JP2021169128A - Screw fastening device and screw fastening method - Google Patents

Abstract

To provide a screw fastening device and a screw fastening method which can improve precision of axial tension of a screw member as much as possible.SOLUTION: During control of fastening a screw member, a virtual seating point is calculated from torque and a rotation angle, for each preset rotation angle specified value, and a most probable virtual seating point is set from a mode value in an approximate normal distribution of the virtual seating points, and control of fastening the screw member is performed until a rotation angle thereof reaches a rotation angle target value on the basis of the virtual seating point, so that high-precision control of fastening the screw member in accordance with a torque tension method is performed.SELECTED DRAWING: Figure 5

Description

The present invention relates to a screw tightening device and a screw tightening method, particularly a screw tightening device and a screw tightening method capable of improving axial force accuracy.

In the automobile manufacturing industry in recent years, for example, with the downsizing required for a vehicle, the rigidity of various members of the vehicle tends to decrease. When such a vehicle member is fastened with a screw member such as a bolt member or a nut member, an important control item for screw tightening is not the tightening torque but the axial force generated in the bolt member. That is, by managing the axial force of the screw member, for example, it is possible to manage the stress related to the deformation of the member to be fastened.

This axial force is a tensile force generated in a shaft member such as a bolt member. For example, a member whose seat surface of a screw member such as a nut or a bolt head is to be tightened (hereinafter referred to as a tightening target member). It is proportional to the tightening rotation angle of the screw member after contact.

Conventionally, most of the screw member tightening is performed according to a tightening method called a torque angle method. In screw tightening by this torque angle method, the seat surface of the screw member comes into contact with the member to be tightened (hereinafter, also referred to as seating), and after a predetermined torque (also referred to as snag torque) is generated, the screw is screwed by a predetermined rotation angle. The member is rotated and tightened. However, in this torque angle method, the rotation angle of the screw member from the seating of the screw member to the generation of snag torque is unknown.

This is because it is difficult to accurately detect the seating of the screw member by monitoring only the tightening torque of the screw member, and the rotation angle of the screw member from the seating of the screw member to the generation of the snag torque is the screw. This is because it changes depending on the coefficient of friction between the seat surface of the member and the member to be tightened. Therefore, in the torque angle method, it is not possible to accurately control the tightening rotation angle of the screw member from the seating of the screw member. As described above, since the axial force of screw tightening is proportional to the tightening rotation angle of the screw member from the seating of the screw member, it is difficult to improve the axial force accuracy of the screw member by the torque angle method.

On the other hand, in the screw tightening method described in Patent Document 1 below, since it is possible to obtain the seating of the screw member, it is possible to manage the tightening rotation angle of the screw member to a predetermined rotation angle. As a result, it is possible to improve the axial force accuracy of the screw member. This screw tightening method is also called a torque tension method. Specifically, after the snag torque is generated, the screw member is tightened by a specified tightening rotation angle, and the difference between the torque at that time and the snag torque and the difference thereof. The torque increase rate with respect to the screw member rotation angle is obtained from the predetermined tightening rotation angle, and the seating point (defined as a virtual seating point) of the screw member is obtained from this torque increase rate and, for example, the snag torque. This makes it possible to obtain the rotation angle of the screw member from the virtual seating point to the generation of the snag torque, so that the screw member tightening rotation angle from the virtual seating point can be set as the rotation angle target value, and as a result. , The axial force accuracy of the screw member can be improved.

Japanese Unexamined Patent Publication No. 62-102978

However, the torque value detected by a torque detector such as a torque transducer is accompanied by noise, and the noise varies in magnitude depending on, for example, the contact state between the seat surface of the screw member and the member to be tightened. In Patent Document 1, when the screw member is tightened by a specified tightening rotation angle after the snag torque is generated, that is, the torque increase rate is obtained from the torque increase amount and the rotation angle increase amount at a specific time point, and a virtual seating point is obtained. Therefore, if noise is added to the detected torque value, the virtual seating point cannot be accurately obtained, and as a result, the axial force accuracy of the screw member cannot be improved.

The present invention has been made in view of the above problems, and an object of the present invention is to provide a screw tightening device and a screw tightening method capable of improving the axial force accuracy of a screw member as much as possible.

In order to achieve the above object, the screw tightening device of the present invention
The rotation angle of the screw member and the torque of screw tightening at the time of screw tightening are detected, the virtual seating point of the seating surface of the screw member is obtained from the torque increase rate of the torque with respect to the rotation angle, and rotation is performed from the virtual seating point. In a screw tightening device that tightens the screw member to an angle target value,
A normal seating point calculation means that calculates the virtual seating point for each preset rotation angle specified value of the screw member and a normal distribution that approximates the distribution of the virtual seating point calculated by the virtual seating point calculating means with a normal distribution. The distribution approximating means, the virtual seating point setting means for setting the virtual seating point corresponding to the most frequent value of the normal distribution of the virtual seating points approximated by the normal distribution approximating means as the most probable virtual seating point, and the virtual seating point. It is characterized by including a control means for controlling the tightening of the screw member based on a virtual seating point set by the point setting means. The mode value in the normal distribution is the median value and also the average value.

According to this configuration, during the tightening control of the screw member, the virtual seating point is calculated from the torque and the rotation angle for each preset rotation angle specified value, and the most frequent value of the approximate normal distribution of the virtual seating point is used. By setting a probable virtual seating point and performing tightening control of the screw member up to the rotation angle target value based on the virtual seating point, the screw member tightening control according to the torque tension method is performed. At this time, since the noise on the detected torque value originally increases or decreases in the vicinity of the torque value to be detected, the virtual seating point set from the mode value of the approximate normal distribution is also included. As the frequency (number of data) increases, the effect of noise is reduced. Therefore, as the screw member tightening control progresses, the virtual seating point set from the mode of the approximate normal distribution improves in accuracy as the most probable virtual seating point, and is based on this highly accurate virtual seating point. By controlling the tightening of the screw member up to the rotation angle target value, a highly accurate torque tension method is achieved, and the axial force accuracy of the screw member can be improved as much as possible.

Further, in another configuration of the present invention, the virtual seating point setting means replaces the mode of the normal distribution of the virtual seating points on the side closest to the mode and having a value larger than the mode. The virtual seating point corresponding to the intersection of the straight line connecting the plurality of virtual seating point data and the straight line connecting the plurality of virtual seating point data on the side closest to the mode and smaller than the mode is the most. It is characterized by setting a certain virtual seating point.

According to this configuration, even if the calculated distribution of virtual seating points cannot be approximated by a normal distribution, those virtual seating points can be made into a histogram, for example, so that the value closest to the mode value of this histogram is used. By setting the intersection of the straight lines connecting the multiple virtual seating points on the large side and the small side of the virtual seating point as the virtual seating point, the mode becomes the mode of the approximate normal distribution as the screw tightening control progresses, as described above. The closest virtual seating point can be set as the most probable virtual seating point. Therefore, even if the arithmetic processing capacity of the screw tightening device is not so high, the screw tightening control according to the torque tension method can be achieved, and thereby the axial force accuracy of the screw member can be improved as much as possible. ..

A further configuration of the present invention is that the control means presets the torque when the torque increase rate with respect to the preset rotation angle of the screw member is less than the preset torque increase rate specified value. It is characterized in that the screw member is switched to the screw member tightening control for rotating the screw member by a predetermined rotation angle set in advance after the torque becomes a specified value.

According to this configuration, after switching the screw member tightening control, the screw member is rotated by a predetermined rotation angle set in advance after reaching a specified torque value, for example, a snag torque, according to the torque angle method. Tightening control is performed. The virtual seating point is obtained from the torque increase rate with respect to the rotation angle of the screw member. The smaller the friction coefficient between the seat surface of the screw member and the member to be tightened, the smaller the torque increase rate. When this torque increase rate is extremely small, the calculation accuracy of the virtual seating point where the rotation angle of the screw member with respect to the tightening target member after seating is 0 is lowered, and as a result, the screw tightening according to the torque tension method is performed. The axial force accuracy of the threaded member due to the attachment control is reduced. Therefore, when the torque increase rate with respect to the rotation angle of the screw member is less than the specified value of the torque increase rate, the screw tightening control according to the torque angle method can ensure the axial force accuracy of the screw member.

Further, another configuration of the present invention detects the rotation angle of the screw member and the torque of screw tightening at the time of screw tightening, and virtually seats the seat surface of the screw member from the torque increase rate of the torque with respect to the rotation angle. In a screw tightening method in which a point is obtained and the screw member is tightened from the virtual seating point to a rotation angle target value.
A normal distribution that approximates the distribution of the virtual seating points calculated in the virtual seating point calculation step for calculating the virtual seating point for each preset rotation angle specified value of the screw member and the virtual seating point calculation step. The distribution approximation step, the virtual seating point setting step that sets the virtual seating point corresponding to the most frequent value of the normal distribution of the virtual seating points approximated by the normal distribution approximation step as the most probable virtual seating point, and the virtual seating point. It is characterized by including a control step for controlling tightening of the screw member based on a virtual seating point set in the point setting step.

According to this configuration, during the tightening control of the screw member, the virtual seating point is calculated from the torque and the rotation angle for each preset rotation angle specified value, and the most frequent value of the approximate normal distribution of the virtual seating point is used. By setting a probable virtual seating point and performing tightening control of the screw member up to the rotation angle target value based on the virtual seating point, the screw member tightening control according to the torque tension method is performed. At this time, since the noise on the detected torque value originally increases or decreases in the vicinity of the torque value to be detected, the virtual seating point set from the mode value of the approximate normal distribution is also included. As the frequency (sampling number) increases, the effect of noise is reduced. Therefore, as the screw member tightening control progresses, the virtual seating point set from the mode of the approximate normal distribution improves in accuracy as the most probable virtual seating point, and is based on this highly accurate virtual seating point. By controlling the tightening of the screw member up to the rotation angle target value, a highly accurate torque tension method is achieved, and the axial force accuracy of the screw member can be improved as much as possible.

As described above, according to the present invention, the influence of noise of the detected torque value is reduced as the screw member tightening control progresses, the most probable virtual seating point is set with high accuracy, and the torque tension with high accuracy is achieved. According to the law, the axial force accuracy of the screw member can be improved as much as possible. As a result, it becomes possible to promote the downsizing of the vehicle without any trouble.

It is a schematic block diagram which shows one Embodiment of the screw tightening apparatus and the screw tightening apparatus to which the screw tightening method of this invention is applied. It is explanatory drawing of the screw tightening method performed by the screw tightening apparatus of FIG. It is explanatory drawing of the torque value detected by the torque detector of FIG. It is explanatory drawing when slip of a screw member occurs by the screw tightening method of FIG. It is a flowchart of the arithmetic processing performed by the control device of FIG. It is explanatory drawing of the operation of the arithmetic processing of FIG. It is explanatory drawing of another example which sets the most probable virtual seating point in the arithmetic processing of FIG. It is explanatory drawing of the conventional screw tightening method.

Hereinafter, an embodiment of the screw tightening device and the screw tightening method of the present invention will be described in detail with reference to the drawings. FIG. 1 is a schematic configuration diagram of a screw tightening device according to this embodiment, and is used, for example, for assembling a vehicle engine. The member configuration of this screw tightening device is the same as or almost the same as that of the conventional screw tightening device. This screw tightening device is, for example, a socket 10 that is fitted on the head of a bolt member or a nut member and rotates the screw member, an electric motor 12 that rotates the socket 10, and a screw member by the electric motor 12. A torque detector 14 such as a torque transducer that detects the tightening torque and a rotation angle sensor 16 that detects the rotation angle of the screw member are provided, and the electric motor 12 is driven by the control device 18. Further, the tightening torque detected by the torque detector 14 and the rotation angle detected by the rotation angle sensor 16 are read by the control device 18.

The control device 18 is configured to include a computer system (not shown) such as a microprocessor. Similar to a well-known computer system, this computer system has, in addition to an arithmetic processing unit having an advanced arithmetic processing function, for example, a storage device for storing a program, reading a sensor signal, and intercommunication with other control devices. It is configured with an input / output device for doing things. Computer systems used in the automobile manufacturing industry include, for example, programmable logic controllers.

Before explaining the screw tightening method performed by the screw tightening device of this embodiment, the screw tightening method by the conventional torque angle method will be described with reference to FIG. The torque angle method, the screw member is seated on the target member tightening, the detected torque value of the tightening preset torque prescribed value, for example, only the snug torque rotation angle phi ₀ with a predetermined clamping after reaching the ST screw Rotate the member to complete the tightening. The detected torque value shall be generated from the time when the screw member is seated on the member to be tightened. For example, assuming that the straight line shown by the solid line in the figure is the average (intermediate value) tightening torque-rotation angle characteristic, this characteristic is, for example, between the seat surface of the screw member and the member to be tightened (hereinafter, seat). It changes according to the coefficient of friction (also referred to as between faces) (of course, the coefficient of friction of the thread also affects it).

For example, when the coefficient of friction between the seat surfaces is large (high μ in the figure) as shown by the alternate long and short dash line in FIG. 8, the rotation angle of the screw member from the seating of the screw member to the generation of snag torque is small. On the other hand, when the friction coefficient between the seat surfaces is small (low μ in the figure) as shown by the alternate long and short dash line in the figure, the rotation angle of the screw member from the seating of the screw member to the generation of snag torque is large. As will be described later, the actual detected torque value is not a simple straight line (curve) as shown in the figure, but is accompanied by fine noise, and when the screw member is seated, the waveform is often disturbed. It is difficult to determine that the screw member is seated. For example, in the case when the friction coefficient larger and a small, if there is an angular difference Δφ on the threaded member rotation angle until snug torque occurs, by rotating the rotation angle phi ₀ only threaded member with the predetermined tightening when completed the tightening screw , The actual rotation angle of the screw member from the seating differs by this angle difference Δφ. Since the axial force of the screw member depends on the tightening rotation angle from the seating of the screw member, if the tightening rotation angle at the completion of tightening differs by the angle difference Δφ, the axial force of both differs accordingly. Therefore, it is difficult to improve the axial force accuracy of the screw member by tightening the screw by the torque angle method.

FIG. 2 is an explanatory diagram of the principle of the screw tightening method performed by the screw tightening device of this embodiment, and specifically, is a screw tightening method by the torque tension method. In the screw tightening method based on this torque tension method, when the screw member is seated on the member to be tightened and the detected torque value of the tightening torque reaches the snag torque ST, the screw member is rotated by the _{initial value φ i for increasing the rotation angle.} Then, the difference value between the detected torque value and the snag torque at that time _{is divided by the rotation angle increase initial value φ i} to obtain the torque increase rate per unit rotation. _{As is clear from the figure, the torque value reached when the screw member is rotated by the initial value φ i, which} increases the rotation angle from the arrival of the snag torque, is when the coefficient of friction between the seat surfaces is large (high μ) and when it is small (low μ). ), But the torque increase rate obtained by dividing the difference value between the detected torque value and the snag torque at that time by the initial value φ _i for increasing the rotation angle reflects the friction coefficient between the seating surfaces, and the friction coefficient is The larger it is, the larger it is, and the smaller it is, the smaller it is.

Therefore, if each tightening torque-rotation angle curve (straight line) is traced back by this torque increase rate, the detected torque value becomes 0, that is, the rotation angle of the screw member with respect to the member to be tightened after sitting is 0. A (straight line) section can be obtained, and this is the virtual seating point. Therefore, graphically, the rotation angle of the screw member from the seating of the screw member to the generation of the snag torque can be obtained from this virtual seating point and, for example, the snag torque. Further, by dividing the snag torque by the torque increase rate, the rotation angle of the screw member from the seating of the screw member to the generation of the snag torque can be directly obtained. _{Since the axial force of the screw member can be defined by the rotation angle target value φ T} from the seating of the screw member, the rotation angle of the screw member from this rotation angle target value φ _T to the generation of the snag torque and the rotation angle increase initial value φ _The rotation angle obtained by subtracting i is the tightening rotation angle of the remaining screw member to be rotated.

However, the torque value detected by the torque detector 14 has high frequency noise. FIG. 3 is an explanatory diagram of a torque value detected during screw tightening control, and as is clear from the figure, the detected torque value fluctuates with fine noise. The screw tightening control by the torque tension method is based on the torque increase rate at a specific time when the screw member is rotated by the _{rotation angle increase initial value φ i from the snag torque generation, so that the detection at that time is detected.} If the torque value includes noise, both the calculated torque increase rate and the virtual seating point become uncertain due to the influence of the noise. However, as can be inferred from FIG. 3, the noise on the detected torque value only increases or decreases in the vicinity of the true value of the torque value that should be originally detected, and the noise is leveled. If so, the detected torque value will be certain. That is, for example, a large number of virtual seating points can be calculated, and the larger the frequency of the distribution, the more probable virtual seating points can be set.

FIG. 4 schematically shows a part of the actual waveform of the tightening torque-rotation angle curve of FIG. 2, for example. The fine waveform in the figure shows the noise on the detected torque value. In this torque curve, a large temporary fluctuation appearing in the tightening torque-rotation angle curve having a large friction coefficient (high μ in the figure) represents the slip of the screw member. If the torque increase rate is calculated during the slip occurrence, the tightening torque-rotation angle curve obtained by the torque increase rate will be as shown by the alternate long and short dash line in the figure. For example, since this is equivalent to that of the tightening torque-rotation angle curve with a small friction coefficient (low μ) shown by the alternate long and short dash line in the figure, the virtual seating point obtained by using the torque increase rate. Is different from the actual virtual seating point. However, since the slip occurrence rate of the screw member is extremely small, even if a virtual seating point different from the actual virtual seating point is calculated due to slipping, the frequency is extremely small in the normal distribution of the calculated virtual seating point, which is substantially the same. It is not counted as a virtual seating point that seems to be probable.

On the other hand, as inferred from FIG. 2, the virtual seating point where the torque increase curve (straight line) is traced back based on the torque increase rate and the rotation angle of the screw member with respect to the tightening target member after seating is 0 is The smaller the torque increase rate, that is, the smaller the coefficient of friction between the seat surfaces, the larger the variation and the lower the calculation accuracy. Further, the smaller the coefficient of friction between the seating surfaces, the smaller _{the rotation angle of the screw member left from the generation of the snag torque to the rotation angle target value φ T.} Therefore, the number of virtual seating points calculated during that period is also small, and the virtual seating is performed. Since the frequency of the mode of the normal distribution of points is also small, the certainty of the virtual seating point set from the mode is also reduced. Therefore, in this embodiment, the torque increase rate when snag torque is generated is obtained, and if this torque increase rate is less than the preset torque increase rate specified value, the torque tension method is stopped and the screw is screwed by the torque angle method. Tighten the member. The torque angle method has higher axial force accuracy of the screw member than the torque tension method in which the virtual seating point of the screw member remains ambiguous. Therefore, the torque increase rate specified value corresponds to the torque increase rate when the friction coefficient between the seat surfaces is so small that it is difficult to accurately obtain the virtual seating point.

FIG. 5 is a flowchart of an arithmetic process performed by the control device 18 in order to control screw tightening of a screw member. This arithmetic processing is performed in the control device 18, for example, 1 deg. It is executed for each rotation angle of the screw member. The rotation angle of the screw member detected by the rotation angle sensor 16 such as the encoder and the torque detected by the torque detector 14 such as the transducer are read into the control device 18 in a very short sampling cycle and stored in the above memory. Stored in the device. Further, in this flowchart, the calculated data is plotted on a distribution map and the plotted data is approximated by a normal distribution, but these processes are actually performed only inside the arithmetic processing unit and the storage device. It may be done in.

This arithmetic processing is started, for example, at the start of tightening of the screw member, and first, in step S1, the detected values of the rotation angle and torque of the screw member are read. Hereinafter, the rotation angle of the screw member is simply referred to as a rotation angle.

Next, the process proceeds to step S2 to determine whether or not the control flag F is in the reset state of 0. If the control flag F is in the reset state of 0, the process proceeds to step S3. If not, the process proceeds to step S3. The process proceeds to step S6.

In step S3, it is determined whether or not the snag torque ST is detected in the detected torque value, and if the snag torque ST is detected, the process proceeds to step S4, and if not, the process returns.

In step S4, the rotation angle when the snag torque ST is detected is set to the rotation angle reference value, and then the process proceeds to step S5.

In step S5, the process proceeds to step S6 after calculating the torque increase rate with respect to the rotation angle and torque stored in the storage device and the rotation angle from the snag torque detection to the detection.

In step S6, it is determined whether or not the torque increase rate calculated in step S5 is equal to or higher than the preset torque increase rate specified value, and if the torque increase rate is equal to or higher than the torque increase rate, step S7 is performed. If not, that is, if the torque increase rate is less than the specified torque increase rate, the process proceeds to step S19. The specified value of the torque increase rate is the torque increase rate when the friction coefficient between the seat surfaces is so small that it is difficult to accurately obtain the virtual seating point.

In step S7, the control flag F is set to 1, and then the process proceeds to step S8.

In step S8, it is determined from the rotation angle reference value _{whether or not the rotation angle has increased by the preset rotation angle increase initial value (φ i} ) or more, and the rotation angle increases by the rotation angle increase initial value or more. If so, the process proceeds to step S9, and if not, the process returns.

In step S9, the rotation angle current value and the rotation angle previous value, and the torque current value and the torque previous value are set. Specifically, the currently detected torque value is used as the current torque value. Further, the rotation angle reference value is set to 0, and the rotation angle from that to the currently detected rotation angle is set as the rotation angle this time value. In addition, the rotation angle of the screw member that is earlier than the current rotation angle (rotation angle current value) by the initial value for increasing the rotation angle is set to the rotation angle previous value, and the rotation angle is detected by the rotation angle of the previous rotation angle. Set the torque value to the previous torque value.

Next, the process proceeds to step S10, including the current rotation angle previous value, and all rotation angles (= rotation) previously set to the rotation angle previous value between the rotation angle previous value and the above rotation angle reference value. After obtaining the rotation angle difference value between the angle previous value) and the rotation angle current value, the process proceeds to step S11.

In step S11, the current torque previous value is included, and all torque values (= torque previous value) set to the torque previous value in the past from the torque previous value to the snag torque are combined with the torque current value. After obtaining the torque difference value, the process proceeds to step S12.

In step S12, the virtual seating point is calculated from all the corresponding rotation angle difference values and torque difference values calculated in steps S10 and S11, and then the process proceeds to step S13. Specifically, all the rotation angle difference values were divided by the corresponding torque difference values to obtain the inverse ratio of the torque increase rate (slope) with respect to the rotation angle, and the inverse ratio of the torque increase rate was multiplied by the current torque value. The value is subtracted from the rotation angle this time to obtain the virtual seating point. Depending on the computing power of the control device 18, the number of virtual seating points obtained for each arithmetic processing may be limited (however, the number of data plotted on the distribution map described later is not limited).

In step S13, the virtual seating points obtained in step S12 are plotted on the distribution map, and then the process proceeds to step S14.

In step S14, the data of the virtual seating points plotted on the distribution map in step S13 is approximated by a normal distribution, and then the process proceeds to step S15. This product distribution approximation of the virtual seating point data can be performed in the above computer system using a well-known application.

In step S15, the most probable virtual seating point in the approximate normal distribution is calculated and set according to individual arithmetic processing (not shown), and then the process proceeds to step S16. In this example, it is assumed that a large number of calculated virtual seating points can be approximated by a normal distribution, so the virtual seating points corresponding to the mode values of this normal distribution are set as the most probable virtual seating points. As is well known, in the normal distribution, the mode is the median and the mean.

In step S16, it is determined whether or not the rotation angle increase amount of the virtual seating point and rotation angle current value set in step S15 _{has reached the rotation angle target value (φ T} ), and the rotation angle increase amount is increased. If the rotation angle target value is reached, the process proceeds to step S17, and if not, the process returns.

In step S17, the process proceeds to step S18 after the screw member tightening control according to the torque tension method is completed.

In step S18, the control flag F is reset to 0 and then returned.

On the other hand, in step S19, the screw tightening control by the torque angle method is performed according to a calculation process (not shown), and then the process proceeds to step S20.

In step S20, it is determined from the rotation angle reference value (snag torque generation rotation angle) whether or not the screw member is tightened by a predetermined rotation angle, and when the screw member is tightened by a predetermined rotation angle from the rotation angle reference value, If not, the process proceeds to step S19.

According to this arithmetic processing, if the rotation angle increases by the initial value of the increase amount of the rotation angle from the generation of the snag torque, for example, 1 deg. For each of the above-mentioned rotation angle specified values set in, the difference value between all the rotation angles set in the past rotation angle previous value and the rotation angle current value, and all the torque values set in the past torque previous value. The virtual seating point is calculated based on all the differences between the torque and the torque current value. The large number of calculated virtual seating points are plotted on the distribution map and further approximated by a normal distribution. As shown in FIG. 6, for example, the data of the virtual seating points approximated by this normal distribution gradually increases as the screw tightening control progresses, and the virtual seating points vary due to the influence of noise on the torque value. Fluctuations in point values also converge as the number of data (frequency) accumulates, reducing the effects of noise. Therefore, as the number of calculated virtual seating point data increases, the certainty of the virtual seating point set from the mode value of the normal distribution also increases. As a result, in the above arithmetic processing, the virtual seating point when the rotation angle target value is achieved is the most probable virtual seating point, and the screw member rotated by the rotation angle target value from this most probable virtual seating point. Axial force accuracy is the highest.

In the above explanation, it is assumed that the arithmetic processing capacity of the control device is high and a large number of calculated virtual seating point data are approximated by a normal distribution. However, when the arithmetic processing capacity of the control device is not so high, , It may be difficult to approximate those many virtual seating point data with a normal distribution. In such a case, for example, the above-mentioned large number of virtual seating point data are represented in a histogram, and as shown in FIG. 7, a plurality of virtuals that are closest to the mode of the histogram and have a value larger than the mode. The intersection of the straight line connecting the seating point data and the straight line connecting the plurality of virtual seating point data whose values are smaller than the mode may be set as the most probable virtual seating point. For example, a well-known least squares method can be used to set a straight line connecting the virtual seating point data.

Further, in the above calculation process, as described above, the torque increase rate when the snag torque is generated (detected) is calculated, and when the torque increase rate is less than the preset torque increase rate specified value, the torque is increased. The screw tightening control is performed by switching from the tension method to the torque angle method. The axial force accuracy of the screw member is higher when the screw member is tightened by the torque angle method than when the screw member is tightened by the torque tension method while the virtual seating point is ambiguous.

As described above, according to the screw tightening device of this embodiment, the virtual seating point is calculated from the torque and the rotation angle for each preset rotation angle specified value during the tightening control of the screw member, and the virtual seating point is calculated. The torque tension method is followed by setting the most probable virtual seating point from the mode of the approximate normal distribution of points and performing tightening control of the screw member to the rotation angle target value based on the virtual seating point. Screw member tightening control is performed. At this time, since the noise on the detected torque value originally increases or decreases in the vicinity of the torque value to be detected, the virtual seating point set from the mode value of the approximate normal distribution is also included. As the frequency (number of data) increases, the effect of noise is reduced. Therefore, as the screw member tightening control progresses, the virtual seating point set from the mode of the approximate normal distribution improves in accuracy as the most probable virtual seating point, and is based on this highly accurate virtual seating point. By controlling the tightening of the screw member up to the rotation angle target value, a highly accurate torque tension method is achieved, and the axial force accuracy of the screw member can be improved as much as possible.

Also, even if the calculated distribution of virtual seating points cannot be approximated by a normal distribution, those virtual seating points can be made into a histogram, for example. By setting the intersection of the straight lines connecting the plurality of virtual seating points on the smaller side as the virtual seating points, the virtual seating points close to the mode of the approximate normal distribution as the screw tightening control progresses, as described above. Can be set as the most probable virtual seating point. Therefore, even if the arithmetic processing capacity of the screw tightening device is not so high, the screw tightening control according to the torque tension method can be achieved, and thereby the axial force accuracy of the screw member can be improved as much as possible. ..

Further, when the torque increase rate with respect to the rotation angle of the screw member is less than the torque increase rate specified value, the screw member tightening control is switched, and after the screw member tightening control is switched, the torque specified value, for example, snag torque is obtained. After that, the screw member is tightened according to the torque angle method, in which the screw member is rotated by a predetermined rotation angle set in advance. As a result, the axial force accuracy of the screw member can be improved as much as possible even if the torque tension method is not as high as the original torque tension method.

Although the screw tightening device and the screw tightening method according to the embodiment have been described above, the present invention is not limited to the configuration described in the above embodiment and varies within the scope of the gist of the present invention. It can be changed. For example, the rotation angle and torque value used to calculate the virtual seating point are not limited to those at the above timing.

10 Socket 12 Electric motor 14 Torque detector 16 Rotation angle sensor 18 Control device

Claims (4)

The rotation angle of the screw member and the torque of screw tightening at the time of screw tightening are detected, the virtual seating point of the seating surface of the screw member is obtained from the torque increase rate of the torque with respect to the rotation angle, and rotation is performed from the virtual seating point. In a screw tightening device that tightens the screw member to an angle target value,
A virtual seating point calculation means for calculating the virtual seating point for each preset rotation angle specified value of the screw member, and
A normal distribution approximating means that approximates the distribution of virtual seating points calculated by the virtual seating point calculating means with a normal distribution, and
A virtual seating point setting means that sets a virtual seating point corresponding to the mode of the normal distribution of the virtual seating points approximated by the normal distribution approximating means as the most probable virtual seating point, and a virtual seating point setting means.
A screw tightening device comprising: a control means for controlling tightening of the screw member based on a virtual seating point set by the virtual seating point setting means. The virtual seating point setting means is a straight line connecting a plurality of virtual seating point data on the side closest to the mode and having a value larger than the mode, instead of the mode of the normal distribution of the virtual seating point. And the virtual seating point corresponding to the intersection with the straight line connecting the plurality of virtual seating point data on the side closest to the mode and smaller than the mode is set as the most probable virtual seating point. The screw tightening device according to claim 1, wherein the screw tightening device is characterized. When the torque increase rate with respect to the preset rotation angle of the screw member is less than the preset torque increase rate specified value, the control means becomes the preset torque specified value. The screw tightening device according to claim 1 or 2, wherein the screw member is switched to a screw member tightening control that rotates the screw member by a predetermined rotation angle set in advance. The rotation angle of the screw member and the torque of screw tightening at the time of screw tightening are detected, the virtual seating point of the seating surface of the screw member is obtained from the torque increase rate of the torque with respect to the rotation angle, and rotation is performed from the virtual seating point. In the screw tightening method for tightening the screw member to the angle target value,
A virtual seating point calculation step for calculating the virtual seating point for each preset rotation angle specified value of the screw member, and a virtual seating point calculation step.
A normal distribution approximation step that approximates the distribution of virtual seating points calculated in the virtual seating point calculation step with a normal distribution, and
A virtual seating point setting step that sets the virtual seating point corresponding to the mode of the normal distribution of the virtual seating points approximated by the normal distribution approximation step as the most probable virtual seating point, and a virtual seating point setting step.
A screw tightening method comprising: a control step for controlling tightening of the screw member based on a virtual seating point set in the virtual seating point setting step.

Images