A small tool numerical control polishing machine tool and an intelligent working method thereof

Abstract

The present application relates to a kind of small tool numerical control polishing machine tool, including machine body;Polishing tool;Driving mechanism, by spindle and polishing tool driving connection;Trimming unit, for on-line trimming of polishing tool;Standard test piece, close to trimming unit and be set on machine body;Interferometer, set on machine body, for the spot drop data of standard test piece;Numerical control module, set in machine body, numerical control module is electrically connected with driving mechanism, trimming unit, interferometer, numerical control module controls trimming unit and carries out trimming work to polishing tool, and after polishing tool completes trimming, it is tried to polish to standard test piece, and then based on the three-dimensional profile data of standard test piece after being collected by interferometer after trying to polish, removal function is updated on-line, and then subsequent compensation polishing work is carried out.The small tool numerical control polishing machine tool can realize polishing tool in situ trimming and removal function automatic calibration integration.The present application also relates to the intelligent working method of the small tool numerical control polishing machine tool.

Description

Technical Field

[0001] This invention relates to the field of optical precision machining equipment technology, specifically to a small-tool CNC polishing machine tool and its intelligent working method and storage medium capable of achieving in-situ trimming and automatic calibration of removal functions, used for the fully automatic calibration and updating of in-situ surface trimming and removal functions of polishing tools during CNC polishing of optical components. Background Technology

[0002] The process determinism of five-axis CNC polishing of small tools relies heavily on the stability of the removal function during machining. The three key parameters of the removal function—peak removal rate, full width at half maximum (FWHM), and volumetric removal rate—determine the material removal speed, the smoothness of the machining trajectory, and the overall machining efficiency, respectively. Any drift in these parameters will cause deviations in the dwell time calculation, resulting in stagnant surface convergence of the component and directly reducing the polishing accuracy.

[0003] The physical and chemical degradation of polishing tools during continuous processing is the main cause of removal function instability, mainly including three types of deterministic degradation mechanisms. The first is glazing degradation, where the polyurethane polishing pad on the polishing tool undergoes plastic creep under contact temperature rise of 5-15℃, normal pressure of 3-50N, and stress of 0.1-1.5MPa. This crushes and closes the micropores on the surface with a diameter of 5-20μm and a porosity of 30%-50%, causing a decrease in material removal rate. The removal rate decreases by 5%-15% for every 60 minutes of cumulative processing. The second is clogging degradation, where glass abrasive debris with a particle size of 0.5-10μm and a Mohs hardness of 5.5-6.5, as well as 10-50 micrometer cerium oxide agglomerates formed by van der Waals forces from 1-3μm primary particles, embed themselves in the micropores of the pad surface. This distorts the removal function from a standard Gaussian shape to an irregular shape, compromising process stability. Thirdly, wear and degradation occur. Processing shear force of 0.03~0.75MPa will cause the pad surface to wear layer by layer, with a wear rate of 0.5~2μm per hour. Long-term wear will cause the pad surface to deform and become uneven, completely losing the symmetry of the removal function.

[0004] Currently, the three main types of calibration methods in the industry include offline disassembly and assembly combined with a dedicated calibration machine and an independent interferometer; industrial camera vision monitoring combined with manual calibration; and online calibration methods adapted from semiconductor planarization processes. However, all of these methods have inherent technical defects and cannot meet the processing requirements of high-precision optical components. The offline disassembly and assembly method combined with a dedicated calibration machine and an independent interferometer suffers from multiple systematic errors. Disassembly and assembly introduces accumulated tool setting errors, and the parameters of the calibration sample do not match those of the actual workpiece. These errors, coupled with human judgment bias and coordinate system isolation, ultimately lead to severely substandard processing accuracy. The industrial camera vision monitoring method combined with manual calibration can only identify macroscopic defects in polishing tools and cannot detect nanoscale glazing loss and efficiency degradation. Furthermore, it lacks a closed-loop linkage mechanism, and human judgment is highly subjective, resulting in poor processing consistency. The online calibration method adapted from semiconductor planarization processes is extremely poorly suited to the working conditions, technical objectives, and processing scale of small tool polishing. Not only is the online calibration mode unsuitable for precision polishing scenarios, but it can also cause secondary glazing defects, making it difficult to meet precision machining requirements.

[0005] In summary, existing polishing and calibration processes generally suffer from technical bottlenecks such as insufficient accuracy and stability, low automation, and poor adaptability to operating conditions. They also fail to address core issues such as removal function drift and surface convergence failure caused by polishing tool degradation. Therefore, there is an urgent need to develop a quantitative closed-loop automated polishing system that integrates in-situ finishing and calibration to overcome existing process limitations and meet the precision polishing requirements of high-precision optical components. Summary of the Invention

[0006] The first technical problem to be solved by the present invention is to provide a small tool CNC polishing machine tool that can realize the integration of in-situ trimming of polishing tools and automatic calibration of removal functions in the above-mentioned prior art.

[0007] The second technical problem to be solved by the present invention is to provide an intelligent working method for a small tool CNC polishing machine tool, which can realize in-situ dressing of polishing tools and automatic calibration of removal functions, thereby improving polishing accuracy and the degree of polishing automation.

[0008] The technical solution adopted by the present invention to solve the first technical problem mentioned above is: a small tool CNC polishing machine tool, comprising...

[0009] Organism;

[0010] Polishing tools;

[0011] The drive mechanism is mounted on the machine body and is connected to the polishing tool via the spindle.

[0012] The dressing unit, mounted on the machine body, is used for online dressing of polishing tools;

[0013] A standard test piece is placed on the machine body near the trimming unit. The standard test piece is made of the same or equivalent material as the optical element workpiece to be processed.

[0014] An interferometer, mounted on the machine body, is used to collect spot fall data of standard specimens;

[0015] The CNC module is located inside the machine body. The CNC module is electrically connected to the drive mechanism, the trimming unit, and the interferometer. The CNC module controls the trimming unit to trim the polishing tool. After the polishing tool is trimmed, the standard specimen is test-polished. Then, based on the three-dimensional contour data of the standard specimen after the test polishing acquired by the interferometer, the removal function is calculated and updated online, and then subsequent compensation polishing work is performed.

[0016] As an improvement, the trimming unit includes a base, an ultrasonic transducer, and a trimming disc, wherein the ultrasonic transducer is integrated into the trimming disc, and the trimming disc is fixed on the base.

[0017] As an improvement, the upper surface of the dressing disc has a diamond abrasive layer, and the ultrasonic transducer is a Langevin piezoelectric ultrasonic transducer.

[0018] As an improvement, the diameter of the dressing disk is 40~80mm, the flatness is greater than or equal to 2μm, and the diamond grain size of the diamond abrasive layer is 15~30μm; the resonant frequency of the ultrasonic transducer is 40kHz±1kHz, and the power is adjustable in the range of 50~200W.

[0019] As an improvement, a positioning pin is provided on the body corresponding to the installation position of the standard test piece, and a positioning hole with interference fit to the positioning pin is provided on the bottom of the standard test piece.

[0020] As an improvement, the clearance between the positioning pin and the positioning hole is 2~5μm.

[0021] As an improvement, the standard specimen has a circular cross-section and is a cylindrical specimen with a diameter of 20-40 mm and a thickness of 10 mm.

[0022] The technical solution adopted by this invention to solve the second technical problem mentioned above is: an intelligent working method for a small tool CNC polishing machine tool as described above, comprising the following steps:

[0023] S1. The CNC module automatically starts in-situ polishing tool dressing and removal function calibration according to preset trigger conditions. The trigger conditions include the cumulative polishing time exceeding a preset time threshold and / or the online drift monitoring of the removal function indicating that the offset exceeds a preset offset threshold.

[0024] S2. The drive mechanism drives the polishing tool to the dressing unit, and the dressing unit performs online dressing of the polishing tool according to the set dressing parameters.

[0025] S3. The drive mechanism drives the polishing tool to move above the standard test piece, and then performs standardized test polishing on the standard test piece according to the preset test polishing parameters, forming a single-point test polishing spot on the surface of the standard test piece;

[0026] S4. The control interferometer collects the three-dimensional contour data of the single-point test-throwing spot. The CNC module performs two-dimensional Gaussian fitting on the three-dimensional contour data, extracts three feature parameters of the removal function: peak removal rate, full width at half maximum (FWHM), and volume removal rate. The extracted feature parameters are compared with the calibration benchmark value to obtain the offset. The graded response decision is executed based on the offset and the feature parameters of the removal function are automatically updated.

[0027] S5. Write the updated removal function feature parameters into the residence time solver parameter table and proceed to the subsequent compensation and polishing work.

[0028] As an improvement, in step S4, the graded response decision is as follows: when the offset is less than 10%, the Preston coefficient is automatically fine-tuned; when the offset is 10% to 30%, a re-repair operation is triggered and the repair time is automatically extended by 50%; after the re-repair operation is completed, steps S3 to S4 are re-executed; when the offset is greater than 30%, a prompt is made to replace the polishing tool.

[0029] As an improvement, the preset threshold for the cumulative polishing time in step S1 is 30~120 minutes;

[0030] The trimming parameters in step S2 are: trimming pressure of 5~15N, trimming speed of 200~1000rpm, trimming time of 10~60 seconds, ultrasonic transducer power of 50~200W, and ultrasonic transducer vibration frequency of 40kHz.

[0031] The test polishing parameters in step S3 are: test polishing pressure of 5~20N, polishing tool speed of 1000rpm, and dwell time of 5~30 seconds.

[0032] Compared with the prior art, the advantages of the present invention are as follows: The small tool CNC polishing machine and its intelligent working method in the present invention change the technical prejudice that the removal function in the field of small tool polishing must be manually calibrated by the operator on a dedicated offline device. In the present invention, the polishing tool is trimmed online by a trimming unit, and the trimmed polishing tool is used to test polish a standard specimen. Then, the trimming effect is calibrated in real time in the form of quantitative data of the removal function parameter offset, so that the trimming parameter is automatically adjusted according to the calibration result, forming a quantitative closed loop that has never been realized in the field of optical surface forming. It integrates in-situ trimming of polishing tool and automatic calibration of removal function, improving polishing accuracy and polishing automation. Attached Figure Description

[0033] Figure 1 This is a schematic diagram of the structure of a small tool CNC polishing machine in an embodiment of the present invention.

[0034] Figure 2 This is a longitudinal structural view of the trimming unit in an embodiment of the present invention.

[0035] Figure 3 This is a longitudinal mounting view of the standard test specimen on the machine body in an embodiment of the present invention.

[0036] Figure 4 This is a flowchart illustrating the intelligent working method of a small-tool CNC polishing machine in an embodiment of the present invention.

[0037] Figure 5 This is a block diagram of the hierarchical response decision logic in an embodiment of the present invention. Detailed Implementation

[0038] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0039] like Figures 1 to 3 As shown, the small-tool CNC polishing machine tool in this embodiment includes a machine body 1, a polishing tool 2, a drive mechanism 3, a dressing unit 4, a standard test piece 5, an interferometer 6, and a CNC module 7. The machine body 1 serves as the supporting foundation for the entire machine tool, providing installation references and motion support for each functional component. The drive mechanism 3 is mounted on the machine body 1 and is connected to the polishing tool 2 via a spindle. The drive mechanism 3 has five-axis linkage capability, enabling it to drive the polishing tool 2 to move along the normal direction of any curved surface in three-dimensional space and control the rotation speed, thereby performing CNC polishing processing on optical component workpieces installed within the working area of ​​the machine body 1.

[0040] Polishing tool 2 is an end effector that directly performs material removal. Its lower end has a polyurethane polishing pad. The polishing pad contacts and rotates with the workpiece surface in the polishing liquid environment, and material removal is achieved by means of the mechanochemical action of abrasive particles in the polishing liquid. During continuous polishing, three types of deterministic degradation occur on the surface of the polishing pad: First, glazing occurs when polyurethane undergoes plastic creep under the combined action of a temperature rise of 5-15°C and a normal pressure of 3-50N (0.1-1.5MPa stress) in the contact area. The micropores with a diameter of 5-20μm and a porosity of 30%-50% are crushed and closed, and the material removal rate decreases by 5%-15% every 60 minutes. Second, clogging occurs when glass abrasives with a particle size of 0.5-10μm and a Mohs hardness of 5.5-6.5, as well as 10-50μm cerium oxide agglomerates formed by the agglomeration of 1-3μm primary particles through van der Waals forces, become embedded in the micropores of the pad surface, causing the removal function to change from a standard Gaussian shape to an irregular shape. Third, wear occurs when the processing shear force of 0.03-0.75MPa causes the pad surface to wear layer by layer at a rate of approximately 0.5-2μm per hour. Long-term wear causes the pad surface to change from a planar surface to a concave or convex surface, resulting in the loss of symmetry of the removal function. All three degradation mechanisms mentioned above will cause the peak removal rate, full width at half maximum (FWHM), and volume removal rate of the removal function to drift, thereby causing dwell time calculation deviations and surface convergence stalls.

[0041] To address the removal function drift problem caused by the degradation of the polishing tool 2, this embodiment incorporates a dressing unit 4 on the machine body 1 for online, in-situ dressing of the polishing tool 2. The dressing unit 4 includes a base 41, an ultrasonic transducer 42, and a dressing disc 43. The base 41 is fixedly mounted at a predetermined position on the edge of the machine body 1's worktable, providing stable support for the ultrasonic transducer 42 and the dressing disc 43. The dressing disc 43 is fixed above the base 41. The upper surface of the dressing disc 43 has a diamond abrasive layer, with the diamond particle size selected within the range of 15-30 μm. A nickel electroforming binder is used to bond the diamond particles to the surface of the dressing disc 43. The diameter of the dressing disc 43 is selected within the range of 40-80 mm, and its surface flatness is better than 2 μm, ensuring uniform and consistent material removal from the polishing pad surface during the dressing process. The ultrasonic transducer 42 is integrated inside the trimming plate 43. In this embodiment, the ultrasonic transducer 42 is a Langevin piezoelectric ultrasonic transducer with a resonant frequency of 40kHz±1kHz and an output power adjustable in the range of 50~200W.

[0042] During dressing, the drive mechanism 3 moves the polishing tool 2 above the dressing disc 43 via the spindle, pressing the polishing tool 2 onto the surface of the dressing disc 43 with a dressing pressure of 5-15N. The polishing tool 2 rotates at a dressing speed of 200-1000rpm, and the dressing time is 10-60 seconds. Simultaneously, the ultrasonic transducer 42 drives the dressing disc 43 to generate high-frequency ultrasonic vibration at a frequency of 40kHz and a power of 50-200W. During the rotational contact of the polishing tool 2, the diamond abrasive layer performs micro-cutting action on the surface of the polishing pad, removing the surface material that has become rheologically closed due to glazing and the agglomerates of abrasive debris clogging the micropores. This allows the surface of the polishing pad to re-expose a fresh porous structure, restoring its ability to carry polishing fluid and remove material. The ultrasonic vibration superimposed on the diamond dressing motion generates micron-level high-frequency impact and relative motion between the diamond particles and the polishing pad. This effectively breaks down and dislodging glass abrasives and cerium oxide aggregates embedded in the micropores, achieving a self-cleaning effect for the dressing disc 43. At the same time, it promotes the reopening of micropores that have been crushed and closed, restoring the porosity of the polyurethane polishing pad. This mechanism avoids the secondary glazing defects caused by continuous contact dressing in traditional online semiconductor planarization dressing solutions.

[0043] The standard test piece 5 is mounted on the machine body 1 near the dressing unit 4. The standard test piece 5 is made of the same or equivalent material as the optical element workpiece to be processed. This material matching requirement is a key factor in ensuring the validity of the calibration results, because the Preston coefficient is directly related to the workpiece material. Only by performing polishing on a material identical or equivalent to the actual processing material can the extracted removal function parameters accurately reflect the tool characteristics under real processing conditions. The standard test piece 5 has a circular cross-section, being a cylindrical sample with a diameter of 20-40 mm and a thickness of 10 mm. A positioning pin 11 is provided on the machine body 1 corresponding to the mounting position of the standard test piece 5. A positioning hole 51 is provided on the bottom of the standard test piece 5, forming an interference fit with the positioning pin 11, with a clearance of 2-5 μm. This precision positioning interface ensures the repeatability of the positioning accuracy of the standard test piece 5 after each installation, making the position of the upper surface of the test piece in the machine tool coordinate system highly deterministic and repeatable. The positional accuracy of the polishing spots is not affected by changes in the test piece clamping. The relative positions of the standard specimen 5 and the dressing disc 43 in the machine tool coordinate system are precisely calibrated at the factory and written into the CNC system parameter table. Based on this, the drive mechanism 3 automatically calculates the motion path of the polishing tool 2 from the dressing disc 43 to the standard specimen 5, and the movement repeatability positioning accuracy is better than ±3μm.

[0044] Interferometer 6 is mounted on the machine body 1 and is used to acquire the three-dimensional contour data of the standard specimen 5 after the test polishing. Interferometer 6 adopts an in-situ interferometric measurement method, directly scanning and measuring the test polishing spots on the surface of the standard specimen 5 after the test polishing is completed. It acquires the three-dimensional contour data of the spots at a resolution of 256×256 pixels or higher and transmits it to the CNC module 7. There is no need to remove the standard specimen 5 from the machine tool and send it to an independent measuring device, thus eliminating the secondary positioning error and time loss caused by disassembly and assembly in offline measurement.

[0045] The CNC module 7 is located inside the machine body 1 and is electrically connected to the drive mechanism 3, the dressing unit 4, and the interferometer 6. The CNC module 7 is the control core of the closed-loop control for the in-situ dressing and automatic calibration of the removal function of the polishing tool 2. Its control logic is as follows: When the trigger condition is met, the CNC module 7 controls the dressing unit 4 to perform in-situ dressing on the polishing tool 2; after dressing, it controls the drive mechanism 3 to move the polishing tool 2 above the standard test piece 5 to perform standardized test polishing; after test polishing, it controls the interferometer 6 to collect the three-dimensional contour data of the test polishing spots; based on the collected contour data, the CNC module 7 calculates and updates the removal function online, writes the updated parameters into the residence time solver, and then controls the subsequent compensation polishing work.

[0046] like Figure 4 As shown, the intelligent working method in this embodiment includes the following steps S1 to S5.

[0047] In step S1, the CNC module 7 automatically starts the in-situ polishing tool 2's dressing and removal function calibration work according to preset trigger conditions. There are two trigger conditions: one is that the cumulative polishing time exceeds a preset time threshold, which can be set within the range of 30-120 minutes; the other is that the online drift monitoring of the removal function indicates an offset exceeding a preset offset threshold. One or both can be used as needed. The cumulative polishing time threshold is set based on the decay rate of the polishing pad's removal function. Under typical process conditions, the removal rate of the polishing material decreases by 5%-15% every 60 minutes of accumulated polishing time on the polyurethane polishing pad. At 30 minutes, the decay is approximately 2.5%-7.5%, close to the 10% fine-tuning threshold. At 120 minutes, the decay has entered the range requiring re-adjustment. Users can flexibly set this range according to the polishing pad material and specific process requirements. Online drift monitoring determines whether the removal function has significantly shifted by comparing the current processing parameters with the model's predicted values ​​in real time. Once the offset exceeds the threshold, it is triggered immediately without waiting for the accumulated time to reach its limit. The two triggering methods complement each other: time-based triggering ensures periodic maintenance, while drift-based triggering provides immediate response.

[0048] In step S2, the drive mechanism 3 drives the polishing tool 2 to the dressing unit 4, where the dressing unit 4 performs online dressing of the polishing tool 2 according to the set dressing parameters. In this embodiment, the dressing parameters are: dressing pressure 5~15N, dressing speed 200~1000rpm, dressing time 10~60 seconds, ultrasonic transducer 42 power 50~200W, and ultrasonic transducer 42 vibration frequency 40kHz. Within these parameter ranges, the diamond abrasive layer combined with high-frequency ultrasonic vibration can effectively restore the porous structure and Gaussian morphology of the removal function on the polishing pad surface. During the dressing process, the CNC module 7 monitors the spindle load and the working status of the ultrasonic transducer 42 in real time to ensure a stable and controllable dressing process.

[0049] In step S3, the drive mechanism 3 drives the polishing tool 2 to move above the standard specimen 5, and then performs standardized test polishing on the standard specimen 5 according to preset test polishing parameters, forming a single-point test polishing spot on the surface of the standard specimen 5. The test polishing parameters are: test polishing pressure 5~20N, polishing tool 2 rotation speed 1000rpm, and dwell time 5~30 seconds. During the test polishing, the polishing tool 2 contacts and rotates along the normal direction of the surface of the standard specimen 5. Since the drive mechanism 3 has the capability of five-axis linkage surface normal control, the contact state between the polishing tool 2 and the specimen during test polishing is completely consistent with that during actual processing of optical components, both contacting along the normal direction. This ensures that the removal function obtained by test polishing is highly consistent with the removal function in the actual polishing process, overcoming the fundamental defect of inconsistent test polishing conditions and actual processing conditions in traditional offline calibration.

[0050] In step S4, the CNC module 7 controls the interferometer 6 to acquire the three-dimensional contour data of the single-point test polishing spots and performs two-dimensional Gaussian fitting on the three-dimensional contour data. The physical basis for selecting the two-dimensional Gaussian model is that the theoretical removal function of the polishing tool 2, derived from the Preston equation and Hertz contact theory, exhibits a rotationally symmetric Gaussian shape under the conditions of a planar workpiece and constant contact force. Therefore, the two-dimensional Gaussian function can accurately describe the contour distribution of the removed spots. During the fitting process, Gaussian model parameters are extracted, and a goodness of fit greater than 0.95 is used as the data quality qualification criterion. When the goodness of fit is lower than 0.95, the system determines that the data quality is unqualified and automatically re-acquires data. After fitting, three characteristic parameters of the removal function are extracted: the peak removal rate determines the material removal speed, which is obtained by converting the Gaussian peak amplitude; the half-height and full width determine the spacing between adjacent trajectories and the path smoothing effect, which is calculated by the standard deviation; and the volume removal rate determines the overall processing efficiency, which is obtained by integrating the Gaussian function over the entire contact area. The CNC module 7 compares the extracted three characteristic parameters with the calibration reference value to calculate the offset and performs a graded response decision based on the offset.

[0051] like Figure 5As shown, the graded response decision is divided into three levels. When the offset is less than 10%, it indicates that the polishing tool 2 is within the normal wear range. The CNC module 7 automatically fine-tunes the Preston coefficient to ensure that the removal function model in the residence time solver is consistent with the current actual state. When the offset is between 10% and 30%, it indicates that the polishing pad has degraded beyond the normal range. The CNC module 7 automatically triggers a re-repair operation and extends the repair time by 50% to more fully restore the surface state of the polishing pad. After the re-repair operation is completed, the third step of trial polishing and the fourth step of calibration are re-executed until the offset returns to an acceptable range. When the offset is greater than 30%, it indicates that the polishing tool 2 has severely degraded and cannot be restored by repair. The CNC module 7 issues a prompt on the operation interface to replace the polishing tool 2, notifying the operator to perform the replacement.

[0052] In step S5, the CNC module 7 writes the updated removal function characteristic parameters into the dwell time solver parameter table and proceeds to the subsequent compensation polishing work. The dwell time solver recalculates the dwell time distribution of each machining trajectory point based on the updated removal function model to ensure that shape convergence remains accurate and controllable even after the state of the polishing tool 2 changes.

[0053] The entire process, from triggering, trimming, trial polishing, measurement, fitting, decision-making to parameter updating, is completed automatically without manual intervention. The total time for a single automatic cycle of trimming, trial polishing, and calibration is less than 10 minutes, including 10-60 seconds for ultrasonic trimming, 2-5 seconds for multi-axis linkage rapid positioning, 5-30 seconds for trial polishing of the speckle, 5-30 seconds for interferometer 6 acquisition and transmission, and 1-5 seconds for Gaussian fitting, grading judgment, and parameter updating. Compared to the traditional offline disassembly and assembly, dedicated trimming machine trimming, and independent interferometer 6 calibration process, which typically takes 30-60 minutes, this embodiment saves 60%-83% of non-machining time, significantly improving the effective utilization rate of the machine tool.

[0054] The technical effects of this embodiment will be explained below with specific processing examples.

[0055] In the processing of small- and medium-diameter optical lenses, the workpiece to be processed is a Φ100mm BK7 spherical lens with a radius of curvature R=200mm, an initial surface shape PV value of 2.5λ, and an RMS value of 0.38λ. The dressing unit 4 uses a Φ60mm dressing disc 43 with a diamond abrasive layer of 20μm grit, nickel electroforming binder, and a flatness of 1.5μm. It integrates a 40kHz Langevin piezoelectric ultrasonic transducer at the bottom, with a power of 100W, a driving voltage of 200V peak-to-peak, and an amplitude of 10μm peak-to-peak. The standard specimen 5 is a Φ25.4mm×10mm BK7 optical glass cylinder, made of the same material as the workpiece. The bottom positioning hole 51 mates with a Φ6mm positioning pin 11 on the machine body 1, with a clearance of 3μm. The system is set to trigger an automatic dressing cycle to calibration every 60 minutes of cumulative polishing.

[0056] Upon initial triggering, the drive mechanism 3 moves the polishing tool 2 to the dressing disc 43, performing ultrasonic dressing with a dressing pressure of 10N, a rotation speed of 500rpm, a dressing time of 30 seconds, and an ultrasonic power of 100W. After dressing, the spindle is raised by 20mm, moved 85.000mm in the X direction, and lowered in the Z direction to contact the specimen, performing single-point polishing with a test polishing pressure of 12N, a rotation speed of 1000rpm, and a dwell time of 15 seconds. The interferometer 6 acquires the three-dimensional contour of the spot at 256×256 pixels. The two-dimensional Gaussian fitting result is a goodness of fit of 0.978, a peak amplitude of 0.312μm, σx=3.42mm, and σy=3.38mm, from which the peak removal rate is extracted to be 1.248μm / min, the full width at half maximum (FWHM) is 8.01mm, and the volume removal rate is 1.51mm³ / min. Compared with the most recent calibration baseline values ​​of 1.312 μm / min peak removal rate and 7.85 mm full width at half maximum (FWHM) width, the peak removal rate shifted by -4.9% and the FWHM width shifted by +2.0%, both less than 10%. The system automatically fine-tuned the Preston coefficient and then wrote it into the residence time solver. The entire process took 7 minutes and 38 seconds.

[0057] During the subsequent six rounds of compensation polishing, the system automatically triggered three trimming-to-calibration cycles at cumulative polishing minutes of 60, 120, and 180 minutes, with Preston coefficient fine-tuning amounts of -3.8%, -7.2%, and -12.5%, respectively. At the third trigger, the offset reached the 10%~30% level, and the system automatically extended the trimming time to 45 seconds and performed a re-trimming. After the re-trimming, a trial polishing and calibration were performed, and the coefficient recovered to -6.1%. The final surface shape converged to a PV value of 0.058λ (λ / 17) and an RMS value of 0.009λ (λ / 111), with a total processing time of 4.5 hours, of which the three automatic trimming-to-calibration cycles took approximately 23 minutes. Compared to the traditional offline trimming-to-calibration method, which requires 30 minutes per cycle and 1.5 hours of non-processing time for three cycles, this embodiment saves approximately 67% of non-processing time.

[0058] In the long-cycle machining of large-diameter aspherical optical components, the workpiece to be machined is a Φ350mm microcrystalline glass aspherical mirror with a vertex curvature radius R=2400mm and a quadratic constant K=-1.2, with an estimated total machining time of 18 hours. To meet the requirements of long-cycle, high-precision machining, the diameter of the dressing disc 43 was increased to Φ80mm, and the diamond abrasive layer particle size was reduced to 15μm to obtain a more uniform dressing texture and a surface roughness of less than 0.5nm. The ultrasonic power was increased to 150W, and the amplitude was increased to a peak-to-peak value of 15μm to enhance the self-cleaning effect. The standard specimen 5 is equipped with two material options: microcrystalline glass and fused silica. Automatic switching is achieved through a multi-station rotary standard specimen 5 holder on the machine body 1, with an indexing repeatability of ±3μm. The CNC module 7 automatically selects the standard specimen 5 with the corresponding material based on the current workpiece material code. The removal function fitting is extended to a seven-parameter rotational asymmetric Gaussian model including rotation angle parameters to be compatible with the non-rotational symmetric removal function caused by uneven edge trimming of large-diameter tools. The graded response threshold is tightened according to the requirements of high-precision machining: when the offset is less than 8%, a fine adjustment coefficient is applied; when the offset is 8% to 25%, a repair is performed; and when the offset is greater than 25%, a tool replacement is prompted.

[0059] A removal function drift model was established based on the wear characteristics of polyurethane polishing pads on microcrystalline glass. During the 18-hour processing run, it was expected to automatically trigger 7 trimming-to-calibration cycles, approximately every 2.5 hours. The peak removal rate of the removal function was expected to gradually decrease from the initial 1.85 μm / min to 1.52 μm / min, a total decrease of 17.8%, while the full width at half maximum (FWHM) was expected to gradually expand from 12.3 mm to 14.1 mm, a total increase of 14.6%. After 7 automatic graded response processes, the 4th and 6th cycles triggered automatic re-trimming at 10%~30% settings. The Preston coefficients were automatically updated after each calibration, and the surface shape steadily converged from the initial PV value of 5.2λ to 0.065λ (λ / 15), and the RMS value converged from 0.92λ to 0.011λ (λ / 91). In contrast, under the same conditions, a machining process using a fixed removal function model without any in-situ trimming and calibration was simulated. After the fourth round, the shape converged but began to diverge, and the residual PV value rebounded from 0.8λ to 2.1λ. After the eighth round, the PV value rebounded to above 0.5λ. This indicates that for long-cycle, large-diameter machining, in-situ trimming to calibration closed loop is an indispensable key guarantee.

[0060] The small-tool CNC polishing machine and its intelligent working method in this invention change the technical bias in the existing small-tool polishing field that the removal function must be manually calibrated by the operator on a dedicated offline device. In this invention, the polishing tool 2 is trimmed online by the trimming unit 4, and the standard test piece 5 is test-polished by the trimmed polishing tool 2. Then, the trimming effect is calibrated in real time in the form of quantitative data of the removal function parameter offset, so that the trimming parameter is automatically adjusted according to the calibration result, forming a quantitative closed loop that has never been realized in the field of optical surface forming. It integrates in-situ trimming of polishing tool 2 and automatic calibration of removal function, improving polishing accuracy and polishing automation.

[0061] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. For those skilled in the art, various modifications and improvements can be made without departing from the concept of the present invention, such as equivalent substitutions, additions, deletions, or recombinations of the technical features in the above embodiments to form new technical solutions; these all fall within the scope of protection of the present invention. Therefore, the patent protection scope of the present invention should be determined by the scope defined in the appended claims.

[0062] The specification and claims of this invention use terms indicating direction, such as "front," "rear," "upper," "lower," "left," "right," "side," "top," and "bottom," to describe various exemplary structural parts and elements of the invention. However, these terms are used herein merely for ease of explanation and are determined based on the exemplary orientations shown in the accompanying drawings. Since the embodiments disclosed in this invention can be arranged in different orientations, these terms indicating direction are for illustrative purposes only and should not be considered as limitations. For example, "upper" and "lower" are not necessarily limited to directions opposite to or consistent with the direction of gravity.

Claims

1. A small-tool CNC polishing machine tool, comprising: Body (1); Polishing tools (2); The drive mechanism (3) is mounted on the machine body (1) and is connected to the polishing tool (2) via the spindle; Its features are: Also includes A trimming unit (4) is installed on the machine body (1) and is used to trim the polishing tool (2) online. A standard test piece (5) is placed on the machine body (1) near the trimming unit (4). The standard test piece (5) is made of the same or equivalent material as the optical element workpiece to be processed. Interferometer (6), mounted on the body (1), is used to collect spot fall data of standard specimen (5); The numerical control module (7) is installed inside the machine body (1). The numerical control module (7) is electrically connected to the drive mechanism (3), the trimming unit (4), and the interferometer (6). The numerical control module (7) controls the trimming unit (4) to trim the polishing tool (2). After the polishing tool (2) is trimmed, the standard specimen (5) is test-polished. Then, based on the three-dimensional contour data of the standard specimen (5) after the test polishing collected by the interferometer (6), the online update removal function is calculated, and then subsequent compensation polishing work is carried out.

2. The small tool CNC polishing machine tool according to claim 1, characterized in that: The trimming unit (4) includes a base (41), an ultrasonic transducer (42), and a trimming disc (43). The ultrasonic transducer (42) is integrated into the trimming disc (43), and the trimming disc (43) is fixed on the base (41).

3. The small tool CNC polishing machine tool according to claim 2, characterized in that: The upper surface of the dressing disc (43) has a diamond abrasive layer, and the ultrasonic transducer (42) is a Langevin piezoelectric ultrasonic transducer.

4. The small tool CNC polishing machine tool according to claim 3, characterized in that: The diameter of the dressing disc (43) is 40~80mm, the flatness is greater than or equal to 2μm, and the diamond grain size of the diamond abrasive layer is 15~30μm; the resonant frequency of the ultrasonic transducer (42) is 40kHz±1kHz, and the power is adjustable in the range of 50~200W.

5. The small tool CNC polishing machine tool according to any one of claims 1 to 4, characterized in that: The body (1) is provided with a positioning pin (11) corresponding to the installation position of the standard test piece (5), and the bottom of the standard test piece (5) is provided with a positioning hole (51) that is interference fit with the positioning pin (11).

6. The small tool CNC polishing machine tool according to claim 5, characterized in that: The clearance between the positioning pin (11) and the positioning hole (51) is 2~5μm.

7. The small tool CNC polishing machine tool according to any one of claims 1 to 4, characterized in that: The standard specimen (5) has a circular cross-section and is a cylindrical specimen with a diameter of 20-40 mm and a thickness of 10 mm.

8. An intelligent working method for a small tool CNC polishing machine tool as described in any one of claims 1 to 7, characterized in that: Includes the following steps: S1. The CNC module (7) automatically starts the in-situ polishing tool (2) trimming work and removal function calibration work according to the preset trigger conditions. The trigger conditions include the cumulative polishing time exceeding the preset time threshold and / or the online drift monitoring of the removal function indicating that the offset exceeds the preset offset threshold. S2. The drive mechanism (3) drives the polishing tool (2) to move to the dressing unit (4), and uses the dressing unit (4) to dress the polishing tool (2) online according to the set dressing parameters; S3. The driving mechanism (3) drives the polishing tool (2) to move above the standard test piece (5), and then performs standardized test polishing work on the standard test piece (5) according to the preset test polishing parameters, forming a single-point test polishing spot on the surface of the standard test piece (5). S4. The control interferometer (6) collects the three-dimensional contour data of the single-point test-throwing spot. The numerical control module (7) performs two-dimensional Gaussian fitting on the three-dimensional contour data, extracts three feature parameters of the removal function: peak removal rate, full width at half maximum and volume removal rate, and compares the extracted feature parameters with the calibration benchmark value to obtain the offset. Based on the offset, it performs hierarchical response decision and automatically updates the feature parameters of the removal function. S5. Write the updated removal function feature parameters into the residence time solver parameter table and proceed to the subsequent compensation and polishing work.

9. The working method according to claim 8, characterized in that: In step S4, the graded response decision is as follows: when the offset is less than 10%, the Preston coefficient is automatically fine-tuned; when the offset is 10% to 30%, the re-repair work is triggered and the repair time is automatically extended by 50%; after the re-repair work is completed, steps S3 to S4 are re-executed; when the offset is greater than 30%, the polishing tool is prompted to be replaced (2).

10. The working method according to claim 8, characterized in that: The preset threshold for the cumulative polishing time mentioned in step S1 is 30~120 minutes; The trimming parameters in step S2 are: trimming pressure is 5~15N, trimming speed is 200~1000rpm, trimming time is 10~60 seconds, ultrasonic transducer (42) power is 50~200W, and ultrasonic transducer (42) vibration frequency is 40kHz. The test polishing parameters in step S3 are: test polishing pressure Wie5~20N, polishing tool (2) speed of 1000rpm, and dwell time of 5~30 seconds.

Images