A mandrel in-place detection machining method

Abstract

The application discloses a mandrel in-situ detection processing method, comprising the following steps: using a roundness instrument to measure the real profile data of the generatrix L of the mandrel; using the shaft control time grinding processing method to make a removal function; measuring the profile data of the generatrix L in the horizontal and vertical directions respectively, calculating the components of the coaxiality error in the horizontal and vertical directions according to the real profile data, and compensating the coaxiality error; using the three-point method cylindrical appearance in-situ measurement and reconstruction technology to reconstruct the cylindrical appearance; using the result obtained by measuring the removal function to solve the residence time in combination with the cylindrical appearance obtained by reconstruction; based on the solved residence time, using the shaft control time grinding processing method to determine the modification of the mandrel; in-situ measuring the surface shape of the mandrel after the modification is completed, and if the surface shape meets the technical index requirement, the process is ended and exited. The application aims to realize the in-situ processing and detection integrated process of the mandrel and improve the processing convergence precision of the mandrel.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of mandrel processing, and particularly relates to a mandrel in-situ detection processing method. BACKGROUND

[0002] The air static pressure spindle is the basis of the contemporary high-end manufacturing industry, and is the technical support for space exploration, information technology development and weapon equipment reform. The ultra-precision machine tool with the air static pressure spindle is known as the "industrial mother machine" of modern manufacturing industry. The mandrel is the core component of the air static pressure spindle, and the manufacturing precision of the mandrel directly affects the rotation accuracy of the air static pressure spindle. According to the principle of computational fluid dynamics and the error homogenization principle, generally only when the roundness of the mandrel is better than 0.1 μm and the cylindricity is better than 1.0 μm, the rotation error of the air static pressure spindle can reach the level of 15-20 nm. If a higher rotation accuracy air static pressure spindle is wanted, the surface accuracy of the mandrel should also be higher.

[0003] The processing technology of the mandrel part generally includes traditional precision turning, traditional precision grinding, manual polishing and time-controlled grinding process.

[0004] The traditional precision turning refers to using the turning method to process the high-precision mandrel, and the processing tool is a turning tool.

[0005] The traditional precision grinding refers to using the grinding method to process the high-precision mandrel, and the processing tool is a grinding wheel, a grinding belt or the like.

[0006] The manual polishing refers to that after the surface accuracy of the workpiece is measured by the engineers using measuring instruments, the mandrel surface is polished by manual grinding, the high points on the mandrel surface are removed manually, and the measurement iteration is continuously performed until the accuracy requirement is met. The manual polishing is highly dependent on the experience and technical level of the engineers, the polishing efficiency is low, the environmental adaptability is poor, the polishing product quality is uneven, and the manual polishing is not suitable for large-scale industrial production.

[0007] The controlled time grinding technology is based on the deterministic machining idea. The controlled time grinding machine tool is composed of spindle, tailstock, controlled time grinding device and numerical control system. The controlled time grinding device mainly includes cylinder, abrasive belt, abrasive belt wheel, contact wheel, guide wheel and so on. The contact wheel which tightens the abrasive belt is the machining tool of the controlled time grinding device. The outer ring of the contact wheel is made of rubber material to increase the contact area in the grinding process and play a buffering role in the grinding process. In the grinding process, the abrasive belt is continuously updated on the outer ring of the contact wheel. The cylinder provides pressure to make the contact wheel and the mandrel contact. The tool drives the abrasive belt to vibrate along the axis direction of the mandrel to achieve the purpose of removing material. The spindle drives the ground mandrel to rotate in the opposite direction of the abrasive belt update direction to realize the machining of different areas of the mandrel surface. The pulse iteration method is used to calculate the residence time of different positions on the mandrel surface. The material to be removed at the high point of the surface error is more, so the residence time is longer. On the contrary, the residence time of the low point of the surface error is relatively shorter. The machining program is generated by using the algorithm written by the computer, and the linkage of each axis of the machine tool is controlled to realize the controlled time grinding machining of the mandrel. A 100mm high and 100mm diameter steel shaft is subjected to controlled time grinding machining. After multiple iterations, the cylindricity can be converged from 6.899μm to 2.078μm. The linearity and stability of the controlled time grinding machining process for reshaping high-precision mandrels are verified for the first time. And its improvement technology makes the cylindricity of the mandrel machined by the controlled time grinding machine tool optimal to 0.15μm, and the cylindricity optimal to 0.35μm. But the size of the mandrel machined by the controlled time grinding machine tool is 100mm in diameter and only 35mm in height, which is small in size and low in difficulty in realizing the surface accuracy. In addition, the analysis of the measurement results is manually removed for some error points. The actual cylindricity and roundness data cannot reach 0.35μm and 0.15μm.

[0008] Generally, the controlled time grinding process has the highest machining precision among the three main machining methods of the mandrel part, but the controlled time grinding process has not further broken through the 0.5 mu m cylindricity of the large-size mandrel part. Through the analysis on the principle of the controlled time grinding machining process, it is not difficult to find that the main reason is that the offline measurement introduces a large system error. Offline measurement refers to that after the machining is completed, the workpiece is disassembled from the machine tool and transferred to a professional measuring device for measurement, so as to obtain the machining precision of the workpiece, if the precision does not meet the requirements, the workpiece needs to be clamped again, and the workpiece is machined again according to the measurement data. The disassembly and installation process of the workpiece in the offline measurement process introduces the secondary clamping error, thereby affecting the authenticity of the measurement result and the accuracy of the error compensation. At the same time, the difference between the machining environment and the measurement environment also causes the uncertainty of the compensation and modification precision. In contrast to offline measurement, in-situ measurement does not disassemble the workpiece, but uses the original structure of the machine tool to build a measuring device to measure the surface error of the mandrel. The in-situ measurement method not only avoids the secondary clamping error and the influence of environmental changes, but also improves the machining efficiency. Therefore, for the mandrel, how to realize the in-situ detection and machining of the mandrel with high precision has become a key technical problem to be solved. SUMMARY

[0009] The technical problem solved by the present application: In view of the above problems of the prior art, a mandrel in-situ detection and machining method is provided, and the present application aims to realize the in-situ detection and machining process of the mandrel and improve the machining convergence precision of the mandrel.

[0010] In order to solve the above technical problems, the technical scheme adopted by the present application is:

[0011] A mandrel in-situ detection and machining method, comprising:

[0012] S101, cleaning the surface of the mandrel to be machined and the shaft for making the removal function;

[0013] S102, installing the mandrel on the roundness instrument and centering and leveling, using the roundness instrument to measure the profile data of the mandrel generatrix L, taking the profile data of the mandrel generatrix L as the true profile data of the mandrel generatrix L; installing the shaft for making the removal function on the machine tool, and using the controlled time grinding machining method to make the removal function;

[0014] S103, disassembling the shaft for making the removal function, measuring the removal function using the roundness instrument, and installing the mandrel on the machine tool;

[0015] S104, using the successive two-point method to measure the profile data of the generatrix L in the horizontal and vertical directions respectively, and calculating the components of the coaxiality error in the horizontal and vertical directions according to the true profile data of the generatrix L measured by the roundness instrument;

[0016] S105, the coaxiality error is compensated by using the three-sensor collected cylindrical profile data of the mandrel;

[0017] S106, the cylindrical profile is reconstructed by using the three-point method cylindrical profile in-situ measurement and reconstruction technology;

[0018] S107, the residence time is calculated by combining the result obtained by the measurement removal function with the cylindrical profile obtained by the reconstruction;

[0019] S108, based on the calculated residence time, the mandrel is deterministically shaped by using the time control grinding processing method;

[0020] S109, the surface profile of the mandrel after the shaping is completed is measured in-situ, if the surface profile meets the technical index requirements, the process is ended and exited; otherwise, the process is continued by jumping to step S105.

[0021] Optionally, step S104 comprises:

[0022] S201, when the mandrel is in a static state, the distance measuring sensors A and B with an interval distance s are horizontally installed on a slide plate of a machine tool, the slide plate is used to drag the distance measuring sensors A and B to perform a feeding motion along the axial direction of the mandrel; the distance measuring sensors A and B are zeroed to eliminate the influence of zeroing errors on data collection; the distance measuring sensors A and B are scanned along the generatrix L of the mandrel in the horizontal x direction; the collected data simultaneously contain the component L x (x) of the linear motion error of the slide plate in the horizontal x direction, the real profile data M(x) of the generatrix L, and the component T x (x) of the coaxiality error in the horizontal x direction; the component L x (x) of the linear motion error of the slide plate in the horizontal x direction is separated by using the successive two-point method to obtain the sum M(x)+T x (x) of the profile data of the generatrix L in the horizontal x direction and the coaxiality error; when the mandrel is in a static state, the distance measuring sensors A and B are vertically installed on the slide plate; the distance measuring sensors A and B are zeroed to eliminate the influence of zeroing errors on data collection; the distance measuring sensors A and B are scanned along the generatrix L of the mandrel in the vertical y direction; the collected data simultaneously contain the component L y (x) of the linear motion error of the slide plate in the vertical y direction, the real profile data M(x) of the generatrix L, and the component T y (x) of the coaxiality error in the vertical y direction; the component L y (x) of the linear motion error of the slide plate in the vertical y direction is separated by using the successive two-point method to obtain the sum M(x)+T y(x);

[0023] S202, sum the contour data of the generatrix L in the horizontal x direction with the coaxiality error M(x)+T x Subtracting the profile data M(x) of the generatrix L from (x) yields the coaxiality error T in the horizontal x-direction. x The sum of the profile data of the generatrix L in the vertical y-direction and the coaxiality error M(x)+T y Subtracting the profile data M(x) of the generatrix L from (x) yields the coaxiality error T in the vertical y-direction. y .

[0024] Optionally, in step S201, the successive two-point method is used to calculate the component L of the linear motion error of the slide in the horizontal x-direction. x Separate (x) to obtain the contour data of the generatrix L in the horizontal x direction and the sum of the coaxiality error M(x)+T. x (x) includes: setting the initial true contour data M(0) = 0 for the busbar L, and during the scanning process of the range sensors A and B, taking the distance s between the range sensors A and B as the step size, calculating the horizontal x-direction component L of the linear motion error of the slide that drags the range sensors A and B in each step. x (x), the actual contour data M(x) of the busbar L, and the component T of the coaxiality error in the horizontal x direction. x The data d collected by the ranging sensor A of (x) Ax (x) and the data d collected by the ranging sensor B Bx The difference (x) is taken as the change in the true contour data of the busbar L, ΔM(x), so that when the end point in the horizontal x direction is reached, the sum of the true contour data of the busbar L in the horizontal x direction and the coaxiality error is obtained as M(x) + T. x (x).

[0025] Optionally, in step S201, the vertical y-direction component L of the linear motion error of the slide is calculated using the successive two-point method. y Separate (x) to obtain the contour data of the generatrix L in the vertical y direction and the sum of the coaxiality error M(x)+T. y (x) includes: setting the initial true contour data M(0) = 0 for the generatrix L, and during the scanning process of the range sensors A and B, taking the distance s between the range sensors A and B as the step size, calculating the vertical y-direction component L of the linear motion error of the slide that drags the range sensors A and B in each step. y (x), the actual contour data of the busbar L M(x), and the component of the coaxiality error in the vertical y-direction T. y The data d collected by the ranging sensor A of (x) Ax (x) and the data d collected by the ranging sensor BBx (x) the difference between the real profile data of the generatrix L and the sum M(x) + T of the real profile data of the generatrix L and the coaxiality error in the horizontal x direction at the end of the scan in the horizontal x direction y (x).

[0026] Optionally, step S105 comprises:

[0027] S301, based on the three-point method, using three distance measuring sensors A, C and D located in the radial plane of the mandrel to measure a plurality of measured sections of the mandrel in the axial direction, wherein the included angle between the axis of the distance measuring sensor A and the axis of the distance measuring sensor C is φ, the included angle between the axis of the distance measuring sensor A and the axis of the distance measuring sensor D is ψ, and the coordinates of any point in the column profile data of the mandrel are determined according to the following formula to achieve the acquisition of the column profile data: j (i),v j (i),w j (i)].

[0028]

[0029] In the above formula, u j (i) is the U coordinate value of the i-th sampling point [u j (i),v j (i),w j (i)] on the section j in the machine tool absolute coordinate system O-UVW, v j (i) is the V coordinate value of the i-th sampling point [u j (i),v j (i),w j (i)] on the section j in the machine tool absolute coordinate system O-UVW, w j (i) is the W coordinate value of the i-th sampling point [u j (i),v j (i),w j (i)] on the section j in the machine tool absolute coordinate system O-UVW, A 1j is the horizontal coordinate of the least squares center of the measured section j, B 1j is the vertical coordinate of the least squares center of the measured section j, R0 is the design radius of the mandrel, R 2j (i) is the roundness error of the section j, A 0j is the difference between the average radius actually measured on the section j and the design radius, i is the i-th sampling point on the section j, and N is the total number of sampling points on the section j, (u 0j ,v 0j ) is the measurement coordinate system O j -X j Yj Z j Position of the origin in the absolute coordinate system O-UVW, Δw is the distance between adjacent sections;

[0030] S302, the coaxiality error is compensated by using the collected cylindrical profile data, including: first, calculating the component size of the coaxiality error on the installation angle of the ranging sensor A, the ranging sensor C and the ranging sensor D; then, subtracting the component size of the coaxiality error on the installation angle of the corresponding ranging sensor from the collected data of the ranging sensor A, the ranging sensor C and the ranging sensor D, so as to compensate the coaxiality error in the position measurement process.

[0031] Optionally, step S106 comprises:

[0032] S401, for the cylindrical profile data compensated by the coaxiality error, the roundness data of each section of the mandrel is calculated according to the three-point error separation principle;

[0033] S402, the center position and the average radius of each section of the mandrel are calculated according to the least square principle fitting;

[0034] S403, the position of the origin of the measurement coordinate system of each section in the absolute coordinate system of the machine tool is calculated;

[0035] S404, according to the roundness data, the center position and the average radius of each section of the mandrel, and the position of the origin of the measurement coordinate system in the absolute coordinate system, the cylindrical appearance in the absolute coordinate system is reconstructed.

[0036] Optionally, the result obtained by using the measurement removal function in step S107 is combined with the cylindrical appearance obtained by reconstruction to calculate the residence time, which comprises:

[0037] S501, an experimental shaft with the same diameter size and material as the mandrel to be machined is selected as the removal function shaft; the removal function shaft is installed on the on-time grinding machine tool, the on-time grinding processing device is installed and debugged, and the processing parameters are set, including the initial radius of the abrasive belt roll, the single-layer abrasive belt thickness, the abrasive belt updating speed, the tool head vibration frequency, etc.; then, the removal function shaft is fixed and subjected to on-time grinding processing for a certain time, and then it is disassembled and measured by using a roundness instrument to measure the surface shape after processing, and the measurement result representing the processing effect after adding a certain time is obtained as the removal function, and the removal amount distribution of each part of the removal function is obtained by measurement;

[0038] S502, the mandrel to be machined is installed on the on-time grinding machine tool, and the in-situ measurement and reconstruction of the cylindrical appearance of the mandrel are performed by using the in-situ measurement device;

[0039] S503, the cylindrical appearance of the mandrel is measured and reconstructed in-situ, the cylindrical error distribution of the mandrel surface is identified, the machining removal effect of the removal function is calculated at each position of the mandrel surface by using the pulse iteration method, and the machining residence time of the mandrel surface under the corresponding cylindrical error is obtained.

[0040] Optionally, the step S108 includes the following when the mandrel is deterministically shaped by using the controlled grinding.

[0041] S601, after the in-situ measurement and reconstruction of the cylindrical appearance of the mandrel to be machined are completed, the controlled grinding machining device and the in-situ measurement device are installed and debugged, and the machining parameters are set, which are consistent with the machining removal function.

[0042] S602, the machining code is generated according to the machining residence time of each position of the mandrel surface, and the controlled grinding device is driven to perform the controlled grinding machining.

[0043] S603, the in-situ measurement and reconstruction of the cylindrical appearance of the machined mandrel are performed again, the cylindrical error value of each position of the mandrel surface is identified, and it is judged whether the cylindrical error of the mandrel meets the accuracy requirement; if yes, the compensation machining is not needed, and the machining process is ended; if not, the residence time required for the compensation machining of each position of the mandrel surface is calculated by using the pulse iteration method, and the step S102 is jumped to.

[0044] Optionally, when the machining parameters are set in the step S601, the set machining parameters include part or all of the following: the initial radius of the abrasive belt, the thickness of the single layer of the abrasive belt, the updating speed of the abrasive belt, and the vibration frequency of the tool head.

[0045] In addition, the application also provides an aerostatic spindle, which comprises an aerostatic spindle body provided with a mandrel, and the mandrel is prepared by using the mandrel in-situ detection machining method.

[0046] Compared with the prior art, the present application has the following advantages: the present application comprises using a roundness measuring instrument to measure the profile data of the mandrel generatrix L, which is used as the true profile data of the mandrel generatrix L; an axis for making a removal function is installed on a machine tool, and a controlled time grinding machining method is used to make the removal function; the profile data of the mandrel generatrix L is measured in the horizontal and vertical directions respectively using a successive two-point method, and the components of the coaxiality error in the horizontal and vertical directions are calculated according to the true profile data of the mandrel generatrix L measured by the roundness measuring instrument; the coaxiality error is compensated using three-sensor acquisition of the cylindrical profile data of the mandrel; the three-point method cylindrical appearance in-situ measurement and reconstruction technology is used to reconstruct the cylindrical appearance; the residence time is calculated by combining the result obtained by measuring the removal function with the cylindrical appearance obtained by reconstruction; based on the calculated residence time, the mandrel is deterministically shaped using the controlled time grinding machining method; the surface shape of the mandrel after shaping is measured in-situ, and if the surface shape meets the technical index requirements, the process is ended and exited; otherwise, the process is iterated, which can realize in-situ machining of the mandrel cylindrical part, improve the machining convergence precision of the mandrel, provide a new solution for high-precision machining of the mandrel part, and realize precision evolution of high-precision machining using a low-precision machine tool. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 It is a basic flowchart of the embodiment method of the present application.

[0048] Figure 2 It is a schematic diagram of the coaxiality error in the embodiment of the present application.

[0049] Figure 3 It is a schematic diagram of measuring the true profile data of the mandrel generatrix L in the embodiment of the present application.

[0050] Figure 4 It is a model schematic diagram of measuring the horizontal coaxiality error component in the embodiment of the present application.

[0051] Figure 5 It is a model schematic diagram of measuring the vertical coaxiality error component in the embodiment of the present application.

[0052] Figure 6 It is a model schematic diagram of cross-section measurement by the three-point method in the embodiment of the present application.

[0053] Figure 7 It is a machining model schematic diagram of controlled time grinding in the embodiment of the present application.

[0054] Figure 8 It is a sand belt grinding contact model schematic diagram in the embodiment of the present application.

[0055] Figure 9 It is a sand belt grinding equivalent model schematic diagram in the embodiment of the present application.

[0056] Figure 10 Figure 4 is a schematic diagram of an equivalent model of a contact area between a grinding belt and a mandrel in a vertical meridian direction in an embodiment of the present application.

[0057] Figure 11 Figure 5 is a schematic diagram of an experimental structure of a processing device in an embodiment of the present application.

[0058] Figure 12 Figure 6 is a schematic diagram of real profile data of a generatrix L measured in an embodiment of the present application.

[0059] Figure 13 Figure 7 is a schematic diagram of coaxiality error data measured in an embodiment of the present application.

[0060] Figure 14 Figure 8 is a measurement result of a removal function in an embodiment of the present application.

[0061] Figure 15 Figure 9 is an extraction result of a removal function in an embodiment of the present application.

[0062] Figure 16 Figure 10 is a cylindrical topography reconstruction diagram in an embodiment of the present application.

[0063] Figure 17 Figure 11 is a two-dimensional development diagram of an in-situ measured cylindrical topography in an embodiment of the present application.

[0064] Figure 18 Figure 12 is a calculated residence time distribution in an embodiment of the present application.

[0065] Figure 19 Figure 13 is a simulated post-processing face shape error in an embodiment of the present application.

[0066] Figure 20 Figure 14 is a post-processing face shape measurement result in an embodiment of the present application.

[0067] Figure 21 Figure 15 is an in-situ measurement result of a post-processing face shape in an embodiment of the present application.

[0068] Figure 22 Figure 16 is a post-processing face shape roundness gauge measurement result in an embodiment of the present application. DETAILED DESCRIPTION

[0069] The machine tool used for processing the mandrel to be processed in the embodiment includes a spindle, a tailstock, an in-situ measurement device, and a time-controlled grinding device. The spindle is an air static pressure spindle, and a Morse point is installed at the spindle end. The tailstock is a pneumatic dead point tailstock. The mandrel to be processed is clamped and fixed by the spindle point and the tailstock point, and can be synchronously rotated with the spindle under the action of a fork. The in-situ measurement device is fixed on one side of the mandrel to be processed, and can move in the axial direction of the mandrel to collect different cross-sectional data or generatrix data. The time-controlled grinding device is fixed on the other side of the mandrel to be processed.

[0070] As Figure 1 shown, the mandrel in-situ detection processing method of the embodiment comprises:

[0071] S101, surface cleaning of the mandrel to be processed and the shaft for making the removal function;

[0072] S102, mounting the mandrel to the roundness instrument and centering and leveling, using the roundness instrument to measure the profile data of the mandrel generatrix L, which is used as the true profile data of the mandrel generatrix L; mounting the shaft for making the removal function to the machine tool, and using the controlled time grinding processing method to make the removal function;

[0073] S103, dismounting the shaft for making the removal function, measuring the removal function using the roundness instrument, and mounting the mandrel to the machine tool;

[0074] S104, using the successive two-point method to measure the profile data of the generatrix L in the horizontal and vertical directions respectively, and calculating the components of the coaxiality error in the horizontal and vertical directions according to the true profile data of the generatrix L measured by the roundness instrument;

[0075] S105, using three-sensor to collect the cylindrical profile data of the mandrel to compensate the coaxiality error;

[0076] S106, reconstructing the cylindrical appearance using the three-point method cylindrical appearance in-situ measurement and reconstruction technology;

[0077] S107, using the results obtained by measuring the removal function to solve the residence time combined with the cylindrical appearance obtained by reconstruction;

[0078] S108, based on the residence time obtained by solving, using the controlled time grinding processing method to determine the profile modification of the mandrel;

[0079] S109, in-situ measuring the surface shape of the mandrel after profile modification, if the surface shape meets the technical index requirements, then end and exit; otherwise, jump to step S105 to continue processing.

[0080] From the geometric structure of the machine tool, it can be known that the data collected by each sensor probe during in-situ measurement includes the surface error data of the mandrel, the radial rotation error of the main shaft, the installation eccentricity error of the main shaft center, the linear motion error of the slide plate, the zero error of the sensor, and the coaxiality error of the main shaft and the tailstock. Since part of the error is separated as known existing method, the mandrel in-situ detection processing method of the embodiment focuses on the coaxiality error compensation, cylindrical appearance reconstruction and compensation processing.

[0081] The coaxiality error is the error of the tailstock center (O') deviating from the main shaft axis X and Y, as Figure 2is shown, wherein O is the position of the main shaft center. The coaxiality error can be divided into two components Δx and Δy in horizontal and vertical directions. The existence of the coaxiality error will directly result in the inclination of the mandrel axis, reducing the measurement accuracy. The embodiment realizes effective compensation of the coaxiality error by means of coaxiality error pre-measurement and compensation of the roundness instrument. As shown in Figure 3 , the profile data M of the generatrix L of the mandrel is obtained by using the roundness instrument first. Since the roundness instrument can accurately center and level the air floating table on which the mandrel is placed, and the linear motion error of the measuring arm thereof is extremely small. Therefore, it can be considered that the profile data M collected by the roundness instrument does not contain the error components such as the deflection of the mandrel, and the profile data M is the true profile data of the generatrix L of the mandrel collected by the roundness instrument.

[0082] In the embodiment, the step S104 comprises: S201, in the case that the mandrel is in a static state, the distance measuring sensors A and B with a spacing distance s are horizontally installed on the slide plate of the machine tool, the slide plate is used to drag the distance measuring sensors A and B to perform feeding motion along the axial direction of the mandrel; the distance measuring sensors A and B are zeroed to eliminate the influence of zeroing error on data collection; the distance measuring sensors A and B are scanned along the generatrix L of the mandrel in the horizontal x direction; the collected data contains the component L x (x) of the linear motion error of the slide plate in the horizontal x direction, the true profile data M(x) of the generatrix L, and the component T x (x) of the coaxiality error in the horizontal x direction; the component L x (x) of the linear motion error of the slide plate in the horizontal x direction is separated by using the successive two-point method, to obtain the sum M(x)+T x (x) of the profile data of the generatrix L in the horizontal x direction and the coaxiality error; in the case that the mandrel is in a static state, the distance measuring sensors A and B are vertically installed on the slide plate; the distance measuring sensors A and B are zeroed to eliminate the influence of zeroing error on data collection; the distance measuring sensors A and B are scanned along the generatrix L of the mandrel in the vertical y direction; the collected data contains the component L y (x) of the linear motion error of the slide plate in the vertical y direction, the true profile data M(x) of the generatrix L, and the component T y (x) of the coaxiality error in the vertical y direction; the component L y (x) of the linear motion error of the slide plate in the vertical y direction is separated by using the successive two-point method, to obtain the sum M(x)+T y (x) of the profile data of the generatrix L in the vertical y direction and the coaxiality error; S202, the profile data M(x) of the generatrix L in the horizontal x direction is subtracted from the sum M(x)+T x (x) of the profile data of the generatrix L in the horizontal x direction and the coaxiality error, to obtain the coaxiality error Tx ; the profile data of the busbar L in the vertical y direction is subtracted from the sum M(x)+T y (x) of the profile data of the busbar L and the concentricity error to obtain the concentricity error T y .

[0083] In step S201 of the embodiment, the linear motion error component L x (x) of the slide plate in the horizontal x direction is separated by using the successive two-point method to obtain the profile data of the busbar L in the horizontal x direction and the sum M(x)+T x (x) of the profile data of the busbar L and the concentricity error. x (x) includes: letting the initial real profile data of the busbar L be M(0)=0, and in the scanning process of the distance sensors A and B, taking the distance s between the distance sensors A and B as a step, calculating the data d x (x) and d Ax (x) collected by the distance sensor B as the data collected by the distance sensor A, and taking the difference between the data d Bx (x) collected by the distance sensor B as the change ΔM(x) of the real profile data of the busbar L, so as to obtain the real profile data of the busbar L in the horizontal x direction and the sum M(x)+T x (x) of the profile data of the busbar L and the concentricity error at the end point of the scanning in the horizontal x direction. Figure 4 Fig. 1 is a schematic diagram of a model for measuring the horizontal concentricity error component in the embodiment, and the data of the busbar L is collected by using the sensor A and the sensor B. The data collected by the sensor A and the sensor B can be expressed by the following mathematical equations:

[0084]

[0085] In the above equations, d Ax (x) and d Bx (x) represent the data collected by the sensor A and the sensor B in the horizontal direction respectively; L x (x) represents the horizontal component of the linear motion error of the slide plate; M(x) represents the real profile data of the busbar L; T x (x) represents the horizontal component of the concentricity error; C Ax and C Bx represent the zeroing errors of the sensor A and the sensor B when collecting the data in the horizontal direction respectively; and s represents the interval between the sensor A and the sensor B. Subtracting the above two equations, we have:

[0086] ΔM(x) = M(x+s)+T x (x+s)-M(x)-Tx (x)=d Bx (x)-d Ax (x)+e,

[0087] Where, e = C Ax -C Bx This is the difference in zero-adjustment errors between the two sensors. Since zero-adjustment operations need to be performed on sensors A and B before measurement, the difference in zero-adjustment errors e can be ignored. Let M(0) = 0, then the sum of the profile data of the generatrix and the coaxiality error M(x) + T can be recursively derived at intervals of s. x (x).

[0088] In step S201 of this embodiment, the vertical y-direction component L of the linear motion error of the slide is determined using the successive two-point method. y Separate (x) to obtain the contour data of the generatrix L in the vertical y direction and the sum of the coaxiality error M(x)+T. y (x) includes: setting the initial true contour data M(0) = 0 for the generatrix L, and during the scanning process of the range sensors A and B, taking the distance s between the range sensors A and B as the step size, calculating the vertical y-direction component L of the linear motion error of the slide that drags the range sensors A and B in each step. y (x), the actual contour data of the busbar L M(x), and the component of the coaxiality error in the vertical y-direction T. y The data d collected by the ranging sensor A of (x) Ax (x) and the data d collected by the ranging sensor B Bx The difference (x) is taken as the change in the true contour data of the busbar L, ΔM(x), so that when the end point in the horizontal x direction is reached, the sum of the true contour data of the busbar L in the horizontal x direction and the coaxiality error is obtained as M(x) + T. y (x). Figure 5 This is a schematic diagram of the model for measuring the coaxiality error component in the vertical direction in this embodiment. Similarly, when measuring in the x direction, the sum of the profile data of the generatrix with intervals of s and the coaxiality error, M(x)+T, can be obtained. y (x). Since the true contour data M(x) of the mandrel generatrix L has been obtained by measuring with a roundness tester, the components T of the coaxiality error in the horizontal and vertical directions can be calculated. x and T y .

[0089] Existing techniques have been used to study the in-situ measurement and reconstruction method of the three-point cylindrical topography under single-end fixed conditions. The model of the cross-sectional profile can be represented as:

[0090]

[0091] where j is the number of the measured section; θ is the angle corresponding to each point on the profile; V j (θ) is the profile function of section j; A 0j represents the error of the radius of each section; represents the center position of each cross section; represents the roundness error of each cross section, A kj and B kj are the sine and cosine coefficients of the Fourier series expansion of the profile of section j; a l1 and b l1 are the coefficients of the Legendre polynomial fitting; P l (w j ) represents the Legendre polynomial; l represents the order of the Legendre polynomial. 1j represents and R 2j represents The double-end fixing of the mandrel is fixed relative to the single-end fixing, and the data collected by the sensor increases the coaxiality error. Therefore, when reconstructing the cylindrical profile, the coaxiality error in the data collected by the sensor needs to be removed in advance.

[0092] The step S105 in the embodiment includes:

[0093] S301, based on the three-point method, using three distance measuring sensors A, C and D located on the radial plane of the mandrel to measure a plurality of measured sections of the mandrel in the axial direction, wherein the angle between the axis of the distance measuring sensor A and the axis of the distance measuring sensor C is φ, and the angle between the axis of the distance measuring sensor A and the axis of the distance measuring sensor D is ψ, as shown in Figure 6 ω1, ω2, ω3, …, ω j-1 , ω j are the numbers of the measured sections; the coordinates [u j (i), v j (i), w j (i)] of any point in the cylindrical profile data of the mandrel are determined according to the following formula to realize the collection of the cylindrical profile data:

[0094]

[0095] In the above formula, u j (i) is the U coordinate value of the i-th sampling point [u j (i), v j (i), w j (i)] on the section j in the absolute coordinate system O-UVW of the machine tool, v j (i) is the V coordinate value of the i-th sampling point [u j (i), v j (i), w j(i) V coordinate value in machine tool absolute coordinate system O-UVW, w j (i) is the i-th sampling point on the cross section j [u j (i), v j (i), w j (i) W coordinate value in machine tool absolute coordinate system O-UVW, A 1j is the horizontal coordinate of the least square center of the measured cross section j, B 1j is the vertical coordinate of the least square center of the measured cross section j, R0 is the design radius of the mandrel, R 2j (i) is the roundness error of the cross section j, A 0j is the difference between the average radius actually measured on the cross section j and the design radius, i is the i-th sampling point on the cross section j, N is the total number of sampling points on the cross section j, (u 0j , v 0j ) is the position of the origin in the measurement coordinate system O j -X j Y j Z j origin in absolute coordinate system O-UVW, Δw is the distance between adjacent cross sections;

[0096] S302, the coaxiality error is compensated by using the collected cylindrical profile data, including: first, calculating the component size of the coaxiality error on the installation angle of the distance measuring sensor A, the distance measuring sensor C and the distance measuring sensor D; then subtracting the component size of the coaxiality error on the corresponding installation angle of the distance measuring sensor A, the distance measuring sensor C and the distance measuring sensor D from the data collected by the three distance measuring sensors, so as to compensate the coaxiality error in the position measurement process.

[0097] Referring to Figure 6 , after eliminating the environmental noise in the original data of the sensor, the data collected by the distance measuring sensor A, the distance measuring sensor C and the distance measuring sensor D on the cross section ω j can be represented as r Aj (i), r Cj (i) and r Dj (i). Since the collected data is affected by the coaxiality error, it is necessary to eliminate the coaxiality error component in the collected data first. According to the geometric relationship, the following can be obtained:

[0098]

[0099] wherein, T xj represents the horizontal component of the coaxiality error when collecting data on the cross section j, T yj represents the vertical component of the coaxiality error when collecting data on the cross section j. r′ Aj (i) is rAj The reconstruction result of (i), r′ Cj (i) is r Cj The reconstruction result of (i), r′ Dj (i) is r Dj The reconstruction result of (i). Through formula derivation, the coordinates of each point on the mandrel surface [u] can be determined. j (i),v j (i),w j (i)], thus completing the reconstruction of the cylindrical morphology.

[0100] In this embodiment, step S106 includes:

[0101] S401, for the cylindrical profile data after coaxiality error compensation, the roundness data of each section of the mandrel is calculated according to the three-point error separation principle (existing method, see "In Situ Measurement and Reconstruction Technology of Cylindrical Shape of High-Precision Mandrel");

[0102] S402, the center position and average radius of each section of the mandrel are calculated by fitting according to the least squares principle;

[0103] S403, calculate the position of the origin of the measurement coordinate system of each section in the absolute coordinate system of the machine tool;

[0104] S404 reconstructs the cylindrical morphology in the absolute coordinate system based on the roundness data, center position, and average radius of each section of the mandrel, as well as the position of the origin of the measurement coordinate system in the absolute coordinate system.

[0105] Figure 7 This is a schematic diagram of the machining model for time-controlled grinding in this embodiment, where V1 is the rotational speed of the tool driving the abrasive belt, and V2 is the rotational speed of the mandrel. In this embodiment, a time-controlled grinding device is used to perform high-precision mandrel compensation shaping. Time-controlled grinding is a type of deterministic machining. Time-controlled grinding compensation requires calculating the tool dwell time based on the distribution of the surface contour to be compensated and the removal efficiency of the abrasive belt used, thereby achieving deterministic removal of high points and further improving surface accuracy. The specific steps of the compensation machining can be divided into the following: creating a removal function, calculating the dwell time based on the cylindrical shape of the mandrel, and time-controlled grinding shaping.

[0106] Grinding removal is a complex process, and various factors are related. The shape of the grinding tool, material, contact force, motion trajectory, grinding speed, grinding duration, grinding temperature, and the composition of the grinding fluid, etc. will affect the effect of grinding. In order to realize the deterministic grinding of the mandrel, it is necessary to establish a mathematical model of grinding according to the actual grinding conditions. The selected modeling method is based on the Preston hypothesis proposed by Preston in 1972, which uses the following linear equation to describe the grinding process:

[0107]

[0108] In the above formula, t is the grinding time; h(x, y) refers to the material thickness removed at point (x, y) after time t; P(x, y, t) is the instantaneous normal pressure at grinding point (x, y); V(x, y, t) is the relative motion speed of the grinding wheel and the mandrel at the grinding point (x, y); K is the Preston coefficient, which is determined by factors other than pressure P(x, y, t) and speed V(x, y, t).

[0109] It can be seen that the material removal efficiency in the Preston equation is proportional to the normal pressure of the grinding wheel on the mandrel and the relative speed between the grinding wheel and the mandrel. Integrating the above formula, we can get:

[0110]

[0111] In the formula, T is the grinding time in a period.

[0112] Normalizing the above formula, we can get the removal function in the deterministic grinding process:

[0113]

[0114] For convenience of theoretical analysis and calculation, according to the actual grinding condition of the mandrel, the following assumptions can be made: (1) The mandrel surface is steep, and the curvature change in the contact area between the abrasive belt grinding tool and the mandrel surface is small; (2) The pressure generated by the meridian direction abrasive belt grinding tool on the mandrel surface changes little, and can be regarded as constant except for the edge area; (3) When the abrasive belt vibrates along the meridian of the mandrel, the contact force W can be considered constant.

[0115] Under ideal conditions, the contact area between the abrasive belt grinding wheel and the mandrel is a line segment. However, due to the influence of pressure and the material quality of the abrasive belt wheel, the contact area between the abrasive belt and the mandrel surface is approximately rectangular. The contact model of abrasive belt grinding and mandrel is equivalent to the contact model of the outer surfaces of two cylinders, as shown in Figure 8 and Figure 9 Figure 8 is a schematic diagram of the abrasive belt grinding contact model in this embodiment, Figure 9 ​Fig. 1 is a schematic diagram of the equivalent model of the contact area between the abrasive belt and the mandrel in the present embodiment, wherein R1, E1, v1, R2, E2 and v2 represent the Young's modulus of the mandrel, the radius of the mandrel, the Poisson's ratio of the mandrel, the radius of the abrasive belt wheel, the Young's modulus of the abrasive belt wheel and the Poisson's ratio of the abrasive belt wheel, respectively; W1 represents the load per meter; E and R represent the Young's modulus and the radius of the equivalent cylinder, respectively. According to the Hertz contact theory, E, R, R1, E1, v1, R2, E2 and v2 satisfy the following relationship:

[0116]

[0117] In the perpendicular meridian direction, the equivalent model of the contact area between the abrasive belt and the mandrel is shown in Fig. 2, and the half-width b of the contact area is: Figure 10

[0118]

[0119] Figure 10 wherein P h is the maximum pressure of the contact area.

[0120] According to the Hertz pressure formula, the maximum pressure P h of the contact area and the pressure distribution P(x) are respectively:

[0121]

[0122] According to the geometric relationship between the update speed V1 of the abrasive belt and the rotating speed V2 of the mandrel, the relative speed of the grinding point is:

[0123] V = V1 + V2,

[0124] When making the removal function, the same area needs to be processed for a period of time. Assuming that all parameters except the variable time t are time-invariant, a linear relationship with the residence time can be achieved, and for a plane grinding tool, the removal function can be expressed as:

[0125]

[0126] Obviously, at x = 0, the center removal depth h(0) is defined as:

[0127]

[0128] Therefore, for the mandrel to be ground for a time length of t', the relative speed of the grinding point is V, and the center removal depth h(0) can be obtained by using a commercial roundness instrument, and the following Preston coefficient can be obtained:

[0129]

[0130] Thus, the removal function model can be obtained.​

[0131] The result obtained by using the measurement removal function in step S107 of the embodiment is combined with the reconstructed cylindrical topography to calculate the residence time, which includes:

[0132] S501, select an experimental shaft with the same diameter and material as the core shaft to be machined as the removal function shaft; install the removal function shaft on the controlled time grinding machine, install and adjust the controlled time grinding processing device, and set the processing parameters, including the initial radius of the abrasive belt, the thickness of the single layer of abrasive belt, the abrasive belt updating speed, the tool head vibration frequency, etc.; then, perform controlled time grinding processing on the removal function shaft at a certain point for a certain time, disassemble it, and measure the processed surface shape using a roundness instrument. The measurement results obtained represent the processing effect after a certain time, which is used as the removal function. The removal amount distribution of the removal function at each point is obtained by measurement;

[0133] S502, install the core shaft to be machined on the controlled time grinding machine, and use the in-situ measurement device to perform in-situ measurement and reconstruction of the cylindrical topography of the core shaft;

[0134] S503, identify the cylindrical error distribution of the core shaft surface using the in-situ measurement and reconstruction of the cylindrical topography of the core shaft, and calculate the processing residence time of the core shaft surface under the corresponding cylindrical error using the pulse iteration method.

[0135] The pulse iteration method is a known method for calculating the residence time of cylindrical deterministic modification. The basic idea of the pulse iteration method is that if the material removal amount in the polishing area of the removal function is concentrated at a point, it is similar to the form of a removal pulse. Therefore, the residence time of the removal function at the current residence point is equal to the ratio of the error amount removed at that point to the removal pulse. The form of the removal pulse function is defined as:

[0136]

[0137] wherein, is the material removal amount at any point in the processing area; Ω represents the entire area processed by the polishing tool during positioning processing. Thus, the initial processing residence time at any point on the workpiece surface can be calculated as:

[0138] d1(x,y)=H1(x,y) / RP,

[0139] wherein, H1(x,y) represents the initial material removal amount at any point on the workpiece surface.

[0140] The residual error after the initial residence time can be calculated as:

[0141] E1(x, y) = H1(x, y) - R(x, y) d1(x, y),

[0142] If the machining residual error after the initial dwell time is determined to meet the target accuracy requirement, the machining process is ended. If the machining residual error does not meet the target accuracy requirement, the dwell time is calculated again. k (x, y), and the sum of all the dwell times is the dwell time D(x, y) for machining any point (x, y) on the workpiece surface:

[0143]

[0144] The final residual error E(x, y) after machining any point (x, y) on the workpiece surface can be obtained:

[0145] E(x, y) = H1(x, y) - R(x, y) D(x, y).

[0146] The deterministic reshaping of the mandrel in step S108 of the embodiment using the controlled time grinding includes:

[0147] S601, after the in-situ measurement and reconstruction of the cylindrical morphology of the mandrel to be machined are completed, the controlled time grinding machining device and the in-situ measurement device are installed and debugged, and the machining parameters are set. The machining parameters are consistent with the machining removal function axis.

[0148] S602, the machining code is generated according to the machining dwell time of each position on the mandrel surface, and the controlled time grinding device is driven to perform controlled time grinding machining.

[0149] S603, the in-situ measurement and reconstruction of the cylindrical morphology of the machined mandrel are performed again, the cylindrical error values of each position on the mandrel surface are identified, and it is determined whether the cylindrical error of the mandrel meets the accuracy requirement. If the accuracy requirement is met, no compensation machining is needed, and the machining process is ended. If the accuracy requirement is not met, the dwell time required for compensation machining of each position on the mandrel surface is calculated using the pulse iteration method, and the step S102 is jumped to.

[0150] When the machining parameters are set in step S601 of the embodiment, the set machining parameters include part or all of the initial radius of the abrasive belt, the thickness of a single layer of abrasive belt, the abrasive belt updating speed, and the tool head vibration frequency.

[0151] To verify the core shaft in-situ detection processing method of the embodiment, in the embodiment, through the in-situ measurement and compensation modification experiment of the cylindrical appearance, the feasibility and advantages of the core shaft in-situ detection processing method of the embodiment are verified. In this study, the CPL190 / C8-2.0 capacitive micro-displacement sensor produced by the Lion Precision Company is selected for data acquisition, and the range thereof can reach 50 microns, and the resolution thereof is 1 nanometer. The in-situ measurement and compensation modification system is based on a precision grinding machine tool, and a system structure diagram thereof is shown in Figure 11 . The system mainly comprises a grinding machine tool (including a spindle, a tailstock and the like), a core shaft, an in-situ measurement device, a compensation processing device and a computer. The measurement device comprises four sensors, i.e., distance measuring sensors A, B, C and D. The distance measuring sensors A and B are used for coaxial error measurement in step I, and the distance measuring sensors A, C and D are used for in-situ measurement and reconstruction of the cylindrical appearance in step II. The compensation processing device comprises a polishing head composed of a sand belt and a rubber wheel and a sand belt updating device, which are used for processing in step III.

[0152] The range of the sensor of the Talyrond 565PRO roundness instrument is 4 mm, and the resolution thereof is 0.3 nm. The radial measurement accuracy thereof is up to ±10 nm, and the axial measurement accuracy can be ensured by using a cylindrical standard. The commercial roundness instrument can meet the accuracy requirements of the experimental measurement. The measurement object of the Talyrond 565PRO roundness instrument is a core shaft of a gas static pressure spindle, and the working area of the core shaft is a cylinder with a diameter of 90 mm and a height of 180 mm. Before in-situ measurement of the coaxial error, the real profile data of the generatrix L of the core shaft are measured by using the roundness instrument. The generatrix L of the core shaft is measured by using the roundness instrument, and the measurement length is 170 mm. The real profile of the generatrix L of the core shaft is shown in Figure 12 .

[0153] According to Figure 11The system structure diagram is shown, which builds the coaxiality error measurement device. The precision displacement table can ensure the precise adjustment of the position measurement device in the vertical direction (V direction), and the slide plate can make the measuring frame precisely adjust in the horizontal direction (U direction). The temperature sensor (I) is used for measuring the environmental temperature compensation, and the metal brush (II) with good conductivity is used to ground the static electricity generated on the mandrel. Sensors A and B are used to collect coaxiality error data. After centering and cleaning operations, the computer controls the double sensors to collect data on the bus L in the horizontal and vertical directions respectively. The collected data contains the following components: bus L profile data, linear motion error of the slide plate dragging the measuring frame, coaxiality error, and environmental noise. Among them, the profile data of the bus L has been measured by the roundness instrument and is a known quantity; environmental noise can be directly removed by filtering; the linear motion error of the slide plate dragging the measuring frame can be removed using the successive two-point method algorithm. The coaxiality error can be calculated, as shown in Figure 13 .

[0154] An experimental shaft with the same material and size as the mandrel to be processed is selected to make the removal function. Before making the removal function, the processing parameters need to be set: I. The pop-up pressure of the time-controlled grinding cylinder is 0.3 Mpa; II. The distance between the rubber wheel and the mandrel is 10 mm before the cylinder pops up; III. The axial vibration frequency of the rubber wheel is set to 9 Hz; IV. The speed of the abrasive belt is 5 mm / s; V. The processing time of the removal function is set to 30 s. After completing the above settings, the removal function mandrel is installed on the machine tool and processed. The processing method of the removal function is to use the time-controlled grinding device to grind the surface of the removal function shaft for a certain period of time, and then quickly move to the next position to continue the same processing, and repeat multiple times. By observation, the two-dimensional shape of the removal function after grinding can be clearly observed, which is approximately dumbbell-shaped. After processing, the roundness instrument is used to measure the processing position, and the measurement results are shown in Figure 14 , where the dark area represents the low point and the light area represents the high point. Figure 14 The red box selected area is the spatial distribution diagram of the dumbbell-shaped removal function, and it can be found that the red box selected area has three removal functions made under the same conditions. One of them can be used as the removal function for residence time calculation, as shown in Figure 15 .

[0155] In the in-situ measurement and reconstruction experiment of cylindrical morphology, an experimental device was constructed based on the structural diagram of the in-situ measurement and compensation shaping system. Distance sensors A, C, and D were used for data acquisition, and encoder signals were acquired simultaneously. Based on the measured coaxiality error data and the cylindrical morphology reconstruction method, accurate measurement and reconstruction of the mandrel's cylindrical morphology can be achieved. In this experiment, roundness measurements were performed on 15 equally spaced cross-sections, and fitting was performed to reconstruct the cylindrical morphology as follows: Figure 16 As shown, the cylindricity error is 1.140 μm.

[0156] In the compensation and shaping experiment, the residence time was calculated using the extracted removal function and the cylindrical morphology reconstructed from in-situ measurements. First, the cylindrical morphology was unfolded into... Figure 17 The two-dimensional graphic shown is used to determine the target surface accuracy in order to calculate the dwell time and machining trajectory data required for machining. The dwell time is calculated using the pulse iteration method, as shown below. Figure 18 As shown, the total processing time is 45 minutes, and the shape error after simulation processing is as follows. Figure 19 As shown, the residual error after simulation machining is 0.082 μm. Since simulation machining is performed under ideal conditions, it can almost completely eliminate cylindricity error, but this effect cannot be achieved in actual machining. Machining code is written using a computer to drive the machine tool for compensation and shaping. The C-axis is a rotary axis, which drives the mandrel to rotate to machine different positions on the same cross-section of the mandrel; the V-axis is a linear axis, which drives the grinding head to feed along the mandrel axis to machine different cross-sections. After machining, the cylindricity error of the mandrel is measured again using an in-situ measuring device. The distribution of the cylindrical morphology after machining is shown in the figure. Figure 20 As shown. The cylindricity error of the mandrel after machining is 0.568μm.

[0157] Before and after the in-situ measurement and compensation shaping experiments, a Talyrond 565 PRO roundness meter was used to measure the cylindrical morphology verification data to verify the accuracy of the in-situ measurement results. The results of measuring the mandrel cylindrical morphology using the roundness meter are as follows: Figure 21 and Figure 22 As shown, the cylindricity before and after processing is 0.956 μm and 0.549 μm, respectively.

[0158] Comparison Figure 16 , Figure 20 , Figure 21 and Figure 22It can be found that the cylindrical profile distribution obtained by in-situ measurement and roundness instrument measurement is very close before and after compensation modification. In order to more clearly show the change of cylindrical error value, the cylindrical error measurement data before and after processing are counted in Table 1. It can be observed that the cylindrical error measured by in-situ measurement and roundness instrument measurement is basically consistent before and after processing, and the error convergence after processing is very obvious. According to the result of in-situ measurement, the cylindrical error of the mandrel is converged by 0.407 μm by one compensation modification processing, and the error convergence ratio reaches about 42.6%.

[0159] Table 1: Cylindrical error measurement data statistics before and after processing.

[0160] Measurement Method Before Processing / pm After Processing / pm Convergent Difference / pm Convergent Ratio On-site Measurement 1.140 0.568 0.572 50.2％ Roundness Gauge Measurement 0.956 0.549 0.407 42.6％

[0161] The measurement results show that the in-situ measurement method proposed in the present study can realize accurate measurement of cylindrical profile and carry out compensation modification. The in-situ processing and detection integrated process method of mandrel cylindrical parts can realize rapid convergence of mandrel cylindrical error, and make the cylindrical error of the mandrel in the experiment converge from 1.0 μm to 0.5 μm.

[0162] In summary, the in-situ detection and processing method of the mandrel in the present embodiment realizes in-situ measurement and compensation modification of the cylindrical profile of the mandrel, and the cylindrical error of the mandrel reaches 0.5 μm. The process method effectively avoids the influence of secondary clamping error, and can effectively improve the compensation modification efficiency of the mandrel. The coaxial error compensation method, cylindrical profile reconstruction method and removal function extraction method are introduced in detail in the embodiment. The present embodiment provides a new solution for high-precision machining of mandrel parts, and realizes precision evolution of high-precision mandrel parts using low-precision machine tools.

[0163] In addition, the present embodiment also provides an air static pressure spindle, comprising an air static pressure spindle body with a mandrel, wherein the mandrel is prepared by using the in-situ detection and processing method of the mandrel described above.

[0164] The above only describes the preferred embodiments of the present application, and the protection scope of the present application is not limited to the above-mentioned embodiments. Any technical solutions falling within the concept of the present application shall fall within the protection scope of the present application. It should be noted that for ordinary skilled persons in the art, some improvements and decorations without departing from the principles of the present application shall also be considered as the protection scope of the present application.

Claims

1. A mandrel-in-place detection machining method characterized by comprising: Comprise: S101, the surface of the shaft to be processed, the shaft for making removal function is cleaned; S102, the core shaft is installed on the roundness instrument and is centered and leveled, the profile data of the core bus L is measured by using the roundness instrument, which is used as the true profile data of the core bus L; the shaft for making removal function is installed on the machine tool, and the removal function is made by using the time control grinding processing method; S103, disassemble the shaft for making removal function, measure the removal function by using the roundness instrument, and install the core shaft on the machine tool; S104, the profile data of the bus L is measured in horizontal and vertical directions respectively by using the successive two-point method, and the components of the coaxiality error in horizontal and vertical directions are calculated according to the true profile data of the bus L measured by the roundness instrument; S105, the coaxiality error is compensated by using three sensor acquisition core shaft cylindrical profile data; S106, the cylindrical appearance is reconstructed by using three-point method cylindrical appearance in-situ measurement and reconstruction technology; S107, the residence time is solved by combining the results obtained by measuring the removal function with the cylindrical appearance reconstructed; S108, based on the residence time solved, the core shaft is deterministically shaped by using the time control grinding processing method; S109, the surface shape of the core shaft after shaping is measured in-situ, if the surface shape meets the technical index requirements, the process is ended and exited; otherwise, jump to step S105 and continue processing; Step S104 comprises: S201, install the distance measuring sensors A and B with a spacing distance s horizontally on the slide plate of the machine tool, which is used to drag the distance measuring sensors A and B to feed along the axis direction of the mandrel; zero the distance measuring sensors A and B to eliminate the influence of zero error on data acquisition; scan the distance measuring sensors A and B along the generatrix L of the mandrel in the horizontal x direction; the obtained data contain the component of the linear motion error of the slide plate dragging the distance measuring sensors A and B in the horizontal x direction , the real profile data of the generatrix L , the component of the coaxial error in the horizontal x direction ; separate the component of the linear motion error of the slide plate in the horizontal x direction using the successive two-point method to obtain the sum of the profile data of the generatrix L in the horizontal x direction and the coaxial error ; install the distance measuring sensors A and B vertically on the slide plate when the mandrel is in a stationary state; zero the distance measuring sensors A and B to eliminate the influence of zero error on data acquisition; scan the distance measuring sensors A and B along the generatrix L of the mandrel in the vertical y direction; the obtained data contain the component of the linear motion error of the slide plate dragging the distance measuring sensors A and B in the vertical y direction , the real profile data of the generatrix L , the component of the coaxial error in the vertical y direction ; separate the component of the linear motion error of the slide plate in the vertical y direction using the successive two-point method to obtain the sum of the profile data of the generatrix L in the vertical y direction and the coaxial error ; S202, add the profile data of the busbar L in the horizontal x direction and the sum of the concentricity errors Subtract the profile data of the busbar L , to obtain the concentricity error in the horizontal x direction ; add the profile data of the busbar L in the vertical y direction and the sum of the concentricity errors Subtract the profile data of the busbar L , to obtain the concentricity error in the vertical y direction .

2. The mandrel-in-place detection machining method according to claim 1, characterized by, The component of the straight line motion error of the slide plate in the horizontal x direction is calculated by using the successive two-point method in step S201 The profile data of the busbar L in the horizontal x direction and the coaxial error are obtained by separating the profile data of the busbar L in the horizontal x direction and the coaxial error The profile data of the busbar L in the horizontal x direction and the coaxial error are obtained by separating the profile data of the busbar L in the horizontal x direction and the coaxial error The profile data of the busbar L in the horizontal x direction and the coaxial error are obtained by separating the profile data of the busbar L in the horizontal x direction and the coaxial error The profile data of the busbar L in the horizontal x direction and the coaxial error are obtained by separating the profile data of the busbar L in the horizontal x direction and the coaxial error The profile data of the busbar L in the horizontal x direction and the coaxial error are obtained by separating the profile data of the busbar L in the horizontal x direction and the coaxial error The profile data of the busbar L in the horizontal x direction and the coaxial error are obtained by separating the profile data of the busbar L in the horizontal x direction and the coaxial error The profile data of the busbar L in the horizontal x direction and the coaxial error are obtained by separating the profile data of the busbar L in the horizontal x direction and the coaxial error The profile data of the busbar L in the horizontal x direction and the coaxial error are obtained by separating the profile data of the busbar L in the horizontal x direction and the coaxial error The profile data of the busbar L in the horizontal x direction and the coaxial error are obtained by separating the profile data of the busbar L in the horizontal x direction and the coaxial error The profile data of the busbar L in the horizontal x direction and the coaxial error are obtained by separating the profile data of the busbar L in the horizontal x direction and the coaxial error 3. The mandrel-in-place detection machining method according to claim 1, characterized by, The component of the straight line motion error of the slide plate in the vertical y direction is calculated by using the successive two-point method in step S201 The profile data of the busbar L in the vertical y direction and the coaxial error are obtained by performing separation The profile data of the busbar L in the vertical y direction and the coaxial error are obtained by performing separation The profile data of the busbar L in the vertical y direction and the coaxial error are obtained by performing separation The profile data of the busbar L in the vertical y direction and the coaxial error are obtained by performing separation The profile data of the busbar L in the vertical y direction and the coaxial error are obtained by performing separation The profile data of the busbar L in the vertical y direction and the coaxial error are obtained by performing separation The profile data of the busbar L in the vertical y direction and the coaxial error are obtained by performing separation The profile data of the busbar L in the vertical y direction and the coaxial error are obtained by performing separation The profile data of the busbar L in the vertical y direction and the coaxial error are obtained by performing separation The profile data of the busbar L in the vertical y direction and the coaxial error are obtained by performing separation 4. The mandrel-in-place detection machining method according to claim 1, characterized by, Step S105 comprises: S301, based on the three-point method, using three ranging sensors A, C and D located in the radial plane of the mandrel to measure a plurality of measured sections of the mandrel in the axial direction, wherein the included angle between the axis of the ranging sensor A and the axis of the ranging sensor C is , the included angle between the axis of the ranging sensor A and the axis of the ranging sensor D is , and the coordinates of any point in the column profile data of the mandrel are determined according to the following formula to achieve the acquisition of the cylindrical profile data: In the above formula, cross section The first sampling points In the absolute coordinate system of the machine tool In Coordinate values cross section The first sampling points In the absolute coordinate system of the machine tool In Coordinate values cross section The first sampling points In the absolute coordinate system of the machine tool In Coordinate values The cross section being measured The x-coordinate of the least squares circle center. The cross section being measured The ordinate of the least squares circle center. The design radius of the mandrel. cross section Roundness error, cross section The difference between the actual measured average radius and the design radius. cross section Upper One sampling point, cross section Total number of sampling points on To measure the cross section being measured Time measurement coordinate system The origin is in the absolute coordinate system The position in the middle, It is the distance between adjacent sections; S302, the coaxiality error is compensated by using the acquired cylindrical profile data, including: first, calculating the component size of the coaxiality error on the installation angle of the ranging sensor A, the ranging sensor C and the ranging sensor D; then subtracting the component size of the coaxiality error on the corresponding ranging sensor installation angle from the data collected by the ranging sensor A, the ranging sensor C and the ranging sensor D, so as to compensate the coaxiality error in the in-situ measurement process.

5. The mandrel-in-place detection machining method according to claim 1, characterized by, Step S106 comprises: S401, for the cylindrical profile data after coaxiality error compensation, the roundness data of each section of the core shaft is calculated according to the three-point method error separation principle; S402, the center position and average radius of each section of the core shaft are calculated by fitting according to the least square principle; S403, the position of the origin of the measurement coordinate system of each section in the absolute coordinate system of the machine tool is calculated; S404, according to the roundness data, the center position and the average radius of each section of the core shaft, and the position of the origin of the measurement coordinate system in the absolute coordinate system, the cylindrical appearance is reconstructed in the absolute coordinate system.

6. The mandrel-in-place detection machining method according to claim 1, characterized by, The result obtained by measuring the removal function in step S107 is combined with the cylindrical appearance reconstructed to solve the residence time, comprising: S501, select an experimental shaft with the same diameter and material as the core shaft to be machined as the removal function shaft; install the removal function shaft on the controlled time grinding machine, install and adjust the controlled time grinding device, and set the machining parameters, including the initial radius of the abrasive belt, the thickness of the single layer of abrasive belt, the abrasive belt updating speed, the tool head vibration frequency, etc.; then, perform controlled time grinding on the removal function shaft at a certain point for a certain time, disassemble it, and measure the processed surface profile using a roundness gauge; the measurement results representing the machining effect after a certain time are taken as the removal function, and the removal amount distribution of the removal function is obtained by measurement; S502, install the core shaft to be machined on the controlled time grinding machine, and perform in-situ measurement and reconstruction of the cylindrical profile of the core shaft using the in-situ measurement device; S503, identify the cylindrical error distribution of the core shaft surface using the in-situ measurement and reconstruction of the cylindrical profile of the core shaft, calculate the machining residence time of the core shaft surface corresponding to the cylindrical error using the pulse iteration method, and obtain the machining residence time of the core shaft surface corresponding to the cylindrical error.

7. The mandrel-in-place detection machining method according to claim 1, characterized by, The step S108 includes the following steps when using controlled time grinding to perform deterministic shaping of the core shaft: S601, after completing the in-situ measurement and reconstruction of the cylindrical profile of the core shaft to be machined, continue to install and adjust the controlled time grinding device and the in-situ measurement device, and set the machining parameters, which are consistent with those when machining the removal function shaft; S602, generate the machining code according to the machining residence time of each position on the core shaft surface, and drive the controlled time grinding device to perform controlled time grinding; S603, perform in-situ measurement and reconstruction of the cylindrical profile of the machined core shaft again, identify the cylindrical error value of each position on the core shaft surface, and determine whether the cylindrical error of the core shaft meets the accuracy requirement; if it meets the accuracy requirement, no compensation machining is needed, and the machining process is completed; if it does not meet the accuracy requirement, continue to calculate the residence time required for compensation machining at each position on the core shaft surface using the pulse iteration method, and jump to step S102.

8. The mandrel-in-place detection machining method according to claim 7, wherein When setting the machining parameters in step S601, the machining parameters include some or all of the initial radius of the abrasive belt, the thickness of the single layer of abrasive belt, the abrasive belt updating speed, and the tool head vibration frequency.

9. An aerostatic spindle comprising an aerostatic spindle body with a mandrel, characterized in that The core shaft is prepared by the in-situ detection and machining method of the core shaft according to any one of claims 1-8.

Images