Prediction device, prediction method, and grinding system

Abstract

To provide a predictive device that predicts the shape of the machined surface of a workpiece processed by a grinding device. [Solution] The prediction device comprises an acquisition unit that acquires vibration data indicating the vibration of the grinding wheel during non-machining mode operation in which the workpiece spindle to which the workpiece is mounted and the spindle to which the grinding wheel is mounted are operated without the grinding wheel contacting the workpiece; a calculation processing unit that performs a simulation; and an output unit. The calculation processing unit calculates the grinding resistance using the machining conditions which are the conditions for the simulation, calculates the grinding wheel clearance amount using the calculated grinding resistance and predetermined rigidity of the grinding wheel and the workpiece, calculates the grinding wheel trajectory indicating the position of the grinding wheel in the time axis using the grinding wheel clearance amount and the vibration data when operated under the machining conditions, calculates the shape data of the machined surface of the workpiece predicted using the grinding wheel trajectory, and the output unit outputs the shape data of the machined surface of the workpiece.

Description

Technical Field

[0001] The present disclosure relates to a prediction device, a prediction method, and a grinding system.

Background Art

[0002] Conventionally, in machine tools, attempts have been made to predict the undulations that occur on the machined surface due to the vibration of the tool. For example, the prediction device of Patent Document 1 predicts the undulation shape of the machined surface using data indicating the vibration during machining of a rotating tool that performs cutting.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, the prediction method of Patent Document 1 is a prediction regarding a cutting device and is difficult to apply to a grinding device. Therefore, a technique for predicting the undulation shape in a grinding device is required.

Means for Solving the Problems

[0005] The present disclosure can be realized in the following forms.

[0006] (1) According to a first embodiment of the present disclosure, a prediction device is provided for predicting the shape of the machined surface of a workpiece processed by a grinding device. The prediction device includes an acquisition unit that acquires vibration data indicating the vibration of the grinding wheel in a non-machining mode operation which operates the workpiece spindle to which the workpiece is mounted and the spindle to which the grinding wheel is mounted without the grinding wheel contacting the workpiece; an arithmetic processing unit that performs a simulation; and an output unit. The arithmetic processing unit calculates the grinding resistance by the simulation using the machining conditions used in the simulation, calculates the grinding wheel relief amount indicating the displacement of the position of the grinding wheel after it has moved due to the grinding resistance applied to it, relative to the command position of the grinding wheel under the machining conditions, using the calculated grinding resistance and predetermined rigidity of the grinding wheel and the workpiece, calculates the grinding wheel trajectory indicating the position of the grinding wheel in the time axis using the grinding wheel relief amount and the vibration data, calculates the shape data of the machined surface of the workpiece predicted by the simulation using the grinding wheel trajectory, and the output unit outputs the shape data of the machined surface of the workpiece. In this configuration, the shape data of a machined surface with undulation can be predicted using vibration data from the non-machining mode. (2) In the prediction device of the above form, the grinding wheel is a screw-shaped grinding wheel, and in the output processing routine as a simulation, the calculation processing unit generates a grinding wheel blade cross-sectional point cloud that defines the cross-sectional shape of the convex blade of the screw-shaped grinding wheel in the axial cross-section of the screw-shaped grinding wheel, and by arranging the grinding wheel blade cross-sectional point cloud in the direction in which the convex blade of the screw-shaped grinding wheel extends, it generates a grinding wheel model that is a three-dimensional model of the convex blade of the screw-shaped grinding wheel, and the workpiece definition point cloud on the surface of the workpiece, and each of the workpiece definition point clouds A workpiece model is generated defined by a group of line segments extending radially from this point as one end, and in a coordinate system based on the workpiece model, the grinding wheel definition point interval Se(i,tn) is the interval between two adjacent grinding wheel cross-section points Pt(i,tn) and Pt(i+1,tn) that constitute the grinding wheel cross-section point group at time tn (where n is an ordinal number), and the grinding wheel definition point interval Se(i,When calculating the grinding resistance, the following steps are performed: Determine the area through which the grinding wheel definition point interval is passed, represented by tn+1), perform an inclusion check to determine whether each workpiece definition point in the group of workpiece definition points constituting the workpiece model is located inside the area through which the grinding wheel definition point interval is passed when viewed from the direction of the line segment with the workpiece definition point being calculated as one end, calculate the intersection point between the target line segment with the workpiece definition point determined to be located inside as one end of the inclusion check and the surface constituting the area through which the grinding wheel definition point interval is passed, change the coordinates of the workpiece definition point located at one end of the target line segment to the coordinates of the intersection point, and calculate the grinding resistance. In this configuration, the grinding resistance parameter used to calculate the grinding resistance is calculated, a region having a boundary centered on one of the workpiece definition points on the surface of the workpiece and passing through the midpoints of adjacent workpiece definition points is defined as a mesh, the grinding resistance is calculated in the mesh containing the modified workpiece definition point among a plurality of meshes using the grinding resistance parameter, and when calculating the shape data of the machined surface of the workpiece using the grinding wheel trajectory, the calculated grinding wheel trajectory is reflected in the area passing through the grinding wheel definition point interval and the output processing routine is performed, the workpiece model defined by the modified workpiece definition point is calculated, and the shape data of the machined surface of the workpiece is calculated using the calculated workpiece model. According to this configuration, the grinding resistance can be calculated accurately by calculating the grinding resistance parameter for each mesh. In addition, by reflecting the grinding wheel trajectory in the area passing through the grinding wheel definition point interval, the shape of the machined surface reflecting vibration data can be predicted. (3) According to a second embodiment of the present disclosure, a prediction method is provided for predicting the shape of the waviness of a machined surface processed by a grinding device. This prediction method calculates grinding resistance by the simulation using the processing conditions used in the simulation, calculates a grinding wheel relief amount which indicates the amount of displacement of the position of the grinding wheel after it has moved due to the grinding resistance applied to it, relative to the command position of the grinding wheel under the processing conditions, using the calculated grinding resistance and predetermined rigidity of the grinding wheel and the workpiece, calculates a grinding wheel trajectory which indicates the position of the grinding wheel in the time axis using vibration data which indicates the vibration of the grinding wheel in the operation of the workpiece spindle to which the workpiece is mounted and the spindle to which the grinding wheel is mounted, without the grinding wheel contacting the workpiece, and the grinding wheel relief amount, and calculates shape data of the machined surface of the workpiece predicted by the simulation using the grinding wheel trajectory. (4) According to a third embodiment of the present disclosure, a grinding system is provided comprising a grinding device and a prediction device. The grinding device comprises a workpiece spindle to which a workpiece is attached, a spindle to which a grinding wheel for grinding the workpiece is attached, a detector for detecting vibrations of the grinding wheel, and a control unit, wherein the control unit operates in either a non-processing mode in which the workpiece spindle and the spindle are operated without the grinding wheel contacting the workpiece, or a processing mode in which the workpiece spindle and the spindle are operated with the grinding wheel in contact with the workpiece, and the prediction device comprises an acquisition unit for acquiring vibration data indicating vibrations of the grinding wheel in the non-processing mode in which the workpiece spindle and the spindle are operated without the grinding wheel contacting the workpiece, a calculation processing unit for performing simulations, and an output unit. The calculation processing unit calculates the grinding resistance by the simulation using the processing conditions used in the simulation, and using the calculated grinding resistance and the predetermined rigidity of the grinding wheel and the workpiece, it calculates the grinding wheel relief amount, which indicates the amount of displacement of the position of the grinding wheel after it has moved due to the grinding resistance applied to it, relative to the command position of the grinding wheel under the processing conditions, and using the grinding wheel relief amount and the vibration data, it calculates the grinding wheel trajectory, which indicates the position of the grinding wheel in the time axis, and calculates the shape data of the machined surface of the workpiece predicted by the simulation using the grinding wheel trajectory, and the output unit outputs the shape data of the machined surface of the workpiece. This disclosure can be implemented in various forms, and in addition to the above forms, it can also be implemented in the form of a prediction program or the like. [Brief explanation of the drawing]

[0007] [Figure 1] This is an explanatory diagram showing the general configuration of the grinding system. [Figure 2] This diagram illustrates the positional relationship between the workpiece and the grinding wheel. [Figure 3] This is a block diagram showing the configuration of the prediction device and the grinding device. [Figure 4] This diagram illustrates the machining process in this embodiment. [Figure 5]This is actual measurement data of the waviness of the machined surface of a workpiece after processing. [Figure 6] This is a flowchart showing the procedure for swell prediction processing. [Figure 7] This is the first flowchart showing the procedure for calculating the grinding wheel clearance. [Figure 8] This is the second flowchart showing the procedure for calculating the grinding wheel clearance. [Figure 9] This is the third flowchart illustrating the procedure for calculating the grinding wheel clearance. [Figure 10] This diagram illustrates the method for calculating grinding resistance parameters. [Figure 11] This figure shows the formulas for calculating each vector and the formula for calculating grinding resistance together. [Figure 12] This diagram illustrates the grinding resistance parameter and the method for calculating the grinding resistance. [Figure 13] This is a flowchart showing the procedure for processing higher-order swells. [Modes for carrying out the invention]

[0008] A. First Embodiment: A1. Mechanical configuration of the grinding machine: Figure 1 is an explanatory diagram showing the schematic configuration of the grinding system 1. The grinding system 1 comprises a grinding device 10 and a prediction device 80. The prediction device 80 is communicatively connected to the grinding device 10. The prediction device 80 receives and acquires detection signals from an acceleration sensor 71, which will be described later, transmitted from the grinding device 10. Then, the prediction device 80 uses the acquired detection signals to predict the shape of the machined surface MS of the workpiece W. Figure 1 shows the X, Y, and Z axes, which are orthogonal to each other. The Z and X axes are coordinate axes parallel to the horizontal plane. The Y axis is a coordinate axis parallel to the vertical direction. The arrows representing the X, Y, and Z axes in Figure 1 and the arrows representing the X, Y, and Z axes in other figures point in the same direction. When specifying the direction, the positive direction, which is the direction pointed to by the arrow, is represented as "+", and the negative direction, which is the direction opposite to the direction pointed to by the arrow, is represented as "-", and both positive and negative signs are used in the direction notation.

[0009] The grinding device 10 includes a bed 12, a workpiece spindle 20, a workpiece axis moving mechanism 30, a ball screw 40, a spindle 50, a spindle housing 52, a swing spindle moving device 60, and a control unit 72. The grinding device 10 processes the workpiece W into an external gear by grinding the outer peripheral surface of the cylindrical workpiece W held by the workpiece spindle 20 with a grinding wheel S held by the spindle 50. The grinding device 10 is configured such that the relative position of the grinding wheel S with respect to the workpiece spindle 20 is changed by the swing spindle moving device 60, so that the workpiece W can be processed into a helical gear.

[0010] The bed 12 supports the workpiece spindle 20, the spindle 50, etc. The bed 12 is formed of, for example, cast iron.

[0011] The workpiece axis moving mechanism 30 rotatably supports the workpiece spindle 20. The workpiece spindle 20 rotates around a workpiece axis AW parallel to the Y axis. The workpiece axis moving mechanism 30 moves along the Y axis. The workpiece W is attached to the workpiece spindle 20.

[0012] The spindle 50 is rotatably supported by the spindle housing 52. A grinding wheel S is attached to the tip of the spindle 50. In the present embodiment, the grinding wheel S is a screw-shaped grinding wheel having spiral grinding teeth (not shown). Inside the spindle housing 52, an acceleration sensor 71 as a detector for detecting vibration is attached. The spindle 50 rotates around a first axis A1. The spindle 50 is also called a grinding wheel axis.

[0013] The ball screw 40 moves the spindle housing 52 along the first axis A1. When the spindle housing 52 moves, the spindle 50 supported by the spindle housing 52 also moves in conjunction.

[0014] The swivel spindle moving device 60 has a swivel spindle support portion 61 and a swivel spindle 62. The swivel spindle 62 is rotatably supported by the swivel spindle support portion 61. The swivel spindle 62 rotates around a second axis A2 parallel to the X-axis. The swivel spindle 62 has a swivel head 62a exposed from the swivel spindle support portion 61 and a swivel shaft base portion 62b housed inside the swivel spindle support portion 61. A ball screw 40 is fixed to the swivel head 62a. When the swivel spindle 62 rotates around the second axis A2, the spindle housing 52 and the spindle 50 rotate in联动. Thereby, the angle of the grindstone S with respect to the workpiece W is changed. The swivel spindle 62 is also referred to as a swivel axis.

[0015] The grinding device 10 operates in either a machining mode or a non-machining mode. The machining mode is a mode for grinding the workpiece W. The non-machining mode is a mode for measuring the vibration of the grindstone S in order to predict undulation. The machining mode is a mode in which the grindstone S is brought into contact with the workpiece W and the workpiece spindle 20 and the spindle 50 are operated. The non-machining mode is a mode in which the workpiece spindle 20 and the spindle 50 are operated without bringing the grindstone S into contact with the workpiece W.

[0016] FIG. 2 is a diagram for explaining the positional relationship between the workpiece W and the grindstone S. In the machining process in the machining mode, the relative position of the grindstone S with respect to the workpiece W is changed in synchronization with the rotation cycle of the workpiece spindle 20, and machining is performed. As described above, the workpiece W rotates around the workpiece axis AW. The grindstone S rotates around the first axis A1. The rotation angle of the swivel spindle 62 with respect to the reference position is also referred to as the swivel angle. In the present embodiment, the maximum value of the swivel angle is about several tens of degrees.

[0017] To machine a workpiece W into a helical gear where the tooth trace TL is inclined with respect to the axis of the workpiece W, the slewing spindle 62 is rotated before machining so that the first axis A1 of the spindle 50 is inclined with respect to the Z axis. Then, the slewing spindle 62 is fixed to the slewing spindle support 61 so that it does not rotate. With the workpiece W set to the first position PO1 in the Y direction, the grinding wheel S and the workpiece W are rotated so that the grinding teeth of the grinding wheel S and the teeth GH of the workpiece W mesh together. Once machining of all the teeth GH of the workpiece W at the first position PO1 is complete, the workpiece W is moved from the first position PO1 to the second position PO2. During the process of moving the workpiece W to the second position PO2, the workpiece W is rotated so that the grinding teeth of the grinding wheel S and the teeth GH of the workpiece W mesh together. Then, similar to the first position PO1, the grinding wheel S and the workpiece W rotate so that the grinding teeth of the grinding wheel S and the teeth GH of the workpiece W engage, and the teeth GH at the second position PO2 are machined. In this way, the teeth GH of the workpiece W are machined by repeating the movement of the workpiece W along the Y direction and the rotation of the grinding wheel S and the workpiece W.

[0018] A2. Electrical configuration of the prediction device and the grinding device: Figure 3 is a block diagram showing the configuration of the prediction device 80 and the grinding device 10. As shown in Figure 3, the prediction device 80 is configured as a computer having a processor 81, a storage unit 82, an input / output interface 83, and a communication interface 84. The storage unit 82 is implemented using memory such as RAM or ROM. The processor 81 and the storage unit 82 are connected via a bus for communication.

[0019] The input / output interface 83 facilitates the exchange of information between the processor 81 and devices such as a display device, storage medium, and keyboard (not shown). The communication interface 84 is an interface for communicating with external devices such as the grinding machine 10.

[0020] The processor 81 has, as functional units, an acquisition unit 87, a calculation processing unit 88 that performs simulations, and an output unit 89.

[0021] The acquisition unit 87 creates vibration data 95 using the detection values ​​from the acceleration sensor 71 transmitted from the grinding device 10 and stores it in the storage unit 82. The vibration data 95 is data that associates time with the displacement of the grinding wheel S. The vibration data 95 is data for three axes: the X axis, Y axis, and Z axis. The acquisition unit 87 also acquires information to be used in the simulation described later. Specifically, the acquisition unit 87 acquires information from other computers, for example, via the communication interface 84. The information used in the simulation includes the final shape of the workpiece W, i.e., the target tooth profile information of the gear G, the material shape of the workpiece W, the outer diameter, axial width, number of grooves of the grinding wheel S, and the mechanical configuration of the grinding device 10.

[0022] The output unit 89 outputs the results of calculations performed by the calculation processing unit 88, such as the shape of the workpiece model Mw and the machined surface MS, which will be described later. Specifically, the output unit 89 stores information indicating the shape of the workpiece model Mw and the machined surface MS as a set of data in the storage unit 82. A set of data is, for example, data in a format that allows the workpiece model Mw to be displayed three-dimensionally on a display device (not shown), or data that allows a graph showing the relationship between the machining time and the shape of the machined surface MS, which will be described later. Each functional unit is realized by executing a program stored in the storage unit 82. The prediction device 80 may also be configured to include electrical circuits such as an ASIC (application specific integrated circuit). The storage unit 82 stores the workpiece model Mw, grinding wheel model Mt, vibration data 95, and grinding wheel relief amount data 96, which will be used in the simulation described later.

[0023] In addition to the above configuration, the grinding apparatus 10 includes a control unit 72 and a communication interface 73. The communication interface 73 is an interface for communicating with external devices such as a prediction device 80. The control unit 72 is configured as a computer having a processor and a memory unit. The memory unit is implemented as memory such as RAM or ROM. The control unit 72 controls various parts such as the workpiece spindle 20. The control unit 72 may also include electrical circuits such as an ASIC.

[0024] The acceleration sensor 71 detects accelerations in the X, Y, and Z axes and transmits the detected values ​​to the control unit 72. Note that the first axis A1 is not necessarily parallel to the Z axis shown in Figure 1 due to the rotation of the pivot axis 62; therefore, the Z axis detected by the acceleration sensor 71 refers to the axis along the first axis A1. The control unit 72 transmits the detected values ​​from the acceleration sensor 71 to the prediction device 80 via the communication interface 73.

[0025] A3. Explanation of surface waviness: Figure 4 is a diagram illustrating the machining process in this embodiment. Figure 4 shows one tooth GH of a gear G, which is a workpiece W, after machining. As shown in Figure 4, in this embodiment, the position of the grinding wheel S is changed six times along the tooth trace direction from the first period PE1 to the sixth period PE6. The amount of material removed in the first period PE1 is called the first material removal amount d1. Similarly, the amount of material removed in each period from the second period PE2 to the sixth period PE6 is called the second material removal amount d2 to the sixth material removal amount d6. Because the amount of material removed from the first material removal amount d1 to the sixth material removal amount d6 varies, undulation occurs on the machined surface MS. In the following description, when the first period PE1 to the sixth period PE6 are not distinguished, they are simply referred to as period PE.

[0026] In the case of a helical gear where the machined workpiece W constitutes a worm gear, the waviness of the machined surface MS is a cause of noise when the worm gear is used. Therefore, in this embodiment, the waviness of the machined surface MS is predicted using simulation. This makes it possible to estimate the waviness before actually performing grinding. Furthermore, currently, appropriate machining conditions to suppress waviness are determined through trial and error. Therefore, by predicting the waviness, it is possible to determine machining conditions that can suppress waviness more efficiently compared to the method of determining machining conditions through trial and error.

[0027] Figure 5 shows measured data of the waviness of the machined surface MS of the workpiece W after machining, measured using a displacement sensor. The horizontal axis of Figure 5 represents the position along the tooth trace line LI, as shown in Figure 4, along the tooth trace direction. The vertical axis of Figure 5 represents the waviness on the machined surface MS. As shown in Figure 5, it is known that waviness includes "low-order waviness" with long periods and "high-order waviness" with short periods. Low-order waviness is also called "waviness," and high-order waviness is also called "surface roughness." The inventors have found that low-order waviness correlates with grinding resistance, and high-order waviness correlates with the vibration of the grinding wheel S. The inventors have also found that this vibration shows a similar trend (although the ratio may differ) both during machining (machining mode) and non-machining (non-machining mode). In the simulation described next, measured vibration data is used to predict high-order waviness.

[0028] A4. Prediction of machined surface shape using simulation: Figure 6 is a flowchart showing the procedure for swell prediction processing using simulation. The prediction method is realized by performing the swell prediction processing. The higher-order swell processing in step S103 shown in Figure 6 is a processing step to reflect the above-mentioned higher-order swell in the simulation. As described above, it has been found that higher-order swell correlates with the vibration of the grinding wheel S. Therefore, in this embodiment, the shape of the machined surface MS with reproduced higher-order swell is predicted by reflecting the measured vibration data of the grinding wheel S in the simulation.

[0029] As shown in Figure 6, in step S101, the arithmetic processing unit 88 performs a subroutine, which is the grinding wheel relief amount calculation process.

[0030] Figures 7 to 9 are flowcharts showing the procedure for calculating the grinding wheel clearance. The processing steps from step S1 to step S29 shown in Figures 7 to 9 are the same as the processing steps from step S1 to step S29 disclosed in Japanese Patent Application Publication No. 2023-33825. Therefore, in this application, the same reference numerals are used for the same processing steps as in Japanese Patent Application Publication No. 2023-33825, and detailed explanations are omitted as appropriate.

[0031] The simulation is performed using the information acquired by the acquisition unit 87. In addition to the information mentioned above, the information for the simulation includes NC command values. The NC command values ​​include the peripheral speed of the workpiece W, the peripheral speed of the grinding wheel S, the speed at which the workpiece W moves along the workpiece axis AW, the speed at which the grinding wheel S moves along the first axis A1, and the position and angle of the grinding wheel S relative to the workpiece W. In the following description, the NC command values, the information on the shape and material of the grinding wheel S mentioned above, and the target tooth profile information of the gear G may be collectively referred to as machining conditions.

[0032] As described above, in actual machining, both the workpiece W and the grinding wheel S rotate. In contrast, the simulation in this embodiment is performed assuming that the position of the workpiece W is fixed and only the position of the grinding wheel S changes. In other words, the simulation is performed assuming that the grinding wheel S rotates around the first axis A1 while rotating around the workpiece axis AW around the workpiece W, which does not displace. In the simulation, the calculation processing unit 88 uses data on the commanded position of the grinding wheel S relative to the workpiece W, which is created using NC command values ​​and correlated with the elapsed time in the actual machining process.

[0033] In step S1 shown in Figure 7, the arithmetic processing unit 88 generates a target tooth profile cross-sectional point group W_tar that defines the cross-sectional shape of the target tooth profile of the gear G in the cross-section perpendicular to the axis of the gear G (Xw-Yw plane). In this embodiment, the gear G is a helical gear. In step S2, the arithmetic processing unit 88 generates a grinding wheel blade cross-sectional point group Pt_g. The grinding wheel blade cross-sectional point group Pt_g defines the cross-sectional shape of the convex blade of the grinding wheel in the axial cross-section of the grinding wheel S, which is a screw-shaped grinding wheel. The grinding wheel blade cross-sectional point group Pt_g is composed of a plurality of grinding wheel blade cross-sectional points Pt. Each of the plurality of grinding wheel blade cross-sectional points Pt is uniquely numbered. To identify each of the plurality of grinding wheel blade cross-sectional points Pt, the grinding wheel blade cross-sectional point Pt(i) is used, where "i" is the assigned number.

[0034] In step S3, the calculation processing unit 88 generates a grinding wheel model Mt. The grinding wheel model Mt is a three-dimensional model created by arranging the grinding wheel blade cross-section point group Pt_g in the direction of extension of the convex blade of the screw-shaped grinding wheel S. Here, each of the multiple grinding wheel blade cross-section point group Pt_g that make up the grinding wheel model Mt is uniquely numbered. To identify each of the multiple grinding wheel blade cross-section point group Pt_g, the grinding wheel blade cross-section point group Pt_g(m) is used, where "m" is the assigned number.

[0035] In step S4, the arithmetic processing unit 88 generates a workpiece model Mw. The workpiece model Mw is defined by a group of workpiece definition points Pw_g and a group of line segments LS_g on the surface of the workpiece W. Each line segment LS in the group of line segments LS_g is a line segment that extends radially across the workpiece W, with each point in the group of workpiece definition points Pw_g as one end. In the following description, one end of each line segment, which is a point in the group of workpiece definition points Pw_g, may be called the start point, and the other end may be called the end point. The group of workpiece definition points Pw_g is composed of multiple workpiece definition points Pw. Each of the multiple workpiece definition points Pw is uniquely numbered. To identify each of the multiple workpiece definition points Pw, the workpiece definition point Pw(q) is used, where "q" is the assigned number.

[0036] In this embodiment, the shape of the workpiece model Mw created in step S4 is cylindrical. The workpiece model Mw is updated during the simulation process. Specifically, in step S29, described later, the workpiece model Mw is updated by updating the positions of the workpiece definition points Pw(q) that constitute it. The storage unit 82 may be configured to store only the latest workpiece model Mw among the workpiece models Mw that are updated as needed, or it may be configured to store workpiece models Mw associated with the elapsed time each time they are updated, that is, to store multiple workpiece models Mw that show the work in progress.

[0037] In step S5, the arithmetic processing unit 88 generates a workpiece division area Ad(h). "h" is a variable representing a unique number assigned to each of the multiple workpiece division areas Ad. The workpiece division area Ad(h) is an area obtained by dividing the workpiece model Mw so that it includes multiple workpiece definition points Pw(q), which are part of the workpiece definition point cloud Pw_g.

[0038] In step S6, the arithmetic processing unit 88 sets the time counter n (where n is an ordinal number) to its initial value of "0". The time counter n is a variable related to time. The time counter n is associated with the elapsed time in the actual machining process, and the value of time that elapses when the time counter n increases by 1 is predetermined by the settings. The time that elapses when the time counter n increases by 1 is also called the "step time".

[0039] In step S7, the arithmetic processing unit 88 determines whether the time counter n is 1 less than the predetermined maximum value n_max of the time counter n. If in step S7 it is determined that the time counter n is not "n_max-1", the arithmetic processing unit 88 increments the time counter n by 1 in step S8 in order to perform calculations for the next time.

[0040] In step S9, the arithmetic processing unit 88 sets the workpiece disassembly area number h (where h is an ordinal number) to its initial value of "0". In step S10, the arithmetic processing unit 88 determines whether the workpiece disassembly area number h is the maximum value h_max. If it is determined in step S10 that the workpiece disassembly area number h is the maximum value h_max, the arithmetic processing unit 88 returns to step S7 because it has completed the calculation for all workpiece division areas Ad(h) at the time being calculated.

[0041] If the arithmetic processing unit 88 determines in step S7 that the time counter n is "n_max-1", then in step S30, the output unit 89 outputs the workpiece model Mw. Here, if the determination in step S7 is "YES", it means that the processing steps from steps S11 to S40, which will be described next, have been completed and all calculations for creating the final workpiece model Mw have been completed. After steps S30 to S42 are performed, this subroutine, the "grinding wheel relief amount calculation process", is terminated and the processing steps are returned to the "undulation prediction process (Figure 6)". Details of the processing steps from step S30 onward will be described later.

[0042] After performing step S10 in Figure 7, in step S11 shown in Figure 8, the arithmetic processing unit 88 increments the workpiece decomposition area number h by 1 in order to perform calculations for the next workpiece division area Ad(h). In step S12, the arithmetic processing unit 88 sets the grinding wheel cross-section point group number m (where m is an ordinal number) to its initial value of "0". The grinding wheel cross-section point group number m is a variable that represents the number "m" of the grinding wheel cross-section point group Pt_g(m).

[0043] In step S13, the calculation processing unit 88 determines whether the grinding wheel blade cross-section point group number m is the maximum value m_max. If, in step S13, it is determined that the grinding wheel blade cross-section point group number m is the maximum value m_max, the calculation processing unit 88 returns to step S10 shown in Figure 7 in order to perform calculations for the next workpiece division area Ad(h).

[0044] In step S13, if it is determined that the grinding wheel blade cross-section point group number m is not the maximum value m_max, in step S14, the arithmetic processing unit 88 increments the grinding wheel blade cross-section point group number m by 1. In step S15, the arithmetic processing unit 88 sets the grinding wheel blade cross-section point number i to its initial value of "0". The grinding wheel blade cross-section point number i is a variable that indicates the number "i" of the grinding wheel blade cross-section point Pt(i). In step S16, the arithmetic processing unit 88 determines whether the grinding wheel blade cross-section point number i is 1 less than the maximum value i_max for the grinding wheel blade cross-section point number i. In step S16, if it is determined that the grinding wheel blade cross-section point number i is "i_max-1", the arithmetic processing unit 88 returns to step S13 because it has completed the calculation for all grinding wheel blade cross-section points Pt(i) in the grinding wheel blade cross-section point group Pt_g(m) that is being calculated.

[0045] If, in step S16, it is determined that the grinding wheel blade section point number i is not "i_max-1", then in step S17, the calculation processing unit 88 increments the grinding wheel blade section point number i by 1. In step S18, the calculation processing unit 88 calculates the grinding wheel definition point intervals Se(i,tn) and Se(i,tn+1). The grinding wheel definition point interval Se(i,tn) is the interval between two adjacent grinding wheel blade section points Pt(i,tn) and Pt(i+1,tn) that constitute the grinding wheel blade section point group Pt_g(tn) at time tn in a coordinate system based on the workpiece model Mw. The grinding wheel blade section point group Pt_g(tn) means the grinding wheel blade section point group Pt_g at time tn. The grinding wheel blade section point Pt(i,tn) means the i-th grinding wheel blade section point Pt of the grinding wheel blade section point group Pt_g at time tn.

[0046] In step S19, the arithmetic processing unit 88 determines the grinding wheel definition point interval passing area At(i.tn). The grinding wheel definition point interval passing area At(i.tn) is the area represented by the grinding wheel definition point interval Se(i,tn) at time tn and the grinding wheel definition point interval Se(i,tn+1) at time tn+1 after Δt. The grinding wheel definition point interval passing area At(i,tn) corresponds to the area that the grinding wheel definition point interval Se(i) passes through between time tn and time tn+1.

[0047] In step S20, the calculation processing unit 88 performs an interference check between the workpiece division area Ad(h) and the grinding wheel definition point section passing area At(i,tn).

[0048] In step S21 of Figure 9, the arithmetic processing unit 88 determines whether or not interference occurred based on the results of the interference check. If it is determined that there is no interference in step S21, the arithmetic processing unit 88 returns to step S17 shown in Figure 8.

[0049] If interference is detected in step S21, in step S22 the arithmetic processing unit 88 sets the initial value of the workpiece definition point number q within the workpiece division area Ad(h) being calculated to "0". In step 23, the arithmetic processing unit 88 determines whether the workpiece definition point number q is at its maximum value q_max. If it is determined in step 23 that the workpiece definition point number q is at its maximum value q_max, the arithmetic processing unit 88 has finished the calculation for all workpiece definition points Pw(q) within the workpiece division area Ad(h) being calculated, and returns to step S16 in Figure 8 for the next processing step.

[0050] In step 23 of Figure 9, if it is determined that the workpiece definition point number q is not the maximum value q_max, then in step S24, the calculation processing unit 88 increments the workpiece definition point number q by 1. In step S25, the calculation processing unit 88 performs an inclusion check to determine whether the workpiece definition point Pw(q) is located inside the grinding wheel definition point section passing area At(i,tn) when viewed from the extension direction of the line segment LS as the target line segment. In step S26, the calculation processing unit 88 determines, based on the inclusion check, whether the workpiece definition point Pw(q) is located inside the grinding wheel definition point section passing area At(i,tn), i.e., whether it is included. If it is determined in step S26 that it is not included, the calculation processing unit 88 returns to step S23 as there is no need to perform step S27.

[0051] If it is determined in step S26 that the area is included, in step S27 the arithmetic processing unit 88 calculates the intersection point Px between the grinding wheel definition point section passing area At(i,tn) and the line segment LS(q). The line segment LS(q) is the line segment LS that starts from the workpiece definition point Pw(q). In step S28, the arithmetic processing unit 88 determines whether the line segment length, which is the length of the line segment LS, has been shortened when the workpiece definition point Pw(q) is changed to the intersection point Px. Specifically, the arithmetic processing unit 88 determines whether the line segment length, which is the length of the line segment LS that starts from the intersection point Px, is shorter than the line segment length of the line segment LS that starts from the workpiece definition point Pw(q). If it is determined in step S28 that the line segment length has not been shortened, the arithmetic processing unit 88 does not need to perform step S29, and returns to step S23.

[0052] If it is determined in step S28 that the line segment length has become shorter, in step S29 the arithmetic processing unit 88 changes the coordinates of the workpiece definition point Pw(q) to the coordinates of the intersection point Px, and changes the line segment length of the line segment LS(q) being calculated to the length of the line segment LS starting from the intersection point Px.

[0053] In step S40, the calculation processing unit 88 calculates the grinding resistance parameter. Figures 10 to 12 are used to calculate the grinding resistance parameter. Figures 10 and 12 are diagrams illustrating the method for calculating the grinding resistance parameter. Figure 11 is a diagram that shows together the formulas used to calculate the main force component vector ft and the back force component vector fn, which will be described later, and the formula used to calculate the grinding resistance F.

[0054] The grinding resistance parameter is a parameter used to calculate the grinding resistance F. The type of grinding resistance parameter varies depending on the grinding model applied. In this embodiment, the case where the grinding resistance is calculated using the horizontal axis surface grinding model shown in Figure 10(a) is used as an example. Note that the applicable surface grinding model is not limited to the horizontal axis surface grinding model shown in Figure 10(a).

[0055] As shown in Figure 10(a), the horizontal-axis surface grinding model used in this embodiment includes at least the depth of cut Δ, the thickness b of the workpiece W, and the table speed v as grinding resistance parameters. In the horizontal-axis surface grinding model, the principal force vector ft and the back force vector fn are defined as shown in Figure 10(a).

[0056] In this embodiment, the grinding resistance parameter is calculated for each mesh at each time step. The grinding resistance parameter is stored in the storage unit 82 in association with the processing time associated with the time counter n. The processing time is the time at which the time counter n is, for example, "0" is the start time. Here, a mesh is defined as an area centered on the workpiece definition point Pw, as shown in Figure 10(b). The boundary between two adjacent meshes is defined by the midpoint of the two adjacent workpiece definition points Pw. That is, the boundary line between two adjacent meshes is a line passing through the midpoint. The surface of the workpiece model Mw is divided into multiple meshes.

[0057] The principal force vector ft and the back force vector fn are expressed using equation (1) shown in Figure 11 (1), using the specific grinding resistance Cp, table speed v, depth of cut (depth direction) Δ, thickness b, grinding wheel peripheral speed V, friction coefficient μ, and grinding wheel cutting edge tip half-angle α.

[0058] In this embodiment, when applying the horizontal axis surface grinding model, the relative velocity vector is used as the principal force component vector ft. The relative velocity vector is a vector defined as shown in Figure 10(c). In this embodiment, the back force component vector fn is defined as a vector perpendicular to both the grinding wheel cross-section vector and the relative velocity vector. The white circles in Figure 10(c) represent four adjacent grinding wheel cutting edge cross-section points Pt(i) at time t. The black circles represent three adjacent workpiece definition points Pw(q) at time t.

[0059] The relative velocity vector is calculated for each workpiece definition point Pw(q) in the workpiece model Mw. In this embodiment, the relative velocity vector used is the velocity vector when the workpiece definition point Pw(q) moves while the grinding wheel model Mt is fixed. The back force component vector fn, which is calculated using the relative velocity vector, is also calculated for each workpiece definition point Pw(q). The grinding wheel cross-section vector for calculating the back force component vector fn is a vector on the grinding wheel axis cross-section that includes the workpiece definition point Pw, and can be calculated from the grinding wheel cross-section vector at time tn and tn+1, respectively.

[0060] In this embodiment, "vΔb" in equation (1) shown in Figure 11(1) is defined as the removal volume Vcut [mm²], which is the removal volume per unit time. 3 Use [ / s]. Equation (2), shown in Figure 11 (2), is obtained by substituting "vΔb" with "Vcut" in equation (1). Note that in equation (2), parameters common to equation (1) are omitted as they are described in Figure 11 (1).

[0061] The removal volume Vcut will be explained using Figure 10(d). Figure 10(d) shows one mesh. The removal volume Vcut is the volume of the region enclosed by the mesh containing the workpiece definition point Pw(q) before its coordinates are changed to the coordinates of the intersection point Px in step S29, and the mesh containing the workpiece definition point Pw(q) after the change. The distance between the mesh before the change and the mesh after the change is called the removal length Lcut. The area of ​​the mesh before the change is called the first mesh area Sm1. The area of ​​the mesh after the change is called the second mesh area Sm2. In this embodiment, the mesh area Smesh, which is the median between the first mesh area Sm1 and the second mesh area Sm2, is used to calculate the removal volume Vcut. Specifically, the upper equation of equation (3) shown in Figure 11(3) is used to calculate the removal volume Vcut. The lower equation of equation (3) is used to calculate the mesh area Smesh.

[0062] As explained above, the principal force vector ft and the back force vector fn can be calculated using equations (2) and (3). Note that the other parameters included in equation (2), namely the specific grinding resistance Cp, the friction coefficient μ, and the half-angle α of the grinding wheel cutting edge, are used according to the types of grinding wheel S and workpiece W being simulated, for example, values ​​entered by the user. The friction coefficient μ is determined, for example, by the type of grinding wheel S. The half-angle α of the grinding wheel cutting edge can be determined using the conical grinding wheel model shown in Figure 12(a).

[0063] In step S40 of Figure 9, the parameters included in equations (2) and (3) are stored in the storage unit 82 as grinding resistance parameters, associated with the value of time counter n or the processing time.

[0064] As described above, in step S30 of Figure 7, the output unit 89 outputs the workpiece model Mw. The workpiece model Mw output here is the workpiece model Mw defined by the workpiece definition point cloud Pw_g that was processed by the simulated grinding process, i.e., the last modified workpiece model. In this embodiment, in addition to the final workpiece model Mw, workpiece model Mw associated with the processing time during the simulated grinding process is also output.

[0065] The simulation may be performed for all stages of the machining process, or for a single period, such as the first period PE1 shown in Figure 4. Furthermore, it may be performed by specifying a position in the depth direction that intersects the tooth trace direction of the gear G.

[0066] In step S41, the output unit 89 outputs grinding resistance parameters. The grinding resistance parameters output to the storage unit 82 are, in detail, a set of grinding resistance parameters associated with the machining time and the workpiece definition point Pw(q).

[0067] In step S42, the calculation unit 88 calculates the grinding resistance F using the grinding resistance parameter. The calculation unit 88 calculates the grinding resistance F for each workpiece division area Ad(h). The formula for calculating the grinding resistance F is as shown in equation (4) in Figure 11. As shown in equation (4), the grinding resistance F is calculated for each of the mutually orthogonal x, y, and z axes. In equation (4), "Fx" means the grinding resistance F in the x axis, "Fy" means the grinding resistance F in the y axis, and "Fz" means the grinding resistance F in the z axis. In equation (4), "fn,q" means the back force vector fn for the workpiece definition point Pw(q). In equation (4), "ft,q" means the main force vector ft for the workpiece definition point Pw(q). In equation (4), "vx", "vy", and "vz" are the respective velocities obtained by decomposing the relative velocity between the workpiece W and the grinding wheel S into x, y, and z axes, respectively.

[0068] The x, y, and z axes of the grinding resistance F correspond to the x, y, and z axes output by the acceleration sensor 71.

[0069] The arithmetic processing unit 88 adjusts the grinding resistance F for each workpiece division area Ad(h) calculated using equation (4) for all workpiece division areas Ad(h) with the same time counter value n. In other words, the arithmetic processing unit 88 calculates the grinding resistance F for the x-axis, y-axis, and z-axis for each machining time.

[0070] In step S43, the calculation processing unit 88 calculates the grinding wheel relief amount [mm]. In this embodiment, the grinding wheel relief amount is calculated using a rigidity model. The grinding wheel relief amount is the amount of displacement of the grinding wheel S due to the grinding resistance F applied from the workpiece W to the grinding wheel S when the grinding wheel S applies grinding force to the workpiece W.

[0071] Figure 12(b) shows a rigidity model. As shown in Figure 12(b), the grinding wheel S applies a grinding force to the workpiece W. In response, the grinding wheel S receives a grinding resistance F from the workpiece W. The grinding wheel S is attached to the spindle 50. When the grinding wheel S is subjected to an external force, the amount of displacement of the grinding wheel S is determined by the resistance component to the external force and the damper component that mitigates the external force. Similarly, the workpiece W is attached to the workpiece spindle 20. When the workpiece W is subjected to an external force, the amount of displacement of the workpiece W is determined by the resistance component and the damper component. Note that the rectangular shape shown in Figure 12(b) represents the damper component.

[0072] When the static stiffness [N / mm] that indicates the magnitude of the displacement of the grinding wheel S when subjected to an external force is defined as "k1", and the static stiffness that indicates the magnitude of the displacement of the workpiece W when subjected to an external force is defined as "k2", the "grinding wheel relief amount", which is the displacement of the grinding wheel S due to the external force, can be expressed by equation (5). Static stiffness k1 and static stiffness k2 can be determined in advance by experimentation or other means. In this embodiment, static stiffness k1 and static stiffness k2 are included in the machining conditions. In this application, as shown in equation (5) in Figure 12(b), the displacement when the external force is grinding resistance F is called the grinding wheel relief amount. In this embodiment, the positive directions of the X, Y, and Z axes are set to coincide with the direction in which the grinding wheel S moves away from the workpiece W. Furthermore, low-order undulation has a low frequency and the movement of the grinding wheel S and workpiece W is slow, so static stiffness can be used as the stiffness. Depending on the processing and the state of the waviness, dynamic stiffness can also be used.

[0073] As described above, in step S42, one grinding resistance F value is calculated for one axis for each machining time. Then, in step S43, one grinding wheel clearance value is calculated for one axis for each machining time. When the calculation processing unit 88 calculates the grinding wheel clearance, it outputs the calculated grinding wheel clearance. That is, the calculation processing unit 88 stores the grinding wheel clearance data 96, which is the calculated grinding wheel clearance as a single set of data, in the storage unit 82.

[0074] After performing step S43, the arithmetic processing unit 88 terminates this subroutine and returns to the swell prediction process shown in Figure 6.

[0075] In step S102 of Figure 6, the arithmetic processing unit 88 performs interpolation. This process is to match the time interval of the grinding wheel escape amount data with the time interval of the vibration data 95, which will be described later. As described above, since the vibration data 95 is used to reproduce higher-order undulations, the time interval of the vibration data 95 is set to be shorter than the time interval of the simulation. Specifically, the arithmetic processing unit 88 uses the grinding wheel escape amount data associated with time to create grinding wheel escape amount data with finer time intervals by linear interpolation. In this embodiment, the time interval in the simulation, that is, the step time that elapses as the time counter n increases by 1, is approximately 0.1 ms or more and less than 0.5 ms. The sampling time interval of the vibration data 95, which will be described later, is approximately 0.5 ms.

[0076] In step S103, the arithmetic processing unit 88 performs a subroutine called higher-order wave processing. Figure 13 is a flowchart showing the procedure for higher-order wave processing.

[0077] In step S50 of Figure 13, the acquisition unit 87 acquires acceleration data, which is data output from the acceleration sensor 71, in non-machining mode under the same machining conditions as the simulation. In this embodiment, acceleration data for the X, Y, and Z axes are acquired. In non-machining mode, the same machining conditions as the simulation specifically mean that conditions other than the distance of the grinding wheel S to the workpiece W, such as the peripheral speed of the grinding wheel S and the peripheral speed of the workpiece W, match the machining conditions.

[0078] In this embodiment, acceleration data for the period corresponding to one tooth groove in one period PE, which is simulated, is acquired as data used in the simulation.

[0079] In step S51, the acquisition unit 87 converts the time-axis acceleration data into frequency-axis acceleration data using the Fast Fourier Transform. In step S52, the acquisition unit 87 integrates the frequency-axis acceleration data twice to create frequency-axis vibration data 95. In step S53, the acquisition unit 87 performs an inverse Fast Fourier Transform on the frequency-axis vibration data 95 to create the time-axis acquisition unit 87 and stores the created time-axis vibration data in the storage unit 82. Note that the vibration data may be filtered to remove high-frequency and low-frequency components, which are noise.

[0080] When the acquisition unit 87 completes step S53, it terminates the high-speed swell processing subroutine and returns to the swell prediction processing step shown in Figure 6.

[0081] In this embodiment, as shown in Figure 6, step S103 is performed after step S101. In other embodiments, step S101 may be performed after step S103, or steps S103 and S101 may be performed simultaneously in parallel.

[0082] In step S104, the acquisition unit 87 acquires vibration data 95 stored in the storage unit 82. In another embodiment, the acquisition unit 87 may acquire separately created vibration data 95 via the communication interface 84.

[0083] In step S105, the acquisition unit 87 acquires the grinding wheel relief amount data 96 stored in the storage unit 82.

[0084] In step S106, the calculation processing unit 88 uses the grinding wheel relief amount data 96 and vibration data 95 to calculate the grinding wheel trajectory, which indicates the position of the grinding wheel S in the time axis. Specifically, the grinding wheel trajectory is calculated for each machining time using the following equation (6). Grinding wheel trajectory (t) = NC command value (t) - grinding wheel relief amount (t) + grinding wheel vibration (t) ... (6) In equation (6), "(t)" represents the value at the machining time. Specifically, the NC command value in equation (6) indicates the commanded position of the grinding wheel S relative to the workpiece W. Grinding wheel vibration refers to the displacement of the grinding wheel S relative to the reference position, as shown by the vibration data 95. Note that the positive directions of the X, Y, and Z axes of the vibration data 95 correspond to the direction in which the grinding wheel S moves away from the workpiece W. The grinding wheel trajectory (t) is calculated for each of the X, Y, and Z axes, and then summed up to calculate one grinding wheel trajectory for each machining time. The grinding wheel trajectory is data in the same format as the NC command value.

[0085] In step S107, the arithmetic processing unit 88 performs the workpiece model output processing, which is a subroutine. The workpiece model output processing is a processing routine that excludes steps S40 to S43 from the grinding wheel relief amount calculation processing shown in Figures 7 to 9. For this reason, a detailed flowchart of the workpiece model output processing is omitted. The grinding wheel trajectory calculated in step S106 is used as the NC command value for the machining conditions used in the workpiece model output processing. Specifically, the grinding wheel trajectory is reflected in the grinding wheel definition point section Se in step S18 and the grinding wheel definition point section passing area At in step S19 in Figure 8. This makes it possible to output a workpiece model Mw that reflects the grinding wheel relief amount and vibration data 95. Note that the step time that elapses as the time counter n increases by 1 in the workpiece model output processing is set to match the sampling time interval of the vibration data 95. By using the grinding wheel trajectory calculated in step S106, efficient machining simulations can be performed, and the calculation of the shape of the machined surface MS in step S108 of Figure 6 can be performed with high accuracy and efficiency.

[0086] In step S108 of Figure 6, the calculation processing unit 88 calculates the shape data of the machined surface MS of the workpiece W predicted using the grinding wheel trajectory. In this embodiment, the shape data of the machined surface MS corresponding to one tooth groove in one period PE is calculated. Specifically, the calculation processing unit 88 calculates the shape data of the machined surface MS using the workpiece model Mw output in step S107. Specifically, the shape data of the machined surface MS is the data of multiple workpiece definition points Pw that constitute one surface constituting one tooth groove in the workpiece model Mw. In other words, the shape data of the machined surface MS is data that can draw the shape of the machined surface MS in three-dimensional space. Note that the shape data created in step S108 is not limited to data that can draw a plane, but may also be one-dimensional data along the tooth trace direction at a desired position, or one-dimensional data along the depth direction. Specifically, one-dimensional data is data that shows the correspondence between a position along one direction and the workpiece definition point Pw.

[0087] In step S109, the output unit 89 outputs the calculated shape data of the machined surface MS and terminates this processing routine. The output shape data can be used to understand the waviness of the machined surface MS. The shape data of the machined surface MS is useful for developing techniques to reduce waviness.

[0088] Steps S1 to S30 of the grinding wheel clearance calculation process are also called the output processing routine.

[0089] According to the embodiment described above, the prediction device 80 comprises an acquisition unit 87, a calculation processing unit 88, and an output unit 89. The calculation processing unit 88 calculates the grinding wheel relief amount using the grinding resistance calculated using static stiffness, and calculates the grinding wheel trajectory using the grinding wheel relief amount and vibration data. Then, it calculates the shape data of the machined surface MS of the workpiece W predicted by simulation using the grinding wheel trajectory. As a result, the simulation is performed reflecting the grinding resistance correlated with low-order undulation and the vibration data correlated with high-order undulation. Therefore, it is possible to calculate the shape data of the predicted machined surface MS in which both low-order and high-order undulation are reproduced.

[0090] Furthermore, by identifying the characteristics of the swell (nature, cause, etc.), calculations for higher-order swells require a large amount of data and therefore a long calculation time due to the high frequency. However, by efficiently calculating higher-order swells using vibration data, the calculation time for simulations can be shortened. In addition, by using vibration data of the grinding wheel S in non-processing mode, it becomes unnecessary to use the actual workpiece W to check the processing state, and the time and effort required for actual processing can be eliminated. Then, by simulating lower-order swells and higher-order swells using vibration data, it becomes possible to predict the shape of the processed surface MS in a relatively short simulation time without processing the workpiece W.

[0091] Furthermore, the calculation processing unit 88 calculates the grinding resistance F and the shape of the machined surface MS using geometric simulation according to the procedure described above. By calculating the grinding resistance parameter for each mesh, the grinding resistance F can be calculated with high accuracy. In addition, by reflecting the grinding wheel trajectory in the grinding wheel definition point section passing area At rear, the shape of the machined surface MS reflecting the vibration data 95 can be predicted.

[0092] B. Other embodiments: (B1) In the above embodiment, an acceleration sensor 71 is provided as an example of a detector for detecting vibrations of the grinding wheel S. The detector is not limited to an acceleration sensor 71, and vibrations may be detected using the current value of the rotary motor for the spindle 50 or the current value of the motor for the ball screw 40. In this embodiment, the acceleration sensor 71 is located inside the spindle housing 52, but the location of the acceleration sensor 71 is not limited to this. For example, the acceleration sensor 71 may be mounted on the workpiece axis AW. Also, the detector is not limited to one, and multiple detectors may be used to detect vibrations of the grinding wheel S.

[0093] (B2) In the above embodiment, shape data of the machined surface MS constituting one tooth groove of one period PE is calculated. In another embodiment, data corresponding to all period PE for one tooth groove may be created. In this case, vibration data 95 for each period PE corresponding to the tooth groove to be simulated is obtained.

[0094] (B3) In the above embodiment, shape data of the machined surface MS constituting one tooth groove of one period PE is calculated. In other forms, data for all tooth grooves of all teeth corresponding to all period PEs or one period PE may be created. The method for creating this data will be described below. The grinding resistance F for each workpiece division area Ad(h) calculated using equation (4) is associated with the machining time. Therefore, the calculation processing unit 88 converts the data so that the grinding resistance F associated with the machining time is associated with the rotation angle of the workpiece W. More specifically, since the simulation is performed assuming that the workpiece W is fixed, the data for the grinding resistance F is converted so that the grinding resistance F is associated with the rotation angle of the workpiece W. The calculation processing unit 88 uses the peripheral speed of the workpiece W and the machining time to convert the data so that the rotation speed of the workpiece W around the workpiece axis AW is associated with the grinding resistance F. Furthermore, the simulation is performed on one tooth groove of the gear G. Therefore, when calculating the grinding resistance F for all tooth grooves of gear G, the data for the grinding resistance F for one tooth groove is duplicated and created for the number of teeth on gear GH. In the machining process, in addition to the case where the teeth of the workpiece W and the teeth of the grinding wheel S mesh in a one-to-one correspondence, there are also cases where the teeth of the workpiece W and the teeth of the grinding wheel S mesh in a two-to-one correspondence, for example. This correspondence is determined by the specifications of gear G and grinding wheel S. In this embodiment, in order to reproduce the case where the teeth of the workpiece W and the teeth of the grinding wheel S mesh in a two-to-one correspondence, for example, the data for one tooth groove, to which the rotation angle of the workpiece W and the grinding resistance F are associated, is adjusted so that it overlaps in the range of angles where the two-to-one correspondence meshes, according to the circular pitch of gear G. Then, for the overlapping data portions, the grinding resistance F is adjusted for the X, Y, and Z axes, respectively. This allows for the calculation of three grinding resistance F values ​​for the X, Y, and Z axes, respectively, corresponding to one rotation of the workpiece W.

[0095] (B4) In the above embodiment, when calculating the grinding wheel trajectory in step S106, both the grinding wheel relief amount and the grinding wheel vibration are used, but either the grinding wheel relief amount or the grinding wheel vibration may be used alone. When the grinding wheel relief amount is used, low-order swells can be predicted, and when the grinding wheel vibration is used, high-order swells can be predicted.

[0096] (B5) In the above embodiment, vibration data 95 in the non-processing mode is used, but vibration data 95 in the processing mode may also be used. Also, in the above embodiment, vibration data 95 of the grinding wheel S is used, but vibration data of the workpiece W may also be used in equation (6). In either case, the accuracy of the simulation can be improved, that is, shape data that is closer to the actual shape of the processed surface MS can be calculated. Furthermore, by balancing the grinding wheel (adjusting to eliminate the deviation of the rotation center) or dressing it, or by changing the processing conditions (performed before processing, etc.) based on the magnitude and period of the wobble of the processed surface MS, it is possible to suppress the wobble.

[0097] This disclosure is not limited to the embodiments described above, and can be implemented in various configurations without departing from its spirit. For example, the technical features of the embodiments corresponding to the technical features in each form described in the summary of the invention can be replaced or combined as appropriate in order to solve some or all of the above-described problems, or to achieve some or all of the above-described effects. Furthermore, if a technical feature is not described as essential in this specification, it can be deleted as appropriate. [Explanation of symbols]

[0098] 1...Grinding system, 10...Grinding device, 12...Bed, 20...Workpiece spindle, 30...Workpiece axis movement mechanism, 40...Ball screw, 50...Spindle, 52...Spindle housing, 60...Swivel spindle movement device, 61...Swivel spindle support, 62...Swivel spindle, 62a...Swivel axis head, 62b...Swivel axis base, 71...Accelerometer, 72...Control unit, 73...Communication interface, 80...Prediction device, 81...Processor, 82...Storage unit, 83...Input / output interface, 84...Communication interface, 87...Acquisition unit, 88...Calculation processing unit, 89...Output unit, 95...Vibration data, 96...Grinding wheel relief amount data, A1...First axis, AW...Workpiece axis, MS...Machined surface, S...Grinding wheel

Claims

1. A predictive device for predicting the shape of the machined surface of a workpiece processed by a grinding device, An acquisition unit that acquires vibration data indicating the vibration of the grinding wheel during non-machining mode operation, which operates the workpiece spindle to which the workpiece is mounted and the spindle to which the grinding wheel is mounted, without the grinding wheel contacting the workpiece. The calculation processing unit that performs the simulation, It includes an output section, The aforementioned arithmetic processing unit is The grinding resistance is calculated by the simulation using the processing conditions used in the simulation. Using the calculated grinding resistance and the predetermined rigidity of the grinding wheel and the workpiece, the grinding wheel relief amount is calculated, which indicates the amount of displacement of the grinding wheel's position after it has moved due to the grinding resistance applied to it, relative to the commanded position of the grinding wheel under the processing conditions. Using the aforementioned grinding wheel relief amount and the vibration data, the grinding wheel trajectory, which indicates the position of the grinding wheel in the time axis, is calculated. The shape data of the machined surface of the workpiece, predicted by the simulation using the grinding wheel trajectory, is calculated. The output unit is a prediction device that outputs shape data of the machined surface of the workpiece.

2. A prediction device according to claim 1, The aforementioned grinding wheel is a threaded grinding wheel, The aforementioned arithmetic processing unit, in the output processing routine for the simulation, A point cloud of grinding wheel blade cross-sections is generated in the axial cross-section of the aforementioned threaded grinding wheel, defining the cross-sectional shape of the convex blade of the threaded grinding wheel. By arranging the cross-sectional point group of the grinding wheel blade in the direction in which the convex blade of the screw-shaped grinding wheel extends, a grinding wheel model is generated that represents the convex blade of the screw-shaped grinding wheel in three dimensions. A workpiece model is generated, defined by a group of workpiece-defining points on the surface of the workpiece, and a group of line segments extending radially from each of the points in the workpiece-defining point group. In a coordinate system based on the workpiece model, the grinding wheel definition point interval area is determined, which is represented by the grinding wheel definition point interval Se(i, tn), which is the interval between two adjacent grinding wheel cross-section points Pt(i, tn) and Pt(i+1, tn) that constitute the grinding wheel cross-section point group at time tn (where n is an ordinal number), and the grinding wheel definition point interval Se(i, tn+1), which is the interval between two adjacent grinding wheel cross-section points Pt(i, tn+1) and Pt(i+1, tn+1) at time tn+1, which is Δt after time tn. An inclusion check is performed to determine whether each workpiece definition point in the workpiece definition point group constituting the workpiece model is located within the area through which the grinding wheel definition point section passes, when viewed from the direction of a line segment with the workpiece definition point being calculated as one end. In the inclusion check, the intersection point of the target line segment having one end as the workpiece definition point determined to be located inside, and the surface constituting the area through which the grinding wheel definition point section passes is calculated. The workpiece definition point located at one end of the aforementioned target line segment is changed to the coordinates of the intersection point, When calculating the grinding resistance, After performing the output processing routine described above, for each workpiece definition point, the grinding resistance parameter used to calculate the grinding resistance in the surface grinding model is calculated using the distance between the workpiece definition point before the change and the workpiece definition point after the change. A region is defined as a mesh, centered on one of the workpiece definition points on the surface of the workpiece and having a boundary passing through the midpoints of adjacent workpiece definition points. The grinding resistance is calculated using the grinding resistance parameter in one of the multiple meshes that includes the modified workpiece definition point. When calculating the shape data of the machined surface of the workpiece using the grinding wheel trajectory, The calculated grinding wheel trajectory is reflected in the area through which the grinding wheel definition point section passes, and the output processing routine is performed to calculate the workpiece model defined by the modified workpiece definition points. A prediction device that calculates shape data of the machined surface of a workpiece using the calculated workpiece model.

3. A prediction method for predicting the shape of undulations on a machined surface processed by a grinding device, The grinding resistance is calculated by the simulation using the processing conditions used in the simulation. Using the calculated grinding resistance and the predetermined rigidity of the grinding wheel and the workpiece, the grinding wheel relief amount is calculated, which indicates the amount of displacement of the grinding wheel's position after it has moved due to the grinding resistance applied to it, relative to the commanded position of the grinding wheel under the processing conditions. Using vibration data indicating the vibration of the grinding wheel during non-machining mode operation, in which the workpiece spindle to which the workpiece is mounted and the spindle to which the grinding wheel is mounted are operated without the grinding wheel contacting the workpiece, and the amount of relief of the grinding wheel, the grinding wheel trajectory indicating the position of the grinding wheel in the time axis is calculated. A prediction method for calculating shape data of the machined surface of a workpiece, which is predicted by the simulation using the grinding wheel trajectory.

4. A grinding system comprising a grinding device and a prediction device, The aforementioned grinding device, The workpiece spindle to which the workpiece is attached, A spindle to which a grinding wheel for grinding the aforementioned workpiece is attached, A detector for detecting the vibration of the grinding wheel, It has a control unit and The control unit operates in either a non-machining mode, in which the workpiece spindle and the spindle are operated without the grinding wheel contacting the workpiece, or a machining mode, in which the grinding wheel contacts the workpiece and the spindle is operated. The prediction device is An acquisition unit that acquires vibration data indicating the vibration of the grinding wheel during non-machining mode operation in which the workpiece spindle and the spindle are operated without contacting the workpiece, The calculation processing unit that performs the simulation, It includes an output section, The aforementioned arithmetic processing unit is The grinding resistance is calculated by the simulation using the processing conditions used in the simulation. Using the calculated grinding resistance and the predetermined rigidity of the grinding wheel and the workpiece, the grinding wheel relief amount is calculated, which indicates the amount of displacement of the grinding wheel's position after it has moved due to the grinding resistance applied to it, relative to the commanded position of the grinding wheel under the processing conditions. Using the aforementioned grinding wheel relief amount and the vibration data, the grinding wheel trajectory, which indicates the position of the grinding wheel in the time axis, is calculated. The shape data of the machined surface of the workpiece, predicted by the simulation using the grinding wheel trajectory, is calculated. The output unit is a grinding system that outputs shape data of the machined surface of the workpiece.

Images