second.

[0038] The specific implementation process is as follows: The controller has an internal sampling timer that periodically generates an interrupt signal according to a preset sampling period. The sampling period is set according to the tool type and machining method, with 0.5 to 1 second for roughing scenarios and 2 to 5 seconds for finishing scenarios. When an interrupt signal is triggered, the controller first determines whether it is currently in a valid cutting period. The determination method is as follows: The controller queries the CNC system for the currently executed machining code through the digital communication interface. If the returned G code belongs to cutting feed instructions such as G01 linear interpolation, G02 clockwise circular interpolation, or G03 counterclockwise circular interpolation, it is determined to be a valid cutting period; if it belongs to non-cutting instructions such as G00 rapid positioning or G04 pause, it is determined to be an idle travel period. If it is an idle travel period, the controller skips this sampling, does not calculate the instantaneous wear rate, and waits for the next interrupt signal.

[0039] If a valid cutting period is determined, the controller reads the current actual distance value from the infrared sensor via the analog input interface. The infrared sensor outputs an analog voltage signal proportional to the detection distance; for example, 0 to 10 volts corresponds to a detection range of 0 to 200 mm. The controller converts the analog voltage into a digital signal via an analog-to-digital converter. The controller subtracts the current distance value from the distance value of the previous sampling period stored in memory, and then divides the difference by the sampling period to obtain the instantaneous wear rate. The calculation formula is: Instantaneous wear rate equals the difference between the distance value of the previous period and the distance value of the current period, divided by the sampling period. When the calculation result is positive, it indicates that the tool has worn; if the calculation result is negative, it indicates that the distance has increased abnormally, and the controller sets the value to zero to eliminate interference from factors such as thermal expansion. After the calculation is completed, the controller stores the current distance value in memory, overwriting the distance value of the previous period, and simultaneously stores the instantaneous wear rate in the wear rate data queue, which stores data from the most recent 50 to 200 sampling periods.

[0040] Step S2: Input the instantaneous wear rate into the prediction model and output the predicted average wear rate. The prediction model refers to a mathematical algorithm module that can predict future trends based on historical time series data. Its self-calibration means that the model automatically adjusts its internal parameters to approximate the true value after receiving new data.

[0041] In this embodiment, the prediction model employs an exponential smoothing model. Internally, the model maintains a sequence of smoothed values. Whenever a new instantaneous wear rate is input, the model weights the new data with the smoothed value from the previous period using a preset smoothing coefficient to obtain the smoothed value for the current period. The smoothing coefficient is set between 0.2 and 0.3 to achieve a balance between response speed and curve smoothness. After smoothing, the model fits a trend line based on several recent smoothed values ​​and extrapolates the slope of this trend line forward to the midpoint of the next processing path segment, obtaining the predicted average wear rate. The next processing path segment is the next continuous processing trajectory segment to be executed in the exponential control system. The predicted average wear rate is a scalar value, measured in micrometers per second.

[0042] Step S3: Obtain the predetermined execution time, calculate the estimated total wear, and plan the micro-compensation sequence. The controller obtains the programmed endpoint coordinates and programmed feed rate of a future machining path segment from the CNC system via a digital communication interface. For straight path segments, the controller calculates the Euclidean distance from the start point to the end point as the geometric length; for circular path segments, the controller calculates the product of the arc radius and the central angle in radians as the geometric length. Dividing the geometric length by the programmed feed rate yields the predetermined execution time. Multiplying the predetermined execution time by the predicted average wear rate yields the estimated total wear, in micrometers.

[0043] Next, the controller plans the estimated total wear amount into a micro-compensation sequence. The micro-compensation sequence is a data table containing multiple micro-compensation values ​​and their corresponding timestamps. The planning process is as follows: determine the target step distance for single-step compensation as 1 to 5 micrometers, divide the estimated total wear amount by the target step distance to obtain the total number of compensation steps, and then divide the predetermined execution time by the total number of steps to obtain the inter-step time interval. Starting from time zero, the controller adds a record to the data table every inter-step time interval. Each record contains a timestamp and a micro-compensation value. In the equal-division allocation scheme, each micro-compensation value is equal, equal to the estimated total wear amount divided by the total number of steps. The synchronization of the micro-compensation sequence with the machining feed process means that the release time of each compensation value in the sequence corresponds to the movement position of the interpolation axis of the CNC system, coordinating the compensation action with the actual cutting process of the tool.

[0044] Step S4: Synchronously drive the piezoelectric ceramic to perform step compensation. Simultaneously with the CNC system starting to execute the next machining path segment, the controller starts the compensation timer, with a timing period equal to the step time interval. When the controller detects the rising edge of the path segment execution start flag signal output by the CNC system, it clears the compensation timer and starts counting again.

[0045] When the compensation timer reaches the step interval for the first time, the controller triggers the first compensation interrupt. In the interrupt service routine, the controller retrieves the first micro-compensation value from the micro-compensation sequence data table according to the step counter index, and converts the micro-compensation value into a drive voltage value based on the nominal voltage-displacement characteristic curve of the piezoelectric ceramic. The voltage-displacement characteristic curve is provided by the piezoelectric ceramic manufacturer and stored in the controller parameter storage area. The drive voltage value is equal to the micro-compensation value divided by the nominal displacement sensitivity of the piezoelectric ceramic. The controller synchronously outputs this drive voltage to each group of piezoelectric ceramics through the analog output interface. After being powered on, each group of piezoelectric ceramics undergoes axial elongation deformation, pushing the clamping mechanism to move a small distance along the axial direction towards one side of the workpiece, completing one step compensation. Afterward, the step counter increments by one, and the compensation timer restarts. This process is repeated during the execution of the machining path segment until the step counter reaches the total number of steps or a path segment completion signal is received.

[0046] By combining the above four steps, the compensation action is moved from after machining to during machining. The large one-time compensation amount is decomposed into dozens of micro compensation amounts that are synchronized with the machining feed, so that the actual cutting edge position of the tool is always kept near the preset trajectory. This avoids the decrease in accuracy caused by wear accumulation in the latter half of the machining path segment, and at the same time, it does not produce tool-connecting steps on the workpiece surface.

[0047] In a preferred embodiment, the following steps are also included: Each time the driving voltage is applied, the driving voltage is corrected according to the hysteresis model of the piezoelectric ceramic; Determine whether the corrected driving voltage exceeds the stroke safety threshold of the piezoelectric ceramic; If the limit is exceeded, an alarm will be triggered and compensation will be suspended. If the error does not exceed the limit, the clamping mechanism is driven to make one step compensation movement along the axis direction according to the corrected driving voltage.

[0048] Specifically, piezoelectric ceramics exhibit hysteresis characteristics, meaning that the displacement corresponding to the same voltage value differs during voltage increase and decrease, forming a hysteresis loop. Under open-loop control, without hysteresis correction, the actual displacement may deviate from the commanded displacement by 10% to 15% of the maximum displacement.

[0049] The specific implementation process is as follows: The controller internally stores a hysteresis model of the piezoelectric ceramic. In this embodiment, the Prandtl-Ishlinskii model is adopted. This model describes the hysteresis characteristics as a linear weighted superposition of multiple Play operators with different thresholds and weights. The threshold and weight parameters of each Play operator are predetermined through offline identification experiments and stored in the controller's non-volatile memory. The identification method is as follows: a slow triangular wave scan from zero to the maximum voltage and then back to zero is applied to the piezoelectric ceramic, and voltage and displacement data are recorded simultaneously to obtain a complete hysteresis loop. Then, the parameters of each operator are identified through a least squares fitting algorithm.

[0050] Before applying the drive voltage each time, the controller performs the following operations. First, it reads the current voltage value stored in the voltage status register, which is the actual voltage applied in the previous control cycle. Then, it obtains the target micro-compensation amount to be applied this time. Next, it calls the inverse model of the hysteresis model, taking the current voltage value and the target micro-compensation amount as inputs, and solves for the output drive voltage value through an iterative algorithm. The convergence condition for the iteration is that the deviation between the calculated displacement and the target micro-compensation amount is less than 0.01 micrometers. This drive voltage value is the corrected drive voltage.

[0051] The controller then determines whether the corrected drive voltage exceeds the piezoelectric ceramic's stroke safety threshold. This threshold is set to 85% of the piezoelectric ceramic's rated voltage. If the corrected drive voltage exceeds this threshold, it indicates that the required displacement has exceeded the piezoelectric ceramic's safe operating range, and continued driving may lead to depolarization failure. In this case, the controller displays an alarm message on the human-machine interface, prompting the operator that the tool wear has exceeded the compensation range and the tool needs to be replaced. Simultaneously, it sends a feed hold signal to the CNC system via the digital communication interface to pause machining. If the corrected drive voltage does not exceed the threshold, the controller outputs the corrected drive voltage to each group of piezoelectric ceramics, driving the clamping mechanism to perform a step-compensation movement.

[0052] In a preferred embodiment, the prediction model includes a Kalman filter and a time-varying wear rate predictor connected in series with the Kalman filter; The prediction model performs self-correction based on the new input instantaneous wear rate, including the following steps: The Kalman filter receives the instantaneous wear rate calculated in the current sampling period as the observation value, and at the same time receives the predicted wear rate output by the time-varying wear rate predictor in the previous period as the state prediction value. The observation value and the state prediction value are fused by the Kalman gain to output the corrected wear rate for the current period. The time-varying wear rate predictor takes the corrected wear rate as input, fits an exponential time-varying curve characterizing the wear rate as a function of time, and extrapolates the time curve to the prediction time window corresponding to the next processing path segment on the time axis to obtain the predicted average wear rate. Wherein, the length of the prediction time window is the predetermined execution duration of the future machining path segment, the extrapolation step size of the time-varying wear rate predictor on the time axis is consistent with the sampling period, and the Kalman gain is dynamically adjusted according to the current machining state parameters, which include the spindle speed and feed rate.

[0053] Specifically, the Kalman filter's function is to reduce noise in the instantaneous wear rate calculated in step S1. The specific implementation process is as follows: First, state equations and observation equations are established. The state equation uses a random walk model, representing the true wear rate of the current cycle as the sum of the true wear rate of the previous cycle and the process noise. The process noise variance Q is initially set to 0.01 square micrometers / square second. The observation equation represents the instantaneous wear rate measured by the infrared sensor as the sum of the true wear rate and the measurement noise. The measurement noise variance R is obtained by calibrating the control device before processing by keeping it stationary and continuously collecting distance values ​​from the infrared sensor.

[0054] In each sampling period, the Kalman filter performs recursive calculations: It calculates the prior state estimate for the current period using the corrected wear rate and state equation from the previous period; it adds the error covariance from the previous period to the process noise variance Q to obtain the prior error covariance; it calculates the Kalman gain, which is equal to the prior error covariance divided by the sum of the prior error covariance and the measurement noise variance R, with a value between 0 and 1; it weights and fuses the prior state estimate and the observed value using the Kalman gain to obtain the corrected wear rate for the current period, calculated as the corrected wear rate equal to the prior state estimate plus the Kalman gain multiplied by the difference between the observed value and the prior state estimate; finally, it updates the error covariance for use in the next period.

[0055] In this claim, the Kalman gain is also dynamically adjusted based on machining state parameters, including spindle speed and feed rate. The controller reads the spindle speed and feed rate values ​​from the CNC system in real time and obtains the corresponding measurement noise variance adjustment coefficient by looking up a table. The higher the speed and feed rate, the more severe the cutting vibration, the greater the noise measured by the infrared sensor, and the larger the adjustment coefficient. The controller multiplies the calibrated basic noise variance by the adjustment coefficient to obtain the actual measurement noise variance for the current cycle, and substitutes it into the Kalman gain formula. Thus, under high-noise roughing conditions, the Kalman gain automatically decreases, resulting in a smoother filtering result; under low-noise finishing conditions, the Kalman gain automatically increases, providing a more timely response to actual wear changes.

[0056] The corrected wear rate output from the Kalman filter is fed into a time-varying wear rate predictor. The predictor maintains a fixed-length data window in memory, storing the corrected wear rate for the most recent 20 to 50 sampling periods. Whenever new data is received, the predictor adds it to the window; if the window is full, the oldest data is removed. Then, an exponential time-varying curve is fitted to the data within the window. The expression for the exponential time-varying curve is R(t) = R0 × exp(α × t), where R0 is the initial value and α is the acceleration factor. The fitting method is as follows: first, the natural logarithm of the time and wear rate values ​​of the data points within the window is taken to transform the exponential fit into a linear fit; then, the optimal estimates of ln(R0) and α are obtained using a linear regression formula; finally, the exponential restoration parameters are taken. The fitting calculation is re-executed after each window data update, achieving self-correction of the prediction model.

[0057] Based on the fitted exponential time-varying curve, the time-varying wear rate predictor extrapolates forward along the time axis to the prediction time window corresponding to a future processing path segment. The length of the prediction time window is the predetermined execution duration of that path segment, and the extrapolation step size is consistent with the sampling period. Starting from the current moment, the predictor takes a predicted value along the curve every sampling period, and outputs the arithmetic mean of all predicted values ​​as the predicted average wear rate.

[0058] In a preferred embodiment, the time-varying wear rate predictor further performs the following steps: The instantaneous wear rate calculated in the current sampling period is compared with a preset sudden wear threshold. When the instantaneous wear rate exceeds the sudden wear threshold, a sudden wear event is determined to have occurred. The time-varying wear rate predictor discards the corrected wear rate of the current sampling period, uses the instantaneous wear rate as the current effective wear rate, and refits the time-varying curve to update the predicted average wear rate for the future processing path segment.

[0059] Specifically, the cascaded structure of the Kalman filter and the time-varying wear rate predictor can effectively handle data filtering and trend prediction during normal progressive tool wear. However, in actual machining processes, tools may experience sudden chipping or instantaneous severe wear due to localized hard spots on the workpiece. Such sudden wear events can cause the wear rate to jump significantly higher than normal within a single sampling period. Due to its recursive smoothing characteristics, the Kalman filter exhibits a lag in its response to such abrupt signals. If the abrupt data is still included in the filtering and smoothing process before being input into the time-varying curve fitting, the prediction model will continuously underestimate the true wear rate over several sampling periods, resulting in insufficient compensation. To address the above problem, this embodiment adds a sudden wear diagnosis and response step to the time-varying wear rate predictor, as detailed below.

[0060] A preset sudden wear threshold is pre-stored in the internal parameter storage area of ​​the controller. The method for determining this threshold will be further described in subsequent claims. In this embodiment, the threshold is a positive real number in micrometers per second. During continuous processing, after calculating the instantaneous wear rate in step S1 of each sampling cycle, the instantaneous wear rate is sent to the Kalman filter and also to the time-varying wear rate predictor for sudden wear diagnosis. The time-varying wear rate predictor compares the instantaneous wear rate calculated in the current sampling cycle with the sudden wear threshold. The comparison operation is performed by the arithmetic logic unit of the controller, specifically by subtracting the absolute value of the instantaneous wear rate from the threshold. If the instantaneous wear rate is less than or equal to the sudden wear threshold, it indicates that the wear rate in this cycle is within the normal fluctuation range. The time-varying wear rate predictor, following the aforementioned normal procedure, waits to receive the corrected wear rate output from the Kalman filter before performing time-varying curve fitting and extrapolation prediction.

[0061] If the instantaneous wear rate exceeds the sudden wear threshold, a sudden wear event is determined to have occurred. At this time, the Kalman filter will still output a corrected wear rate according to the normal recursive process. However, due to the smoothing characteristic of the Kalman filter, this corrected wear rate will be significantly lower than the true instantaneous wear rate. If used for time-varying curve fitting, it will cause the fitted curve to shift downwards, resulting in a lower predicted average wear rate. Therefore, the time-varying wear rate predictor performs a discard operation: the corrected wear rate output by the Kalman filter in this cycle is directly discarded and not included in the data window. Simultaneously, the time-varying wear rate predictor directly uses the original instantaneous wear rate calculated in step S1 of this cycle as the current effective wear rate. The current effective wear rate refers to the actual data point used to participate in the time-varying curve fitting. The time-varying wear rate predictor places this current effective wear rate into the data window. According to the window management rules described in claim 3, if the window is full, the oldest data is removed, and then the updated data within the window is used to re-execute the exponential time-varying curve fitting calculation. After fitting is completed, the updated predicted average wear rate is obtained by extrapolation based on the new time-varying curve, and this updated value is output to step S3 for use. Because the refitted curve directly incorporates high wear rate data points from sudden wear, the growth trend of the fitted curve will be immediately adjusted upwards, making the predicted average wear rate of subsequent machining path segments more accurately reflect the actual wear state of the tool after the sudden change.

[0062] This implementation method adds a sudden wear diagnosis and response step, enabling the prediction model to bypass the smoothing delay of the Kalman filter when faced with abnormal wear events such as sudden tool chipping. It quickly reflects the abrupt change information in the prediction results, thus resolving the contradiction between the Kalman filter's delayed response to abrupt signals and the need for rapid compensation for sudden wear. This gives the entire online compensation method a rapid adaptive capability to sudden changes in tool wear state.

[0063] In a preferred embodiment, planning the estimated total wear amount as a set of micro-compensation sequences synchronized with the machining feed process includes the following steps: Obtain the curvature distribution information of the machining trajectory corresponding to the future machining path segment and the feed speed information of each sub-segment; Based on the curvature distribution information, the future processing path segment is divided into at least one high curvature sub-segment and at least one low curvature sub-segment; According to the preset allocation weight, the estimated total wear amount is non-equally allocated to each sub-segment, wherein the unit length compensation amount allocated to the high curvature sub-segment is greater than the unit length compensation amount allocated to the low curvature sub-segment. Within each sub-segment, based on the feed rate variation curve of that sub-segment, the total compensation amount allocated to that sub-segment is weighted and distributed to each micro-compensation amount according to the feed rate ratio, so that the higher the feed rate, the larger the compensation amount per unit time is allocated, and finally the micro-compensation sequence synchronized with the machining feed process is generated.

[0064] Specifically, as previously described, the generation of the micro-compensation sequence can employ an equal-division allocation scheme. However, this scheme ignores the influence of the geometric and kinematic characteristics of the machining trajectory on the tool wear rate. In actual cutting, when machining high-curvature trajectory segments, the contact arc length between the tool and the workpiece increases, the cutting force rises, and the wear per unit cutting length is significantly higher than in straight segments. Simultaneously, in regions with higher feed rates, the tool cuts a larger volume of material per unit time, resulting in a correspondingly higher wear rate. If equal-division allocation is still used, it will lead to insufficient compensation in high-curvature and high-speed feed regions, while excessive compensation in straight and low-speed regions, resulting in a spatial distribution of compensation effect inconsistent with actual tool wear. This embodiment addresses these issues by introducing a non-equal-division allocation scheme based on curvature and feed rate information. The specific implementation is as follows.

[0065] The first step involves the controller acquiring the curvature distribution information of the machining trajectory corresponding to a future machining path segment, as well as the feed rate information of each sub-segment. The controller reads the trajectory data of this path segment from the pre-read buffer of the CNC system via a digital communication interface. The pre-read buffer of the CNC system stores several machining codes to be executed and their parsing results. The curvature distribution information is acquired based on the trajectory type; for straight trajectory segments, the CNC system returns a trajectory type identifier as a straight line, and the controller directly sets the curvature value to zero. For circular arc trajectory segments, the CNC system returns a trajectory type identifier as an arc, along with the radius parameter of the arc. There are two methods to obtain the radius of the arc: If the CNC system machining code uses the R instruction to directly specify the arc radius (code format G02 or G03 with R value), the controller reads the value after the R instruction as the arc radius. If the CNC system machining code uses the I, J, K instructions to specify the incremental coordinates of the arc center relative to the starting point (code format G02 or G03 with I, J, K values), the controller reads the I, J, and K values ​​and calculates the square root of I² + J² + K² to obtain the arc radius. Here, I corresponds to the offset of the center in the X-axis direction relative to the arc starting point, J corresponds to the offset of the center in the Y-axis direction relative to the arc starting point, and K corresponds to the offset of the center in the Z-axis direction relative to the arc starting point. The controller takes the reciprocal of the arc radius to obtain the curvature value of the arc segment, in millimeters. For complex trajectory segments composed of spline curves or continuous micro-straight lines, the pre-read buffer of the CNC system stores a sequence of discrete interpolation point coordinates generated by the interpolator. Each interpolation point contains three-dimensional coordinate values ​​and the corresponding arc length parameter. The controller traverses the coordinates of three consecutive adjacent interpolation points and calculates the circumcircle radius of the local position using the three-point circle method: first, the perpendicular bisector of the line connecting the first and second points is calculated; then, the perpendicular bisector of the line connecting the second and third points is calculated. The foot of the perpendicular between the intersection of the two perpendicular bisectors and the tangent to the trajectory at the second point is the center of the circumcircle, and the distance from the center of the circumcircle to the three points is the circumcircle radius. The controller takes the reciprocal of the circumcircle radius as an approximate curvature value for the second point. This process is repeated to obtain a sequence of curvature values ​​for each discrete sampling position on the entire complex trajectory segment; this sequence is the curvature distribution information. The feed rate information is obtained by the controller reading the value after the F command from the machining code as the programmed feed rate, in millimeters per minute. If there are multiple F instructions within the machining path segment, the programmed feed rates for each sub-segment may differ. For CNC systems that support acceleration and deceleration control, the actual feed rate will accelerate and decelerate at the beginning and end of the path segment. At this time, the controller reads the actual feed rate curve output by the interpolator from the CNC system's pre-read buffer. This curve describes the actual feed rate value corresponding to each interpolation position within the path segment.

[0066] The second step involves the controller dividing the machining path segment into at least one high-curvature sub-segment and at least one low-curvature sub-segment based on curvature distribution information. This division is based on a preset curvature threshold value. The physical meaning of the curvature threshold value is the critical curvature value that distinguishes between high-curvature and low-curvature regions, and it is set based on the tool diameter. The smaller the tool diameter, the greater the change in the tool's wrap angle when machining contours with the same radius of curvature, resulting in more severe fluctuations in cutting force and a more significant wear acceleration effect. Therefore, the curvature threshold value should be set as small as possible, meaning it more sensitively includes medium-curvature regions within the high-curvature sub-segment. For a 10 mm diameter end mill, the curvature threshold value can be set to 0.01 per millimeter, corresponding to a curvature radius of 100 mm; that is, arc segments with a curvature radius less than 100 mm are considered high-curvature regions. For a 20 mm diameter end mill, the curvature threshold value can be set to 0.02 per millimeter, corresponding to a curvature radius of 50 mm. The specific values ​​of the curvature boundary are calibrated experimentally before processing and stored in the controller's parameter storage area. The controller traverses each sampling point in the curvature distribution information sequence, comparing the curvature value of each sampling point with the curvature boundary value. Consecutive sampling points with curvature values ​​greater than the curvature boundary value are grouped into a high-curvature sub-segment, while consecutive sampling points with curvature values ​​less than or equal to the curvature boundary value are grouped into a low-curvature sub-segment. Each sub-segment contains at least one sampling point, and the boundaries between sub-segments are marked by the arc length positions of the start and end points of that sub-segment on the processing path segment.

[0067] The third step involves the controller distributing the estimated total wear amount unequally to each sub-segment according to preset weights. The preset weights represent the proportion of additional compensation required per unit length for the high-curvature sub-segment compared to the low-curvature sub-segment, and their values ​​are obtained through experimental calibration. The calibration method is as follows: Under the same cutting conditions as the actual application scenario of this control method—i.e., the same tool type, workpiece material, spindle speed, and average feed rate—a 100mm high-curvature arc segment and a 100mm low-curvature straight segment are test-cut. After machining, the actual change in tool length before and after the two machining operations is measured using a tool pre-adjustment device. The wear amount caused by machining the arc segment is divided by the wear amount caused by machining the straight segment to obtain the weighting coefficient. For example, after three repeated experiments and taking the average, if the wear amount of the arc segment is 13 micrometers and the wear amount of the straight segment is 10 micrometers, then the weighting of the high-curvature sub-segment relative to the low-curvature sub-segment is 1.3. The allocation process is as follows: First, calculate the geometric length of each segment. The geometric length of a segment is equal to the sum of the arc lengths between the sampling points contained within that segment. For straight segments, the length is obtained directly from the Euclidean distance between the starting and ending coordinates. For circular arc segments, the length is obtained by multiplying the central angle and the radius. For complex curve segments, the arc length is approximated by the cumulative sum of the chord lengths between adjacent sampling points. The total length is obtained by summing the lengths of all segments. The estimated total wear is initially allocated to each segment according to the proportion of each segment's length to the total length, resulting in the initial compensation amount for each segment. Then, the initial compensation amount for each high-curvature segment is multiplied by the corresponding allocation weight coefficient to obtain the final compensation amount for the high-curvature segment. Any excess in the total compensation amount due to the increased compensation amount for high-curvature segments is equally deducted from each low-curvature segment according to the proportion of the initial compensation amount for each low-curvature segment, ensuring that the sum of the final compensation amounts for each segment still equals the estimated total wear. Thus, the unit length compensation allocated to high curvature segments is greater than the unit length compensation allocated to low curvature segments.

[0068] The fourth step involves allocating the micro-compensation amount to each micro-compensation position within each sub-segment. The controller, based on the feed rate variation curve corresponding to the sub-segment, weights the total compensation amount allocated to that sub-segment according to the feed rate ratio and distributes it to each micro-compensation amount. Specifically, the sub-segment is discretized along the machining path into several micro-compensation positions, with the discretization density matching the CNC system's interpolation cycle. Typically, one micro-compensation position is set every 1 to 5 millimeters of path length. For each micro-compensation position, the controller reads the corresponding feed rate value. The formula for calculating the micro-compensation amount at that position is: the micro-compensation amount equals the total compensation amount of the sub-segment multiplied by the feed rate value at that position, then divided by the sum of the feed rate values ​​of all micro-compensation positions within that sub-segment. Thus, positions with higher feed rates receive larger micro-compensation amounts, resulting in a greater amount of compensation released per unit time, matching the increased tool wear caused by high-speed cutting per unit time. Finally, all micro-compensation positions are arranged in arc length order to generate a micro-compensation sequence synchronized with the machining feed process.

[0069] In a preferred embodiment, after obtaining the curvature distribution information of the machining trajectory corresponding to the future machining path segment and the feed speed information of each sub-segment, the method further includes the following steps: Based on the curvature distribution information, calculate the total curvature change of the future processing path segment; The total curvature change is compared with a preset curvature threshold. The step of dividing the future processing path segment into at least one high-curvature sub-segment and at least one low-curvature sub-segment based on the curvature distribution information includes the following steps: If the total curvature change is greater than or equal to the preset curvature threshold, then based on the curvature distribution information, the future processing path segment is divided into at least one high curvature sub-segment and at least one low curvature sub-segment.

[0070] Specifically, after acquiring curvature distribution information and feed rate information, a curvature threshold judgment step is added. Previously, a method was given to divide the machining path segment into high-curvature and low-curvature sub-segments. However, in actual production, there are many scenarios involving purely straight-line cutting or near-straight-line cutting, such as large-area cutting in planar milling and axial cutting in external cylindrical turning. In these scenarios, the curvature of the entire machining path segment is either zero or extremely small, with no clear distinction between high-curvature and low-curvature regions. Performing curvature segmentation in these situations would have no practical engineering significance and would waste the controller's computational resources. This implementation adds a pre-judgment step before curvature segmentation, setting reasonable start conditions for the segmentation operation. The specific implementation is as follows.

[0071] After acquiring the curvature distribution information of a future machining path segment, the controller first calculates the total curvature change of that segment before determining whether to perform curvature segmentation. The total curvature change is a scalar value used to measure the overall bending degree of the entire path segment. The calculation method is as follows: the controller iterates through all sampling points in the curvature distribution information sequence, takes the absolute value of the curvature value for each sampling point, and then sums all the absolute values ​​to obtain the total curvature change. It should be noted that taking the absolute value of curvature here is to prevent positive and negative curvatures from canceling each other out. In some CNC systems, the curvature signs of clockwise and counterclockwise arcs may be opposite. If directly summed algebraically, the total curvature change of a path segment containing both arcs might incorrectly approach zero. By taking the absolute value, each arc contributes positively regardless of its direction, and the total curvature change truly reflects the overall bending degree of the path segment.

[0072] The controller then compares the calculated total curvature change with a preset curvature threshold. The preset curvature threshold is an empirical constant stored in the controller's parameter storage area, used to determine whether the path segment has sufficiently significant curvature characteristics to warrant segmentation by curvature. The preset curvature threshold is set based on the following: when the total curvature change is extremely small, it indicates that the entire path segment is almost a straight line, and curvature segmentation is meaningless. In this embodiment, the curvature threshold is set to 0.001 per millimeter. The physical meaning of this value is: if there is only one arc with a curvature radius of 1000 millimeters in the entire path segment, and the length of this arc accounts for only a small proportion of the total length of the path segment, then the total curvature change will be less than this threshold. A curvature radius of 1000 millimeters is very close to a straight line in actual machining, and the tool wear rate on this arc segment is almost no different from that on a straight segment.

[0073] After comparison, the controller determines the subsequent processing path based on the comparison results. If the total curvature change is greater than or equal to the preset curvature threshold, it indicates that there is a section with a large degree of curvature within the path segment, and dividing it into high-curvature and low-curvature sub-segments based on curvature is of practical significance. At this time, the controller divides the processing path segment into at least one high-curvature sub-segment and at least one low-curvature sub-segment according to the curvature distribution information, as described above. Subsequently, it performs non-equal allocation according to preset allocation weights, and then generates a micro-compensation sequence within each sub-segment by weighting the allocation according to the feed rate ratio.

[0074] In a preferred embodiment, after comparing the total curvature change with a preset curvature threshold, the method further includes the following steps: If the total curvature change is less than the preset curvature threshold, it is determined to be an approximately straight-line cutting scenario. The estimated total wear is then weighted and distributed according to the feed rate ratio throughout the entire future machining path segment to generate the micro-compensation sequence.

[0075] Furthermore, when the total curvature change is determined to be less than a preset curvature threshold, the machining path segment is classified as an approximately straight-line cutting scenario. In this scenario, the curvature of the entire path segment is almost zero or minimal, and the difference in tool wear rate at different locations mainly originates not from curvature changes, but from changes in feed rate. The specific implementation method is as follows.

[0076] When the controller determines that the total curvature change is less than the preset curvature threshold, it skips the step of dividing the machining path segment into high-curvature and low-curvature sub-segments based on curvature distribution information, and also skips the step of non-equal allocation of each sub-segment according to preset weights. The controller directly adopts a feed rate proportional weighted allocation scheme to allocate the estimated total wear amount in one go over the entire future machining path segment.

[0077] The specific execution method of feed rate proportional weighted allocation is as follows: The controller discretizes the machining path segment along the machining direction into several micro-compensation positions, with the discretization density matching the interpolation cycle of the CNC system. For each micro-compensation position, the controller reads the corresponding feed rate value based on the feed rate variation curve obtained from the CNC system. The micro-compensation amount for each micro-compensation position is equal to the estimated total wear multiplied by the feed rate value at that position, and then divided by the sum of the feed rate values ​​for all micro-compensation positions. Thus, within the entire approximately straight path segment, positions with higher feed rates receive larger micro-compensation amounts, and positions with lower feed rates receive smaller micro-compensation amounts. The release density of the micro-compensation amount is proportional to the volume of material cut by the tool per unit time. Finally, a micro-compensation sequence synchronized with the machining feed process is generated.

[0078] If the feed rate remains constant throughout the entire path segment in this approximately linear cutting scenario, i.e., the feed rate variation curve is a horizontal straight line, then the feed rate values ​​at each micro-compensation position are equal. The aforementioned weighted allocation formula degenerates into an equal allocation, with each micro-compensation amount being equal, and the micro-compensation sequence being consistent with the sequence generated by the equal allocation scheme. This degenerate result is self-consistent, ensuring the logical consistency of the control method under different scenarios.

[0079] This constitutes a complete mechanism for determining curvature scenarios and selecting allocation strategies. When the curvature of the machining trajectory is sufficiently large, a joint allocation scheme of curvature and feed rate is executed; when the machining trajectory is approximately a straight line, it reverts to an allocation scheme that relies solely on feed rate. This dual-mode allocation strategy makes the entire control method applicable to different types of machining scenarios, such as milling contour machining and turning end face machining, and eliminates boundary blind spots where allocation logic fails due to unavailable curvature information or meaningless curvature itself.

[0080] In a preferred embodiment, driving the clamping mechanism to perform a step-compensation movement along the axial direction according to the corrected driving voltage includes the following steps: Within the time window of one step compensation movement, an initial driving voltage is applied to each group of piezoelectric ceramics. The initial driving voltage is calculated based on the micro-compensation amount and the nominal voltage-displacement curve of the piezoelectric ceramics. After the initial driving voltage is applied, the actual displacement of the cutting edge is obtained by the infrared sensor; The actual displacement is compared with the target micro-compensation amount of this step compensation to calculate the displacement deviation; The displacement deviation is input to a proportional-integral controller, and the proportional-integral controller outputs a compensation voltage correction amount; The compensation voltage correction is superimposed on the initial driving voltage and applied again to each group of piezoelectric ceramics as the corrected driving voltage to form a closed-loop displacement correction for a single step compensation movement.

[0081] Specifically, in the first step, within the time window of one step compensation movement, the controller applies an initial driving voltage to each group of piezoelectric ceramics. The time window of one step compensation movement refers to the time interval from the start of the current compensation to the start of the next compensation, and its length is equal to the step time interval of the micro-compensation sequence. Before applying the initial driving voltage, the controller first calculates the initial driving voltage based on the target micro-compensation amount of this step compensation and the nominal voltage-displacement curve of the piezoelectric ceramics. The nominal voltage-displacement curve is a static characteristic curve provided by the piezoelectric ceramic manufacturer, describing the correspondence between the driving voltage and the output displacement under quasi-static conditions, and is stored in the controller parameter storage area. The initial driving voltage is calculated by dividing the target micro-compensation amount by the nominal displacement sensitivity of the piezoelectric ceramic to obtain the corresponding voltage value. The controller uses this voltage value as the initial driving voltage and outputs it to each group of piezoelectric ceramics through the analog output interface. After being powered on, each group of piezoelectric ceramics begins to undergo axial elongation deformation. This deformation is transmitted to the clamping mechanism through the wedge block transmission mechanism inside the device, causing the clamping mechanism to begin moving along the axial direction.

[0082] The second step involves the controller acquiring the actual displacement of the clamping mechanism's front end via an infrared sensor after the initial drive voltage is applied. In this step, the controller reads the current distance value output by the infrared sensor and subtracts it from the distance reference value recorded before the start of this step compensation. This difference is used as the actual displacement of the clamping mechanism's front end. The distance reference value is the distance value read from the infrared sensor and stored in memory by the controller just moments before the initial drive voltage is applied.

[0083] Third, the controller compares the measured actual displacement with the target micro-compensation value for this step compensation and calculates the displacement deviation. The displacement deviation equals the target micro-compensation value minus the actual displacement, in micrometers. If the displacement deviation is positive, it means the actual displacement is insufficient to reach the target value, and the driving voltage needs to be increased to allow the piezoelectric ceramic to continue elongating. If the displacement deviation is negative, it means the actual displacement has exceeded the target value, and the driving voltage needs to be decreased to allow the piezoelectric ceramic to retract appropriately. If the absolute value of the displacement deviation is less than a preset allowable error (in this embodiment, the allowable error is set to 0.1 micrometers), then the step compensation is considered complete, and no further correction is performed.

[0084] The fourth step involves the controller inputting the displacement deviation into the proportional-integral (PI) controller. The PI controller is a closed-loop feedback algorithm module within the controller, its function being to calculate the appropriate correction voltage value based on the input deviation signal. The mathematical expression for the PI controller is: the output compensation voltage correction equals the proportional coefficient multiplied by the current displacement deviation, plus the integral coefficient multiplied by the integral value of the displacement deviation over time. The proportional coefficient represents the controller's immediate response strength to the current deviation; in this embodiment, it is set to 0.5 to 1.0 volts per micrometer, with the specific value determined by the reciprocal of the voltage-displacement sensitivity of the piezoelectric ceramic. The integral coefficient represents the controller's ability to correct historical accumulated deviations, used to eliminate steady-state residual deviations caused by creep or friction; in this embodiment, it is set to 0.05 to 0.2 volts per microsecond. The controller calculates the PI controller's output value once per control cycle, the control cycle being equal to the controller's interrupt cycle, typically 1 millisecond or less.

[0085] In the fifth step, the controller superimposes the compensation voltage correction output by the PI controller with the initial drive voltage calculated in the first step, resulting in the corrected drive voltage. The controller then applies the corrected drive voltage back to each group of piezoelectric ceramics, causing the piezoelectric ceramics to adjust their elongation according to the corrected voltage value, thus driving the clamping mechanism to move further to approach the target micro-compensation amount. Steps two through five can be executed multiple times within the time window of a single step compensation movement, forming a closed-loop displacement correction for that step compensation movement. The termination condition for the loop is that the absolute value of the displacement deviation is less than the preset allowable error, or the preset maximum number of correction loops is reached. In this embodiment, the maximum number of correction loops is set to 10 to prevent infinite loops due to sensor failure or mechanism jamming. When the termination condition is met, the current step compensation movement is completed, and the controller enters the waiting state for the next compensation.

[0086] In a preferred embodiment, the sudden wear threshold is determined according to the following steps: Before the continuous machining process begins, the clamping device is controlled to perform a pre-run during an empty stroke without contacting the workpiece, and during the pre-run, multiple reference distance values ​​are collected by the infrared sensor; The standard deviation of the measurement noise of the infrared sensor under no-travel conditions is calculated based on the fluctuation range of multiple reference distance values. The sudden wear threshold is set to N times the standard deviation of the measurement noise, where N is an integer between 3 and 6, so that the sudden wear threshold can distinguish between normal measurement noise and real sudden wear signals.

[0087] Specifically, the first step involves the controller controlling the clamping device to perform a pre-run during its idle stroke before the continuous machining process begins, i.e., before the tool contacts the workpiece. The pre-run process is as follows: the controller sends a pre-run program segment instruction to the CNC system via a digital communication interface. This program segment controls the machine tool spindle to rotate the clamping device at the normal machining speed, while simultaneously controlling the clamping device to move along a predetermined trajectory at the normal machining feed rate. However, no workpiece is installed on this trajectory, or a safe distance is maintained from the workpiece, ensuring that the tool does not contact any object during the entire pre-run. The pre-run duration is set to 5 to 10 seconds, which is sufficient to collect a adequate number of infrared sensor data samples for statistical analysis.

[0088] The second step involves the controller periodically acquiring multiple reference distance values ​​from the infrared sensor at a set normal sampling period. The reference distance value refers to the measured distance between the front end of the clamping mechanism and the machine tool's workpiece clamping tool, detected by the infrared sensor when the tool is not in contact with the workpiece. Since the tool does not wear during the pre-run, the reference distance value is theoretically constant. However, due to electronic noise inherent in the infrared sensor, coupled with minor vibrations from the machine tool's rotation and slight interference from cutting oil mist in the environment, the reference distance value will fluctuate within a very small range. The controller stores all reference distance values ​​acquired during the pre-run into a temporary array. The number of data points acquired is determined by dividing the pre-run duration by the sampling period, typically ranging from 10 to 50.

[0089] The third step involves the controller calculating the standard deviation of the infrared sensor's measurement noise under no-travel conditions based on the fluctuation amplitude of multiple collected reference distance values. The standard deviation is calculated as follows: First, the arithmetic mean of all reference distance values ​​is calculated. Then, for each reference distance value, its deviation from the mean is calculated, and the deviation is squared. Next, the squares of all deviations are summed, and this summation is divided by the number of reference distance values ​​minus one to obtain the sample variance. Finally, the square root of the variance is taken to obtain the standard deviation of the measurement noise, in micrometers.

[0090] Fourth, the controller sets the sudden wear threshold to N times the standard deviation of the measurement noise, where N is an integer between 3 and 6. In this embodiment, N is set to 5. The physical basis for the value of N comes from the Laida criterion in statistics, which states that the probability of a measured value falling outside the mean by plus or minus 3 standard deviations under a normal distribution is approximately 0.27%, and the probability of falling outside the mean by 6 standard deviations is far less than one in a million. Setting N between 3 and 6 means that only when the abnormal jump in the instantaneous wear rate exceeds 3 to 6 times the normal fluctuation range of the measurement noise will it be judged as a sudden wear signal. In this way, normal data fluctuations caused by infrared sensor measurement noise are almost impossible to trigger false alarms of sudden wear, while the sudden change in wear rate caused by actual tool chipping or hard point impact is usually much greater than 6 times the standard deviation of the measurement noise and can be reliably detected. The specific value of N can be selected according to the tolerance of the machining scenario for false alarms. For finishing scenarios, tool wear has a significant impact on machining accuracy and requires more sensitive detection of sudden wear; N can be a smaller value of 3 or 4. For rough machining scenarios, the machining vibration is large and the sensor noise itself is high. To reduce false alarms, N can be taken as a large value of 5 or 6. After the pre-run is completed, the controller stores the calculated sudden wear threshold in the parameter storage area, and uses this threshold for sudden wear diagnosis throughout the continuous machining process.

[0091] In a preferred embodiment, obtaining the actual displacement of the cutting tool tip using the infrared sensor includes the following steps: Within the time window of the single step compensation movement, the working mode of the infrared sensor is switched from wear detection mode to displacement feedback mode; In the displacement feedback mode, the sampling period of the infrared sensor is shortened, and the actual displacement of the tool tip is measured using the infrared sensor reading at the moment the initial driving voltage is applied as the zero reference. After a step-compensation movement is completed, the infrared sensor is restored to the wear detection mode, and the actual distance value is acquired again at the originally set sampling period.

[0092] Specifically, the first step involves switching the infrared sensor's operating mode from wear detection mode to displacement feedback mode at the start of a step compensation movement time window, i.e., before applying the initial drive voltage. The mode switch is implemented as follows: the controller internally modifies the infrared sensor's status flag from a value representing wear detection mode to a value representing displacement feedback mode. The controller determines the processing path for the infrared sensor data based on the status flag value. In wear detection mode, the acquired distance values ​​are stored in the wear rate calculation queue. In displacement feedback mode, the acquired distance values ​​are stored in the displacement feedback data buffer.

[0093] The second step, in displacement feedback mode, is to shorten the sampling period of the infrared sensor. The reason for shortening the sampling period is that the sampling period in wear detection mode is too long and cannot meet the real-time feedback requirements of closed-loop displacement correction. The sampling period in displacement feedback mode is set to 1 to 10 milliseconds, with the specific value depending on the maximum sampling frequency of the infrared sensor and the analog-to-digital conversion speed of the controller. The operation to shorten the sampling period is as follows: the controller resets the timing period of the sampling timer, changing it from the original period value of wear detection mode to the period value of displacement feedback mode.

[0094] Simultaneously, the controller establishes a zero-position reference in displacement feedback mode. This zero-position reference is established as follows: moment before applying the initial drive voltage, the controller reads a distance value from the infrared sensor and stores this value in memory as the zero-position reference. The physical meaning of the zero-position reference is the initial position of the clamping mechanism's front end before the start of this step compensation movement. Thereafter, each distance value read in displacement feedback mode is subtracted from the zero-position reference; this difference represents the actual displacement of the clamping mechanism's front end since the start of this compensation. The controller then provides this actual displacement to the closed-loop displacement correction process.

[0095] Thirdly, after one step compensation movement is completed, i.e., after the termination condition of closed-loop displacement correction is met, the controller restores the infrared sensor's operating mode to wear detection mode. The restoration operation includes: the controller changing the infrared sensor's status flag back to the value representing wear detection mode, restoring the sampling timer's timing period to the original sampling period value of the wear detection mode, and clearing the data in the displacement feedback data buffer. Afterward, the infrared sensor continues to periodically acquire the actual distance value between the front end of the clamping mechanism and the workpiece clamping tool at the originally set sampling period, for use in step S1 to calculate the instantaneous wear rate.

[0096] It should be noted that during the brief period of displacement feedback mode, the wear rate calculation queue pauses receiving new distance data because the infrared sensor is temporarily used for displacement feedback measurement. Since the duration of displacement feedback mode is extremely short, typically a portion of the step time interval (tens to hundreds of milliseconds), while the sampling period for wear detection is usually on the order of seconds, the time occupied by displacement feedback mode is much shorter than a normal sampling period. Therefore, it does not substantially affect the continuity of wear rate calculation.

[0097] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the inventive concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. An adaptive progressive online compensation control method for tool wear in CNC machine tools, applied to a tool wear self-compensating clamping device, the device comprising: The housing has a connecting part for connecting the machine tool spindle; A clamping mechanism is slidably mounted inside the housing along the axial direction of the housing and is used to clamp the cutting tool; An infrared sensor, fixedly installed, is used to detect the actual distance value representing the distance between the front end of the clamping mechanism and the machine tool workpiece clamping tool; Multiple sets of piezoelectric ceramics are fixedly installed and configured to drive the clamping mechanism to generate compensated movement along the axial direction by deformation upon power-on. The controller is electrically connected to both the infrared sensor and the piezoelectric ceramic sensor. The control method is characterized by being executed by the controller and includes the following steps: During continuous processing, the actual distance value is periodically acquired from the infrared sensor, and the instantaneous wear rate is calculated based on the difference between two consecutive actual distance values. Multiple instantaneous wear rates obtained through continuous calculation are input into the prediction model built into the controller in real time. The prediction model performs self-correction based on the newly input instantaneous wear rate and outputs the predicted average wear rate corresponding to a future processing path segment. Obtain the predetermined execution time of the future processing path segment, calculate the estimated total wear based on the predetermined execution time and the predicted average wear rate, and plan the estimated total wear as a set of micro-compensation sequences synchronized with the processing feed process; During the execution of the future machining path segment by the CNC system, according to the micro-compensation sequence, a driving voltage corresponding to each micro-compensation amount is synchronously applied to each group of piezoelectric ceramics, causing each group of piezoelectric ceramics to produce a series of step-by-step axial elongation deformations, thereby driving the clamping mechanism to perform multiple step-by-step compensation movements along the axial direction.

2. The adaptive progressive online compensation control method for CNC machine tool tool wear according to claim 1, characterized in that, It also includes the following steps: Each time the driving voltage is applied, the driving voltage is corrected according to the hysteresis model of the piezoelectric ceramic; Determine whether the corrected driving voltage exceeds the stroke safety threshold of the piezoelectric ceramic; If the limit is exceeded, an alarm will be triggered and compensation will be suspended. If the error does not exceed the limit, the clamping mechanism is driven to make one step compensation movement along the axis direction according to the corrected driving voltage.

3. The adaptive progressive online compensation control method for CNC machine tool tool wear according to claim 1, characterized in that, The prediction model includes a Kalman filter and a time-varying wear rate predictor connected in series with the Kalman filter; The prediction model performs self-correction based on the new input instantaneous wear rate, including the following steps: The Kalman filter receives the instantaneous wear rate calculated in the current sampling period as the observation value, and at the same time receives the predicted wear rate output by the time-varying wear rate predictor in the previous period as the state prediction value. The observation value and the state prediction value are fused by the Kalman gain to output the corrected wear rate for the current period. The time-varying wear rate predictor takes the corrected wear rate as input, fits an exponential time-varying curve characterizing the wear rate as a function of time, and extrapolates the time curve to the prediction time window corresponding to the next processing path segment on the time axis to obtain the predicted average wear rate. Wherein, the length of the prediction time window is the predetermined execution duration of the future machining path segment, the extrapolation step size of the time-varying wear rate predictor on the time axis is consistent with the sampling period, and the Kalman gain is dynamically adjusted according to the current machining state parameters, which include the spindle speed and feed rate.

4. The adaptive progressive online compensation control method for CNC machine tool tool wear according to claim 3, characterized in that, The time-varying wear rate predictor also performs the following steps: The instantaneous wear rate calculated in the current sampling period is compared with a preset sudden wear threshold. When the instantaneous wear rate exceeds the sudden wear threshold, a sudden wear event is determined to have occurred. The time-varying wear rate predictor discards the corrected wear rate of the current sampling period, uses the instantaneous wear rate as the current effective wear rate, and refits the time-varying curve to update the predicted average wear rate for the future processing path segment.

5. The adaptive progressive online compensation control method for CNC machine tool tool wear according to claim 1, characterized in that, The step of planning the estimated total wear amount into a set of micro-compensation sequences synchronized with the machining feed process includes the following steps: Obtain the curvature distribution information of the machining trajectory corresponding to the future machining path segment and the feed speed information of each sub-segment; Based on the curvature distribution information, the future processing path segment is divided into at least one high curvature sub-segment and at least one low curvature sub-segment; According to the preset allocation weight, the estimated total wear amount is non-equally allocated to each sub-segment, wherein the unit length compensation amount allocated to the high curvature sub-segment is greater than the unit length compensation amount allocated to the low curvature sub-segment. Within each sub-segment, based on the feed rate variation curve of that sub-segment, the total compensation amount allocated to that sub-segment is weighted and distributed to each micro-compensation amount according to the feed rate ratio, so that the higher the feed rate, the larger the compensation amount per unit time is allocated, and finally the micro-compensation sequence synchronized with the machining feed process is generated.

6. The adaptive progressive online compensation control method for CNC machine tool tool wear according to claim 5, characterized in that, After obtaining the curvature distribution information of the machining trajectory corresponding to the future machining path segment and the feed speed information of each sub-segment, the method further includes the following steps: Based on the curvature distribution information, calculate the total curvature change of the future processing path segment; The total curvature change is compared with a preset curvature threshold. The step of dividing the future processing path segment into at least one high-curvature sub-segment and at least one low-curvature sub-segment based on the curvature distribution information includes the following steps: If the total curvature change is greater than or equal to the preset curvature threshold, then based on the curvature distribution information, the future processing path segment is divided into at least one high curvature sub-segment and at least one low curvature sub-segment.

7. The adaptive progressive online compensation control method for CNC machine tool tool wear according to claim 6, characterized in that, After comparing the total curvature change with a preset curvature threshold, the method further includes the following steps: If the total curvature change is less than the preset curvature threshold, it is determined to be an approximately straight-line cutting scenario. The estimated total wear is then weighted and distributed according to the feed rate ratio throughout the entire future machining path segment to generate the micro-compensation sequence.

8. The adaptive progressive online compensation control method for CNC machine tool tool wear according to claim 2, characterized in that, The step of driving the clamping mechanism to perform a step-compensation movement along the axial direction according to the corrected driving voltage includes the following steps: Within the time window of one step compensation movement, an initial driving voltage is applied to each group of piezoelectric ceramics. The initial driving voltage is calculated based on the micro-compensation amount and the nominal voltage-displacement curve of the piezoelectric ceramics. After the initial driving voltage is applied, the actual displacement of the front end of the clamping mechanism is obtained by the infrared sensor; The actual displacement is compared with the target micro-compensation amount of this step compensation to calculate the displacement deviation; The displacement deviation is input to a proportional-integral controller, and the proportional-integral controller outputs a compensation voltage correction amount; The compensation voltage correction is superimposed on the initial driving voltage and applied again to each group of piezoelectric ceramics as the corrected driving voltage to form a closed-loop displacement correction for a single step compensation movement.

9. The adaptive progressive online compensation control method for CNC machine tool tool wear according to claim 4, characterized in that, The sudden wear threshold is determined according to the following steps: Before the continuous machining process begins, the clamping device is controlled to perform a pre-run during an empty stroke without contacting the workpiece, and during the pre-run, multiple reference distance values ​​are collected by the infrared sensor; The standard deviation of the measurement noise of the infrared sensor under no-travel conditions is calculated based on the fluctuation range of multiple reference distance values. The sudden wear threshold is set to N times the standard deviation of the measurement noise, where N is an integer between 3 and 6, so that the sudden wear threshold can distinguish between normal measurement noise and real sudden wear signals.

10. The adaptive progressive online compensation control method for CNC machine tool tool wear according to claim 8, characterized in that, The step of obtaining the actual displacement of the cutting tool tip using the infrared sensor includes the following steps: Within the time window of the single step compensation movement, the working mode of the infrared sensor is switched from wear detection mode to displacement feedback mode; In the displacement feedback mode, the sampling period of the infrared sensor is shortened, and the actual displacement of the front end of the clamping mechanism is measured with the infrared sensor reading at the moment the initial driving voltage is applied as the zero reference. After a step-compensation movement is completed, the infrared sensor is restored to the wear detection mode, and the actual distance value is acquired again at the originally set sampling period.

Images