Hemispherical resonator precision intelligent machining method, system, and medium

Abstract

This invention relates to a precision intelligent machining method, system, and medium for hemispherical resonators. The method includes: constructing a primary digital twin model of the machining equipment; performing surface sampling and contour fitting on the rough-machined hemispherical resonator to establish a feed rate mapping model and plan the grinding wheel trajectory; performing on-machine shaping and closed-loop calibration of the grinding wheel based on the grinding wheel trajectory; correcting and optimizing the primary digital twin model to obtain a high-precision digital twin model; performing machining simulation, state monitoring, and error prediction based on the high-precision digital twin model, feeding back the error prediction results to the control system to adaptively adjust machining parameters; performing on-machine accuracy testing and evaluation on the finished product, comparing and verifying the test data with the model, and realizing closed-loop iterative optimization of the entire machining process. This invention can effectively avoid the breakage and cracking of hard and brittle thin-walled workpieces, realize simulation, monitoring, error prediction, and adaptive parameter adjustment of the entire machining process, and achieve ultra-precision intelligent machining.

Description

Technical Field

[0001] This invention relates to the field of precision manufacturing and intelligent machining technology, specifically to a precision intelligent machining method, system, and medium for hemispherical harmonic oscillators. Background Technology

[0002] A hemispherical resonant gyroscope (HRG), a high-precision inertial navigation device, uses a hemispherical resonator—its core sensing element—which is typically made of high-purity fused silica. This material has extremely high hardness and brittleness, and very low fracture toughness. In the actual manufacturing process of the hemispherical resonator, the surface profile of the rough-machined workpiece deviates significantly from the theoretically designed profile due to factors such as machine tool precision, tool wear, and the release of internal stress in the material. The machining allowance is unevenly distributed and difficult to quantify precisely, making precise control of the grinding process impossible.

[0003] Traditional precision grinding processes typically employ fixed feed rates and preset machining paths. When dealing with workpiece surfaces exhibiting abrupt changes in allowance, these processes are prone to defects such as chipping and micro-cracks due to rapid variations in grinding force, severely impacting the machining accuracy and yield of hemispherical resonators. Furthermore, to ensure machining safety, existing processes generally employ conservative machining parameters, resulting in low machining efficiency and difficulty in meeting the demands of mass production with high precision.

[0004] Furthermore, traditional processing methods lack the ability to monitor the processing process in real time, predict errors, and dynamically adjust them, making it difficult to balance the processing accuracy, surface quality, yield, and production efficiency of hemispherical harmonic oscillators. There is an urgent need for a high-precision, intelligent, and adaptive precision processing method to solve the above-mentioned technical problems. Summary of the Invention

[0005] In view of the above-mentioned defects or deficiencies in the prior art, the present invention proposes a precision intelligent machining method, system and storage medium for hemispherical resonators.

[0006] This invention provides a precision intelligent machining method for hemispherical resonators, comprising: constructing a primary digital twin model of the machining equipment; performing rough machining on the hemispherical resonator, and performing surface sampling and contour fitting on the rough-machined hemispherical resonator, calculating the machining allowance based on the actual contour, establishing a feed rate mapping model and planning the grinding wheel trajectory; performing precision grinding on the hemispherical resonator based on the grinding wheel trajectory, while simultaneously performing in-machine shaping and closed-loop calibration of the grinding wheel; collecting multi-source data during the grinding and shaping process, correcting and optimizing the primary digital twin model to obtain a high-precision digital twin model; performing machining simulation, state monitoring and error prediction based on the high-precision digital twin model, feeding back the error prediction results to the control system to adaptively adjust the machining parameters; performing in-machine accuracy detection and evaluation on the finished product, comparing and verifying the detection data with the model, realizing closed-loop iterative optimization of the entire machining process, and completing the high-precision intelligent machining of the hemispherical resonator.

[0007] In one optional implementation, the hemispherical harmonic oscillator is rough-machined, and the rough-machined hemispherical harmonic oscillator undergoes surface sampling and contour fitting. Machining allowance is calculated based on the actual contour, a feed rate mapping model is established, and a grinding wheel trajectory is planned. Specifically, this includes: discretizing the rough-machined hemispherical harmonic oscillator to obtain a set of surface points; performing curve fitting on the set of surface points to generate a true workpiece contour curve; calculating the radial machining allowance at each point based on the theoretical contour curve and the true workpiece contour curve; establishing a feed rate mapping model based on the radial machining allowance at each point; and forming a grinding wheel trajectory based on the feed rate mapping model.

[0008] In one alternative implementation, laser non-contact measurement is used to acquire point cloud data of the workpiece surface, and the contour curve is fitted using the least squares method to generate a real workpiece contour curve.

[0009] In one optional implementation, establishing the feed rate mapping model specifically involves: using the machining allowance as input and the grinding wheel feed rate as output, establishing a nonlinear mapping relationship and performing real-time matching.

[0010] In one optional implementation, the grinding wheel trajectory is generated by interpolation based on the feed rate mapping model and the spherical coordinates of the hemispherical harmonic oscillator, and the trajectory points correspond one-to-one with the machining allowance.

[0011] In one alternative implementation, the on-machine dressing includes using a diamond roller to dress the grinding wheel profile online, so that the grinding wheel profile always maintains the theoretical spherical accuracy.

[0012] In one alternative implementation, the multi-source data includes grinding wheel wear data, machining vibration data, contour error data, feed displacement data, and temperature drift data.

[0013] In one alternative implementation, in-machine accuracy testing includes online evaluation of the roundness, sphericity, wall thickness uniformity, and surface roughness of the hemispherical harmonic oscillator.

[0014] This invention also proposes a precision intelligent machining system for hemispherical harmonic oscillators, comprising: a digital twin modeling unit for constructing a primary digital twin model of the machining equipment; a roughing execution unit for roughing the hemispherical harmonic oscillator; a data acquisition and trajectory planning unit for surface sampling and contour fitting of the rough-machined hemispherical harmonic oscillator, calculating the machining allowance based on the actual contour, and establishing a feed rate mapping model to plan the grinding wheel trajectory; a precision grinding execution unit for precision grinding of the hemispherical harmonic oscillator according to the grinding wheel trajectory, while simultaneously performing on-machine shaping and closed-loop calibration of the grinding wheel; a model optimization unit for acquiring multi-source data during the grinding and shaping process, correcting and optimizing the primary digital twin model to obtain a high-precision digital twin model; an adaptive control unit for using the high-precision digital twin model for machining simulation, state monitoring, and error prediction, and feeding back the error prediction results to the control system to adaptively adjust the machining parameters; and an on-machine detection and closed-loop iteration unit for on-machine accuracy detection and evaluation of the machined product, comparing and verifying the detection data with the high-precision digital twin model, and realizing closed-loop iterative optimization of the entire machining process.

[0015] The present invention also proposes a storage medium, characterized in that it stores a computer program, which, when executed by a processor, implements any of the precision intelligent machining methods for hemispherical harmonic oscillators described in the present invention.

[0016] The beneficial technical effects of this invention are as follows: by constructing a primary digital twin model and combining it with on-machine measurement, contour fitting, and adaptive trajectory planning, the machining allowance of the hemispherical harmonic oscillator can be accurately quantified and the feed rate can be dynamically matched, effectively avoiding the breakage and cracking of hard and brittle thin-walled workpieces. At the same time, the grinding wheel is used for online dressing to ensure grinding accuracy. Then, a high-precision digital twin model is formed through multi-source data iterative optimization, realizing simulation, monitoring, error prediction, and adaptive parameter adjustment of the entire machining process. Finally, combined with on-machine detection, a closed-loop iterative optimization is formed, which significantly improves the machining accuracy, surface quality, yield, and machining efficiency of the hemispherical harmonic oscillator as a whole, realizing ultra-precision intelligent machining. Attached Figure Description

[0017] 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:

[0018] Figure 1 A flowchart illustrating a precision intelligent machining method for a hemispherical harmonic oscillator provided in an embodiment of the present invention;

[0019] Figure 2 A flowchart of a precision intelligent machining method for a hemispherical harmonic oscillator provided in another embodiment of the present invention;

[0020] Figure 3A flowchart of the grinding wheel on-machine shaping and closed-loop calibration method for the precision intelligent machining method of hemispherical harmonic oscillators provided in an embodiment of the present invention;

[0021] Figure 4 A schematic diagram of the fitting curve for the precision intelligent machining method of a hemispherical harmonic oscillator provided in an embodiment of the present invention;

[0022] Figure 5 A digital twin architecture diagram of the physical entity-virtual model-service platform for the precision intelligent machining method of hemispherical harmonic oscillators provided in an embodiment of the present invention;

[0023] Figure 6 The overall flowchart of a precision intelligent machining method for a hemispherical harmonic oscillator provided in another embodiment of the present invention is shown. Detailed Implementation

[0024] In the manufacturing process chain of hemispherical harmonic oscillators, the transition from rough machining using the forming method to ultra-precision grinding presents significant technological challenges, primarily in the following aspects:

[0025] 1. Uncertainty and uneven distribution of machining allowance: Due to limitations in the positioning accuracy of roughing machine tools, nonlinear wear of diamond tools, and the release of internal stress in quartz materials, the surface morphology of the rough-machined hemispherical blank is complex, often exhibiting an irregular "hill"-like allowance distribution. The actual profile deviates significantly from the theoretical design profile, and quantitative description is lacking.

[0026] 2. Machining Risks of Hard and Brittle Thin-Walled Parts: Traditional CNC machining programs (G-code) typically employ fixed toolpaths and constant feed rates. When the grinding wheel cuts into a "hill" area with abrupt changes in allowance, the instantaneous grinding force increases dramatically. For quartz thin-walled parts with wall thicknesses only in the millimeter range, this impact load can easily induce brittle fracture or subsurface microcrack propagation, directly leading to the scrapping of high-value workpieces.

[0027] 3. The contradiction between processing efficiency and safety: To avoid the risk of breakage, existing processes are forced to adopt a conservative strategy, that is, setting an extremely low safe feed rate based on the maximum estimated allowance. This results in the machine tool being in an inefficient "empty cutting" state for a long time in the "flat zone" where the allowance is small or non-existent, which seriously restricts the improvement of production efficiency.

[0028] The following is in conjunction with the appendix Figures 1 to 6 The present application will be further described in detail with reference to the embodiments. It is understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention. It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0029] 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.

[0030] Please refer to Figure 1 and Figure 2 This invention provides a precision intelligent machining method for a hemispherical harmonic oscillator, comprising the following steps:

[0031] Step S101: Construct a primary digital twin model of the processing equipment.

[0032] Step S103: Roughly machine the hemispherical resonator, and perform surface sampling and contour fitting on the rough-machined hemispherical resonator. Calculate the machining allowance based on the actual contour, establish a feed rate mapping model, and plan the grinding wheel trajectory.

[0033] Step S105: Precision grinding of the hemispherical harmonic oscillator based on the grinding wheel trajectory, while simultaneously performing on-machine dressing and closed-loop calibration of the grinding wheel;

[0034] Step S107: Collect multi-source data during the grinding and shaping process, correct and optimize the primary digital twin model to obtain a high-precision digital twin model;

[0035] Step S109: Perform machining simulation, status monitoring and error prediction based on a high-precision digital twin model, and feed the error prediction results back to the control system to adaptively adjust the machining parameters;

[0036] Step S1011: Perform in-machine accuracy testing and evaluation on the finished product, compare and verify the test data with the model, realize closed-loop iterative optimization of the entire processing process, and complete the high-precision intelligent processing of the hemispherical harmonic oscillator.

[0037] This invention is based on an adaptive grinding process for in-machine testing, which establishes a full-field digital characterization of the hemispherical harmonic oscillator margin by integrating a high-precision contact probe on a precision grinding machine.

[0038] Before dressing the grinding wheel, the machining equipment (grinding machine, grinding wheel, workpiece) already has an initial digital twin model, that is, a primary digital twin model of the machining equipment, which is built based on CAD drawings, factory parameters, and historical data. This model contains information such as the ideal geometric profile of the grinding wheel, the geometric errors of the machine tool, and dynamic characteristics, and is the basis for subsequent dressing simulation and decision-making.

[0039] By constructing a primary digital twin model of the processing equipment and combining it with in-machine high-precision measurement and surface contour fitting technology, a full-field digital and accurate characterization of the actual shape of the hemispherical harmonic oscillator after rough machining is achieved. This enables the rapid acquisition of the workpiece's true contour and quantitative calculation of the radial machining allowance at each point. It fundamentally solves the technical problems of uneven distribution of rough machining allowance, inaccurate quantification, and difficulty in compensating for contour deviations in traditional processes, providing a reliable data foundation for subsequent adaptive grinding.

[0040] This invention constructs a primary digital twin model of the processing equipment and combines it with in-machine high-precision measurement and surface contour fitting technology to achieve full-field digital and accurate characterization of the actual shape of the hemispherical harmonic oscillator after rough machining. It can quickly obtain the true contour of the workpiece and quantitatively calculate the radial machining allowance at each point. This fundamentally solves the technical problems of uneven distribution of rough machining allowance, inaccurate quantification, and difficulty in compensating for contour deviation in traditional processes, and provides a reliable data foundation for subsequent adaptive grinding.

[0041] This invention constructs a feed rate mapping model based on the machining allowance in real time. It can dynamically generate the optimal grinding wheel trajectory according to the size of the local allowance on the workpiece, achieving adaptive matching of the feed rate during grinding. This avoids fatal defects in fused silica materials, such as edge chipping, breakage, and subsurface crack propagation, caused by a sudden increase in grinding force due to abrupt changes in the allowance. It significantly improves the stability and yield of thin-walled, brittle hemispherical resonators during ultra-precision grinding. Simultaneously, the dynamic adaptive machining method effectively solves the problems of low machining efficiency and long machine tool idle running time caused by traditional fixed feed rates, greatly improving grinding efficiency while ensuring machining safety.

[0042] Simultaneous machining and closed-loop precision calibration of the grinding wheel during grinding enables real-time correction of wheel wear, profile deviations, and equipment motion errors, ensuring the grinding wheel maintains an ideal cutting state and further improving the surface accuracy and quality of the hemispherical harmonic oscillator. By collecting multi-source machining data and continuously refining and iteratively optimizing the primary digital twin model, a high-precision digital twin is formed. This allows for real-time simulation, status monitoring, and error prediction throughout the entire machining process. The prediction results are then fed back to the control system in real-time for adaptive adjustment of machining parameters, giving the machining process intelligent characteristics of self-sensing, self-decision-making, and self-correction.

[0043] Finally, through precision testing and model data comparison verification, a closed-loop iterative mechanism of "modeling-processing-inspection-optimization-reprocessing" is formed, continuously improving the accuracy of virtual-real mapping and processing control, so that the roundness, sphericity, wall thickness uniformity, and surface roughness of the hemispherical resonator all meet the requirements of ultra-precision machining. This invention achieves an organic unity of high precision, high stability, high efficiency, and high intelligence, and is particularly suitable for the mass production of hard, brittle, thin-walled hemispherical resonator devices such as fused silica.

[0044] Further, based on step S103, the hemispherical harmonic oscillator is rough-machined, and the surface of the rough-machined hemispherical harmonic oscillator is sampled and its contour is fitted. The machining allowance is calculated based on the actual contour, a feed rate mapping model is established, and the grinding wheel trajectory is planned. Specifically, the steps include the following:

[0045] Step S1031: Discretely sample the rough-machined hemispherical harmonic oscillator to obtain a surface point set.

[0046] Step S1033: Perform curve fitting on the surface point set to generate the true workpiece contour curve.

[0047] Step S1035: Calculate the radial machining allowance at each point based on the theoretical contour curve and the actual workpiece contour curve.

[0048] Step S1037: Establish a feed rate mapping model based on the radial machining allowance at each point.

[0049] Step S1039: Form the grinding wheel trajectory based on the feed rate mapping model.

[0050] This invention incorporates a multi-level intelligent processing mechanism in the precision grinding control process to further enhance machining reliability and control accuracy. Firstly, an intelligent screening mechanism is established. Through preset contour deviation thresholds and rapid detection logic for key feature points, the rough-machined workpiece is pre-detected before grinding begins. This automatically identifies and removes unqualified blanks such as those with overcutting or excessive deformation, preventing defective products from flowing into subsequent processing steps and reducing ineffective processing losses and equipment usage.

[0051] Meanwhile, a high-density discrete sampling and high-precision contour reconstruction strategy is adopted to perform micron-level discrete sampling along the Z-axis direction of the hemispherical harmonic oscillator generatrix, with the sampling interval set to 0.02 mm, in order to obtain a high-density workpiece surface point set P(x,z).

[0052] At the machining execution level, a parametric macro program reconstruction and adaptive control mode are adopted, abandoning the fixed path and fixed parameter G-code program in traditional machining, and developing an intelligent macro program architecture driven by process variables in real time. Among them, the geometric adaptive control module uses the fitted actual contour radius R as a dynamic variable input to the system, driving the grinding wheel trajectory to adjust in real time according to the actual shape of the workpiece, realizing precise envelope matching between the grinding wheel and the workpiece surface, and ensuring uniform and stable grinding removal.

[0053] This invention, by setting up an intelligent screening mechanism and a rapid detection logic for key feature points, can automatically remove overcut waste and out-of-tolerance blanks before processing, avoiding ineffective processing and resource waste, reducing material costs and equipment energy consumption, and improving the economy and stability of the overall process.

[0054] By replacing the traditional fixed G-code with a parametric intelligent macro program, the machining logic can be made more flexible and intelligent. Through geometric adaptive control, the grinding wheel trajectory can be dynamically enveloped according to the actual contour of the workpiece, which can effectively solve problems such as grinding impact and surface deviation caused by uneven allowance, and significantly improve the machining accuracy and surface quality of the hemispherical harmonic oscillator.

[0055] Among them, laser non-contact measurement is used to acquire point cloud data of the workpiece surface, and the least squares method is used to fit the contour curve to generate the real workpiece contour curve.

[0056] By combining least squares filtering with B-spline interpolation, measurement noise and outliers are effectively eliminated, and the surface profile curve R=f(Z) representing the actual shape of the workpiece is accurately fitted and generated. The radial machining allowance Δ(Z) is calculated point by point based on the theoretical profile and the actual profile, so as to realize the full-field quantitative representation of the allowance distribution.

[0057] By employing a high-density micron-level discrete sampling along the hemispherical generatrix combined with a contour reconstruction algorithm using least squares and B-spline interpolation, the true shape of the workpiece can be accurately restored while suppressing measurement noise. This enables precise calculation of machining allowances and full-field digital characterization, providing reliable data support for adaptive grinding.

[0058] The feed rate mapping model is established by taking the machining allowance as input and the grinding wheel feed rate as output, establishing a nonlinear mapping relationship and performing real-time matching.

[0059] Using the radial machining allowance corresponding to each sampling point of the hemispherical harmonic oscillator as the input variable and the grinding wheel feed rate as the output variable, and combining the grinding removal characteristics of fused silica (a hard and brittle material), the dynamic response characteristics of the machine tool, and the cutting performance of the grinding wheel, a nonlinear mapping relationship between the machining allowance and the feed rate is constructed through process experiments and data fitting. This mapping relationship is then embedded into the machining control system to achieve real-time matching and dynamic allocation of the machining allowance and the grinding wheel feed rate. During the grinding process, the control system automatically outputs the optimal feed rate according to the real-time acquired local machining allowance and the nonlinear mapping model, achieving adaptive adjustment for low-speed grinding in areas with large machining allowances and high-speed grinding in areas with small machining allowances.

[0060] By constructing a nonlinear mapping model with machining allowance as input and grinding wheel feed rate as output, the optimal grinding parameters can be dynamically matched according to the local material removal amount of the workpiece. This fundamentally solves the problems of sudden grinding force changes and large processing impact caused by traditional constant feed rate, effectively avoids edge chipping, breakage and subsurface crack defects in fused silica hemispherical resonators, and greatly improves the processing safety and yield of thin-walled hard and brittle components.

[0061] The real-time matching mechanism enables adaptive speed adjustment during the processing, maximizing material removal efficiency while ensuring processing accuracy and surface quality. It also eliminates machine tool idle cutting and inefficient running periods, significantly shortens the single-piece processing cycle of the hemispherical harmonic oscillator, and improves overall processing efficiency.

[0062] This nonlinear mapping model closely matches actual processing conditions and material properties, offering high control precision and fast response speed. It can be highly coordinated with digital twins, on-machine inspection, trajectory planning, and other processes, further enhancing the intelligence and adaptability of the entire processing flow and ensuring stable compliance with surface accuracy and wall thickness uniformity standards.

[0063] Furthermore, the grinding wheel trajectory is generated by interpolation based on the feed rate mapping model and the spherical coordinates of the hemispherical harmonic oscillator, and the trajectory points correspond one-to-one with the machining allowance.

[0064] The grinding wheel trajectory is generated with high precision based on the feed rate mapping model and the standard spherical coordinates of the hemispherical harmonic oscillator. During trajectory planning, the theoretical spherical polar coordinates and generatrix coordinates of the hemispherical harmonic oscillator are used as the reference. The radial machining allowance and adaptive feed rate corresponding to each sampling point are used as the trajectory interpolation input parameters. The spherical interpolation algorithm calculates the center coordinates and attitude angle of the grinding wheel point by point, so that the generated trajectory points correspond one-to-one with the machining allowance at each position on the workpiece surface and are accurately matched. Finally, an adaptive grinding trajectory that fits the actual shape of the workpiece is formed, ensuring that the grinding wheel and the spherical surface of the harmonic oscillator maintain a stable normal contact state during the grinding process.

[0065] By combining the feed rate mapping model with the spherical coordinate interpolation algorithm to generate an adaptive grinding wheel trajectory, the grinding path can be made to strictly fit the real spherical shape of the hemispherical harmonic oscillator, and the trajectory points and machining allowances can be matched one-to-one. This solves the problem of grinding error accumulation caused by uneven distribution of allowances after roughing from the machining path level, and significantly improves the spherical contour accuracy and wall thickness uniformity of the hemispherical harmonic oscillator.

[0066] The dedicated interpolation method based on spherical coordinates better matches the rotational symmetry characteristics of the hemispherical harmonic oscillator, which can effectively improve the smoothness of the trajectory and the accuracy of the motion. It avoids problems such as trajectory abrupt changes and motion impact caused by traditional rectangular coordinate interpolation, reduces vibration and impact loads during the grinding process, and protects the hard and brittle fused silica material from defects such as edge chipping and cracks.

[0067] The trajectory points correspond one-to-one with the machining allowance, enabling precise removal and uniform grinding of local allowances, avoiding over-cutting and under-cutting. At the same time, with the adaptive feed speed, it realizes intelligent grinding that "slows down when the allowance is large and speeds up when the allowance is small", which greatly improves the processing efficiency while ensuring the processing quality, and achieves a synergistic unity of high precision and high efficiency.

[0068] Physically adaptive (variable speed grinding) establishes a mapping model between feed rate F and machining allowance Δ(Z). In the "hilly area" where the allowance is greater than the threshold, the feed rate is automatically reduced to Flow to reduce grinding force; in the "flat area" where the allowance approaches zero, the feed rate is automatically increased to Fhigh to eliminate idle time, achieving high-efficiency grinding with constant or quasi-constant load.

[0069] The aforementioned adaptive grinding process based on in-machine measurement profoundly combines the material removal mechanism of quartz glass with the macro-programming function of the CNC system. By "using measurement instead of exploration and using calculation to optimize grinding," it transforms traditional blind processing into deterministic intelligent processing. This not only solves the current problem of breakage caused by uneven allowance, but also represents a crucial step towards intelligent manufacturing of hemispherical harmonic oscillators.

[0070] like Figure 2 As shown, the hemispherical resonator hard and brittle thin-walled component provided in this embodiment is based on an adaptive precision grinding process using in-machine measurement, and its specific implementation is as follows:

[0071] First, the rough-machined hemispherical resonator workpiece is clamped and machined, and a high-precision probe is integrated into the machining equipment for in-machine inspection. Rapid measurement of key feature points determines whether the workpiece allowance meets machining requirements; if the allowance is insufficient, it is deemed scrap and rejected; if the allowance is sufficient, high-density discrete sampling is performed along the Z-axis at intervals of ΔZ = 0.02 mm to obtain a set of discrete points on the workpiece surface. .

[0072] The next step is data modeling and macro program reconstruction: the sampled discrete point set is smoothed and denoised, and the contour is reconstructed through curve fitting to obtain the workpiece's true contour curve R=f(Z); the normal machining allowance Δ(Z) is calculated point by point based on the theoretical contour and the true contour; based on this, a parameterized macro program is established, and geometric correction variables (#R_FIT) and allowance model data are generated to provide data support for subsequent adaptive grinding.

[0073] During the adaptive grinding cycle execution phase, the machining cycle is started in a Z-axis layered manner: the current machining point's allowance status is judged in real time. If it is determined to be a "hill" area (area with large allowance), a low feed rate strategy A is adopted; if it is determined to be a "flat" area (area with small allowance), a high feed rate strategy B is adopted. Then, the grinding action is executed, and the grinding wheel movement is controlled by the G01 X#R_FITZ#Z_CURRENT F#FEED_RATE command. After the current layer is completed, the Z-axis stepping is performed, and the cycle is repeated until the entire workpiece is machined.

[0074] The on-machine dressing and closed-loop calibration technology effectively avoids the secondary clamping deviation of traditional offline dressing by integrating dressing devices, precise positioning and calibration, and real-time error feedback compensation, thus ensuring the stability of grinding accuracy.

[0075] Integrated dressing device and precise positioning calibration: A dedicated grinding wheel dressing device is integrated into the side of the grinding table, realizing the integrated layout of dressing function and grinding process, simplifying the machining process. A high-precision probe of the machine tool is used to calibrate the position of the dressing tool. Through fitting and optimization of multiple sets of sampling data, the reference position of the dressing tool in the machine tool coordinate system is accurately determined, providing a reliable positioning basis for subsequent dressing motion control.

[0076] In-situ dressing: Using the calibrated dressing tool position as a reference, the CNC system drives the grinding wheel to move along a preset trajectory, completing the in-situ dressing operation in the machine tool coordinate system. This method eliminates the need to disassemble and transfer the grinding wheel, avoiding positioning deviations caused by secondary clamping of the grinding wheel during offline dressing, and ensuring coordinate consistency between the dressing action and subsequent grinding actions.

[0077] Re-measurement feedback and error compensation closed-loop control: After the shaping operation is completed, the actual diameter and profile of the grinding wheel are re-measured with high precision using the same probe, and the deviation data between the actual geometric parameters and the ideal parameters are collected. The error data is fed back to the CNC system in real time. The system automatically generates correction commands through a preset compensation algorithm to adjust the grinding wheel motion parameters to offset the deviation, forming a closed-loop control link of "calibration-shaping-re-measurement-compensation".

[0078] The aforementioned technology, through integrated design, precise calibration, and a closed-loop feedback mechanism, fundamentally eliminates the secondary clamping eccentricity error of traditional offline wheel dressing, significantly improving the accuracy and consistency of grinding wheel dressing, and providing core technological support for high-precision grinding. The framework of this method is as follows: Figure 3 As shown.

[0079] The machining process includes using diamond rollers to dress the grinding wheel profile online, ensuring that the grinding wheel profile always maintains the theoretical spherical accuracy.

[0080] The dressing process of the grinding wheel in the machine does not require disassembly of the grinding wheel and can be completed directly on the machining station. By acquiring the current contour dimensions, curvature, and wear of the grinding wheel in real time through the machine detection unit, the detection data is compared and calculated with the theoretical spherical contour. The diamond roller is driven to precisely dress the working surface of the grinding wheel according to the preset dressing path, continuously correcting the grinding wheel wear, changes in the blunt radius, and contour distortion. This ensures that the grinding wheel maintains the theoretical spherical accuracy, cutting edge sharpness, and geometric consistency throughout the entire grinding cycle, guaranteeing a high degree of fit between the grinding envelope and the theoretical profile of the hemispherical harmonic oscillator.

[0081] The use of diamond rollers enables online grinding wheel dressing, allowing the dressing operation to be completed without stopping the machine to disassemble and install the grinding wheel. This significantly reduces auxiliary time and clamping errors, greatly improves the continuity of processing and overall production efficiency, and meets the needs of continuous and batch precision machining of hemispherical harmonic oscillators.

[0082] By maintaining the theoretical spherical accuracy of the grinding wheel through machine repair, problems such as out-of-tolerance resonator profile and uneven surface texture caused by grinding wheel wear and profile distortion can be effectively avoided, significantly improving the profile accuracy, sphericity and wall thickness uniformity of the hemispherical resonator.

[0083] Maintaining the grinding wheel's sharpness and ideal geometry reduces grinding force and heat accumulation, decreasing the probability of microcracks, edge chipping, and subsurface damage in fused silica, thus significantly improving finished product yield and processing stability.

[0084] By forming a closed-loop collaborative control between machining and grinding, the entire process of "machining-inspection-machining-remachining" is fully automated, further improving the system's intelligence level and machining reliability.

[0085] The multi-source data includes grinding wheel wear data, machining vibration data, contour error data, feed displacement data, and temperature drift data.

[0086] In-machine precision testing includes online evaluation of the roundness, sphericity, wall thickness uniformity, and surface roughness of the hemispherical harmonic oscillator.

[0087] Based on digital twin virtual development and debugging, a digital twin of the processing equipment is constructed, adopting an open architecture of "CNC production line + AI module".

[0088] The system utilizes the EtherCAT high-speed bus to collect heterogeneous data from multiple sources, such as current, position, and speed, and then uses big data analysis to uncover key performance indicators.

[0089] A digital twin driven by a hybrid mechanism model (dynamic equations) and a data model (neural network) is established to achieve real-time prediction of cutting force, profile error and chatter.

[0090] Virtual debugging of CNC code is performed in a virtual space to verify the logical correctness and collision safety of the adaptive macro program, achieving "virtual-real synchronization and virtual-to-real control".

[0091] This invention also proposes a precision intelligent machining system for hemispherical harmonic oscillators, comprising:

[0092] Digital twin modeling unit, used to build a primary digital twin model of processing equipment.

[0093] The roughing execution unit is used for roughing the hemispherical harmonic oscillator.

[0094] The data acquisition and trajectory planning unit is used to perform surface sampling and contour fitting on the rough-machined hemispherical harmonic oscillator, calculate the machining allowance based on the actual contour, and establish a feed rate mapping model to plan the grinding wheel trajectory.

[0095] The precision grinding actuator is used to perform precision grinding on the hemispherical harmonic oscillator according to the grinding wheel trajectory, while simultaneously performing on-machine dressing and closed-loop calibration of the grinding wheel.

[0096] The model optimization unit is used to collect multi-source data during the grinding and shaping process, and to correct and optimize the primary digital twin model to obtain a high-precision digital twin model.

[0097] The adaptive control unit is used to perform machining simulation, status monitoring and error prediction using a high-precision digital twin model, and feeds back the error prediction results to the control system to adaptively adjust the machining parameters.

[0098] The in-machine inspection and closed-loop iteration unit is used to perform in-machine accuracy inspection and evaluation on the finished product, compare and verify the inspection data with a high-precision digital twin model, and realize closed-loop iterative optimization of the entire processing process.

[0099] This invention proposes an ultra-precision intelligent machining technology for hemispherical harmonic oscillators based on in-machine testing. Through in-machine accuracy testing of the hemispherical harmonic oscillators and in-machine wheel shaping, it intelligently plans the machining allowance for ultra-precision processes, providing crucial technical support for achieving high-precision machining. This invention also proposes a modular virtual development method for ultra-precision grinding machining centers based on digital twin modeling and virtual debugging technology. Through virtual debugging and modular optimization design, it lays the foundation for further development of ultra-precision CNC machining centers.

[0100] Example 1: Adaptive precision grinding process for hemispherical harmonic oscillators.

[0101] This embodiment provides an adaptive precision grinding method for hemispherical harmonic oscillators. The equipment used is an ultra-precision grinding machine equipped with a high-precision contact probe (such as the Renishaw OMP series) and an open CNC system. The data modeling and decision-making steps are completed on the host computer (PC), and the machining execution steps are completed on the slave computer (CNC).

[0102] Step S1: Workpiece Selection and Intelligent Decision Making

[0103] After the rough-machined hemispherical resonator workpiece is clamped on the grinding machine table, the system calls the probe detection subroutine to quickly measure 3-5 key feature points at the top of the hemisphere and near the equator.

[0104] The system's preset criterion is: if the measured radius R of any point is... meas <(R final +△ safe ), where Rfinal For the final design radius of the workpiece, △ safe This is the preset minimum safety margin. When this criterion is met, the workpiece is determined to be a defective blank, the system automatically outputs a "raw material scrap" alarm signal, and skips subsequent processing steps to avoid ineffective processing and waste of equipment resources.

[0105] The specific control logic is as follows:

[0106] 1. Parameter conversion: Convert the minimum safety margin Δ in the diameter direction. safe Converted to radial safety margin, safe = △ safe / 2, and calculate the minimum safe radius minSafeRadius} = R final +safe.

[0107] 2. Data Traversal and Statistics: Traverse all measurement points, calculate the actual radius of each point, determine the minimum radius minRadius, the maximum radius maxRadius, and the measurement point index corresponding to the minimum radius, and count the number of "danger points" with radii smaller than the minimum safe radius.

[0108] 3. Alarm and Handling: If the minimum radius is less than the minimum safety radius, the system outputs a scrap alarm, identifies the cause of the alarm (insufficient raw material size, insufficient safety margin, etc.), and provides handling suggestions (stop processing, check measuring equipment, replace with qualified billet, etc.) to ensure the safety and economy of the processing flow.

[0109] Step S2: High-resolution topography reconstruction. Perform a full contour scan on the qualified workpiece.

[0110] 1. Sampling strategy: Along the Z-axis (rotation axis), the sampling step size ΔZ = 0.02 mm is set. The probe is triggered along the generatrix, and the system records the machine tool coordinates (Xi, Zi).

[0111] 2. Data cleaning: The moving average filtering algorithm is used to remove singularities caused by surface dust or electromagnetic interference.

[0112] 3. Margin Calculation: Let the set of points on the actual trajectory be P(x, z), then the radial margin at any position is:

[0113] .

[0114] 4. Mathematical Modeling: The discrete point set is fitted using a cubic B-spline algorithm to generate a continuous and smooth actual contour function X=ffit(Z). For example... Figure 4 As shown.

[0115] Step S3: Generation and execution of parameterized macro programs

[0116] Based on the aforementioned contour fitting model and allowance calculation model, the system automatically generates and calls the parametric machining macro program to achieve adaptive control of the grinding process.

[0117] First, define the processing variables:

[0118] #100 is the current Z-axis coordinate;

[0119] #101 is the coordinate of the Z-axis machining endpoint;

[0120] #102 is the fitting radius corresponding to the current Z position, which is read from the table by the algorithm;

[0121] #103 represents the machining allowance corresponding to the current position;

[0122] #104 is the preset margin threshold;

[0123] #105 is the dynamic adaptive feed rate.

[0124] During machining, #100 is assigned as the starting coordinate of the Z-axis, and cyclic machining is performed as long as #100 is greater than or equal to the ending coordinate of the Z-axis #101.

[0125] 1. Read the fitted radius #102 corresponding to the current Z coordinate, and obtain the current machining allowance #103 by subtracting the theoretical radius from the fitted radius.

[0126] 2. Adaptive feed rate determination based on machining allowance:

[0127] If the machining allowance #103 is greater than the allowance threshold #104, it is determined to be a high load area, i.e. a hilly area. A flexible grinding strategy is adopted, and the feed rate #105 is set to low speed.

[0128] If the machining allowance #103 is less than or equal to the allowance threshold #104, it is determined to be a low-load area, i.e. a flat area. An efficient machining strategy is adopted, and the feed rate #105 is set to high speed.

[0129] 3. Perform geometric adaptive grinding, using the fitted radius #102 as the target position on the X-axis, and control the grinding wheel to complete the grinding along X#102 and Z#100 at a feed rate of #105 according to the G01 linear interpolation command.

[0130] 4. Decrease the Z-axis coordinate #100 by a fixed step size and proceed to the next processing layer until the cycle ends.

[0131] In actual machining, the relevant variable values ​​are calculated by the host computer algorithm to generate a variable table, which is then transmitted to the CNC system via DNC transmission or direct assignment.

[0132] Example 2: In-situ dressing technology for grinding wheels

[0133] To eliminate the runout error caused by secondary clamping of the grinding wheel, this invention adopts a closed-loop on-machine dressing scheme:

[0134] Reference calibration: First, a high-precision contact probe is used to perform contact measurement on the diamond dressing roller (or dressing pen) rigidly mounted on the machine tool worktable to obtain the precise position (X, Z) of the dressing tool in the machine tool coordinate system.

[0135] Dressing Cycle: The grinding wheel moves to the preset dressing station, and then the system calls the built-in dressing cycle module. The CNC system controls the relative motion trajectory between the grinding wheel and the dressing tool based on the preset ideal profile of the grinding wheel (such as an arc or a specific curve).

[0136] Closed-loop verification: After the dressing is completed, the probe scans and measures the grinding wheel again to calculate the shape error. If the error exceeds the tolerance zone, the system automatically calculates the compensation value and re-executes the finishing process to ensure that the actual profile of the dressed grinding wheel closely matches the ideal design profile within the allowable tolerance range.

[0137] Example 3: Digital Twin and Virtual Debugging System

[0138] This embodiment constructs a digital twin architecture based on "physical entity - virtual model - service platform" (see attached document). Figure 5 ).

[0139] Data mapping: Through the open API of the CNC system and external sensors, the spindle current (characterizing cutting force), servo axis position deviation (characterizing contour error) and vibration signals are collected in real time.

[0140] Model-driven:

[0141] Mechanism model: Establish the dynamic equation of the feed system and analyze the frequency domain response of the system using the transfer function G(s).

[0142] ;

[0143] Time-varying mass matrix;

[0144] The mass changes dynamically with the machining process, including the equivalent mass of the grinding wheel, spindle, and workpiece, as well as the change in workpiece mass due to material removal.

[0145] : Acceleration vector

[0146] The vibration acceleration representing the contact point between the grinding wheel and the workpiece is a key indicator for determining the occurrence of chatter.

[0147] Time-varying damping matrix

[0148] It includes structural damping, grinding fluid damping, and process damping, and its value changes dynamically with the grinding depth and the wear state of the grinding wheel.

[0149] Velocity vector

[0150] The relative speed between the grinding wheel and the workpiece directly affects the grinding force and surface roughness.

[0151] Time-varying stiffness matrix

[0152] It reflects the dynamic changes in the contact stiffness between the grinding wheel and the workpiece, as well as the stiffness of the thin-walled structure of the workpiece, with material removal and grinding wheel wear.

[0153] Displacement vector

[0154] The relative displacement between the grinding wheel and the workpiece is a core state quantity for controlling machining accuracy (such as sphericity and wall thickness uniformity).

[0155] External excitation force vector

[0156] The main disturbance source is the grinding force, which is determined by the depth of cut, feed rate, and material properties.

[0157] Time-delay feedback force vector

[0158] The time delay effects corresponding to the in-machine measurement and adaptive control links, such as probe sampling delay, signal transmission delay, macro program execution delay, etc., are represented by \tau, which is the total time delay of the control loop.

[0159] By solving this equation, the system stability under different grinding parameters can be analyzed, the critical conditions for chatter occurrence can be predicted, and a theoretical basis can be provided for optimizing feed rate and grinding wheel speed.

[0160] Based on real-time acquired vibration signals ( , The displacement error X(t) is inferred from the model and the numerical control system is driven to perform dynamic compensation, thereby improving the surface accuracy of the hemispherical harmonic oscillator.

[0161] Considering time delay term More robust adaptive control algorithms can be designed to compensate for measurement and control delays, thereby improving the response speed and stability of the "on-machine measurement-adaptive grinding" closed-loop system.

[0162] Combination Figure 6As shown, a deep neural network (DNN) is used to train historical processing data to establish a nonlinear mapping relationship between "feed pose and natural frequency".

[0163] Virtual Commissioning: Before physical machining, the adaptive macro program is imported into the virtual CNC system (X1). Virtual simulation verifies the entire machine tool motion process, identifies collision and interference risks, and checks the smoothness of speed change logic. An integrated cutting force prediction model monitors the evolution of cutting forces in real time, comparing it with the safety threshold calibrated by the material strength. If the threshold is exceeded, the system automatically initiates adaptive parameter optimization, dynamically adjusting key parameters such as cutting speed and feed rate, and simultaneously correcting macro program logic flaws. This effectively avoids the risk of breakage during quartz part machining, highlighting the core value of digital twin technology and ensuring machining stability.

[0164] The beneficial effects of this invention are as follows:

[0165] 1. Enhanced safety: The adaptive speed control strategy completely eliminates the impact of grinding force caused by sudden changes in allowance, significantly reducing the risk of breakage of hard and brittle thin-walled parts.

[0166] 2. Leap in efficiency: Intelligent identification and reduction of "empty tool" time, with an expected overall processing efficiency increase of over 30%.

[0167] 3. Precision assurance: The grinding wheel's machining shape and closed-loop error compensation mechanism ensure micron-level contour machining accuracy.

[0168] 4. Intelligent Upgrade: The introduction of digital twin and AI technologies makes the processing process predictable and explainable, laying the foundation for realizing unmanned, lights-out factories.

[0169] 5. Strong hybrid programming capability: Breaking through the limitations of single G-code, it is compatible with hybrid programming interfaces of high-level languages ​​(such as Python / C++), providing core support for real-time control of machine tools by complex algorithms such as neural networks.

[0170] The present invention also proposes a storage medium storing a computer program, which, when executed by a processor, implements a precision intelligent machining method for hemispherical harmonic oscillators.

[0171] In the above embodiments, the storage medium may include, but is not limited to, various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0172] Optionally, the processor executes the method steps described in the above embodiments based on the program code stored in the storage medium.

[0173] Specific examples in this embodiment can be found in the examples described in the above embodiments and optional implementations, and will not be repeated here.

[0174] Obviously, those skilled in the art should understand that the modules or steps of the present disclosure described above can be implemented using general-purpose computing devices, which can be centralized on a single computing device or distributed on a network of multiple computing devices.

[0175] Alternatively, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computing device. Furthermore, in some cases, the steps shown or described can be performed in a different order than those described herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, this disclosure is not limited to any particular combination of hardware and software.

[0176] 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. A method for precision intelligent machining of a hemispherical resonator, characterized in that, include: Construct a basic digital twin model of the processing equipment; The hemispherical harmonic oscillator is rough-machined, and the surface of the rough-machined hemispherical harmonic oscillator is sampled and the contour is fitted. The machining allowance is calculated based on the actual contour, a feed rate mapping model is established, and the grinding wheel trajectory is planned. The hemispherical harmonic oscillator is precision ground based on the aforementioned grinding wheel trajectory, while the grinding wheel is simultaneously subjected to in-machine dressing and closed-loop calibration. Multi-source data from the grinding and shaping process are collected to correct and optimize the initial digital twin model, resulting in a high-precision digital twin model. Based on the high-precision digital twin model, processing simulation, status monitoring and error prediction are performed, and the error prediction results are fed back to the control system to adaptively adjust the processing parameters. The finished product undergoes in-machine precision testing and evaluation. The test data is compared and verified with the model to achieve closed-loop iterative optimization of the entire processing process, thus completing the high-precision intelligent processing of the hemispherical harmonic oscillator.

2. The hemispherical resonator precision smart machining method of claim 1, wherein, The process involves rough machining of the hemispherical harmonic oscillator, followed by surface sampling and contour fitting of the rough-machined oscillator. Machining allowances are calculated based on the actual contours. A feed rate mapping model is established, and the grinding wheel trajectory is planned. Specifically, this includes: Discrete sampling is performed on the rough-machined hemispherical harmonic oscillator to obtain the surface point set; Curve fitting is performed on the surface point set to generate a real workpiece contour curve; The radial machining allowance at each point is calculated based on the theoretical contour curve and the actual workpiece contour curve. A feed rate mapping model is established based on the radial machining allowance at each point; The grinding wheel trajectory is formed based on the feed rate mapping model.

3. The hemispherical resonator precision smart machining method of claim 2, wherein, Point cloud data of the workpiece surface is acquired by laser non-contact measurement, and the contour curve is fitted by the least squares method to generate the real workpiece contour curve.

4. The hemispherical resonator precision smart machining method of claim 1, wherein, The feed rate mapping model is established by taking the machining allowance as input and the grinding wheel feed rate as output, establishing a nonlinear mapping relationship and performing real-time matching.

5. The hemispherical resonator precision smart machining method of claim 1, wherein, The grinding wheel trajectory is generated by interpolation based on the feed speed mapping model and the spherical coordinates of the hemispherical harmonic oscillator, and the trajectory points correspond one-to-one with the machining allowance.

6. The hemispherical resonator precision smart machining method of claim 1, wherein, The in-machine dressing process includes using diamond rollers to dress the profile of the grinding wheel online, so that the profile of the grinding wheel always maintains the theoretical spherical accuracy.

7. The hemispherical resonator precision smart machining method of claim 1, wherein, The multi-source data includes grinding wheel wear data, machining vibration data, contour error data, feed displacement data, and temperature drift data.

8. The hemispherical resonator precision smart machining method of claim 1, wherein, In-machine precision testing includes online evaluation of the roundness, sphericity, wall thickness uniformity, and surface roughness of the hemispherical harmonic oscillator.

9. A precision smart machining system of a hemispherical resonator, characterized in that, include: Digital twin modeling unit, used to build a primary digital twin model of processing equipment; The roughing execution unit is used for roughing the hemispherical harmonic oscillator; The data acquisition and trajectory planning unit is used to perform surface sampling and contour fitting on the rough-machined hemispherical harmonic oscillator, calculate the machining allowance based on the actual contour, and establish a feed rate mapping model to plan the grinding wheel trajectory. The precision grinding execution unit is used to perform precision grinding on the hemispherical harmonic oscillator according to the grinding wheel trajectory, and at the same time to perform on-machine dressing and closed-loop calibration of the grinding wheel. The model optimization unit is used to collect multi-source data during the grinding and shaping process, and to correct and optimize the primary digital twin model to obtain a high-precision digital twin model. An adaptive control unit is used to perform machining simulation, status monitoring and error prediction using the high-precision digital twin model, and to feed back the error prediction results to the control system to adaptively adjust the machining parameters. The in-machine inspection and closed-loop iteration unit is used to perform in-machine accuracy inspection and evaluation on the finished product, compare and verify the inspection data with a high-precision digital twin model, and realize closed-loop iterative optimization of the entire processing process.

10. A computer-readable storage medium, characterized in that, The device contains a computer program that, when executed by a processor, implements the precision intelligent machining method for a hemispherical harmonic oscillator as described in any one of claims 1 to 8.

Images