JP2018124996A - Numerical control device and control method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a numerical control device and a control method capable of improving processing quality even with a traveling route in which small blocks continue.SOLUTION: A numerical control device creates a speed profile along a route obtained by interpolating command points commanded by an NC program. The numerical control device specifies the focused command point from among the command points and calculates an influenced section before and after the focused command point (S26). The numerical control device divides the influenced section and calculates a command interpolation point (S27). The numerical control device processes the command interpolation point with a travel average filter (S28) and calculates an operation interpolation point. The numerical control device calculates, as an error, a distance between the focused command point and the closest operation interpolation point (S29). When the error exceeds a target error, the numerical control device calculates a deceleration rate and updates the speed profile based on the influenced section and the deceleration rate. The numerical control device decelerates at the influenced section including the command point, and thus acceleration and deceleration are not frequently repeated even with a route in which small blocks continue.SELECTED DRAWING: Figure 10

Description

  The present invention relates to a numerical control device and a control method.

  A numerical control device of a machine tool performs machining by moving a tool or a work material along a path commanded by an NC program. At that time, the numerical control device automatically performs smooth acceleration / deceleration and performs acceleration / deceleration control for suppressing vibration of the machine tool. When acceleration / deceleration control is performed, the machining cycle is extended once the block is stopped at the joint between the blocks. Therefore, the numerical control apparatus adopts a control method that starts acceleration of a new block during deceleration of the previous block, thereby shortening the machining cycle. This method has a problem that an inward error occurs at the corner of the joint. The numerical control device disclosed in Patent Document 1 adjusts the overlap time in which the deceleration of the previous block and the acceleration of the new block overlap at the joint portion of the block by delaying the start of the block movement command. The inward error is due to the overlap time. Therefore, the numerical controller can adjust the inward error within the allowable error.

Japanese Patent Laying-Open No. 2006-133888

  The numerical control device described above has a problem in that the quality of the machined surface deteriorates because the corner portion is frequently decelerated and the block portion is accelerated particularly in a smooth movement path in which minute blocks are continuous.

  An object of the present invention is to provide a numerical control device and a control method capable of improving machining quality even on a movement path in which minute blocks are continuous.

  The numerical control device according to claim 1 of the present invention creates a speed profile including information on distance and speed between each command point along a path obtained by interpolating a plurality of command points commanded by the NC program. In the numerical control apparatus including a control unit that controls the movement of the tool or the work material based on the result of the acceleration / deceleration processing by the filter, the control unit pays attention to the command points in the path. A designation means for designating a command point as an attention command point, an interval designation means for designating a time interval based on a time constant of the filter before and after the attention command point designated by the designation means, and the interval designation means A dividing unit that divides the specified section at a predetermined time; a processing unit that processes the section divided by the dividing unit with a filter having a time constant equivalent to the filter; and Error calculating means for calculating an error from the path at the target command point in the section processed in step (i), and a determining means for determining whether the error calculated by the error calculating means exceeds a preset specified value; When the determining means determines that the error exceeds the specified value, a calculating means for calculating a deceleration rate based on the specified value and the error, the section specified by the section specifying means, and the calculating means Update means for updating the speed profile based on the calculated deceleration rate is provided. Since the numerical control device does not decelerate only at the command point but decelerates in the section including the command point, the acceleration / deceleration is not repeated frequently even in a route in which minute blocks continue. Therefore, the numerical control device can improve the quality of the work surface of the work material. The numerical control apparatus can comprehensively suppress errors caused by acceleration / deceleration in the corner portion, the arc portion, and the curved portion in the path by sequentially specifying a plurality of command points for each command point in the path.

  The section of the numerical control device according to claim 2 includes a small section that moves at least half the total time constant of the filter before the target command point, and after the target command point, It is good to include the same small section. Therefore, since the numerical control apparatus can update the speed profile in the section that affects the error at the command point, the error at the command point can be effectively suppressed.

  The calculation means of the numerical control device according to claim 3 may calculate a ratio obtained by dividing the specified value by the error as the deceleration rate. Therefore, the numerical control device can easily calculate the deceleration rate.

  The numerical control device according to claim 4, wherein the error determination means for determining whether or not the error when moving in the section at the deceleration rate calculated by the calculation means is within an allowable range with respect to the specified value; When the error determining unit determines that the error is outside the allowable range, the error determining unit may include a changing unit that changes the deceleration rate so that the error is within the allowable range. Since the numerical control device changes the deceleration rate so that the error falls within the allowable range, it does not decelerate too much. Therefore, the numerical control device can complete the machining of the work material in a short machining cycle. Since there is no shortage of deceleration, it is more certain that the error does not exceed the allowable range.

  The numerical control device according to claim 5 is an interpolation determining unit that determines whether a block between the command points in the path is a linear interpolation or a circular interpolation, and the interpolation determining unit includes: An additional means for adding a virtual command point to the circular interpolation portion when it is determined that the circular interpolation is performed, and the designation means determines that the interpolation determination means determines that the block is the linear interpolation When the end point of the block is designated as the target command point, and the interpolation determining unit determines that the block is the circular interpolation, the virtual command point added by the adding unit, the end point of the block, May be designated as the attention command points. When the block in the path is circular interpolation, an inward error due to the curvature of the circular arc portion can be taken into consideration by adding a virtual command point to the circular interpolation portion. Therefore, even in the case of circular interpolation, the inward error at the virtual command point, that is, the inward error with respect to the circular arc can be suppressed to a specified value or less. Therefore, the quality of the work surface of the work material can be improved even in the circular interpolation portion.

  The updating means of the numerical control device according to claim 6 sets a deceleration section that moves at the deceleration rate calculated by the calculating means to the section when the attention command point is an end point of the block. When the attention command point is the virtual command point, it is preferable to include setting means and second setting means for setting the deceleration section at least over the entire area of the circular interpolation. When the target command point is a virtual command point added on the circular arc, the numerical control device sets the deceleration section over the entire circular interpolation. Therefore, the numerical control device can take into account an inward error due to the curvature of the arc portion.

  The numerical control device according to claim 7, the recalculation unit that recalculates the error when the update unit updates the speed profile and then moves based on the speed profile, and the recalculation unit recalculates. Re-determination means for re-determining whether the error exceeds the specified value, and when the re-determination means determines that the error exceeds the specified value, the calculation means, and the update Each process by the means may be executed again. Therefore, the numerical control device can surely suppress the error from the path below the specified value.

  The numerical control device according to claim 8, the recalculation unit that recalculates the error when the update unit updates the speed profile and then moves based on the speed profile, and the recalculation unit recalculates Re-determination means for re-determining whether the error exceeds the specified value, and when the re-determination means determines that the error exceeds the specified value, the error is equal to or less than the specified value. Until it becomes, it is good to repeatedly perform each process by the said calculation means and the said update means. Therefore, the numerical control device can surely suppress the error from the path below the specified value.

  A control method according to claim 9 of the present invention creates a speed profile including information on distance and speed between each command point along a path obtained by interpolating a plurality of command points commanded by the NC program. In the deceleration control method of the numerical control device for controlling the movement of the tool or the work material based on the result of the acceleration / deceleration processing by the filter, the command point to be noticed is selected from the command points in the path. A designation step that designates a time interval based on a time constant of the filter before and after the target command point designated in the designation step, and a predetermined interval that is designated in the interval designation step. A division step of dividing by time, a processing step of processing the section divided by the division step with a filter having a time constant equivalent to the filter, and the section processed in the processing step. In the error calculating step of calculating an error from the route at the attention command point, the determining step of determining whether the error calculated in the error calculating step exceeds a preset specified value, and the determining step, When it is determined that the error exceeds the specified value, a calculation step of calculating a deceleration rate based on the specified value and the error, the interval specified in the interval specifying step, and the deceleration rate calculated in the calculating step And an update step for updating the speed profile. Therefore, the numerical control device can obtain the effect of claim 1 by executing the above steps.

FIG. 3 is a block diagram showing an electrical configuration of the numerical control device 30 and the machine tool 1. The figure which processed the moving speed with the two-stage moving average filter. The figure of program path | route R1 comprised by linear interpolation block B1 and B2, and path | route R2 after a filter process. The figure which shows the change of the moving speeds 1 and 2 when moving the path | route R2 shown in FIG. It is a figure of path | route R4 when not decelerating program path | route R3, and path | route R5 when a fully automatic deceleration process is carried out. It is a figure which shows the speed block when not decelerating, and the speed block when a fully automatic deceleration process is carried out. The flowchart of the main process. The conceptual diagram of the block information 20. FIG. The flowchart of a fully automatic deceleration process. Flow chart of deceleration calculation processing. The figure which shows the circular arc center command point set to circular interpolation. The flowchart of a circular arc center command point calculation process. The figure which shows the attention command point in the program path | route, an influence start point, and an influence end point. The flowchart of an influence area calculation process. The figure which shows the influence area (T / 2) before an instruction | indication command point, and an influence start point. The figure which shows the influence area (T / 2) after an attention command point, and an influence end point. The flowchart of an error calculation process. The figure which shows the difference | error of a command point and a search point. The figure which shows the closest point from the command point in case the angle which a vector makes is an acute angle. The figure which shows the vector p which looked at the motion interpolation point with the _2nd smallest error from the motion interpolation point with the smallest error, and the vector ε2 when the command point was seen from the motion interpolation point with the smallest error. It is a figure which shows the closest point from the command point in FIG. 20, and error (epsilon). The figure which shows the closest point from the command point in case the angle which a vector makes is an obtuse angle. It is a flowchart of an error suppression process. The figure which shows the method of approaching an error to a target error by linear interpolation. The flowchart of a speed profile update process. The figure which shows the influence area when the C3 command point is designated as the target command point. The chart which shows the speed profile of C3 command point. The figure which shows the influence area when the C4 command point is designated as the target command point. The chart which shows the speed profile of C4 command point. A chart in which the speed profiles of C3 command point and C4 command point are superimposed. The conceptual diagram of the speed profile data which integrated each speed profile of C3 command point and C4 command point. The figure which set the deceleration area of the circular arc center command point to the whole circular arc. The chart which shows the velocity profile of a circular arc center command point. The movement locus in which the error of the next command point becomes large as a result of deceleration to suppress the error of the command point of interest and its enlarged view. The flowchart of an error check process. The figure which added the virtual command point to the path | route comprised by block B1 and B2. The figure which shows the speed profile produced | generated by the deceleration calculation process. The figure which shows the speed profile produced | generated by the error confirmation process. The figure which summarized two speed profile data. The figure which shows the speed profile data after a substitution process. The figure which shows the program path | route used for simulation. The figure which shows the speed profile which is a simulation result. The figure which shows the moving speed (after filter process) which is a simulation result. The figure which shows the movement path | route which is a simulation result. The figure which shows the error which is a simulation result. The figure which shows the program path | route in which a micro block continues after a long block. The flowchart of a speed division number adjustment process. The figure which shows the speed profile before and behind speed division number adjustment processing.

  Embodiments of the present invention will be described below. FIG. 1 shows an example of the configuration of a machine tool targeted by the present invention. A numerical control device 30 shown in FIG. 1 controls cutting of a workpiece (not shown) held on the upper surface of a table (not shown) by controlling the axial movement of the machine tool 1. The left-right direction, the front-rear direction, and the vertical direction of the machine tool 1 are an X-axis direction, a Y-axis direction, and a Z-axis direction, respectively.

  The configuration of the machine tool 1 will be described with reference to FIG. The machine tool 1 is a vertical machine tool that performs cutting by moving a tool mounted on a spindle in the X-axis direction, the Y-axis direction, and the Z-axis direction with respect to a work material held on a table upper surface, for example. The machine tool 1 includes a spindle mechanism, a spindle movement mechanism, a tool changer, and the like (not shown). The spindle mechanism includes a spindle motor 52, and rotates the spindle on which a tool is mounted. The spindle moving mechanism includes a Z-axis motor 51, an X-axis motor 53, and a Y-axis motor 54, and moves the spindle relative to the work material supported on the upper surface of a table (not shown) in each of the XYZ axial directions. . The tool changer includes a magazine motor 55, drives a tool magazine (not shown) that stores a plurality of tools, and exchanges the tool mounted on the spindle with another tool. The machine tool 1 further includes an operation panel 10. The operation panel 10 includes an input unit 24 and a display unit 25. The input unit 24 is a device for performing various inputs, instructions, settings, and the like. The display unit 25 is a device that displays various screens. The operation panel 10 is connected to the input / output unit 35 of the numerical controller 30. The Z-axis motor 51 includes an encoder 51B. The spindle motor 52 includes an encoder 52B. The X-axis motor 53 includes an encoder 53B. The Y-axis motor 54 includes an encoder 54B. The magazine motor 55 includes an encoder 55B. The encoders 51B to 55B are connected to drive circuits 51A to 55A, which will be described later, of the numerical control device 30, respectively.

  The electrical configuration of the numerical control device 30 will be described with reference to FIG. The numerical control device 30 includes a CPU 31, a ROM 32, a RAM 33, a storage device 34, an input / output unit 35, drive circuits 51A to 55A, and the like. The CPU 31 performs overall control of the numerical control device 30. The ROM 32 stores various programs such as a main program. The RAM 33 stores various data during execution of various processes. The storage device 34 is a nonvolatile memory. The input / output unit 35 is connected to the operation panel 10. The drive circuits 51A to 55A are servo amplifiers. The drive circuit 51A is connected to the Z-axis motor 51 and the encoder 51B. The drive circuit 52A is connected to the spindle motor 52 and the encoder 52B. The drive circuit 53A is connected to the X-axis motor 53 and the encoder 53B. The drive circuit 54A is connected to the Y-axis motor 54 and the encoder 54B. The drive circuit 55A is connected to the magazine motor 55 and the encoder 55B. The drive circuits 51A to 55A receive drive signals described later from the CPU 31, and output drive currents (pulses) to the corresponding motors 51 to 55, respectively. The drive circuits 51A to 55A receive feedback signals from the encoders 51B to 55B and perform position and speed feedback control. The drive circuits 51A to 55A may be, for example, FPGA circuits.

  With reference to FIG. 2, the moving average filter and time constant used for acceleration / deceleration processing will be described. The numerical controller 30 calculates a target position, a moving distance, a moving speed, a moving time, and the like for each of the X-axis, Y-axis, and Z-axis based on the interpolation command of the NC program. The interpolation command is a control command used when moving the axis at the moving speed designated by the address, and is, for example, a linear interpolation command, a circular interpolation command, or the like. The numerical controller 30 performs post-interpolation acceleration / deceleration processing. The post-interpolation acceleration / deceleration process is a process of smoothing a speed change by passing a moving average filter (FIR filter) at least twice through the calculated movement speed for each axis.

Figure 2 shows the results of the moving velocity for moving from x ₁ to x ₂ in the X-axis direction, and treated twice with moving average filter. The acceleration / deceleration time constant of the moving average filter (hereinafter referred to as the time constant) corresponds to the number of samples that the moving average filter averages. For example, when the sampling time is 1 msec and the time constant of the moving average filter is 10 msec, the moving average filter uses the average of up to ten commands including the current interpolation command as the current output. The time constant of the moving average filter of the first stage (FIR1) _{t 1,} the time constant of the moving average filter of the second stage (FIR2) and _{t 2.}

As a result of processing the moving speed with a two-stage moving average filter (FIR1, FIR2), the change in acceleration becomes less than a certain value, so the tool slowly increases the speed by t ₁ + t ₂ and reaches the maximum speed, and then , T ₁ + t ₂ , and slowly stop and stop. Therefore, the numerical control device 30 can absorb a sudden change in the moving speed by processing the moving speed of each axis with a plurality of moving average filters, thereby suppressing the vibration of the machine tool 1 and the maximum torque required for the operation. it can.

  With reference to FIG. 3 and FIG. 4, an inward error occurring at a corner in the movement path will be described. The path R1 shown in FIG. 3 is formed by two linear interpolation commands (G1 command), and the linear interpolation block B2 is connected at the command point P next to the linear interpolation block B1. Block B1 is linear interpolation in the X-axis direction, and block B2 is linear interpolation in the Y-axis direction. The inter-block angle β is 90 °. Therefore, the command point P forms a corner. FIG. 4 is a diagram in which the moving speed (speed 1) of the block B1 and the moving speed (speed 2) of the block B2 when moving along the route R1 in FIG. 3 are processed by a two-stage moving average filter (FIR1, 2). It is.

When speed 1 begins to decelerate from the maximum speed, speed 2 begins to accelerate. In the portion where the speed 1 and the speed 2 overlap, the tool moves inward around the command point P that is a corner of the path R1 (see the path R2). When speed 1 starts decelerating from the maximum speed, the tool is located at A1. A2 is the point where the command point P and the route R2 are closest. The distance from the command point P to A2 is an inward error (hereinafter referred to as an error). When speed 1 decelerates from maximum speed to 0, the tool is located at A3. A2 is a position where the error is maximum with respect to the route R1. A3 is a position where the error is eliminated. Time to A1~A3 until the error is eliminated occurs when corresponds to the sum of the constants _{t 1} and _{t 2.} As an example of a method for suppressing an error occurring in a corner to a target error, there is a method of decelerating before the corner. By decelerating before the corner, the overlap time of the speed 1 and the speed 2 can be reduced, so that the error can be suppressed to the target error. However, as will be described later, the method of decelerating before the corner causes a problem in a path in which minute blocks are continuous.

  With reference to FIG. 5 and FIG. 6, the outline of the present invention and an inner loop error that occurs in a path in which minute blocks are continuous will be described. A route R3 shown in FIG. 5 (1) is a program route based on the NC program. The path R3 includes a plurality of micro blocks (G1 blocks) B1 to B6 having a short block length. Block and block seams are command points. Micro block B1 is between command points C0 and C1, micro block B2 is between command points C1 and C2, micro block B3 is between command points C2 and C3, micro block B4 is between command points C3 and C4, micro block B5 is between command points C4 and C5, and micro block B6 is between command points C5 and C6.

  FIG. 6 (1) is a speed profile obtained by processing the moving speed when moving along the route R3 in FIG. 5 (1) with a two-stage moving average filter. The speed profile is a set distance and speed (cutting feed speed) of each block. The moving speed after the filtering process gradually increases to reach the maximum speed, and the micro blocks B1 to B6 move at the maximum speed, and then slowly decrease and stop. A route R4 shown in FIG. 5 (1) is a route after filtering that has moved at the moving speed shown in FIG. 6 (1). In the portion where the micro blocks B1 to B6 are continuous, the axis is greatly moved inward. For example, since there is the next command point C2 immediately after the command point C1, the next command point C2 is passed without the error occurring at the command point C1 being eliminated. Then, since the command points C3, C4,... Sequentially pass, the route R4 moves largely inward with respect to the route R3. The separation distance between the command points C0 to C6 and the route R4 corresponds to an error. In order to suppress such an error to the target error, as described above, when the method of decelerating at the joint part of the block is adopted, in the path where the minute block continues, the block joint part is decelerated and the next block is accelerated. Will be repeated frequently. If frequent acceleration / deceleration is performed, machine vibration may occur and the quality of the machined surface may deteriorate. Further, since the cutting traces (tool marks and cutter marks) on the processed surface change depending on the feed speed, if acceleration / deceleration is repeated, streaks due to the cutting traces are generated and the quality of the processed surface is deteriorated.

  In the present embodiment, a fully automatic deceleration process (see FIG. 9), which will be described later, is executed in order to suppress an error occurring in a path along which micro blocks are continuous to a predetermined target error or less. The fully automatic deceleration process is a process for obtaining an affected section that affects an error for each command point, calculating a deceleration rate of the affected section, and determining an optimum speed profile. The influence section is a section before and after the command point, and is a small section that moves in half the time constant in the present embodiment.

  The speed profile shown in FIG. 6 (2) is obtained by performing a fully automatic deceleration process on the speed profile shown in FIG. 6 (1). The movement speed after the fully automatic deceleration process is different from the movement speed obtained by only the filter process, and each of the command points C0 to C6 is changed during the movement of the minute blocks B1 to B6 after increasing the speed to reach the maximum speed. Slow down slowly according to the affected section. That is, in the portion where the micro blocks B1 to B6 are continuous, the moving speed does not frequently repeat deceleration and acceleration. After passing through block B6, it accelerates to the maximum speed again, and then decelerates and stops. A route R5 illustrated in FIG. 5B is a movement route after the fully automatic deceleration process. The route R5 has a smaller error from the route R3 than the route R4. Therefore, this embodiment can suppress an error that occurs in a path in which minute blocks are continuous.

  The main process will be described with reference to FIG. When the operator designates an NC program on the operation panel 10 and inputs an execution operation, the CPU 31 calls the main program from the ROM 32 and executes this processing. The CPU 31 calls and interprets the designated NC program from the storage device 34 (S1). The CPU 31 generates a plurality of movement commands based on the interpretation of the NC program (S2). The movement command specifies the position (coordinate value) for moving the axis. An example of the movement command is a linear interpolation command and a circular interpolation command. The position specified by the movement command is the command point. There is a block between the command points. The path connecting each command point is a program path.

  The CPU 31 creates block information 20 based on the generated plurality of movement commands (S3). As shown in FIG. 8, the block information 20 includes a block number and axis movement information. The block number is the order of blocks constituting the program path. The axis movement information is information related to the axis movement in each block, and includes various information such as a movement vector, a block length, a movement speed, and a movement time. The block information 20 is stored in the RAM 33. The movement vector is a direction vector of the block.

  The CPU 31 determines whether or not the fully automatic deceleration mode is on (S4). The fully automatic deceleration mode is a mode in which the movement path error is automatically adjusted to be equal to or less than a preset target error by decelerating in the affected section of each command point. The fully automatic deceleration mode can be set, for example, with the input unit 24 of the operation panel 10. When the fully automatic deceleration mode is on (S4: YES), the CPU 31 executes a fully automatic deceleration process (see FIG. 9) (S5). After the fully automatic deceleration process is executed, the CPU 31 ends this process. When the fully automatic deceleration mode is off (S4: NO), the CPU 31 executes normal processing (S6). In the normal process, the moving speed is filtered based on the block information 20 stored in the RAM 33, and the acceleration / deceleration control is performed for the axis movement. After executing the normal process, the CPU 31 ends this process.

  The fully automatic deceleration process will be described with reference to FIGS. In the present embodiment, it is necessary to determine a speed profile before executing a movement command for a certain block. In order to determine the speed profile, it is necessary to execute a later-described deceleration calculation process (S11) and an error confirmation process (S12) of the fully automatic deceleration process shown in FIG. Therefore, the CPU 31 needs to prefetch the block information 20 when executing the block movement command. Prefetching is to read in advance a block to be executed in the future. In the deceleration calculation process, each command point has a past influence block before the command point and a future influence block after the command point so that the path error at each command point is within the target range. The speed is changed. That is, the speed profile is updated. The influence block is a block within the range that affects the error of each command point. As will be described later, the influence block exists in the range of half the time constant before and after the command point, respectively. It is a block. In the error confirmation process, it is confirmed that the path error at each command point is within the target range for the speed profile updated by the deceleration calculation process. If the path error is not within the target range, the speed is checked again. Update the profile. In the deceleration calculation process, as described above, the speed of each command point and the block within the range of influence before and after the command point are updated. Therefore, the error check process is a command point that performs the deceleration calculation process to prevent mutual influence. This must be done for blocks outside the range of effects. That is, in the deceleration calculation process and the error check process, the speeds of the past influence block and the future influence block are changed. Therefore, when executing the block movement command, at least after the time constant doubled. It is necessary to read ahead for the blocks to be processed up to.

  Constants S and G are used for prefetching. S is the number of prefetched blocks, and G is the number of affected blocks. Since the number of influence blocks is determined by the sum of the time constants, the worst condition is the situation where the sum of the time constants is the largest in a system having a variable time constant. As the worst condition, for example, a case where the total time constant is 6000 msec (maximum set value) is considered. In addition, when the processing cycle (sample time) of the CPU 31 is 1 msec, the shortest block that can be processed by the CPU 31 is a block that completes moving in 1 msec. When all the blocks are blocks that have been moved in a sample time of 1 msec, G takes the maximum value. When the time constant is 6000 msec which is the maximum setting value, G is 6000/2 = 3000 blocks. As described above, since the number of blocks required for prefetching is the number of blocks that are processed before the time constant is doubled, the CPU 31 needs to prefetch 4 × 3000 = 12000 blocks. As shown in FIG. 8, when k is an error check block number, the deceleration calculation block number is k + 2G, the prefetch block number is k + 3G, and the execution block number is k−G. The deceleration calculation block number is a block number for executing a deceleration calculation process. The error check block number is a block number for executing error check processing. The execution block number is a block number to be executed.

  The process flow from the initial state will be described with reference to FIG. The CPU 31 reads blocks B (1) to B (3G) from the block information 20 (see FIG. 8) stored in the RAM 33 (S7). As described above, G in this embodiment is 3000. Therefore, the CPU 31 reads block B (1) to block B (9000). The CPU 31 executes the deceleration calculation process for the read blocks B (1) to B (2G) (S8). The CPU 31 sets 1 to the error check block number k (S9).

  The deceleration calculation process will be described with reference to FIG. Since k + 2G = 6001, the CPU 31 may pay attention to the block B (6001). The CPU 31 determines whether the movement command of the block B (6001) of interest is a linear interpolation command (G1 command) or a circular interpolation command (G2 command) (S22). When the command is a linear interpolation command (S22: linear interpolation), the CPU 31 sets the end point of the block B (6001) as a target command point (S25). If the command is an arc interpolation command (S22: arc interpolation), the CPU 31 executes an arc center command point calculation process (S23).

  The arc center command point calculation process will be described with reference to FIGS. In the fully automatic deceleration process, an error at the target command point is calculated, and a moving speed that makes the error equal to or less than the target error is obtained. However, for example, as shown in FIG. 11, in the circular interpolation command for interpolating between the points A and B with a circular arc, only the error at the command point at the arc start point and the error at the command point at the arc end point is suppressed. The error may exceed the target error in the middle of the arc. In this embodiment, a virtual command point (hereinafter referred to as an arc center command point) is created in the arc center portion, and errors due to the curvature of the arc can be suppressed by suppressing errors in the arc center command points. . The arc center command point is a point at which the movement distance when the command point of the arc end point is circularly interpolated from the command point of the arc start point is halved.

The affected section calculation process will be described with reference to FIGS. In the program path shown in FIG. 13, the command point of interest is the i-th command point pos _i . The meanings of symbols used in this calculation process are as follows.
L _i : i-th block length (mm)
F _i : i-th block speed (mm / min)
Nc _i : Number of command interpolation points (real number) of i-th block (number)
・ N _block : Maximum number of prefetched commands (pieces)
Dt: Sample time (msec)
・ T: Total time constant (msec)
Posx _i , posy _i , posz _i : command point of i-th and i + 1-th joint posx _is , posy _is , posz _is : the earliest command point that affects the error (mm)
_{· Posx ie, posy ie, posz} ie: the slowest command point impact on the error (mm)
• i _s : earliest command number that affects error • i _e : latest command number that affects error • posx _sti , posy _sti , posz _sti : influence start point for i-th command point (mm)
Posx _eni , posy _eni , posz _eni : Influence end point for the i-th command point (mm)
Note that the sample time dt corresponds to a processing cycle.

As shown in FIG. 14, the CPU 31 obtains a command interpolation point by a real number from the moving distance and the moving speed for each block (S51). The CPU 31 calculates, for example, the command interpolation point (real number) of the i-th block by the following formula.
Nc _i = (L _i × 60 × 1000 / F _i ) / dt

  Next, the CPU 31 obtains the command number of the earliest command point that affects the target command point (S52). The earliest command point that affects the target command point is the command point with the lowest command number among the command points existing in the range from the target command point to the time point that is half the time constant total. is there.

  Next, the CPU 31 obtains the command number of the slowest command point that affects the target command point (S53). The slowest command point that affects the target command point is the command point with the smallest command number among the command points that exist outside the range from the target command point to the point when half the time constant has elapsed. is there.

The CPU 31 obtains the influence start point (S54). The influence start point is the start position of the section that affects the error of the target block. As shown in FIG. 15, i-th impact starting point _{pos sti} of command point pos _i, before than the command point pos _i, and the distance the distance from the command point pos _i corresponds to half the total time constant It becomes the position.

The CPU 31 obtains the influence end point (S55). The influence end point is the end position of the section that affects the error of the target block. As shown in FIG. 16, i-th impact end point _{pos eni} the command point pos _{i is} the distance later than the command point pos _i, the and the distance from the command point pos _i corresponds to half the total time constant It becomes the position. CPU31 complete | finishes this process and advances a process to S27 of FIG. The CPU 31 executes a command interpolation point calculation process (S27). The command interpolation point calculation process is a process for obtaining a command interpolation point between the influence start point and the influence end point. The command interpolation point is obtained by dividing each block for each sample time, and can be calculated from the moving speed of each block and the sample time.

CPU31 performs a filter process (S28). In the filter process, the motion interpolation point is obtained by processing the command interpolation point calculated in the process of S27 with a moving average filter. Two or more moving average filters are generally used. For example, filter processing is performed in three stages using three moving average filters having different cutting feed time constants. The meanings of symbols used in this calculation process are as follows.
• px _i , py _i , pz _i : Command interpolation point (mm)
Prx _i , pry _i , prz _i : motion interpolation points (mm)
T ₁ : Time constant 1 (msec)
T ₂ : Time constant 2 (msec)
T ₃ : Time constant 3 (msec)
Dt: Sample time (msec)
Each time constant 1 to 3 is an integral multiple of the sample time.

The calculation formula of the moving average filter of the i-th interpolation point is as follows.

Note that j = 1 to 3. The CPU 31 adds the same value as the start point and the end point to the command interpolation point by (t _j / dt) before and after the start point of the command interpolation point in order to make the start point and end point of the command interpolation point and the motion interpolation point coincide. calculate. The CPU 31 executes error calculation processing (S29).

The error calculation process will be described with reference to FIGS. As shown in FIG. 18, the error is the one that minimizes the distance between the command point and the motion interpolation point, and is obtained as a single solution. The meanings of symbols used in this calculation process are as follows.
_{· Posx i, posy i, posz} i: i -th block and the (i + 1) -th command point _· prx _i seam _block, pry _{i, prz} i: operating interpolation point (mm)
I _sp : search point ε ₁ : distance (mm) between the motion interpolation point immediately before the search point and the command point
・ Ε ₂ : Distance between search point and command point (mm)
· Epsilon _3: distance command point and operating interpolation point after one of the search points (mm)
・ Ε: Error (mm)
-Vector p: Vector in which the motion interpolation point with the second smallest error is seen from the motion interpolation point with the smallest error-Vector ε ₂ : Vector in which the command point is seen from the motion interpolation point with the smallest error

As shown in FIG. 17, the CPU 31 uses the motion interpolation point as a search point (see FIG. 18), and calculates the error between the search point and the target command point using the following [Equation 2] (S71). Note that the initial search point is an operation interpolation point corresponding to the command interpolation point closest to the target command point.

The CPU 31 calculates an error between the motion interpolation point before and after the search point and the target command point (S72). The CPU 31 calculates the error between the motion interpolation point immediately before the search point and the target command point by the following [Equation 3].

The CPU 31 calculates the error between the motion interpolation point immediately after the search point and the target command point by the following [Equation 4].

CPU31 compares the operating interpolation points before and after the search point and search point (S73), the error epsilon ₂ is minimal or determine the search point (S74). When the error ε _{2 of the} search point is the minimum (S74: YES), that is, when ε ₂ ≦ ε ₁ and ε ₂ ≦ ε ₃ , the process proceeds to S78 described later. If the error search point epsilon ₂ is not the minimum (S74: NO), the error epsilon ₁ of operating interpolation point before a search point to minimize or determined (S75). When the error ε ₁ is the smallest (S75: YES), that is, when ε ₁ <ε ₂ , the CPU 31 moves the search point to the previous position by setting i _sp = i _sp −1 (S76). ) And return to S71 to repeat the process. When the error ε ₃ of the motion interpolation point immediately after the search point is the smallest (S75: NO), that is, when ε ₃ <ε ₂ , the CPU 31 searches by setting i _sp = i _sp +1. The point is moved back by one (S77), and the process returns to S71 to repeat the process.

When the error ε _{2 of the} search point is minimized (S74: YES), the CPU 31 searches for a point having the second smallest error among the motion interpolation points before and after the search point, as shown in FIGS. The line segment connecting the points and the line segment connecting the search points and the command points are calculated as follows (S78).
・ When ε ₁ ≦ ε ₃

・ When ε ₃ <ε ₁

The CPU 31 determines whether or not the angle formed by the calculated vector is an acute angle (the inner product is positive) (S79). As shown in FIG. 20, when the angle formed by the vector is an acute angle (S79: YES), as shown in FIG. 21, the CPU 31 determines the closest point from the command point based on the three-square theorem as follows: Calculation is performed using Equation 7] (S80). The CPU 31 calculates an error from the command point and the closest approach point obtained by calculation (S81). By performing interpolation in this way, the closest point can be obtained with higher accuracy even when the angle formed by the vector is an acute angle.

As shown in FIG. 22, when the angle formed by the vector is an obtuse angle (S79: NO), the perpendicular foot (the closest point) to the straight line connecting the search point and the motion interpolation point with the second smallest error is extrapolated Since it becomes a point, the error may be far from the ideal value. Thus CPU31 determines the error epsilon ₂ obtained in S74 to the error epsilon (S82). The error ε is stored in the RAM 33. CPU31 performs an error suppression process (S30).

The error suppression processing will be described with reference to FIGS. This process assumes that the error and speed are in a linear relationship, calculates the deceleration rate by interpolation processing, and adjusts the moving speed according to the deceleration rate, thereby matching the error with the target error. It is. This embodiment will be described as the k-th error suppression process when the target command point is the end point of the i-th block. The meanings of symbols used in this processing are as follows.
・ Ε: Error (mm)
・ Ε _k : Error of the target command point recalculated k times (mm)
Ε _r : target error (mm)
Ε _t : Error convergence tolerance (mm)
・_Rk : Deceleration rate recalculated k times (-)
_Fik : Speed recalculated k times (mm / min)
I _s : earliest command that affects error (-)
I _e : Slowest command (−) that affects error
N _block : Maximum number of read-ahead blocks (-)

As shown in FIG. 23, CPU 31, the error epsilon obtained by the error calculation process, determines whether less than the target error ε _r (S91). If the error epsilon is equal to or lower than the target error _{ε r (S91: YES),} CPU31 terminates the present process, the process proceeds to S31 which will be described later in FIG. 10. When the error ε is larger than the target error ε _r (S91: NO), the CPU 31 calculates the deceleration rate R _k using the following formula (S92).
· _{R k} = ε _r / ε

CPU31 calculates the speed of an influence area with the following numerical formula (S93).
・ F _ik = F _i × R _k
However, i _s ≦ i ≦ i _e .

The CPU 31 recalculates the affected section at the determined moving speed (S94). The calculation method is the same as the calculation method in the influence interval calculation process shown in FIG. The CPU 31 recalculates the command interpolation point (S95). The CPU 31 executes the command interpolation point calculation process again. The CPU 31 recalculates the motion interpolation point using the obtained command interpolation point (S96). The calculation method is the same as the calculation method in the filter processing of S28 of FIG. The CPU 31 recalculates the error ε _k by using the obtained motion interpolation point and command point (S97). The calculation method is the same as the calculation method in the error calculation process of FIG.

The CPU 31 determines whether or not the error obtained by the calculation satisfies the error condition (S98). The error condition of this embodiment is a condition in which the difference obtained by subtracting the error from the target error is 0 or more and satisfies the error convergence tolerance (0 ≦ ε _r −ε _k ≦ ε _t ). When the error satisfies the error condition (S98: YES), the CPU 31 determines the deceleration rate, the influence start point, and the influence end point to the values obtained in S92 and S94, respectively (S99). The CPU 31 returns to the flowchart of FIG. 10 and advances the process to S31 described later.

If the error does not satisfy the error condition (S98: NO), for example, as shown in FIG. 24, since the error epsilon _b is larger than the target error epsilon _r, so that the target error epsilon _r, CPU 31 changes the speed reduction ratio There is a need to. Although the error epsilon _a is smaller than the target error epsilon _r, the difference obtained by subtracting the error epsilon _a from the target error epsilon _r is greater than the error convergence tolerance epsilon _t. In this case, since the error from the path is sufficiently small, the machining quality can be ensured, but since the deceleration rate becomes small, the speed becomes slow and the machining cycle is extended. Thus, as the error epsilon _a is the error convergence within tolerance epsilon _t, CPU 31 is required to change the deceleration rate. The CPU 31 changes the deceleration rate so that the error ε _k becomes the target error ε _r by, for example, linear interpolation (S100). In the present embodiment, it is assumed that the relationship between the error and the deceleration rate is linear. For example, when the error in the deceleration rate R _b is ε _b and the error when the deceleration rate R _b obtained by performing the deceleration process again is ε _a , the CPU 31 sets the deceleration rate R _{k + 1} as follows: Calculate with [Equation 8]. In general, the relationship between the error and the deceleration rate is not linear, but the deceleration rate that satisfies the error convergence tolerance can be obtained by repeating a plurality of trials. In the present embodiment, when the deceleration rate is obtained by linear interpolation, the target error is a value obtained by subtracting half of the error convergence allowable value ε _t from the target error ε _r . This is for efficiently obtaining a deceleration rate that satisfies the error convergence tolerance.

However, when ε _k <ε _r and ε _a <ε _k , ε _a = ε _k and R _a = R _k . On the other hand, when ε _k > ε _r and ε _b > ε _k , ε _b = ε _ik and R _b = R _k . The CPU 31 returns to S93 and repeats the above processing to recalculate the error. The initial value of epsilon _a is 0, the initial value of epsilon _b is epsilon, the initial value of _{R a} is 0, the initial value of _{R b} is 1.

  When the error calculated by the recalculation satisfies the error condition (S98: YES), the CPU 31 determines the deceleration rate, the influence start point, and the influence end point (S99). Since the error is equal to or less than the target error and within the range of the allowable error value, this embodiment can complete the machining in a short machining cycle while improving the machining quality. The CPU 31 executes a speed profile update process (S31).

  The speed profile update process will be described with reference to FIGS. In the present embodiment, a case where the speed profile of the program path shown in FIG. 26 is created will be described as an example. As shown in FIG. 25, for example, when the command point C3 is set as the target command point, the CPU 31 determines whether the command point C3 is an arc center command point and the moving time of the arc is shorter than the total time constant (S101). ). Since the command point C3 is the block end point (S101: NO), the CPU 131 sets the affected zone as the deceleration zone (S102). The deceleration zone is a zone where the vehicle is decelerated at the deceleration rate determined by the error suppression process of FIG. The CPU 31 determines the speed of the target block based on the influence start point, the influence end point, and the deceleration rate for the command point C3 determined in the error suppression process (S104), and generates a speed profile (S105).

FIG. 27 shows a speed profile with the command point C3 as the target command point. The influence start point of the command point C3 is located in the middle of the block length L2 between C1 and C2. The influence end point of the command point C3 is located in the middle of the block length L5 of C4 to C5. The deceleration section is an influence section from the influence start point to the influence end point. The moving speed of the block corresponding to the block length L1~L6 and F, if the deceleration rate determined by the error suppression process is R _3, the moving speed of the impact start point to effect end point, the F × R ₃ . The generated speed profile is stored in the RAM 33.

When the speed profile of the command point C3 is generated, the CPU 31 updates the speed profile by integrating the generated speed profile and the speed profile previously generated and stored in the RAM 33 (S106). When the update of the speed profile at the command point C3 is completed, as shown in FIG. 28, the CPU 31 sets the next command point C4 as the target command point, and creates a speed profile in the same manner as the command point C3. FIG. 29 shows a speed profile with the command point C4 as the target command point. When the command point C4 is set as the target command point, the influence start point of the command point C4 is located in the middle of the block length L3 between C2 and C3. The influence end point of the command point C4 is located in the middle of the block length L6 of C5 to C6. The deceleration section is an influence section from the influence start point to the influence end point. Thus the speed of the speed block corresponding to the block length L1~L6 and F, if the deceleration rate determined by the error suppression process is R _4, the moving speed of the impact start point to effect end point, the F × R ₄ Become. The generated speed profile is stored in the RAM 33.

FIG. 30 shows a superposition of the speed profile at the command point C3 and the speed profile at the command point C4. According to these two speed profiles, there are overlapping sections with different moving speeds. In the overlapping section, the CPU 31 selects the one with the lower moving speed (see the dark gray portion in FIG. 30). Therefore, the moving speed is decelerated from F to F × R ₃ in the middle of the block length L2, and further decelerated from F × R ₃ to F × R ₄ in the middle of the block length L3. Thereafter, the moving speed is accelerated from F × R ₄ to F in the middle of the block length L6. In this way, the speed profiles are integrated.

  FIG. 31 is speed profile data of the integration result of FIG. The speed profile data includes commands, block lengths, speeds, speed division numbers, distances, deceleration rates, and the like, and these various data are stored in association with each other. The program paths shown in FIGS. 26 and 28 are configured by blocks B1 to B6 that are commanded by movement commands 1 to 6. Block length L1 for move command 1, block length L2 for move command 2, block length L3 for move command 3, block length L4 for move command 4, block length L5 for move command 5, The block length L6 is stored for the movement command 6. The movement speed is a speed designated by the movement command, and the movement speeds 1 to 6 are all the same movement speed F.

  The speed division number is a number for dividing the movement speed in each of the blocks B1 to B6. As shown in FIG. 30, the moving speed is divided at block lengths L2, L3, and L6. Therefore, the speed division number is 1 in the block lengths L2, L3, and L6, and is 0 because the other divisions are not performed. The distance is the distance from the start point of the block to the speed change point as seen. For example, for the block length L2, the distances p2 and L2 are stored. The deceleration rate is the rate at which the speed is reduced. For example, in the block length L2, the deceleration rate = 1 for the distance p2, and the deceleration rate = R3 is stored for the distance L2. The CPU 31 stores the speed profile data in the RAM 33, returns to the flowchart of FIG. 10, and advances the process to S32.

  The velocity profile of the arc center command point will be described with reference to FIGS. The program path shown in FIG. 32 is a path in which a circular interpolation command (G2 command) is sandwiched between two linear interpolation commands (G1 command). In the path of circular interpolation, there is a problem that an error occurs due to the curvature of the circular arc. Therefore, the CPU 31 adds the arc center command point to the arc center as described above. As shown in FIG. 25, when the command point of interest is an arc center command point and the movement time of the arc is equal to or greater than the total of the time constants (S101: YES), the CPU 31 sets the entire arc as a deceleration zone (S103). ). If the moving time of the arc is shorter than the sum of the time constants and the affected section exceeds the entire arc, the CPU 31 may set the affected section as a deceleration section as a rule.

  FIG. 33 is a velocity profile with the circular arc center command point shown in FIG. 32 as the target command point. When the movement speed of the block corresponding to the G1 and G2 commands is F and the deceleration rate determined by the error suppression process is R, the movement speed of the block corresponding to the G2 command for the entire arc is F × R. In the circular interpolation, in addition to the path error that occurs at the command point at the end of the circular interpolation block, an error due to the curvature of the circular arc occurs in the entire area of the circular arc. That is, the path error caused by the curvature is the same at any arc position that is more than half of the total time constant from the command point at the end of the circular interpolation block. In this embodiment, for example, an arc center command point is added as a representative point. Then, deceleration is performed over the entire arc so that the position error becomes the target error. Therefore, this embodiment can consider not only the error of the command point but also the error caused by the curvature of the arc part of the arc interpolation.

  After updating the speed profile (S106 in FIG. 25), the CPU 31 determines whether or not the command point of interest is an arc center command point (S32). When the target command point is the arc center command point (S32: YES), the CPU 31 sets the end point of the circular interpolation block as the target command point (S25) and repeats the above processing. When the command point of interest is not the arc center command point but the block end point (S32: NO), the CPU 31 sets the error check block number k = 1 (S9) and reads the block B (k + 3G) (S10). CPU31 performs the deceleration calculation process of block B (k + 2G) (S11). The CPU 31 pays attention to the block B (k) and executes an error confirmation process (S12).

  The error check process will be described with reference to FIGS. The deceleration calculation process shown in FIG. 10 is an algorithm that decelerates in order to reduce the error of the target command point, and does not consider errors other than the target command point. FIG. 34 (1) is a chart comparing the program path, the one obtained by performing deceleration calculation processing on each command point of the program path, and the one not subjected to deceleration calculation processing. FIG. 34 (2) is an enlarged view of the W1 region shown in FIG. 34 (1). As a result of executing the deceleration calculation process to suppress the error at the target command point, the error at other command points may exceed the target error. The error p in the figure is an example in which the error becomes larger than when the deceleration process is not performed because the deceleration calculation process is executed to reduce the error in the W2 region. This is based on the assumption that, in this algorithm, the error decreases when deceleration is performed, but the program path and motion interpolation point are changed by changing the movement direction (acceleration direction) as shown in FIG. This is because, in a trajectory where the paths connecting the two crosses, the error may increase by decelerating. In order to solve this problem, the CPU 31 executes an error check process. In the error check process, the same process as the deceleration calculation process is performed again on the block that has been subjected to the deceleration calculation process.

In the example shown in FIG. 8, the error check block number is k = 3001. As shown in FIG. 35, the CPU 31 executes a process similar to the deceleration calculation process of FIG. 10 for the focused block (S111). However, the block after the speed profile is generated is obtained by dividing the block before the deceleration calculation process by the number of speed divisions, and the movement speed changes in one block. Therefore, the CPU 31 cannot use the above mathematical formula used in the deceleration calculation process as it is. In the present embodiment, the block after the speed profile is generated is a virtual block that is divided for each joint where the moving speed changes, and the point between the virtual block and the joint of the virtual block is a virtual command point. The CPU 31 can use the above formula by setting the virtual block and the virtual command point as the block and the command point.

As shown in FIG. 36 (1), for example, a substantially L-shaped moving path composed of two blocks B1 and B2 is assumed. Seam blocks B1 and B2 is a command point _{C n.} The block length of the block B1 is L1, and the block length of the block B2 is L2. As a result of the deceleration calculation process (see FIG. 10) for the command point Cn, in the speed profile shown in FIG. 36 (2), the block B1 is divided into two blocks B1-1 and B1-2, and the block B2 is divided into the block B2 -1 and B2-2. The moving speed is decelerated at the distance p11 at the block length L1, and then accelerated at the distance p21 at the block length L2.

  As shown in FIG. 36 (3), the CPU 31 sets two joints whose speeds change in the block lengths L1 and L2 as virtual command points. Blocks B1-1, B1-2, B2-1, and B2-2 are virtual blocks. By processing the generated virtual block and virtual command point as a block and command point, the CPU 31 can use the above formula. Since the virtual command point is not the command point at which the error is to be suppressed, the CPU 31 executes the error calculation process (see S29 in FIG. 10) during the deceleration calculation process performed in the error check process in FIG. It does not have to be. Therefore, the CPU 31 executes the deceleration calculation process (S111). As a result, even if the error at the other command point becomes larger than the target error, the error check process is performed to obtain another command point that has become larger. The error can be suppressed and the speed profile can be updated again. The CPU 31 executes a replacement process (S112) and returns to the flowchart of FIG.

  The contents of the replacement process will be described with reference to FIGS. As described above, the speed profile is updated by the error confirmation process for each virtual block. However, the speed profile data is output to the drive circuits 51A, 53A, 54A (see FIG. 1) for controlling the axis movement for each block, not for each virtual block. Therefore, the CPU 31 needs to replace the speed profile data calculated in the virtual block with the speed division number, distance, and deceleration rate of the block. The CPU 31 executes replacement processing to convert it into speed profile data for each block.

The speed division number, distance, and deceleration rate of the block used in this process are defined as follows.
Number of speed divisions of block: Total value of (number of speed divisions for each virtual block at the time of error check process + 1) −1
Block distance: Distance from the start point of the block to the speed change point seen from the block Note that the speed change point includes the end point of the block.
Block deceleration rate: product of the deceleration rate obtained by the deceleration calculation process and the deceleration rate obtained by the error check process

  FIG. 37 is an example of a speed profile obtained by the deceleration calculation process. The speed profile is for a path composed of four blocks B1 to B4. The block length of the block B1 is L1, the block length of the block B2 is L2, the block length of the block B3 is L3, and the block length of the block B4 is L4. R3 and R4 are deceleration rates. The moving speed varies with the block lengths L2, L3, and L4. In the middle of the block length L2, the speed is reduced from the speed F to the speed F × R3. In the middle of the block length L3, the speed is further reduced from the speed F × R3 to the speed F × R4. In the middle of the block length L4, the speed is accelerated from the speed F × R4 to the speed F. The block B2 is divided into two blocks B2-1 and B2-2. The block B3 is also divided into two blocks B3-1 and B3-2. The block B4 is also divided into two blocks B4-1 and B4-2. Block B1 is not divided.

  FIG. 38 shows an example of the speed profile obtained by the error check process, and the same process as the deceleration calculation process is performed again on the block after the speed profile generation of FIG. 37 as described above. Even in this speed profile, the speed changes with the block lengths L2, L3, and L4, respectively, but the movement speed further changes in stages in the blocks B3-1 and B3-2. R3 to R7 are deceleration rates. In block B3-1, the speed is reduced from F * R3 to F * R3 * R5 (= F * R4 * R6), and in block B3-2, the speed is accelerated from F * R3 * R5 to F * R4 * R7.

  FIG. 39 summarizes the data of each speed profile of FIG. 37 and FIG. In the speed profile data generated by the deceleration calculation process, the speed division number of block number 1 (B1) is 0, the distance is L1, and the deceleration rate is 1. The number of speed divisions of block number 2 (B2) is 1. Since the distances are the distances from the start point of the block B2 to the blocks B2-1 and B2-2, they are p2 and L2. The deceleration rate of block B2-1 is 1, and the deceleration rate of block B2-2 is R3. The number of speed divisions of block number 3 (B3) is 1. The distances are p3 and L3 because they are the distances from the start point of the block B3 to the blocks B3-1 and B3-2. The deceleration rate of block B3-1 is R3, and the deceleration rate of block B3-2 is R4. The number of speed divisions of block number 4 (B4) is 1. Since the distances are the respective distances from the start point of the block B4 to the blocks B4-1 and B4-2, they are p4 and L4. The deceleration rate of block B4-1 is R4, and the deceleration rate of block B4-2 is 1.

  In the speed profile data generated by the error check process, the speed division number of block number 1 (B1) is 0, the distance is L1, and the deceleration rate is 1. Of the block number 2 (B2), the speed division number of the block B2-1 is 0, the distance is p2, and the deceleration rate is 1. Since the speed division number of the block B2-2 is 0 and the distance is the distance from the virtual command point of the block B2, L2-p2 and the deceleration rate are 1. Of the block number 3 (B3), the speed division number of the block B3-1 is 1. Since the block B3-1 is further divided into two, the distance of the block B3-1 is the distance from the starting point of the block B3-1 to the two speed change points, and therefore, p31 and p3. The deceleration rate is 1 and R5. The number of speed divisions in the block B3-2 is 1. Since the block B3-2 is further divided into two, the distance of the block B3-2 is the distance from the starting point of the block B3-2 to the two speed change points, and therefore p32 and L3-p3. The deceleration rate is R6 and R7. Of the block number 4 (B4), the speed division number of the block B4-1 is 0, the distance is p4, and the deceleration rate is 1. Since the speed division number of the block B4-2 is 0 and the distance is the distance from the virtual command point of the block B4, L4-p4 and the deceleration rate is 1.

  The CPU 31 replaces the speed profile data generated by the error confirmation process with the speed division number, distance, and deceleration rate of the block. FIG. 40 shows speed profile data after the replacement process. The deceleration rate after the replacement process is the product of the deceleration rate obtained by the deceleration calculation process and the deceleration rate obtained by the error confirmation process. Of the block number 2 (B2), the distance of the block B2-2 is replaced with the distance L2 from the start point of the block B2 from the distance L2-p2. Of the block number 3 (B3), the distance of the block B3-2 is replaced with the distance p3 + p32 and the distance L3 from the starting point of the block B3 from the distance p32 and the distance L3-p3. Of the block number 4 (B4), the distance of the block B4-2 is replaced by the distance p4 and the distance L4 from the starting point of the block B4 from the distance L4-p4. In this way, the speed profile data for each virtual block is converted into speed profile data for each block. The replaced speed profile data is stored in the RAM 33.

  The CPU 31 returns to the flowchart of FIG. If k> G (S13: YES), the error check process for the first block has been completed, so block B (k-G) is executed (S14). When k ≦ G (S13: NO), the error check process for the first block is not completed, so the block is not executed and the process proceeds to the next process. In the example illustrated in FIG. 8, when the error check block number k = 3001, the execution block number is k−G = 3001−3000 = 1. The CPU 31 outputs the speed profile data stored in the RAM 33 to the drive circuits 51A, 53A, 54A for each axis, and executes the movement command for the execution block. The drive circuits 51A, 53A, and 54A execute the axis movement by performing acceleration / deceleration after interpolation based on the axis movement information and the speed profile data of the block information 20. The CPU 31 adds 1 to k (S15).

  The CPU 31 determines whether k + 3G is larger than the total number of blocks (S16). If k + 3G is equal to or less than the total number of blocks (S16: NO), the CPU 31 returns to S10, reads block B (k + 3G), and repeats the process. When k + 3G is larger than the total number of blocks (S16: YES), the CPU 31 determines whether k + 2G is larger than the total number of blocks (S17). When k + 2G is equal to or less than the total number of blocks (S17: NO), the CPU 31 returns to S11, executes the deceleration calculation process for block B (k + 2G), and repeats the process. When k + 2G is larger than the total number of blocks (S17: YES), the CPU 31 determines whether k is larger than the total number of blocks (S18). When k is equal to or less than the total number of blocks (S18: NO), the CPU 31 returns to S12, executes the error check process of the block B (k), and repeats the process. When k is larger than the total number of blocks (S18: YES), the CPU 31 determines whether k-G is larger than the total number of blocks (S19). If k-G is equal to or less than the total number of blocks (S19: NO), the CPU 31 returns to S13, and if k> G (S13: YES), executes block B (k-G) (S14) and repeats the process. When k-G is larger than the total number of blocks (S19: YES), since execution of all the prefetched blocks has been completed, the CPU 31 returns to the flowchart of FIG. 7 and ends the main process.

The simulation result of the fully automatic deceleration process will be described with reference to FIGS. In the present embodiment, a simulation when the fully automatic deceleration process is executed is performed on the program path shown in FIG. As shown in FIG. 41, the program path includes a G1 block that moves by linear interpolation in the X-axis direction, a G1 minute block that moves by curving in a large arc shape in the Y-axis direction from the end point of the G1 block, and a G1 minute The G1 block moves by linear interpolation in the Y-axis direction from the end point of the block. The simulation conditions are as follows.
・ Time constant of moving average filter: 25, 22, 13 (msec)
・ Program command speed: F5000 (mm / min)
・ Target error: 0.100 (mm)

  FIG. 42 (1) shows a speed profile when the fully automatic deceleration process is not performed, and FIG. 42 (2) shows a speed profile when the fully automatic deceleration process is executed. It can be seen that by executing the fully automatic deceleration process, the vehicle is smoothly decelerating around the curved portion connected by a plurality of minute blocks. Also, as shown in FIG. 43, it can be seen that even when the change in the moving speed is seen, the vehicle decelerates much more smoothly than the unprocessed state and accelerates smoothly again. Thus, by executing the fully automatic deceleration process, the speed change can be reduced in addition to the repetition of the deceleration and the acceleration. Thereby, the numerical control apparatus 30 of this embodiment can perform an axial movement stably, without applying a load to the machine tool 1. FIG.

  As shown in FIG. 44, with respect to the curve portion of the G1 minute block in the program path, the movement path after the fully automatic deceleration process is largely shifted toward the program path side compared to the unprocessed area that does not decelerate. From this, it was found that the present embodiment can greatly suppress the error. FIG. 45 shows the error at nine command points forming the curve portion of the program path. In the unprocessed movement path, a large error of about 0.2 to 0.3 mm was generated, but in the movement path after the fully automatic deceleration process, the maximum error value could be suppressed to the target error at each command point. Therefore, the numerical control device 30 of the present embodiment can automatically calculate the optimum moving speed by setting the target error as a parameter in advance. And it was proved that both the improvement of the machining accuracy and the short machining cycle can be achieved because the error at the corners of each command point does not become too large and the speed is not reduced too much.

  The number of speed divisions in one block of the movement path will be described with reference to FIGS. FIG. 46 is an example of a route in which a long block having a long moving distance is followed by a minute block having a short moving distance. The portion where the minute blocks are continuous is bent in an arc shape. If a minute block continues after a long block, there may be a large number of start points of the affected section of each command point of the minute block in the long block. In this case, the speed division number of the long block may increase. As the number of speed divisions in one block increases, the amount of speed profile data output to the drive circuits 51A, 53A, and 54A increases. Therefore, there is a possibility that the processing speed of the CPU 31 is not in time, and that a large amount of memory capacity for storing data must be secured. In the present embodiment, in order to solve this problem, the maximum speed division number of one block is set in advance using parameters. In the present embodiment, when the speed division number of one block exceeds the maximum speed division number, the speed division number adjustment process shown in FIG. 47 is executed to adjust the speed profile so that the speed division number becomes the maximum division number. May be.

  With reference to FIGS. 47 and 48, the speed division number adjustment processing will be described. For example, after executing the speed profile update process shown in FIG. 25, the CPU 31 may read the speed division number adjustment program from the ROM 32 and execute this process. The CPU 31 determines whether or not the speed division number of one block exceeds the maximum speed division number for the speed profile generated by the speed profile update process (S121). The maximum speed division number is stored in the storage device 34 in advance. When the number of speed divisions for one block does not exceed the maximum number of speed divisions (S121: NO), the CPU 31 ends this process.

  FIG. 48 (1) is an example of a velocity profile including a long block in the program path. The A1 section is a section that is affected by the previous block of the long block. Therefore, A1 section is divided | segmented into many blocks, and a moving speed changes repeatedly. The A3 section is a section that is affected by the subsequent block of the long block. Therefore, the A3 section is also divided into a large number of blocks, and the moving speed changes repeatedly. The A2 section, which is between the A1 section and the A3 section, is a section that is not affected by the preceding and following blocks. Therefore, the A2 section is not divided and the moving speed is constant. Since it is divided into a large number in the sections A1 and A3, the number of speed divisions as a whole is large.

  If the speed division number exceeds the maximum speed division number (S121: NO), the CPU 31 equally divides the affected section in one block by the maximum division number (S122). However, when there is an A2 section that is not affected by the preceding and following blocks as in the speed profile shown in FIG. 48 (1), the CPU 31 determines the A1 and A3 sections that are affected by the preceding and following blocks as the maximum speed division number. -1) Divide evenly by the number of 2. Therefore, the maximum number of speed divisions is set to an odd number. The A2 section is not divided. As shown in FIG. 48 (2), the A1 section and the A3 section are each divided into four.

  The CPU 31 selects the smallest moving speed in each divided section and generates a speed profile (S123). As shown in FIG. 48 (2), the speed of each block is adjusted as in the portion shown in gray. In FIG. 48 (2), the boundary line of the section before the division and the boundary line of the section after the division match, but even when the boundary line does not match, the smallest moving speed is selected in each new section similarly. . In the speed profile after adjustment, the number of speed divisions is reduced as compared with the speed profile before adjustment. Therefore, the CPU 31 can reduce the data amount of the speed profile output to the drive circuits 51A, 53A, 54A. Further, since the speed change can be reduced compared to the speed profile before adjustment, the shaft movement can be performed stably and the load applied to the machine tool 1 can also be reduced. The CPU 31 stores the adjusted speed profile in the RAM 33 and ends this process.

  The arc center command point of the above embodiment is an example of the virtual command point of the present invention. The target error is an example of a specified value of the present invention. The CPU 31 is an example of the control means of the present invention. The CPU 31 that executes the processes of S24 and S25 in FIG. 10 is an example of the specifying means of the present invention. The CPU 31 that executes the process of S26 is an example of the section specifying means of the present invention. The CPU 31 that executes the process of S27 is an example of a dividing unit of the present invention. The CPU 31 that executes the process of S28 is an example of the processing means of the present invention. The CPU 31 that executes the process of S29 is an example of the error calculation means of the present invention. The CPU 31 that executes the process of S91 in FIG. 23 is an example of a determination unit of the present invention. The CPU 31 that executes the process of S92 is an example of the calculating means of the present invention. The CPU 31 that executes the process of S31 in FIG. 10 is an example of the updating unit of the present invention.

  The CPU 31 that executes the process of S98 in FIG. 23 is an example of the error determination means of the present invention. The CPU 31 that executes the process of S100 is an example of a changing unit of the present invention. The CPU 31 that executes the process of S22 of FIG. 10 is an example of the interpolation determining means of the present invention. The CPU 31 that executes the processes of S23 and S24 is an example of the adding means of the present invention. The CPPU 31 that executes the process of S102 of FIG. 25 is an example of a first setting unit of the present invention. The CPU 31 that executes the process of S103 is an example of a second setting unit of the present invention.

  The CPU 31 that executes an error calculation process (see S29 in FIG. 10) performed in the deceleration calculation process in S111 in FIG. 35 is an example of a recalculation unit of the present invention. The CPU 31 that executes the process of S91 (see FIG. 23) of the error suppression process performed in the deceleration calculation process of S111 is an example of a redetermination unit of the present invention.

  As described above, the CPU 31 of the numerical controller 30 according to the present embodiment creates a speed profile along a path obtained by interpolating a plurality of command points commanded by the NC program. The speed profile includes distance and speed information between the command points. The CPU 31 controls the axis movement based on the result of performing the process with the moving average filter used for the acceleration / deceleration process on the speed profile. The CPU 31 calculates a target command point from command points in the program path. CPU31 calculates the influence area based on the time constant of the moving average filter used for acceleration / deceleration processing before and after the calculated attention command point. The CPU 31 divides the calculated influence interval by the sample time and calculates a command interpolation point. The CPU 31 processes the command interpolation point obtained by calculation with a moving average filter having a time constant equivalent to that of the moving average filter used for acceleration / deceleration processing, and calculates an operation interpolation point. The CPU 31 calculates an error from the program path at the motion interpolation point processed by the moving average filter. The CPU 31 determines whether or not the error at the motion interpolation point exceeds the target error. When the error exceeds the target error, the CPU 31 calculates the target error and a deceleration rate based on the error. The CPU 31 updates the speed profile based on the affected section and the deceleration rate obtained by calculation. Since the numerical control device 30 does not decelerate only at the command point, but decelerates in the affected section including the command point, the acceleration / deceleration is not repeated frequently even in a path where micro blocks continue. Therefore, the quality of the processed surface of the work material can be improved. By sequentially specifying a plurality of command points for each command point in the route, errors caused by acceleration / deceleration in, for example, a bent corner portion, arc portion, and curved portion in the route can be comprehensively suppressed.

  In the influence interval calculation process (see FIG. 14) of the above-described embodiment, the influence interval is at least two times before and after the target command point. Includes small sections. Therefore, since the speed profile can be updated in the influence section that affects the error at the command point, the error at the command point can be effectively suppressed.

  Since the deceleration rate used first in the error suppression process (see FIG. 23) of the above embodiment is a ratio obtained by dividing the target error by the error, the deceleration rate can be easily calculated.

  In the error suppression process of the above embodiment, the CPU 31 determines whether the difference between the error and the target error when the affected area is moved by the deceleration rate is equal to or less than the error convergence allowable value. The deceleration rate is changed so that the difference between the target error and the target error is less than the error convergence tolerance. Therefore, since it does not decelerate too much, the machining of the work material can be completed in a short machining cycle. Since there is no shortage of deceleration, it is more certain that the difference between the error and the target error does not exceed the error convergence tolerance.

  In the deceleration calculation process (see FIG. 10) of the above embodiment, the CPU 31 determines whether the block in the program path is linear interpolation or circular interpolation. When the block is linear interpolation, the CPU 31 designates the end point of the block as a target command point. When the block is circular interpolation, the CPU 31 designates the circular arc center command point as the target command point. Therefore, the numerical control device 30 can consider an error caused by the curvature of the arc portion. Even in the case of circular interpolation, the error at the circular arc center command point, that is, the error with respect to the circular arc can be suppressed to a predetermined value or less. Therefore, the numerical controller 30 can improve the quality of the work surface of the work material even in the circular interpolation portion.

  In the speed profile update means (see FIG. 25) of the above embodiment, the CPU 31 sets the deceleration section as the influence section when the target command point is the end point of the block. When the target command point is the arc center command point, the deceleration section is set to the entire area of circular interpolation. Therefore, the numerical controller 30 can take into account errors due to the curvature of the arc portion.

  In the above embodiment, the error check process is executed after the speed profile is updated by the deceleration calculation process (see FIG. 9). In the error check process, the CPU 31 calculates the error again, and determines again whether the calculated error exceeds the target error. If the error exceeds the target error, the CPU 31 recalculates the deceleration rate and updates the speed profile again. Therefore, the numerical controller 30 can suppress the error below the target error.

  In the above embodiment, when the error exceeds the target error, the deceleration rate is recalculated and the speed profile is renewed only once. However, the error may be repeatedly performed several times, for example, the error is less than the target error. You may go until

  The present invention is not limited to the above embodiment, and various modifications can be made. In the above embodiment, the time constant of the moving average filter is used as the time constant (acceleration / deceleration time constant) of the motor, but the present invention can also be applied to other filters having a time constant. For example, a low-pass filter, a band-pass filter, or the like can be applied. Also, a weighted moving average filter (general FIR filter) can be used. In the above embodiment, the moving speed is processed by a three-stage moving average filter, but it may be processed by a two-stage or four or more-stage moving average filter. Furthermore, after acceleration / deceleration processing using the time constant is performed, vibration suppression processing using a notch filter having a time constant, an input shaping method, or the like is performed to remove the natural vibration component of the machine from the speed profile. In this case, the motor time constant in the present invention is the sum of the acceleration / deceleration time constant and the vibration control time constant. Further, other processes for generating a speed profile for acceleration / deceleration in a certain time can be used in the same manner as the filter.

  In the above embodiment, the sample time used for the deceleration calculation process (S8, S11 in FIG. 9) and the error confirmation process (S12 in FIG. 9) is the same value as the CPU processing period. However, the purpose is to simplify the calculation process. Thus, a value that is an integral multiple of the CPU processing cycle may be used as the sample time. However, the accuracy of error calculation is reduced and the target error may not be satisfied.

The error confirmation process (S16 in FIG. 9) and the subsequent replacement process during the fully automatic deceleration process of the above embodiment may be omitted for the purpose of simplifying the calculation process. In that case, when executing the block movement command, pre-reading is necessary for the block to be processed before the total time of the time constant.

Of the error suppression processing (see FIG. 23) of the above embodiment, the processing of S93 to S100 may be omitted. The error condition of S98 is that the error is less than or equal to the target error (ε _r ≧ ε _k ), or the difference between the error and the target error satisfies the error convergence tolerance or less (0 ≦ ε _r −ε _k ≦ ε _t ). Other error conditions may be set. For example, the error may be set to be equal to or smaller than the target error (ε _r ≧ ε _k ), or | ε _r −ε _k | ≦ ε _t . However, in consideration of shortening the machining cycle, it is preferable to add a condition that the difference between the error and the target error is equal to or less than the error convergence tolerance.

  In the above-described embodiment, the influence interval before and after the command point is a distance corresponding to half the time of the total time constant. However, it may be an interval including at least a distance corresponding to half the time constant.

  In S23 and S24 of the deceleration calculation processing of the above embodiment, when the target block is circular interpolation, a virtual command point (circular center command point) is created at the center of the circular arc. Virtual command points may be created at other positions. A plurality of virtual command points may be created. The CPU 31 may advance the calculation using the block end point as the target command point without determining whether the target block is linear interpolation or circular interpolation. The arc may be decomposed into blocks of minute lines, and the error may be calculated as a linear interpolation command for each block.

  In the above-described embodiment, the vertical machine tool 1 in which the axis of the main shaft extends in the vertical direction has been described as an example of the machine tool controlled by the numerical control device 30, but a horizontal machine tool in which the axis of the main shaft extends in the horizontal direction is also described. Good. The drive circuits 51A to 55A (see FIG. 1) are provided in the numerical control device 30, but may be provided on the machine tool 1 side.

  Further, the configuration of the machine tool may include a rotation axis such as an A axis, a B axis, and a C axis in addition to the X axis, the Y axis, and the Z axis. Moreover, the configuration of the machine tool may include only a rotating shaft.

  In the above embodiment, the case where the numerical controller 30 performs the fully automatic deceleration process in the main process in FIG. 7 is described. However, the numerical controller 30 performs the fully automatic deceleration process prior to the main process, The result may be stored in the RAM 33 or the storage device 34, and processing may be performed using this result. Further, a similar effect may be obtained by processing the NC program in advance by a computer other than the numerical controller 30 to perform a fully automatic deceleration process and inputting the result to the numerical controller 30.

1 Machine tool 30 Numerical control device 31 CPU

Claims (9)

Based on the result of acceleration / deceleration processing using a filter, creating a speed profile that includes distance and speed information between each command point along the path obtained by interpolating multiple command points commanded by the NC program In a numerical control device comprising a control means for controlling the movement of a tool or work material,
The control means includes
A designation means for designating a command point of interest as a target command point from the command points in the route;
Section specifying means for specifying a time section based on the time constant of the filter before and after the target command point specified by the specifying means;
Dividing means for dividing the section designated by the section designating means at a predetermined time;
Processing means for processing the section divided by the dividing means with a filter having a time constant equivalent to the filter;
Error calculating means for calculating an error from the route at the target command point in the section processed by the processing means;
A determination means for determining whether the error calculated by the error calculation means exceeds a preset specified value;
A calculating means for calculating a deceleration rate based on the specified value and the error when the determining means determines that the error exceeds the specified value;
A numerical control apparatus comprising: an updating unit that updates the speed profile based on the section specified by the section specifying unit and the deceleration rate calculated by the calculating unit. The section includes a small section that moves at least half the total time constant of the filter before the target command point, and includes a similar small section after the target command point. The numerical control apparatus according to claim 1. The calculating means includes
The numerical control apparatus according to claim 1, wherein a ratio obtained by dividing the specified value by the error is calculated as the deceleration rate. Error determining means for determining whether or not the error when moving in the section at the deceleration rate calculated by the calculating means is within an allowable range with respect to the specified value;
The said error determination means is provided with the change means which changes the said deceleration rate so that the said error may be in the said allowable range, when it determines that the said error is outside the said allowable range. The numerical controller according to any one of 1 to 3. Interpolation determining means for determining whether a block between each command point in the path is linear interpolation or circular interpolation;
When the interpolation determining means determines that the block is the circular interpolation, the interpolation determining means includes an additional means for adding a virtual command point to the circular interpolation portion,
The designation means is:
When the interpolation determining means determines that the block is the linear interpolation, the end point of the block is designated as the attention command point,
When the interpolation determining unit determines that the block is the circular interpolation, the virtual command point added by the adding unit and the end point of the block are respectively designated as the target command points. The numerical control device according to claim 1. The updating means includes
When the attention command point is the end point of the block, a first setting unit that sets a deceleration section that moves at the deceleration rate calculated by the calculation unit to the section;
The numerical control apparatus according to claim 5, further comprising: a second setting unit that sets the deceleration section to at least the entire area of the circular interpolation when the attention command point is the virtual command point. Re-calculating means for recalculating the error when moving based on the speed profile after the updating means has updated the speed profile;
Re-determination means for re-determining whether the error calculated again by the re-calculation means exceeds the specified value,
The control means includes
The said re-determination means performs again each process by the said calculation means and the said update means, when it judges that the said error exceeds the said regulation value, The one in any one of Claim 1 to 6 characterized by the above-mentioned. Numerical control unit. Re-calculating means for recalculating the error when moving based on the speed profile after the updating means has updated the speed profile;
Re-determination means for re-determining whether the error calculated again by the re-calculation means exceeds the specified value,
The control means includes
When the re-determination unit determines that the error exceeds the specified value, the processing by the calculating unit and the updating unit is repeatedly executed until the error becomes equal to or less than the specified value. The numerical control device according to claim 1. Based on the result of acceleration / deceleration processing using a filter, creating a speed profile that includes distance and speed information between each command point along the path obtained by interpolating multiple command points commanded by the NC program In the control method of the numerical control device for controlling the movement of the tool or the work material,
A designating step of designating a command point of interest from among the command points in the path as a command point of interest;
An interval designating step for designating an interval of time based on the time constant of the filter before and after the attention command point specified in the specifying step;
A dividing step of dividing the section specified in the section specifying step by a predetermined time;
A processing step of processing the section divided in the dividing step with a filter having a time constant equivalent to the filter;
In the section processed in the processing step, an error calculating step of calculating an error from the route at the attention command point;
A determination step of determining whether the error calculated in the error calculation step exceeds a preset specified value;
In the determining step, when it is determined that the error exceeds the specified value, a calculating step for calculating a deceleration rate based on the specified value and the error;
A control method for a numerical controller, comprising: an update step of updating the speed profile based on the interval specified in the interval specifying step and the deceleration rate calculated in the calculation step.