JPH0679357A - Bending or twisting method and twisting torque loading device - Google Patents

Abstract

PURPOSE:To enable high grade bending or twisting while a spring back amount is precisely predicted regardless of the lot and/or individual difference of a material to be worked. CONSTITUTION:While the load or torque applied on the material to be worked from a loading device and the amount of deformation of bending or twisting coming up with the load or torque is monitored in real time during loading, the elastic coefficient and reversing line of the material to be worked is calculated and the spring back amount of the material is predicted.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention predicts the amount of springback that occurs in the process of bending or twisting an elastic work piece to perform bending or twisting to a target position or shape. The present invention relates to a bending or twisting method to be performed and a twisting torque load device.

[0002]

2. Description of the Related Art Conventionally, when bending or twisting a work piece made of an elastic material, the amount of springback generated in the working process differs depending on the work piece, which makes automation difficult and accurate. It was difficult to predict, and it often relied on the intuition of the worker and the experience of manual work. Further, for example, according to JP-A-63-199025,
By arbitrarily extracting a sample from one lot and grasping the material properties of the sample representing that lot,
We are predicting the springback amount of a work piece limited to a certain lot.

[0003]

However, according to the conventional technique, not only is the data limited to a lot,
It is not possible to adapt to individual differences of the workpiece,
Eventually, the operator had no choice but to rely on the correction work for bending or twisting, which required a great deal of time and cost. The present invention has been made to solve such a problem, and an object thereof is to accurately predict a springback amount regardless of a lot of a workpiece or a difference in individual difference, while achieving high accuracy. To provide a bending or twisting method and a twisting torque load device capable of performing various bending or twisting.

[0004]

In order to achieve the above object, the present invention detects the load or torque applied to a workpiece from a load device and detects the amount of bending or twisting deformation caused by it. The present invention is characterized in that the elastic constant and the restoration straight line of the workpiece are calculated and the springback amount of the workpiece is predicted while monitoring in real time during loading.

In addition, based on the springback amount predicted above, after bending or twisting, the load or torque is temporarily removed to a predetermined amount, and at this time, the return slope is newly monitored, It is characterized by revising the elastic constants obtained by and re-estimating the springback amount of the workpiece.

The unloading point of the load or torque is determined from the springback amount and the target value of the deformation amount predicted by calculating the elastic constant and the restoration straight line, and the target deformation is performed by unloading the load or torque. It is characterized by obtaining.

When a load is used to apply a torque to a workpiece in order to predict the springback amount or obtain a target deformation, the load component and the torque component are detected and the load component is eliminated to remove the load component. Aligning the twist center of the loading device with the twist center of the work piece by applying only pure torque to the work piece to prevent unnecessary deformation of the work piece. Characterize.

When the twist center of the load device is aligned with the twist center of the workpiece, the compliance control function is used as the load device to prevent unnecessary deformation of the workpiece. The feature is that the twist center of the load device is aligned with the twist center of the workpiece.

The configuration of the device for aligning the twist center on the device side with the twist center of the workpiece by using the compliance control function is a voice coil which applies a twist torque.
A gap sensor group for detecting the position and twisting angle of the motor and the voice coil motor including the bit, a bit for grasping the work piece, and the bit and the voice coil motor are held with a low friction force and twisted. Rotation direction of the center axis (Z- axis) and X intersecting with the Z- axis
-With an air or magnetic bearing mechanism that has a degree of freedom in the axis and Y-axis directions, it is possible to simultaneously monitor the load and the amount of deformation associated with it in-process, and to have a centering function that enables compliance control. It is characterized by having done.

The structure of the above-mentioned device includes a load device for applying a torsional torque, a multi-axis load cell for detecting a load component and a torque component, a gap sensor for measuring a torsional deformation, and a workpiece. The gripping bit and the work piece twisting center line X so as to align the work piece twisting center with the twisting center of the load device in order to remove the load component and perform centering X It is also characterized in that it is provided with a translation table that can move in two directions within the -Y plane.

Further, in the method of determining the elastic constant, when the elastic constant is detected while applying a load or a torque to the workpiece, the region from the start of the load to a predetermined load value or torque value is processed. It is characterized in that it is determined in advance in the elastic range of the object, and the elastic constant of the workpiece is determined by calculating the elastic constant using a predetermined method from the data of the load value or torque value and the deformation amount in that area. To do.

In the calculation of the elastic constant, a predetermined threshold value set in the elastic region of the workpiece is changed based on the material characteristics and load history of the workpiece, and the elasticity corresponding to the change of the elastic region of the workpiece is changed. It is characterized by calculating a constant. Further, when calculating the elastic constant, the load or torque is second-order differentiated by the amount of deformation, and the range in which the value is equal to or less than a predetermined value near 0 is set as the elastic constant detection range. The detection range of the elastic constant is defined as a range excluding at least one predetermined range at both ends of the detection range. When detecting the elastic constant, a linear approximation is performed based on data of a load or torque and a deformation amount within a predetermined elastic region, and the inclination thereof is used as the elastic constant. When detecting the elastic constant, the load value or torque value applied to the workpiece is first-order differentiated by the deformation amount, and the first-order average of the first-order differential value data in the preset elastic range is obtained. It is characterized in that the value is obtained and this is used as the elastic constant of the workpiece.

In addition, in a situation where a load or torque is repeatedly applied to a workpiece, when the elastic constant is detected and the springback amount is predicted,
Instead of calculating the elastic constant based on the load or torque and the deformation amount data within the preset area, the slope of the trajectory during unloading is defined as the elastic constant by the hysteresis before the previous load. Characterize.

As a load or torque loading method for predicting the springback amount or obtaining a target deformation, a stepwise load is applied and the bending or twisting deformation amount is monitored at each step. It is characterized in that each load value or each torque value and the value of the deformation amount are accurately measured by going down. When the step load is applied, the load is increased at a constant pitch, and the deformation amount is measured for each step. As a method of the step load, an optimum step width is obtained from the step width by the differential value of the SF curve, and the load or torque is increased at the step width. As a method of the step load of the load or the torque, the step width is changed according to the load value or the torque value, and the load is increased at the step width.

Alternatively, as a method of step load of load or torque, when the predicted final deformation amount approaches a target value to a predetermined value, the step width of load is sufficiently reduced to improve the adjustment accuracy. Characterize.

Further, the present invention is characterized in that when a load or torque is applied stepwise by using a compliance control function as a load device, a viscous force is applied to stabilize the displacement amount quickly.

It is characterized in that the springback amount of the workpiece is predicted by predicting the springback amount or by using a correction coefficient for the elastic constant obtained to obtain the target deformation to obtain a restoration straight line.

As the correction coefficient, the ratio of the inclination of the straight line in the elastic region and the inclination of the restoration line and the offset are used. It is characterized in that the correction coefficient is provided for each direction by combining bending or twisting directions in one direction and the opposite direction. These correction factors are characterized in that they are given for each number of times when the same workpiece is repeatedly bent or twisted.

In the method for obtaining the target deformation, as a method for unloading the load or torque, the load or torque is unloaded at a predetermined step to suppress the vibration of the work piece and to predict the actual final deformation amount. It is characterized by reducing the error of the final deformation amount of.

As a method of unloading the load or torque, a load device having compliance control is used to apply a load or torque to the workpiece, and then a predetermined viscous force is applied to unload the workpiece to cause deformation. It is characterized by shortening the stabilization time of the quantity and shortening the adjustment time of the workpiece. Further, in the method for obtaining the target deformation, after completely removing the load or torque from the unloading point, a predetermined amount of load or torque is further applied to prevent or reduce the aftereffect.

[0021]

With the above structure, according to the present invention, when bending or twisting is performed, the load or torque required for the deformation and the amount of deformation or the twisting angle that accompanies the load or torque are real time. While monitoring, the load-deformation curve or torque-torsion angle characteristic peculiar to the workpiece is detected, the springback amount of the workpiece after removing the load or torque is accurately predicted, and the target position Alternatively, the angle or shape is permanently deformed.

Therefore, the apparatus for carrying out the present invention moves a bit that grips the workpiece by a load device such as a voice coil motor to deform the workpiece so that a multi-axis load cell or the like is used. The sensor detects the load in the unnecessary direction applied to the work piece, quantitatively detects the pure load or torque value in the load direction, and the moving mechanism such as air or magnetic bearing mechanism or linear motion table is detected. In order to prevent unnecessary deformation of the work piece by driving, for example, in order to apply only pure torque to the work piece, the twist center of the work piece and the twist center of the loading device are set. In addition, the amount of deformation or twist angle of the workpiece is detected by a sensor such as a gap sensor, and the control necessary for proper processing is performed by a personal computer from these output values.

[0023]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT A preferred embodiment of the present invention will be described below with reference to the drawings by taking a twisting process as an example. FIG. 1 shows an outline of a device for carrying out the twisting process. The main components of this device are a load device 1 that applies a torsion torque,
A load cell 2 for detecting a torque value, a gap sensor 3 for detecting a torsional deformation, a bit 7 for gripping a workpiece 4, and a jig 8 for holding these components.
It consists of and. Then, the load device 1 applies a twisting torque to the fulcrum 6 while monitoring the target 5 of the workpiece 4 with the gap sensor 3.

(Embodiment 1) In this embodiment, the load or torque applied to the work piece from the loading device and the amount of bending or twisting deformation caused by the load are monitored in real time during the load and the elastic constants are set. The restoration straight line is calculated and the springback amount of the workpiece is predicted. FIG. 2 shows a flowchart of this embodiment. That is, in the execution block 10 in this flow chart, first, the target value X ₀ is set. This X ₀ is
It is a target final torsional deformation amount. When the target is set, the load device 1 is driven to apply the torsional deformation to the workpiece 4. This is done in execution block 12. In the execution block 14 and the execution block 16, the torsion torque and the torsion deformation amount generated in the load device 1 are detected and, as shown in FIG. The locus of the point P having the twist deformation amount X and the torsion torque Y as variables is monitored in real time.

Next, in a predetermined area (preliminarily set in the algorithm), the execution block 18 calculates the gradient θ as shown in FIG.
Then, the restoration straight line A after applying the load is calculated by calculation. In the execution block 22, as shown in FIG. 5, it is judged whether or not the locus of the point P intersects with the restoration straight line A obtained by the execution block 20. The point Q where the locus of the point P intersects the straight line A is defined as the point Q. If the point P is not on the restoration straight line A, the process from the execution block 12 is repeated, and if the point P is on the restoration straight line A, that is, if it is on the point Q, Next execution block 24
Do the following. Then, as shown in FIG. 6, in the execution block 24, the driving of the load device 1 is stopped, and the execution block 26 returns the load device 1 until the load or torque of the load device 1 becomes zero. By performing this series of processes,
It is possible to perform twisting with high accuracy up to the target value X ₀ .

(Embodiment 2) In this embodiment, based on the springback amount predicted as described above, after bending or twisting, up to a predetermined amount (for example, maximum load value or half of maximum torque). The load or torque is once removed, the return gradient is newly monitored at this time, and the elastic constant obtained as described above is reviewed and the springback amount of the workpiece is predicted again.

FIG. 7 shows a flowchart thereof. In this flowchart, the execution blocks 30 to 42 are the execution blocks 10 of FIG.
To execution block 22. In execution block 34 and execution block 36, the torsion torque and the torsion deformation amount generated in the load device 1 are detected,
As shown in FIG. 8, the locus of the point P having two parameters, that is, the torsional deformation amount X and the torsional torque Y as variables, is monitored in real time. Also, execution block 4
In 2, it is judged whether or not the locus of the point P intersects with the restoration straight line A obtained by the execution block 40, as shown in FIG. In this execution block 42, it is judged whether or not the permanent torsional deformation such that the target value becomes X ₀ is achieved by whether or not the locus of the point P intersects with the restoration straight line A. Then, if it is determined that the intersections are made here, the twisting is performed by the same method as described above. However, if they do not intersect, the following measures are taken.

That is, if the point P is not on the restoration straight line A at the stage of the execution block 42, the load amount is increased to a predetermined amount (for example, 1/2 of the maximum torque value) in order to obtain a new gradient θ _1. Then, as shown in FIG. 10, the new gradient θ ₁ is recalculated by executing a series of operations from the execution block 48 to the execution block 50. As a result, by executing the execution block 32 and the subsequent steps based on the new straight line which is the restored straight line B, it is possible to perform the twisting process with high accuracy up to the target value X ₀ as shown in FIG. Becomes

(Embodiment 3) This embodiment is a method of applying a load or torque to a workpiece while automatically adjusting the bending or twisting center by a predetermined loading method. This will be described below with reference to FIG. In FIG. 12, the execution block 52 sets the target value X ₀ . This X ₀ is a target final deformation amount. When the target value is determined, the execution block 54 aligns the twist center of the machine with the twist center of the work piece, and then the execution block 56 applies a load or torque to the work piece in a predetermined manner. And execution block 5
At 8, the amount of deformation or the twist angle is detected as data.
Next, when the execution block 60 determines that the data is within the elastic range, the execution block 62 executes calculation of the elastic constant based on the data. When it is determined that the data has left the elastic range, the process proceeds to execution block 64 to calculate the current deformation expected value X. And
In execution block 66, the degree of approach to the target value is tested, and if the approach distance is not within α (preset in the algorithm), the execution block 54 and subsequent steps are repeated. If the approach distance is within α, execution block 68 is entered. Move and change to a more accurate loading method. In this case, the load or torque is applied at a predetermined small pitch. Next, execution block 70
Is checked whether or not the deformation amount X has approached the target value X ₀ , and if not, the execution block 68 is repeated. The load or torque is stopped, and finally, in execution block 74, the load or torque is returned to 0 by a predetermined method to remove the load. By performing this series of processes, accurate bending or twisting can be performed up to the target value X ₀ .

(Embodiment 4) In the twisting process, normally, it was very difficult to make the twist center of the work piece coincide with the twist center of the load device side. Therefore, an unnecessary load or torque is applied to the work piece, which adversely affects the processing accuracy. The invention of the present embodiment solves this fundamentally, and it becomes possible to apply a pure torque to the workpiece. Therefore, the work piece is not deformed by an unnecessary load component.

This embodiment will be described below with reference to FIG. The process of this embodiment begins with execution block 82.
The target value X ₀ is set by. Next execution block 84
The unnecessary load component loaded on the workpiece is detected by, for example, a multi-axis load cell. At this time, if an unnecessary load component is working, it is judged in the next execution block 86, and it is canceled in the execution block 88. For example, the twist center of the workpiece is set by the linear guide table. The unnecessary load components are moved to 0, and then the execution block 84 and the execution block 86 are repeated to set all unnecessary load components to 0. Next, in execution block 90, a load or torque is applied to the work piece by a predetermined method, and at the same time, execution block 92 detects the amount of deformation or the twist angle. As described above, by repeating the execution block 84 and subsequent steps to apply the load, the twist center that can move with the deformation of the workpiece can always be aligned with the twist center of the twist load device. It is possible. This process allows the work piece to be loaded with pure torque.

(Embodiment 5) Another method of preventing an unnecessary load component or torque from being applied to a workpiece is a method of using a compliance control function on the load device side. Needless to say, this function must be provided on the load device side, but the process of the automatic centering function when this function is used will be described with reference to FIG.

The target value X ₀ is set by the execution block 102 in FIG. Next, the execution block 104 sets the rigidity on the load device side to some extent. This is set to the minimum necessary value necessary for positioning the workpiece in order to clamp it. After the workpiece is clamped in the execution block 106, the rigidity of the load device side is returned to zero in the execution block 108. After that, even if a twisting torque is applied to the workpiece, the center of twisting on the machine side follows the center of twisting of the workpiece, so that the centering function automatically operates at any time.
After that, in the execution block 110, a load or torque is applied to the work piece by a predetermined load method, and the execution block 1
The deformation amount or the twist angle of the work piece due to the load is detected by 12.

By following the above process, it is possible to apply only pure torque to the work piece and prevent unnecessary load components from intervening. As a result, a highly accurate correlation between the torque and the deformation amount can be obtained, and the twisting processing accuracy can be improved.

(Embodiment 6) This embodiment is an invention intended to realize the twisting method of Embodiment 5, and FIG.
It will be explained based on 5. FIG. 15 shows an embodiment of the torsion torque load device. A plurality of voice coil motors 124 for applying a torsional torque are arranged on the normal line of the rotational axis direction of the torsional center. Although FIG. 15 shows an example of three arrangements, the number of voice coil motors or the torque arm length is laid out so that the required performance can be satisfied according to the maximum required torque. Further, a twist central axis (Z- axis) is used for position control of the drive system 126 including the voice coil motor 124 and the bit 125.
The gap sensor group 127 for detecting the moving distance in the XY plane intersecting with the Z-axis and the twist angle around the Z-axis is connected to the personal computer and the high-speed processing element, and is fed back by the position information. Compliance control is possible. The entire drive system 126 in which these devices and the bit 125 are arranged on the same axis is air or magnetic bearing 12
No. 8 is used to float and hold with extremely low friction. This bearing 128
Is capable of not only maintaining extremely low frictional force in the rotation direction of the entire drive system 126 but also freely moving in the XY plane with almost no frictional resistance. After being clamped at 125, the twist center of the drive system 126 can be moved together with the change of the twist center position of the workpiece 4 during the torsion torque load. This mechanism enables the alignment method shown in the fifth embodiment.

(Embodiment 7) The invention of this embodiment is intended to realize the twisting method of Embodiment 4, and will be described with reference to FIG. FIG. 16 shows an embodiment of the torsion torque load device. A rotating table using a drive motor such as a stepping motor is used as the load device 129 corresponding to the load applying unit of the present device. Further, the load device 129 applies X to the workpiece 4.
A multi-axis load cell 130 capable of detecting loads and torques in the Y-axis, Y-axis, and θ-z directions is provided. This multi-axis load cell 130 is connected to a personal computer, and for example, when the workpiece 4 has a torsion center eccentric during the torque load and a load component is generated in the XY plane,
By detecting them with the multi-axis load cell 130 described above and sending a command to the X- and Y- axis linear motion guide table 131 to move so that the value becomes zero or becomes extremely small, it is transmitted through the personal computer. Adjusted.

As described above, even when the work 4 is twisted and the twist center is deviated, this function causes the twist center on the machine side to move with respect to the movement of the work 4. It is possible to follow up at any time, so that it is possible to always apply no load component other than pure torque to the workpiece 4.
With the functions as described above, an automatic centering method is possible in which the linear guide table is moved so as to eliminate the in-plane force and align the twist center of the workpiece with the twist center of the load device.

(Embodiment 8) This embodiment will be described with reference to FIG. In order to give a target amount of bending or twisting deformation to the work piece, it is necessary to detect the springback amount with high accuracy. FIG. 17 is an SF curve showing a typical load-deformation amount relationship at the time of bending or twisting deformation of a workpiece. The prediction of the springback amount of the work piece has a great influence on the accuracy of the deformation amount to the target value X ₀ . In order to make this prediction, it is necessary to detect the elastic constant θ of the work piece, including its inherent difference, in-process. However, the elastic constant θ can be calculated from the correlation data between the load or torque and the amount of deformation.
In order to detect, the slope of the straight line must be obtained from the correlation data of the load or torque and the deformation amount in the elastic region, that is, the region where the correlation between the load or torque and the deformation amount is linear. Therefore, other data, for example, data in the plastic region, that is, the non-linear region, must be excluded. In order to identify this data, as shown in FIG. 17, a region a ₁ in the elastic region of the workpiece is set in advance in the algorithm for obtaining the elastic constant, and the region a _{1 is set.}
Only the correlation data in the above is taken in, the slope of the straight line is obtained by a predetermined method, and the elastic constant θ is calculated. This area a ₁ set in advance can not only be set to an appropriate area in advance by a material test of the work piece, etc., but also can be changed at any time in consideration of factors such as the load direction or the twist direction, or the load history. Optimization is possible at any time.

(Embodiment 9) In Embodiment 8, the elastic constant is a proportional constant in a region where the load or torque of the workpiece and the deformation amount are in a proportional relationship, that is, in the elastic region. The region changes due to work hardening, the Bauschinger effect, and the like, for example, due to the past load history of the workpiece, and is not necessarily the same as the virgin stress state. In order to accurately calculate the elastic constants, data on the plastic region of the work piece when collecting data on the load value or torque value and the amount of deformation, and data on non-linear regions such as slippage on the load device side immediately after the start of loading It is necessary to collect as much data as possible in the linear region while eliminating as much as possible. Therefore, the outline of the present embodiment is a method of eliminating unnecessary data by changing the preset load value or torque value or the range of the deformation amount according to the load history of the workpiece and the variation of the material.

This embodiment will be described below with reference to FIG. Specifically, when the processing direction of the work piece is the same in the first and second times, the work piece undergoes work hardening, so that the elastic region tends to widen as shown in FIG. 18 (a). is there. Therefore, the load value or the torque value, which is the threshold value of the elastic region, can be set higher in the elastic region b ₁ of the curve of the second processing than in the elastic region a ₁ of the curve of the first processing. In the figure, X ₀ is a target value. On the other hand, when the load direction and the twisting direction of the first and second machining are opposite to each other, the workpiece is affected by the Bauschinger effect.
As shown in (b), the elastic region b ₂ of the curve of the second machining
Is lower than the elastic region a ₁ of the curve of the first processing, it is necessary to set the load value or the torque value as the threshold value of the elastic region to be smaller than usual. As described above, the elastic constant can be accurately calculated by changing the setting of the elastic region according to the load history of the workpiece.

(Embodiment 10) The elastic constant can be generally considered as a proportional constant in a region where the load or torque and the amount of deformation are in a proportional relationship. However, as the load or torque increases, the deformation mode gradually shifts from elastic deformation to plastic deformation, and the proportional relationship does not hold. Even in the vicinity of the load starting point, the original proportional relationship may not appear due to insufficient engagement between the jig and the workpiece. Therefore, in order to obtain the elastic constant with high accuracy, it is necessary to set the detection range as wide as possible and not to include the non-linear region such as the plastic deformation region.

In setting the detection range of the elastic constant, the easiest method is to set the range of the load or torque or the deformation amount that the proportional relationship is likely to be established in advance by experiments etc., and set the elasticity from the data within the range. This is a method of obtaining a constant. However, in the case of such a method, there is a possibility that even a non-linear region may be included due to variations in characteristics of individual workpieces. Further, in order to avoid this, it is necessary to make the setting area sufficiently small, but in such a case, there is a drawback that the detection accuracy of the elastic constant becomes poor.

The present invention of this embodiment has been made to solve such a problem, and a concrete embodiment thereof will be described below with reference to FIG. FIG. 19A shows a typical relationship between the load or torque F applied to the work piece and the deformation amount S in bending or twisting deformation. As load or torque is applied, the deformation moves from the initial unstable region c ₀ to the nonlinear plastic deformation region d via the elastic deformation region c having a substantially proportional relationship. FIG. 19 (b)
Is the load or torque F in FIG.
DF / dS differentiated by the above, and the proportional region c appears as a constant value region with a zero slope here.
Further, FIG. 19C shows d ² F / dS ² which is second-order differentiated, and the range in which this value is near zero corresponds to the linear region.

In the above second-order differential curve, it is fine near zero.
Set a small range 0 ± α and d²F / dS ²Is within this range
The region R in is the detection range of the elastic constant. here
Is to detect the range where the linear relationship between F and S holds
Setting allowable range of the second derivative of. R is determined by this
The deformation amount S is within the elastic constant detection range. That is, this range
The load or torque F as shown in FIG.
And the elastic constant as the proportional constant of the deformation amount S, or
The elastic constant is calculated as the differential value dF / dS in FIG.
It

According to such a method, even if the range of the elastic region is unstable due to variations in the characteristics of the workpiece, etc., the elastic region is always set automatically and reliably to the maximum extent according to the individual workpiece. It is possible to improve the detection accuracy of the elastic constant.

(Embodiment 11) According to the method of the embodiment 10, it is possible to automatically set the elastic region according to the characteristics of the individual work pieces, but it is clear from FIG. 19 (c). The detection range R thus obtained always includes the transition region to the initial unstable region and the plastic deformation region at both ends thereof. In order to avoid this, it is necessary to make the setting range α of the second-order differential value as small as possible, but in that case, the elastic region to be taken in is determined to be outside the detection range due to a slight change in the detection data. there is a possibility.

FIG. 20 illustrates a method for solving such a problem. Similar to FIG. 19 (c), the second-order differential value d ² F / dS ^{2 of the} load or torque applied to the workpiece and the deformation amount S are obtained. Shows the relationship. Second derivative d ² F / dS ²
The method is similar to the above method until the range R is determined as a region within 0 ± α, but the relatively small region r near both ends is excluded, and the remaining region R ′ is used to detect the elastic constant. And With such a method, it is possible to exclude a transitional region that deviates from the elastic deformation region, and it is possible to detect the elastic constant with higher accuracy. It should be noted that the excluded areas do not necessarily have to be at both ends, and the excluded areas do not have to have the same size in the case of both ends.

(Example 12) In Example 8, the elastic constant was calculated by the following method. The elastic constant has a proportional relationship between the load or torque and the amount of deformation, and when it is indicated by a straight line, it is considered to be the inclination thereof. As a specific example of the calculation method of the elastic constant, linear approximation is performed by using the least squares method based on the data within the threshold value a ₃ to the threshold value b ₃ within the elastic region set in advance as shown in FIG. , And the slope is the elastic constant.

(Example 13) In Example 8, the elastic constant is calculated by the following method. When a load or torque is applied to the work piece, the correlation between the load or torque and the deformation amount obtained thereby is first-order differentiated by the deformation amount to show the slope of the function indicating the correlation. In particular, if this value is specified within the elastic region of the workpiece, the elastic constant or Young's modulus peculiar to the material of the workpiece is shown. This fact will be described with reference to FIG.

FIG. 22A is an SF curve showing the relationship between the load or torque value of a general material and the amount of deformation. This curve shows the unstable area A at the beginning of load or torque.
_1, the elastic region A ₂ weight deformation load or torque is proportional, material begins to yield, it can be divided into three regions of a region A ₃ entering the plastic region. When this correlation is first-order differentiated by the amount of deformation, a curve as shown in FIG. 22 (b) is obtained. This first-order differential value is used to calculate the elastic constant θ of the workpiece.

Based on the data between the upper and lower threshold values a ₃ and b ₃ that divide the elastic region A ₂ set in advance, the first-order differential value with respect to the deformation amount of the correlation data of the load value or the torque value and the deformation amount is calculated. Then, the primary average value of the data of A _{2 in} the elastic region is calculated. This value becomes the elastic constant θ in the material of the work piece.

At this time, in order to improve the accuracy in the calculation of the elastic constant θ, the maximum value of the first-order differential value in the elastic region A ₂ of FIG. It is also possible to remove the minimum value from the population. With the method as described above, the elastic constant of the workpiece can be calculated in-process with accuracy including its inherent difference.

(Embodiment 14) In the case where the elastic coefficient is to be detected while applying a load or a torque to a workpiece, especially when the load or the torque has to be repeatedly applied, a direction opposite to the previous load direction is applied. Depending on the state of residual stress, the elastic limit may be extremely low when applied to. In this case, the calculation of the elastic constant is extremely difficult because there is no elastic region or almost no elastic region even if a certain region is defined in-process for calculating the elastic constant. For this reason, the slope of the trajectory at the time of unloading with the previous load is stored, this is defined as the elastic constant, and the springback amount is predicted based on that. The above method will be described with reference to FIG.

FIG. 23A shows the relationship between the load or torque of a very general material and the amount of deformation. In this case, even if the load is applied in the opposite direction to the previous load direction, the elastic limit a _{4 in} either direction is the same, and therefore it is considered that the workpiece has no residual stress. In such a case, the elastic region, that is, the linear region clearly appears when the force is applied in either direction. Therefore, the calculation of the elastic constant is relatively easy, which makes it possible to predict the springback amount of the workpiece. However, as shown in FIG. 23 (b), some load history P ₁
Therefore, when the elastic limit a ₅ is extremely low, it is difficult to calculate an accurate elastic constant θ-3 in-process. In such a case, if the inclination θ-2 of the trajectory P ₂ at the time of previous unloading is stored, it is possible to predict the spring bag amount by using it as an elastic constant.

(Embodiment 15) This embodiment is an embodiment of a load or torque loading method of the third embodiment. Generally, when a load or torque is applied to a work piece, the work piece deforms after a certain time delay. Therefore, when the load or torque applied to the workpiece is sequentially increased and the amount of deformation is monitored, an error occurs in the relationship between the obtained load or torque and the amount of deformation. The invention of the present embodiment reduces this error and improves the adjustment accuracy, which will be described below with reference to FIG.

The load or torque is increased by a predetermined amount by the execution block 113 of FIG. Next, the execution block 114 waits for the time required for the amount of deformation to stabilize.
At this time, the waiting time is set to the minimum necessary so that the degree of stability of the deformation amount becomes negligible as compared with the allowable adjustment accuracy. After that, in the execution block 115, the deformation amount is measured. After that, in the execution block 116, it is determined whether to increase the load or the torque based on a predetermined reference, and if the increase is continued, the process returns to the execution block 113. By following the above process, it is possible to apply the load or torque while more accurately measuring the relationship between the load or torque and the deformation amount.

(Embodiment 16) This embodiment is an embodiment of a load method with good measurement accuracy when the method of Embodiment 15 is used as the load or torque load method of Embodiment 3. When step loading is performed in the fifteenth embodiment, as shown in FIG. 25, the deformation amount S ₁ per unit torque F ₁ at low load and the deformation amount S per unit torque F ₁ at high load in the SF curve are shown. _There is a difference between ₂ and the deformation amount per unit torque F ₁ . On the other hand, the amount of deformation per step has a great influence on the adjustment accuracy, so it is necessary to keep it below a certain value. For this reason, the amount of deformation per step must be below a certain value even under high load, and the step width must be made sufficiently small. However, in this case, the number of measurement points at low load becomes unnecessarily large, which causes increase in tact time.

As a solution, the present embodiment and the first embodiment
7, 18 will be described. Description will be given below with reference to FIG. The load or torque is increased by a predetermined amount by the execution block 117 of FIG. Next, the execution block 118 waits for the time required for the amount of deformation to stabilize. afterwards,
In the execution block 119, the deformation amount is measured. After that, in the execution block 120, the step width of the load or the torque is set. The step width of the execution block 120 is a constant value in this case. After that, the process proceeds to the execution block 122, and after determining whether to increase the load or the torque based on a predetermined reference, when the increase continues, the process returns to the execution block 117. By following the above process, the amount of deformation per step can be kept below a certain value even under high load, and it is possible to apply load or torque without increasing the number of measurement points during low load. become.

(Embodiment 17) The invention of this embodiment shows a method of calculating an appropriate step width when the invention of embodiment 15 is used. As shown in FIG. 27, when the load or torque is applied stepwise, the unit load or unit torque F ₁ , F ₁ at low load and high load respectively
The deformation increase amounts S ₁ and S ₂ with respect to are related to increase when the differential value of the SF curve (the slope of the tangent line in the deformation regions S ₁ and S ₂ ) decreases. Therefore, the appropriate step width of the load or torque can be calculated by the differential value of the SF curve. The invention of this embodiment utilizes this.

In increasing the load or torque, FIG.
As shown in FIG. 8, the slope between the two points of the SF curve is obtained from the load or torque value one step before, the amount of deformation, and the respective values at that time, and this is used as an approximately differentiated value. As a function of this value, step width F ₂ of execution block 120 of FIG.
Then, the load or torque is increased by the step width. With the above configuration, it becomes possible to apply a load or torque with an appropriate deformation increase amount S ₃ .

(Embodiment 18) As described in Embodiment 17, at the time of high load, the slope of the SF curve becomes small, and accordingly, the deformation amount per unit load or unit torque increases,
The deformation amount increases for the same step width. As a method of solving this, a method of calculating an appropriate step width based on the slope of the SF curve as in the seventeenth embodiment can be considered.
However, as a simpler method, the fact that the S-F curves are almost the same can be used when the workpieces are limited to the same material and the same shape.

This embodiment utilizes the above-mentioned facts.
Since the -F curve draws almost the same locus, the differential value of the SF curve at the same load or torque is almost the same value, and the step width of the load or torque is changed according to the load or torque value at that time. As an example thereof, it is conceivable that the step width of the execution block 120 in FIG. 26 is a function inversely proportional to the load or torque value at that time. An advantage of this case is that the step time is calculated only by the command value, as compared with the case where the step width is changed by the differential value, so that the calculation time can be reduced. As described above, an appropriate step width can be calculated by changing the load or torque step width according to the load or torque value at that time.

(Embodiment 19) The invention of this embodiment is Embodiment 1
In the method of applying load or torque stepwise as in the invention of 5, the adjustment accuracy is improved.
When the load or torque is applied stepwise, FIG.
As shown in (a), the amount of increase in deformation corresponding to the step amount F ₃ of the load or torque greatly affects the adjustment accuracy. In order to reduce this effect, it is necessary to reduce the load or torque step amount. However, when the step amount is reduced, the number of measurement points increases, and therefore the adjustment time becomes long.
The invention of this embodiment solves this problem and is described below.

FIG. 2 shows that the amount of increase in deformation corresponding to the step amount of load or torque greatly affects the adjustment accuracy.
When the predicted final deformation amount S ₀ approaches the target value X _{0 as shown} in 9 (a), the error δ ₁ may increase. Elsewhere, the impact is small. Therefore, as shown in FIG. 29B, when the predicted final deformation amount S ₀ approaches the target value X ₀ to a predetermined value, the load or torque step amount F ₄ is made sufficiently small. The error δ ₂ becomes small. In other places, the minimum required step amount is used. With the above configuration, it is possible to improve the adjustment accuracy without extending the adjustment time.

(Embodiment 20) The invention of this embodiment is a bending or twisting load device capable of compliance control, and when the load or torque is applied stepwise as in the invention of embodiment 15, the displacement amount is changed. It stabilizes quickly. When a load or torque is applied stepwise to a work piece having rigidity as shown by a curve C in FIG.
As shown by the curve D in (b), it changes with time. The decay time at this time greatly affects the adjustment time and the adjustment accuracy. In the present embodiment, the adjustment time is significantly shortened and the adjustment accuracy is improved by shortening the decay time, which will be described below.

When a load or torque is applied stepwise as shown by the curve D in FIG. 30 (b), the workpiece undergoes damped vibration. If the amount of attenuation at this time is not sufficient, the amount of deformation of the workpiece that has been taken in will include the amplitude at this time, but this may cause an adjustment error. Further, if the decay time is sufficiently long to prevent this, the adjustment accuracy is improved, but the adjustment time is extended. In order to solve these problems, the first derivative of the deformation amount of the work piece obtained by compliance control with respect to time (not shown)
Then, a viscous force is applied to the load device so that the damping state of the workpiece (see FIG. 30 (b)) is minimized as shown by the curve E in FIG. 30 (c). In this case, the viscosity coefficient is obtained in advance by experiments and the value is used.
By the above method, the amount of deformation of the workpiece can be stabilized quickly,
It is possible to improve the adjustment accuracy and shorten the adjustment time.

(Embodiment 21) The invention of this embodiment is Embodiment 3
The amount of springback is predicted by using a correction coefficient for the elastic constant obtained in the invention. In bending or twisting, the slopes of the elastic line under load and the restoring line during load removal are theoretically equal. However, in reality, they may not match due to the work shape, the fixing method, the measurement error, and the like. In such a case, by estimating the springback amount by using the correction coefficient in advance, it is possible to perform more accurate machining.

A method of using the inclination ratio of the straight line and the offset as the correction coefficient will be described with reference to FIG. Let θ1 be the inclination of a straight line in the elastic region obtained by starting the processing and monitoring the load and the displacement. From the preset target value T, the offset B ₀ (correction coefficient) and θ1 and the inclination correction coefficient A ₀ are used to find θ2 = A ₀ · θ1, and the restored straight line is estimated using the obtained θ2. Further processing, point R where the load and displacement curves intersect the estimated restoration line
Remove the load at ₁ (load removal point). Here, the case of A ₀ = 1 or B ₀ = 0 is included.

In the above embodiment, the tendency may differ depending on the bending or twisting direction, such as when the shape of the workpiece is asymmetric. In such a case, as shown in FIG. 32, the bending or twisting directions are combined in one direction and in the opposite direction, and the restoration straight line is estimated for each direction. Thereby, it is possible to perform bending or twisting with higher accuracy. That is, in FIG. 32, θ2 ′ is the inclination of the estimated restoration straight line in which the bending or twisting is in the negative direction, and is calculated by θ2 ′ = A ′ · θ1 ′ similarly to θ2, and A ′ is the inclination in the negative direction. A correction coefficient, B'is a negative offset (correction coefficient), T'is a negative target value, and R'is a negative load removal point.

In high-accuracy bending and twisting processing in which the allowable range of the target value is narrow, there is a case where the accuracy is not satisfied by one-time processing and repeated processing is required twice or three times. In the second and subsequent machining, the tendency may be different due to the influence of the previous machining (residual stress, etc.). In such a case, by having a correction coefficient for each number of times, it is possible to accurately perform the desired bending or twisting processing with a small number of processing times.
Table 1 is an example applied to FIG.

[0071]

[Table 1]

(Embodiment 22) The invention of this embodiment is Embodiment 3
In, it is an invention relating to a method for unloading a load or torque. When the load or torque is released all at once from the state in which the load or torque is applied to the work piece, as shown in FIG.
As shown in (a), the work piece is subjected to an inertial force G in the direction opposite to the direction in which the load or torque has been received by the load device.
Receive. This may affect the error between the actual final deformation amount and the predicted value. This embodiment solves this problem and will be described below.

The inertial force as described above affects the final deformation amount when the inertial force exceeds the elastic limit in the direction opposite to the direction in which the load or torque of the workpiece is being applied. When is small, there is almost no deformation due to this force. Therefore, the load or torque is gradually removed on the restoration straight line from the unloading point Q ₀ toward the target value X ₀ of the deformation amount. For example, as shown in FIG. 33 (b), when unloading a load or torque, the step amount is unloaded stepwise with small step widths of F ₂₁ , F ₂₂ , ... Make sure that little inertial force is applied. By doing as described above, it is possible to reduce the error between the predicted final deformation amount and the actual final deformation amount.

(Embodiment 23) For the reason as described in Embodiment 22, when the load or torque is applied and then the load or torque is released at once, the adjustment accuracy may be adversely affected. Moreover, the adjustment time is extended due to the damping time of the vibration generated at that time. The present embodiment solves these problems.

In the process of unloading after the load or torque is applied to the workpiece by the compliance control,
A predetermined viscous force is applied to the load device by performing first-order differentiation of the restoration amount of the workpiece obtained by compliance control with respect to time. As the proportionality constant at this time, an appropriate value obtained by experiments or the like is used as the viscosity coefficient. As a result, originally, FIG.
The deformation amount that oscillates with the damping time T ₀ as shown in (a) is damped with the short damping time T ₀ as shown in FIG. 34 (b), and the inertial force is not applied to the workpiece, and the adjustment accuracy is improved and the damping is performed. Adjustment time can be shortened by shortening the time.

(Embodiment 24) A workpiece formed by bending or twisting by unloading a load or a torque toward a target value X ₀ of the deformation amount is finished by a post-effect as shown in FIG. The deformation amount S ₄ changes in the direction of the arrow with time. If this quantity is of a non-negligible order compared to the permissible accuracy, the aftereffect becomes a big problem. The invention of this embodiment prevents or reduces this aftereffect.

Residual stress is introduced into the work piece that has been bent or twisted. If the residual stress is sufficiently smaller than the yield stress of the work piece, the residual stress is not semipermanently released and the work piece is maintained in that state. However, if the residual stress is not sufficiently small compared to the yield stress of the work piece, the residual stress is
Gradually released, the work piece is in the form of aftereffect,
Deforms over time. In order to prevent or reduce this, it is considered to release the residual stress in advance.
Therefore, after completely removing the load from the unloading point Q _{0 as} shown in FIG. 36, a predetermined amount of load or torque is further applied, for example, at least once in the opposite direction and the same direction as the load or torque, and the residual stress Urge the liberation of. By this operation,
The final deformation amount S ₅ is obtained. With the above, it is possible to prevent or reduce the aftereffect.

Although the above embodiment has been described with respect to the case where the workpiece is twisted, the invention can be appropriately applied to the case of bending.

[0079]

As described in detail above, according to the present invention, the elastic constant of the work piece is detected in real time, and an appropriate restoration straight line is calculated to calculate the springback amount. Therefore, it is possible to efficiently perform bending or twisting regardless of the lot or individual difference of the workpiece.

[Brief description of drawings]

FIG. 1 is a diagram showing an outline of an apparatus for twisting a workpiece.

FIG. 2 is a first example for applying a twisting process to a workpiece.
It is a figure which shows the flowchart in.

FIG. 3 is a diagram for explaining a twisting process in the first embodiment.

FIG. 4 is a diagram for explaining a twisting process in the first embodiment.

FIG. 5 is a diagram for explaining a twisting process in the first embodiment.

FIG. 6 is a diagram for explaining a twisting process in the first embodiment.

FIG. 7 is a second example for applying a twisting process to a workpiece.
It is a figure which shows the flowchart in.

FIG. 8 is a diagram for explaining a twisting process according to the second embodiment.

FIG. 9 is a diagram for explaining a twisting process in the second embodiment.

FIG. 10 is a diagram for explaining a twisting process according to the second embodiment.

FIG. 11 is a diagram for explaining a twisting process in the second embodiment.

FIG. 12 is a diagram showing a flowchart in a third embodiment for applying a twisting process to a work piece.

FIG. 13 is a diagram illustrating a flowchart of an automatic centering method according to a fourth embodiment.

FIG. 14 is a diagram illustrating a flowchart of an automatic centering method according to a fifth embodiment.

FIG. 15 is a configuration diagram of a torsion torque load device capable of compliance control and having an automatic centering function.

FIG. 16 is a configuration diagram of a torsion torque load device having a self-aligning function.

FIG. 17 is an explanatory diagram of a method of determining an elastic constant.

FIG. 18 (a) is an explanatory view showing a change in elastic region when a load is applied to the work piece a plurality of times, and FIG. 18 (b) is a case where the work piece is loaded so that the load directions are opposite to each other. It is explanatory drawing which shows the change of an elastic region.

19 (a) is a diagram showing a relationship between a force F applied to a workpiece and a deformation amount S, FIG. 19 (b) is a diagram showing changes in its differential value dF / dS, and FIG. 19 (c) is a second-order differential d. from the relationship of ² F / dS ² and the amount of deformation S is an explanatory view for determining the detection range of the elastic constants.

FIG. 20 is a diagram showing setting of a detection range of an elastic constant.

FIG. 21 is an explanatory diagram for obtaining an elastic constant from linear approximation of data of load and deformation amount.

22A is a diagram showing a relationship between a load value and a deformation amount of a general material, and FIG. 22B is a diagram showing a change of its first-order differential value.

23A is a diagram showing a relationship between a load value and a deformation amount of a general material, and FIG. 23B is an explanatory diagram of a method of determining an elastic constant when the elastic limit is extremely low.

FIG. 24 is a flow chart of a load or torque load method according to a fifteenth embodiment.

FIG. 25 is an explanatory diagram of a difference in deformation amount per unit torque pitch.

FIG. 26 is a flow chart of a load or torque load method according to a sixteenth embodiment.

FIG. 27 is an explanatory diagram of a relationship between a deformation increase amount and a tangent line slope with respect to a unit load or a unit torque in a curve indicating a relationship between a load value and a deformation amount.

FIG. 28 is an explanatory diagram of a proper step width and a proper deformation increase amount in a curve showing a relationship between a load value and a deformation amount.

29A is a diagram showing a relationship between a step amount of load or torque and a predicted final deformation amount, and FIG. 29B is a diagram showing a relationship between the step amount and a predicted final deformation amount when the step amount is sufficiently small. Is.

FIG. 30 (a) is a diagram showing a relationship between load or torque and time when the load or torque is applied stepwise,
(B) is a diagram showing the relationship between the amount of deformation and the damping time when a load or torque is applied stepwise, and (c) is a deformation when the load or torque is applied stepwise by applying the viscous force of the loading device. It is a figure which shows the relationship between quantity and decay time.

FIG. 31 is an explanatory diagram of a method of obtaining a spring bag amount by an estimated restoration straight line obtained by using a correction coefficient.

FIG. 32 is an explanatory diagram of a method of obtaining a spring bag amount by an estimated restoration straight line obtained by using a correction coefficient when the bending or twisting direction is different in positive and negative directions.

FIG. 33 (a) is a diagram showing an inertial force when a load or torque is released from a workpiece at one time, and FIG. 33 (b) is an explanatory diagram of a method for unloading the load or torque stepwise.

FIG. 34 (a) is a diagram showing a damping vibration when a load or torque is released from a work piece at a time, and (b) shows a damping vibration of the work piece when a viscous force of a load device is applied. It is a figure.

FIG. 35 is a diagram showing a change in final deformation amount due to aftereffect.

FIG. 36 is an explanatory diagram of a method for releasing residual stress.

[Explanation of symbols]

1,129 Load device 2 Load cell 3 Gap sensor 4 Workpiece 5 Target 6 Support point 7,125,133 Bit 8 Jig 124 Voice coil motor 126 Drive system 127,132 Gap sensor 128 Air or magnetic bearing 130 Multi-axis load cell 131 X-axis, Y-axis guide table A ₀ , B ₀ , A ', B'correction coefficient F ₃ proper step width F ₂ , F ₄ step amount R, R'elastic constant detection range A ₂ , a ₁ , b ₁ , b ₂ Elastic region a ₃ , b ₃ Threshold value r Predetermined range α Allowable width of second derivative value θ, θ2, θ2 ′, θ-1, θ-2, θ-3, elastic constant

[Procedure amendment]

[Submission date] March 9, 1993

[Procedure Amendment 1]

[Document name to be amended] Statement

[Name of item to be corrected] 0032

[Correction method] Change

[Correction content]

(Embodiment 5) As another method for preventing an unnecessary load component or torque from being applied to a workpiece, a compliance control function (elasticity control and
Viscosity control) is available. Needless to say, this function must be provided on the load device side, but the process of the automatic centering function when this function is used will be described with reference to FIG.

[Procedure Amendment 2]

[Document name to be amended] Statement

[Name of item to be corrected] 0072

[Correction method] Change

[Correction content]

(Embodiment 22) The invention of this embodiment is Embodiment 3
In, it is an invention relating to a method for unloading a load or torque. When the load or torque is released all at once from the state in which the load or torque is applied to the work piece, as shown in FIG.
Workpiece as (a) by the load device, the a direction that was subjected to a load or torque until it and the opposite direction of the apparatus
It receives the inertial force G of the work piece . This may affect the error between the actual final deformation amount and the predicted value.
This embodiment solves this problem and will be described below.

[Procedure 3]

[Document name to be amended] Statement

[Correction target item name] 0077

[Correction method] Change

[Correction content]

Residual stress is introduced into the work piece that has been bent or twisted. If the residual stress is sufficiently smaller than the yield stress of the work piece, the residual stress is not semipermanently released and the work piece is maintained in that state. However, if the residual stress is not sufficiently small compared to the yield stress of the work piece, the residual stress is
Gradually released, the work piece is in the form of aftereffect,
Deforms over time. In order to prevent or reduce this, it is considered to release the residual stress in advance.
Therefore, as shown in FIG. 36, after completely removing the load from the unloading point Q ₀ , a predetermined amount of load or torque is further applied, for example, in the opposite direction to the load or torque and at least once in the same direction, and residual stress is applied. Urge the liberation of. By this operation,
The final deformation amount S ₅ is obtained. Ru Ah is possible to prevent or alleviate the aftereffect by the above.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Toru Kuwata 1048, Kadoma, Kadoma, Osaka Prefecture Matsushita Electric Works, Ltd. (72) Ryosuke Maruyama, 1048, Kadoma, Kadoma, Osaka Prefecture, Matsushita Electric Works, Ltd. (72) Inventor Hideki Yamashita 1048 Kadoma, Kadoma City, Osaka Prefecture Matsushita Electric Works, Ltd.

Claims (27)

[Claims] 1. The elasticity of the work piece is monitored while the load or torque applied to the work piece from the load device is detected, and the amount of bending or twisting deformation caused thereby is detected in real time during the load. A bending or twisting method characterized by calculating a constant and a restoration straight line and predicting a springback amount of a workpiece. 2. Based on the springback amount predicted as described above, after bending or twisting, the load or torque is temporarily removed to a predetermined amount, and the return gradient is newly monitored at this time to obtain the value as described above. The bending or twisting method according to claim 1, wherein the elastic constant is reviewed and the springback amount of the workpiece is predicted again. 3. The load or torque unloading point is determined from the springback amount and the target value of the deformation amount predicted by calculating the elastic constant and the restoration straight line, and the load or torque is unloaded to determine the target value. Deformation is obtained.
Alternatively, the bending or twisting method according to claim 2. 4. A load component and a torque component are detected when a torque is applied to a workpiece by using a load device, the load component is eliminated, and only a pure torque is applied to the workpiece. 4. The twist center of the load device is aligned with the twist center of the work by preventing the work from being unnecessarily deformed. Bending or twisting method. 5. A compliance control function is used as a loading device to prevent unnecessary deformation from being given to the workpiece, and to align the twist center of the loading device with the twist center of the workpiece. The bending or twisting method according to claim 4, which is characterized in that. 6. A voice coil motor for applying a torsion torque, a gap sensor group for detecting a position and a twist angle of the voice coil motor including a bit, and a bit for grasping a workpiece. The bit and voice coil motors are held with a low frictional force, and the degree of freedom is provided in the rotational direction of the torsion center axis (Z- axis) and the directions of the X- axis and the Y- axis intersecting the Z- axis. Twisting torque characterized by having an air or magnetic bearing mechanism that it has, capable of simultaneously monitoring the load or torque and the amount of deformation accompanying it in-process, and having the centering function capable of compliance control Load device. 7. A load device for applying a torsion torque, a multi-axis load cell for detecting a load component and a torque component,
A gap sensor for measuring the torsional deformation, a bit for grasping the work piece, a twist center of the work piece for the center of the twist of the work piece in order to remove the load component and perform centering. Twisting torque load device comprising a linear motion table capable of moving the workpiece in two directions in the XY plane intersecting the central axis of the twisting so as to conform to the above. 8. When detecting an elastic constant of a work piece while applying a load or a torque to the work piece, a region from the start of the load to a predetermined load value or torque value is set in advance in the elasticity region of the work piece. 4. The elastic constant of the workpiece is determined by calculating an elastic constant from a data of a load value or a torque value and a deformation amount within the area using a predetermined method. Bending or twisting method described. 9. When calculating the elastic constant, a predetermined threshold value set in the elastic region of the workpiece is changed based on the material characteristics and load history of the workpiece so as to respond to the change of the elastic region of the workpiece. The bending or twisting method according to claim 8, wherein an elastic constant is calculated. 10. The bending or twisting according to claim 8, wherein the load or torque is second-derivatively differentiated by the amount of deformation, and the range in which the value is less than or equal to a predetermined value near 0 is set as the elastic constant detection range. Jig processing method. 11. The bending or twisting method according to claim 10, wherein a range obtained by removing at least one predetermined range at both ends of the detection range is set as the elastic constant detection range. 12. The elastic constant is detected by linearly approximating a load or torque within a predetermined elastic region based on data of a deformation amount and an elastic constant. The bending or twisting method according to item 8. 13. When detecting an elastic constant, a load value or a torque value applied to a workpiece is first-derivatively differentiated by a deformation amount, and a first-order differential value within a preset elastic region is calculated. 9. The bending or twisting method according to claim 8, wherein a first-order average value of the data is obtained and used as an elastic constant of the workpiece. 14. A load or torque within a preset region when the elastic constant is detected to predict the springback amount in a situation where a load or torque is repeatedly applied to a workpiece. Bending or twisting characterized by defining the slope of the trajectory during unloading as the elastic constant based on the hysteresis of the previous load, rather than calculating the elastic constant based on the deformation amount data Method. 15. A method for applying a load or a torque,
The load is applied in steps, and the bending or twisting deformation amount is monitored at each step, whereby each load value or torque value and the deformation amount value are accurately measured. The bending or twisting method according to any one of claims 1 to 3. 16. The bending or twisting process according to claim 15, wherein, when a step load is applied, the load is increased at a constant pitch to measure the amount of deformation for each step. Method. 17. The step load method is characterized in that an optimum step width is obtained from the step width by the differential value of the SF curve, and the load or torque is increased at the step width. Bending or twisting method described. 18. The load or torque step load method, wherein the step width is changed according to the load value or the torque value, and the load is increased at the step width. Bending or twisting method. 19. A method of step load of load or torque, wherein when the predicted final deformation amount approaches a target value to a predetermined value, the step width of load is sufficiently reduced to improve the adjustment accuracy. The bending or twisting method according to claim 15, wherein the bending or twisting method is used. 20. The bending amount according to claim 15, wherein when a load or torque is applied stepwise by using a compliance control function as a load device, a viscous force is applied to stabilize the displacement amount quickly. Or twist processing method. 21. The bending or twisting process according to claim 1, wherein a restoring straight line is obtained by using a correction coefficient for the obtained elastic constant to predict the springback amount of the workpiece. Method. 22. The bending or twisting method according to claim 21, wherein a ratio of a slope of a straight line in an elastic region and a slope of a restored straight line and an offset are used as the correction coefficient. 23. The bending or twisting method according to claim 21, wherein the correction coefficient is provided for each direction by combining bending and twisting directions in one direction and the opposite direction. 24. The bending or twisting process according to claim 21 or 23, wherein when the same workpiece is repeatedly bent or twisted, a correction coefficient is provided for each number of times. Method. 25. As a method of unloading a load or torque, unloading the load or torque in a predetermined step suppresses the vibration of the work piece and reduces the error between the predicted final deformation amount and the actual final deformation amount. The bending or twisting method according to claim 3, wherein the bending or twisting is reduced. 26. After a load or torque is applied to the work piece by a load device having compliance control,
26. The bending or twisting method according to claim 25, wherein the stabilization time of the deformation amount is shortened and the adjustment time of the workpiece is shortened by applying a predetermined viscous force and unloading. 27. The bending or bending according to claim 3, wherein after the load or torque is completely removed from the unloading point, a predetermined amount of load or torque is further applied to prevent or reduce the aftereffect. Twisting method.