JP6349093B2 - Motor control device and correction data creation method in the same - Google Patents

Description

  The present invention is obtained by generating a position detection signal of a rotating shaft based on a signal output from a sensor such as a resolver or an encoder connected to the rotating shaft of the motor, and further correcting the position detection signal with correction data. The present invention relates to a motor control apparatus that controls a motor by feeding back a corrected position signal, and a correction data generation method that generates correction data used in the apparatus.

  In the motor control technology, it is necessary to detect the rotational position of the rotary shaft in order to obtain a positional deviation between the position command and the actual rotational position of the rotary shaft of the motor. One of the position detecting means is a resolver. Here, when a resolver is used, a position correction technique that takes into account the position detection error of the resolver is required. Therefore, for example, in the devices described in Patent Document 1 and Patent Document 2, correction data is stored in a memory in advance, and a position detection signal created based on a signal output from the resolver is corrected with correction data during motor control. Thus, a corrected position signal is created, and the motor is feedback-controlled using the corrected position signal.

Japanese Patent No. 2541169 Japanese Patent No. 5281102

  In Patent Document 1 and Patent Document 2, correction data is created from a position detection signal detected by a resolver. However, the correction data is not necessarily highly accurate, and position detection is performed using the correction data. Even if the signal is corrected, it is difficult to control the motor with high accuracy.

  The present invention has been made in view of the above problems, and can accurately create correction data used in a motor control device that controls a motor based on a signal output from a sensor such as a resolver or an encoder. The purpose is to improve the accuracy of control.

The first aspect of the present invention generates a position detection signal related to the rotational position of the rotating shaft based on a signal output from the sensor corresponding to the rotational angle of the rotating shaft of the motor rotating in accordance with the position command signal. A speed command signal is generated based on the corrected position signal and the position command signal obtained by correcting the position detection signal with the correction data, and the motor driving speed and the speed command signal obtained by differentiating the corrected position signal. Is a correction data generation method for generating correction data in a motor control device that feedback-controls a motor by a current command signal generated based on the motor, and a motor having a driving pattern including a constant speed region in which a rotating shaft rotates at a constant speed rotation time command information a current command signal from the sampling data obtained by sampling during the rotation shaft while driving to rotate one revolution or more at a constant speed region A first step of Tokusuru, by multiplying the similarity ratio on the rotation time of command information and a second step of creating correction data, in the first step, the motor is fed back without correcting the position detection signal with the correction data The DC ripple is removed from the sampling data obtained by sampling the current command signal while rotating at least one revolution in the constant speed region while driving, and the current ripple is obtained as rotation command information. In the second step, the similarity ratio candidate A plurality of similarity ratio candidate values different from each other are prepared, and for each similarity ratio candidate value, a corrected position detection signal is obtained based on data obtained by multiplying the similarity ratio candidate value by a current ripple. A sample obtained by sampling the current command signal while rotating one or more revolutions in the constant speed range while driving the motor by feeding back the position signal Characterized in that calculated as a current ripple amplitude difference between the maximum value and the minimum value of the current ripple from the data by removing the DC component, to set a homothetic ratio candidate value having the minimum value among the plurality of current ripple amplitude as similarity ratio It is said.
According to a second aspect of the present invention, a position detection signal related to the rotational position of the rotating shaft is generated based on a signal output from the sensor in accordance with the rotational angle of the rotating shaft of the motor rotating in accordance with the position command signal. Generate a speed command signal based on the corrected position signal and position command signal obtained by correcting the position detection signal with the correction data, and the motor driving speed and speed obtained by differentiating the corrected position signal A correction data generation method for generating correction data in a motor control device that feedback-controls a motor based on a current command signal generated based on a command signal, the driving pattern including a constant speed region in which a rotating shaft rotates at a constant speed Command during rotation from the sampling data obtained by sampling the current command signal while rotating the motor one or more times in the constant speed range while driving the motor A first step of acquiring information and a second step of generating correction data by multiplying the rotation command information by the similarity ratio, and in the first step, the motor is driven with an outward drive pattern including a constant speed region. While rotating the rotating shaft in the constant speed range, the DC rotation is removed from the positive rotation sampling data obtained by sampling the current command signal during one or more positive rotations in the constant speed range. The reverse rotation current ripple is obtained by removing the DC component from the reverse rotation sampling data obtained by sampling the current command signal while the rotation axis rotates backward or more in the constant speed region while driving the motor with the return path drive pattern including The average current ripple for each rotation angle obtained by calculating the average of the forward rotation current ripple and the reverse rotation current ripple for each rotation angle of the rotation shaft is used as the rotation command information. In the second step, data obtained by preparing a plurality of different similarity ratio candidate values as similarity ratio candidates and multiplying the similarity ratio candidate value by the rotation time command information for each similarity ratio candidate value. A positive rotation current ripple is obtained by feeding back the corrected position signal obtained by correcting the position detection signal based on the feedback and driving the motor back and forth, and removing the DC component from the sampling data obtained by sampling the current command signal during the forward drive. The reverse rotation current ripple is obtained by removing the DC component from the sampling data obtained by sampling the current command signal during the backward drive, and the average of the forward rotation current ripple and the reverse rotation current ripple is calculated for each rotation angle of the rotating shaft. The difference between the maximum and minimum values of the average current ripple for each rotation angle obtained by calculation is calculated as the current ripple amplitude, and multiple current ripples are calculated. It is characterized in that the similarity ratio candidate value that is the minimum value among the amplitudes is set as the similarity ratio.

According to a third aspect of the present invention, a position detection signal related to the rotational position of the rotating shaft is generated based on a signal output from the sensor in accordance with the rotational angle of the rotating shaft of the motor rotating in accordance with the position command signal. Generate a speed command signal based on the corrected position signal and position command signal obtained by correcting the position detection signal with the correction data, and the motor driving speed and speed obtained by differentiating the corrected position signal a motor control apparatus for feedback control of the motor by the current command signal is generated based on the command signal, it is characterized by Ru a correction data creation unit that creates correction data by the correction data generation method.

  In the invention configured as described above, the command signal is sampled while the rotating shaft rotates at least once in the constant speed region. Then, rotation command information is obtained from the sampling data. As will be described in detail later, this rotational command information has a similar relationship with the dynamic actual error, and correction data can be created with high accuracy by multiplying the rotational command information by the similarity ratio. Become. Further, the accuracy of motor control is improved by performing feedback control using the correction data thus created.

  As described above, according to the present invention, the correction data to be used in the motor control device is created by multiplying the rotation command information by the similarity ratio, so that the correction data can be created with high accuracy. It is possible to improve the accuracy of motor control.

It is a figure which shows the structure of the motor control apparatus used in order to acquire the knowledge used as the foundation of this invention, the relationship between the dynamic real error detected by the said apparatus, and a current ripple, and a static real error. It is a figure which shows the electrical constitution of 1st Embodiment of the motor control apparatus concerning this invention. 3 is a flowchart showing a correction data creation operation in the motor control device shown in FIG. 2. It is a figure which shows the rotational speed pattern of the motor in correction data creation in 1st Embodiment. It is a graph which shows an example of a normal rotation current ripple. It is a graph which shows offset adjustment operation | movement. It is a graph which shows an example of similarity ratio matching operation. It is a figure which shows the relationship between the static real error recognized in the apparatus shown to Fig.1 (a), and the dynamic real error at the time of forward rotation. It is a figure which shows the relationship between the static real error recognized in the apparatus shown to Fig.1 (a), and the dynamic real error at the time of reverse rotation. It is a figure which shows the relationship between the dynamic actual error at the time of forward rotation, the dynamic actual error at the time of reverse rotation, and the static actual error when the motor is reciprocated at the same constant speed in the apparatus shown in FIG. It is a flowchart which shows the correction data creation operation | movement in 2nd Embodiment. It is a figure which shows the rotational speed pattern of the motor in correction data creation. It is a graph which shows an example of the current ripple at the time of forward rotation and reverse rotation. It is a graph which shows an example of phase alignment operation. It is a graph which shows an example of similarity ratio matching operation. It is a graph which shows offset adjustment operation | movement. It is a graph which shows the improvement of a static real error by using correction data.

A. Similarity Relationship between Resolver Position Detection Error and Current Command The inventor of the present application conducted various experiments using the motor control device having the configuration shown in FIG. I gained the knowledge that a relationship exists. The motor control device 100 used in the experiment here is a robot controller, and the motor M is rotated by the motor control device 100. As shown in FIG. 1A, a reference position detector E capable of detecting a position with higher accuracy than the resolver Re is attached to the rotating shaft of the motor M. Detection signals from the resolver Re and the reference position detector E are output to the motor control device 100, respectively. In addition, the code | symbol PC in the figure (a) is a personal computer connected to the motor control apparatus 100 via the serial communication interface (RS232C).

FIG. 1B is a diagram showing the relationship between the dynamic actual error and the current command when the resolver Re is used. This experimental result was sampled while rotating the motor M at a constant speed of 30 [rps]. The dynamic actual error in FIG. 5B means the difference between the detection position of the resolver measured in the constant speed operation state and the reference position detector. In this embodiment, the dynamic actual error is determined from the detection position of the resolver Re at each rotation angle. A value obtained by subtracting the detection position of the reference position detector E is used, and the position detection error of the resolver Re is indicated by this dynamic actual error. The current command indicates the AC component of the current command value in the constant speed operation state, that is, the current ripple. In this embodiment, the current value of the current command signal sampled during one rotation section is used as sampling data, A section average is obtained, and a current ripple is obtained by subtracting the section average from the current value of the current command signal in the section. The horizontal axis in FIG. 5B, that is, the rotation angle is obtained by converting the detection position of the reference position detector E into an angle.

  As is clear from FIG. 5B, the dynamic actual error, which is the position detection error of the resolver Re, has a similar relationship with the AC component (current ripple) of the current command value. Therefore, the inventor of the present application can estimate the dynamic actual error from the current ripple, create correction data from the estimation error, and improve the position accuracy by correcting the position detection signal of the resolver Re with the correction data. The knowledge that it was possible to get it. Here, the above knowledge is obtained based on the sampling result sampled while rotating the motor M at a constant speed of 30 [rps], but the speed range including the 30 [rps], specifically The present inventor has confirmed that the same knowledge as described above holds even in the speed range of 15 [rps] to 100 [rps].

  FIG. 6C shows a static actual error. The “static actual error” means a difference between the detection angle of the resolver Re and the reference position detector E measured in a stationary state. Here, the point to be noted in the static actual error shown in FIG. 4C is that the static actual error is zero when the rotation angle is zero. Therefore, the present inventor further considers that the static actual error becomes zero at the rotation angle of zero in addition to estimating the dynamic actual error from the current ripple as described above when generating the correction data. As a result, it has been found that the accuracy of the correction data can be further improved.

  Based on such knowledge, the inventor of the present application obtains a current ripple from sampling data obtained by sampling a command signal while the rotation shaft of the motor M rotates at a constant speed, and multiplies the current ripple by an appropriate similarity ratio. It was concluded that correction data can be created with high accuracy by combining them. Hereinafter, embodiments of the invention for creating correction data using the above knowledge will be described with reference to the drawings.

B. First Embodiment FIG. 2 is a diagram showing an electrical configuration of a first embodiment of a motor control device according to the present invention. The motor control device 1 includes a motor drive unit 2 that drives a motor M based on a position command signal, and a correction data creation unit 3 that creates correction data. In the following, the configuration of the correction data creation unit 3 will be described after the configuration of the motor drive unit 2 is described first.

  In the motor drive unit 2, the position control unit 21 outputs a speed command signal based on a difference between a position command signal given from the outside and a corrected position signal described later. The speed control unit 22 outputs a current command signal based on the difference between the driving speed of the motor M calculated by differentiating the corrected position signal with the differentiator 23 and the speed command signal. This current command signal is output to the correction data creation unit 3 described later and also to the current control unit 24. The current control unit 24 outputs a drive command signal to the motor M via the current sensor 25 based on the current command signal. More specifically, the current sensor 25 detects the drive command signal output from the current control unit 24 and feeds back a current detection signal indicating the current value of the drive command signal to the input side of the current control unit 24. Then, the current control unit 24 outputs a drive command signal based on the difference between the current detection signal and the current command signal. And the motor M which received the drive command signal outputs the torque according to this drive command signal.

  In addition, a resolver Re is connected to the rotation shaft (not shown) of the motor M, and the rotation position of the rotation shaft is detected and an analog signal indicating the rotation position is output. This analog signal is input to the RD converter 26, converted into a digital position detection signal, and output. A signal corrected by adding a correction signal generated by the correction data generating unit 3 described below to the position detection signal is the corrected position signal. As described above, the position control unit 21 and the differentiator 23 receive the corrected signal. Given.

  Next, after describing the configuration of the correction data creation unit 3, a method for creating correction data will be described. The correction data creation unit 3 includes an arithmetic processing unit 31 having an arithmetic function such as a CPU (Central Processing Unit) and a storage unit 32 having four memories 321 to 324. The arithmetic processing unit 31 samples a current command signal output from the speed control unit 22 in accordance with a program stored in advance while the motor M is reciprocatingly driven by the motor driving unit 2 with a predetermined driving pattern. Next, a current ripple acquisition process, an offset adjustment process, and a similarity ratio adjustment process, which will be described in detail, are executed to create correction data. Thus, in the present embodiment, the arithmetic processing unit 31 functions as the current command value sampling unit 311, the current ripple acquisition unit 312, the similarity ratio matching unit 313, and the offset matching unit 314.

  FIG. 3 is a flowchart showing a correction data creation operation in the motor control device shown in FIG. In the present embodiment, when the motor M or the resolver Re is attached or exchanged or when the creation of correction data is instructed by the user, the motor M is incorporated in the apparatus, for example, the motor M is driven by the robot. When used as a source, correction data creation processing described below is executed while the movable part of the robot (moving mechanism part such as a ball screw) is connected to the rotation axis of the motor M.

  In step S1, the distance for moving the robot by operating the motor M by the motor driving unit 2 and the speeds V0 + and V0− in the constant speed region are set. With these settings, when the correction data creation process is performed, the motor M drives the robot and the operation pattern in which the robot operates, that is, the motor M drive pattern by the motor driving unit 2 is determined. In addition, a plurality of operation modes are prepared in advance as the operation mode of the motor drive unit 2, but in the correction data creation process, a correction signal is not first output to the motor drive unit 2, that is, correction by correction data is performed. “No correction operation” is selected to drive the motor M (step S2), and current ripple acquisition processing is executed (steps S3 to S5).

  In this current ripple acquisition process, driving of the motor M by the motor driving unit 2 is started to reciprocate the robot, and current sampling is performed during the forward movement (step S3). That is, when the motor M starts to be driven, for example, as shown in FIG. 4, the motor M rotates forward, and the moving speed of the robot is accelerated to the speed V0 + over a predetermined time, according to the moving distance set in step S <b> 1. After maintaining the constant speed V0 + for the time, deceleration is started and the forward rotation of the motor M is stopped. In this way, moving the robot by rotating the motor M forward is referred to as “outward movement”, and the drive pattern of the motor M for such movement is referred to as “outward drive pattern” and the forward drive pattern. A region where the constant speed V0 + is maintained in is referred to as a “constant speed region”. In addition, at almost the intermediate position of the constant speed region during the forward movement, the current command signal output from the speed control unit 22 of the motor drive unit 2 during one rotation of the motor M is sampled at every rotation angle. The current command value is written in the current sampling memory 321 as “sampling data”.

  In the next step S4, the current command value for one rotation is read from the current sampling memory 321, and the average value is calculated. Then, for each rotation angle, the average value is subtracted from the current command value to obtain a positive rotation current ripple, and the current sampling memory 321 is overwritten. Thus, the positive rotation current ripple that is the AC component of the current command value during the forward movement is acquired (FIG. 5).

  Further, since the static actual error at the rotation angle of zero is zero, offset adjustment is performed (step S5). That is, of the values for each rotation angle of the rotating shaft constituting the current ripple, the value where the rotation angle is zero is used as the offset data, and the offset data is subtracted from the value for each rotation angle (for example, in FIG. 6). And is written in the correction source table in the memory 322 (step S6).

When the above-described current ripple acquisition process is completed and “similarity ratio matching operation” is selected as the operation mode (step S7), the similarity ratio matching process described below is performed from the start timing T1 of the next reciprocating drive (FIG. 4). Are executed (steps S8 to S15). In this similarity ratio matching process, first, various initial settings are performed. In the present embodiment, the initial setting as one of the value obtained by subtracting the minimum value from the maximum value of current ripple one rotation sets a "1" as an initial value of the count value n, that is the minimum value of the current ripple amplitude RWmin is set to a default value (“10000” in this embodiment) (step S8). In the next step S9, the similarity ratio is set to a value (n / 20). For example, when n = 1, the similarity ratio is set to an initial value (1/20).

  After setting the similarity ratio (n / 20), the offset matched current ripple is read from the correction source table memory 322, and a product obtained by multiplying the offset matched current ripple by the similarity ratio is output to the motor drive unit 2 as a correction signal. . Then, while performing correction using the correction signal, the motor driving unit 2 starts driving the motor M to reciprocate the robot, and current sampling is performed during the movement (step S10). The reciprocating movement operation and the current sampling operation are the same as those in the current ripple acquisition process (step S3) except that the correction signal is used for correction, and sampling data for one rotation is stored in the current sampling memory 321. Written. For example, when n = 1, current sampling is performed for one rotation substantially in the middle of the constant speed region of the forward path driving pattern while driving the motor M from timing T1 to timing T2 in FIG.

  When the current sampling is completed, the current ripple is calculated and written into the similarity ratio matching memory 323 (step S11). That is, the current ripple is calculated in the same manner as in step S4. Then, the current ripple when the similarity ratio is set to (n / 20) is written to the similarity ratio matching memory 323.

  In the next step S12, after obtaining the current ripple amplitude from the current ripple for one rotation written in the similarity ratio matching memory 323, it is determined whether or not the current ripple amplitude is less than the minimum value RWmin. Then, when “YES” is determined in step S12, that is, when it is determined that the current ripple amplitude at the count value n is less than the minimum value RWmin, the minimum value RWmin is updated to the current ripple amplitude and the count value n is optimized. After the value Kbest is written in the similarity ratio matching memory 323 (step S13), the process proceeds to step S14. Note that the storage destination of the optimum value Kbest is not limited to the similarity comparison memory 323, and can be written to an arbitrary memory. On the other hand, if “NO” is determined in the step S12, the minimum value RWmin and the optimum value Kbest are not updated, and the process directly proceeds to a step S14.

  After the count value n is incremented by “1” in step S14, it is determined whether or not the count value n is “21” (step S15). Then, while it is determined as “NO” in step S15, that is, the count value n is “20” or less, the process returns to step S9 to repeat the similarity ratio derivation process (steps S9 to S14) that minimizes the current ripple amplitude. . That is, in the present embodiment, (1/20), (2/20),..., (20/20) are prepared as candidates for similarity ratio, and the similarity ratio derivation process is performed for each similarity ratio candidate. In addition, the optimum value Kbest is determined.

On the other hand, when it is determined as "YES" in the step S15, the process proceeds to step S 16 in FIG. In step S 16, it reads out the optimum value Kbest from similarity ratio registration memory 323 to determine the optimal similarity ratio of current ripple amplitude is minimized to (= Kbest / 20). Then, the offset-matched current ripple is read from the correction source table memory 322, and the similarity-matched data is created by multiplying the value for each rotation angle constituting the offset-matched current ripple by the optimum similarity ratio and correcting it. Data is written in the correction table memory 324 (step S 16 ). Thereby, more accurate correction data can be obtained.

  As described above, according to the present embodiment, correction data used in the motor control device 1 is created based on the knowledge described in the section “A. Similarity between position detection error of resolver and current command”. Therefore, the correction data can be created with high accuracy, and the control accuracy of the motor M can be improved.

  In addition, as a method for creating correction data, a method has been proposed in which a reference position detector is attached and an error from a position measured by the reference position detector is used as correction data. A correction facility for preparation is separately required. On the other hand, in the said embodiment, correction data can be created by the original function of the motor control apparatus 1 without using the said correction equipment at all, and own performance can be improved. That is, the motor control device 1 according to the above embodiment has a self-correction function. Further, the correction data can be created with the motor M attached to the robot. Therefore, even when the motor M is replaced or repaired at a factory or the like where the robot is installed, correction data can be created after replacement and repair on the spot.

  Further, in the above embodiment, the correction data is created using the current command signal after the speed command signal is amplified by the speed control unit 22, and the resolution can be sufficiently high, and the corrected position accuracy is achieved by the prior art. Than can be raised. Here, the correction data may be created using the current detection signal instead of the current command signal, and the same effect as the above embodiment can be obtained.

  Thus, in the present embodiment, the resolver Re corresponds to an example of the “sensor” of the present invention. In the above embodiment, the current ripple having been offset-adjusted is obtained by the current ripple acquisition process and the offset adjustment process, and these processes correspond to an example of the “first step” of the present invention. The similarity ratio matching process corresponds to an example of the “second step” in the present invention. In this similarity ratio matching process, data obtained by multiplying the similarity ratio (n / 20) by the offset matched current ripple corresponds to an example of the “correction data candidate” of the present invention, and the operation of acquiring the data is the present invention. The “correction data candidate creation step” of FIG. 5 and an operation of driving the motor M by feedback control with the data correspond to an example of the “motor drive step” of the present invention.

  Further, the offset adjusted current ripple obtained in step S6 corresponds to an example of “rotational command information” of the present invention. As for the “rotational command information” of the present invention, a positive rotation current ripple may be used as the “rotational command information” of the present invention. In this case, the current ripple acquisition process corresponds to an example of the “first step” of the present invention. In addition, since the offset matching process is not executed before the similarity ratio matching process is executed, it is desirable to use the offset matched data obtained by performing the offset matching process on the similarity matched data as the correction data. . In the above embodiment, the positive rotation current ripple is acquired from the sampling data obtained by performing current sampling during the forward rotation, and the offset adjusted current ripple and the similarity ratio completed data are obtained. Instead of the rotation sampling data, sampling data obtained by performing current sampling during reverse rotation may be used. That is, correction data may be created based on the reverse rotation current ripple.

C. Second Embodiment As described above, in the first embodiment, the correction data is generated using the forward rotation current ripple or the reverse rotation current ripple, but the correction data may be generated using both. Thereby, the “rotational command information” of the present invention can be obtained with high accuracy, and the accuracy of the correction data can be further increased. The reason is that the error increasing direction and the phase changing direction at the time of forward rotation and reverse rotation are opposite to each other, and the dynamic actual error can be estimated with higher accuracy from the forward rotation current ripple and the reverse rotation current ripple. Because it can. This is another finding obtained by an experiment by the motor control device having the configuration shown in FIG. 1A, and accuracy can be improved by using this. Hereinafter, after describing the “other knowledge”, a second embodiment of the present invention using the “other knowledge” will be described.

  The inventor of the present application measures the dynamic actual error while rotating the motor M forward at various constant speeds (15 [rps], 30 [rps], 60 [rps]) in the motor control device of FIG. The measurement results are summarized in FIG. 8 together with static actual errors. Further, the dynamic actual error when the motor M was rotated in the reverse direction was also measured, and the measurement results are summarized in FIG. 9 together with the static actual error.

  FIG. 8 is a diagram showing the relationship between the static actual error recognized in the apparatus shown in FIG. 1A and the dynamic actual error during forward rotation. FIG. 9 is a diagram showing the relationship between the static actual error recognized in the apparatus shown in FIG. 1A and the dynamic actual error during reverse rotation. Furthermore, FIG. 10 shows the relationship between the dynamic actual error during forward rotation, the dynamic actual error during reverse rotation, and the static actual error when the motor is reciprocated at the same constant speed in the apparatus shown in FIG. FIG.

  At the time of forward rotation, as shown in FIG. 8, the phase is delayed as the speed increases, and the (−) side error is increased. On the other hand, at the time of reverse rotation, as shown in FIG. 9, since the rotation angle of the horizontal axis is scanned from 360 ° to 0 °, the phase is delayed as the speed increases, and the error on the (+) side is also increased. It has increased. One of the main reasons can be attributed to the RD converter built in the motor control device 100. This “RD converter” is a resolver-digital converter and has a function of converting the output signal of the resolver into digital angle data. Therefore, due to the delay system of the RD converter, it is considered that the error in the detection result of the resolver Re increases and the phase tends to be delayed with respect to the detection result of the reference position detector E without delay.

  Further, when the dynamic actual error when the motor M is normally rotated at a constant speed, for example, 60 [rps], the dynamic actual error when the motor M is reversely rotated at the same speed, and the static actual error are plotted on the same graph, FIG. As shown in Fig. 3, the error increasing direction and the phase changing direction are opposite in the forward and reverse rotations, and the average of the dynamic actual error in the forward rotation and the dynamic actual error in the reverse rotation is the static actual error. Is approximate. That is, while the motor M is reciprocatingly driven at the same constant speed, the dynamic actual error at the time of forward rotation and the dynamic actual error at the time of reverse rotation are obtained, and the values are averaged for each rotation angle. The inventor of the present application has obtained the knowledge that the actual error can be approximately obtained. 8 to 10, only the measurement result of 15 to 60 [rps] is shown, but the same measurement result is obtained for 60 to 100 [rps], and the above knowledge is established.

  Based on such knowledge, the inventor of the present application uses a forward drive pattern including a constant speed region in which the rotation shaft of the motor M rotates forward at a constant speed and a return path drive pattern including a constant speed region in which the motor M rotates backward at the same speed. It was concluded that the correction data can be created with high accuracy from the dynamic actual error during forward rotation and the dynamic actual error during reverse rotation while the motor M is driven reciprocally. Hereinafter, a second embodiment of the invention for creating correction data using the above knowledge will be described with reference to FIGS.

  The configuration of the motor control device 1 in the second embodiment is basically the same as that of the first embodiment, and only part of the processing executed by the arithmetic processing unit 31 is different. That is, the arithmetic processing unit 31 of the correction data generating unit 3 is output from the speed control unit 22 according to a program stored in advance while the motor M is reciprocatingly driven with a predetermined drive pattern by the motor driving unit 2. The current command signal is sampled, and correction data is created by executing a phase matching process, a similarity ratio matching process, and an offset matching process, which will be described in detail below. Thus, in the present embodiment, the arithmetic processing unit 31 functions as the current command value sampling unit 311, the phase matching unit 315, the similarity ratio matching unit 313, and the offset matching unit 314.

  FIG. 11 is a flowchart showing the correction data creation operation in the second embodiment. Also in this embodiment, as in the first embodiment, when the motor M or resolver Re is attached or exchanged, or when the creation of correction data is instructed by the user, the motor M is incorporated in the apparatus. When the motor M is used as a driving source for the robot, for example, the correction data creation process described below is executed while the robot's movable part (moving mechanism part such as a ball screw) is connected to the rotating shaft of the motor M. Is done.

Similar to the first embodiment, the distance for moving the robot and the speeds V0 + and V0− in the constant speed region are set in step S21. However, in the second embodiment, | V0 + | = | V0− |
Is set to hold. Subsequently, “no correction operation” is selected as the operation mode of the motor drive unit 2 (step S22), and phase matching processing is executed (steps S23 to S26).

  In this phase matching processing, the motor M is started to be driven by the motor drive unit 2 to reciprocate the robot, and sampling of forward rotation and reverse rotation is performed during the movement (step S23). In other words, when the motor M starts to be driven, as shown in FIG. 12, for example, the motor M rotates forward, and the moving speed of the robot is accelerated to the speed V0 + over a predetermined time, and according to the moving distance set in step S21. After maintaining the constant speed V0 + for the time, deceleration is started and the forward rotation of the motor M is stopped. In this way, the current command signal output from the speed control unit 22 of the motor drive unit 2 during one rotation of the motor M is sampled at a constant time interval at a substantially intermediate position in the constant speed region during driving with the forward drive pattern. Then, the current command value for each rotation angle is written in the current sampling memory 321 as “positive rotation sampling data”.

  In the next step S24, normal rotation sampling data is read from the current sampling memory 321, and an average value of current command values for one rotation is calculated. Then, for each rotation angle, the average value is subtracted from the current command value to obtain a positive rotation current ripple, and the current sampling memory 321 is overwritten. Thus, the positive rotation current ripple that is the AC component of the current command value during the forward movement is acquired (solid line in FIG. 13).

  On the reverse rotation side, the reverse rotation current ripple is obtained in the same manner as the normal rotation side (step S25). That is, the reverse rotation sampling data is read from the current sampling memory 321, the average value of the current command value for one rotation is calculated, the reverse rotation current ripple is obtained by subtracting the average value from the current command value for each rotation angle, The current sampling memory 321 is overwritten. Thus, the reverse rotation current ripple that is the AC component of the current command value during the backward movement is acquired (dotted line in FIG. 13).

  When the normal rotation current ripple and the reverse rotation current ripple are obtained, the normal rotation current ripple and the reverse rotation current ripple are read from the current sampling memory 321 for each rotation angle, and the average value of these values is calculated. Then, the average value (= (forward rotation current ripple + reverse rotation current ripple) / 2) is written as an average current ripple in the correction source table in the memory 322 (step S26).

  When the phase matching process described above is completed (timing T1 in FIG. 12), “similarity matching operation” is selected as the operation mode (step S27), and the similarity matching process similar to that of the first embodiment is executed (step S27). Steps S28 to S36). Here, the point in which the similarity ratio matching process in the second embodiment is greatly different from that in the first embodiment is that the similarity ratio matching is performed on the average current ripple. That is, in the similarity ratio matching process of the second embodiment, various initial settings are performed as in the first embodiment (step S28). In the next step S29, the similarity ratio is set to a value (n / 20). For example, when n = 1, the similarity ratio is set to an initial value (1/20).

  After setting the similarity ratio (n / 20), the average current ripple is read from the correction source table memory 322, and a value obtained by multiplying the similarity ratio by the average current ripple for each rotation angle is output to the motor drive unit 2 as a correction signal. Then, while performing the correction by the correction signal, the driving of the motor M by the motor driving unit 2 is started to reciprocate the robot, and current sampling is performed during the movement (step S30). Except for the point of performing correction by the correction signal, the reciprocating movement operation, forward rotation and reverse rotation current sampling operations are the same as those in the phase matching process (step S23), and the forward rotation for one rotation is performed. Sampling data and reverse rotation sampling data for one rotation are written in the current sampling memory 321. For example, when n = 1, the motor M is driven with a continuous forward drive pattern and a backward drive pattern between timing T1 and timing T2 in FIG. Sampling is performed.

  When the current sampling is completed, the current ripple is calculated and written to the similarity ratio matching memory 323 (step S31). That is, the forward rotation current ripple is calculated in the same manner as in step S4, the reverse rotation current ripple is calculated in the same manner as in step S5, and the average current ripple is calculated in the same manner as in step S6. Then, the average current ripple when the similarity ratio is set to (n / 20) is written to the similarity ratio matching memory 323.

  In the next step S32, after obtaining the current ripple amplitude from the average current ripple for one rotation written in the similarity ratio matching memory 323, it is determined whether or not the current ripple amplitude is less than the minimum value RWmin. Then, when “YES” is determined in step S32, that is, when it is determined that the current ripple amplitude at the count value n is less than the minimum value RWmin, the minimum value RWmin is updated to the current ripple amplitude and the count value n is optimized. After the value Kbest is written in the similarity ratio matching memory 323 (step S33), the process proceeds to step S34. Note that the storage destination of the optimum value Kbest is not limited to the similarity comparison memory 323, and can be written to an arbitrary memory. On the other hand, when “NO” is determined in the step S32, the minimum value RWmin and the optimum value Kbest are not updated, and the process directly proceeds to a step S34.

After the count value n is incremented by “1” in step S34, it is determined whether or not the count value n is “21” (step S35). Then, "NO" in the step S35, that is, between the judges that the count value n is equal to or less than "20", the derivation of similarity ratio as the current ripple amplitude minimum returns to Step S 2 9 (step S29～S35) repeat. That is, also in this embodiment, (1/20), (2/20), ..., (20/20) are prepared as candidates for similarity ratio, and the similarity ratio derivation process is performed for each candidate for similarity ratio. After that, the optimum value Kbest is determined.

  On the other hand, if “YES” is determined in the step S35, the optimum value Kbest is read from the similarity ratio matching memory 323, and the optimum similarity ratio (= Kbest / 20) at which the current ripple amplitude is minimized is determined. Then, the average current ripple is read from the correction source table memory 322, and a value obtained by multiplying the average current ripple for each rotation angle by the optimum similarity ratio is written in the correction table memory 324 as similarity ratio completed data (step S36).

  In this way, by adjusting the average current ripple with the optimum similarity ratio, similar ratio matched data (solid line in FIG. 15) approximated to the static actual error (dotted line in FIG. 15) is obtained. As described above, the similarity ratio-matched data is offset from the static actual error, and the static actual error when the rotation angle is 0 ° is zero as shown by the dotted line in FIG. Finally, offset adjustment is performed (step S37). That is, the similarity-matched data with the rotation angle of zero is subtracted from the similarity-matched data for each rotation angle, and the offset-matched data is used as final correction data (for example, a solid line in FIG. 16) to the correction table memory 324. Overwrite to. Thereby, more accurate correction data can be obtained.

  As described above, according to the second embodiment, in addition to the knowledge described in the above-mentioned section “A. Similar relationship between position detection error of resolver and current command”, the above “other knowledge” is considered. Since the correction data used by the motor control device 1 is created, not only the same effect as the first embodiment can be obtained, but also the correction data can be created with higher accuracy, and the control accuracy of the motor M is improved. Can be further improved.

  As described above, in the second embodiment, the average current ripple obtained in step S26 corresponds to an example of “rotational command information” of the present invention. Here, as in the first embodiment, the offset-adjusted current ripple obtained by performing the offset adjustment processing on the average current ripple may be used as “rotational command information” of the present invention. In this case, step S37 is not necessary, and the similarity-matched data calculated in step S36 can be used as correction data.

  In the second embodiment, the phase matching process corresponds to an example of the “first step” of the present invention. The similarity ratio matching process corresponds to an example of the “second step” in the present invention. In this similarity ratio matching process, data obtained by multiplying the similarity ratio (n / 20) by the average current ripple corresponds to an example of the “correction data candidate” of the present invention, and the operation of acquiring the data is “ An operation of driving the motor M by performing feedback control with the data corresponds to an example of the “correction data candidate creation step” and corresponds to an example of the “motor drive step” of the present invention.

C. Others The present invention is not limited to the above-described embodiment, and various modifications can be made to the above-described one without departing from the gist thereof. For example, in the above embodiment, the present invention is applied to the motor control device 1 that controls the motor M using a signal output from the resolver Re. However, the present invention is applicable to position detection sensors other than the resolver Re, such as an encoder. The present invention can also be applied to a motor control device used as a “sensor”.

  In the above embodiment, the sampling period of the current command signal is one rotation of the motor M, but may be sampled exceeding one rotation. For example, when the current command signal for m rotations is sampled, m sampling data may be averaged for each rotation angle to derive “forward rotation current ripple” or “reverse rotation current ripple”.

  In the above embodiment, as shown in FIGS. 3 and 11, the pair of sampling operations are repeated 20 times while switching the similarity ratio candidates during the similarity ratio matching process. However, the number of repetitions (similarity ratio candidates) The number) is not limited to this and is arbitrary.

  Furthermore, although the case where the motor M is used as a driving source of the robot has been described in the above embodiment, the application target of the present invention is not limited to the technique for controlling the motor attached to the robot. That is, the present invention can also be applied to a motor control device that controls a motor attached to an industrial device such as a surface mounter, a component inspection device, or a semiconductor manufacturing device, in which a movable part is driven by a motor.

  The present invention can be applied to a correction data generation technique for generating correction data for correcting a signal output from a sensor connected to a rotating shaft of a motor and a motor control technique for controlling a motor using the correction data. It is.

DESCRIPTION OF SYMBOLS 1 ... Motor control apparatus 3 ... Correction data creation part 31 ... Operation processing part M ... Motor Re ... Resolver

Claims (7)

A position detection signal related to the rotation position of the rotary shaft is generated based on a signal output from the sensor corresponding to the rotation angle of the rotary shaft of the motor that rotates in accordance with the position command signal, and the position detection signal is corrected. A speed command signal is generated based on the corrected position signal obtained by correcting with data and the position command signal, and the motor driving speed and the speed command signal obtained by differentiating the corrected position signal are obtained. A correction data creating method for creating the correction data in a motor control device that feedback-controls the motor by a current command signal generated based on the current command signal,
Sampling obtained by sampling the current command signal while driving the motor with a driving pattern including a constant speed region in which the rotating shaft rotates at a constant speed while the rotating shaft rotates one or more times in the constant speed region. A first step of obtaining rotational command information from the data;
A second step of creating the correction data by multiplying the rotation command information by a similarity ratio ;
In the first step, the position detection signal is fed back without being corrected by the correction data, and is obtained by sampling the current command signal during one rotation or more in the constant speed region while driving the motor. DC component is removed from the sampling data, and the current ripple is obtained as the rotation command information.
In the second step, a plurality of different similarity ratio candidate values are prepared as candidates for the similarity ratio, and data obtained by multiplying the similarity ratio candidate value by the current ripple for each similarity ratio candidate value Sampling data obtained by sampling the current command signal while rotating one or more revolutions in the constant speed region while driving the motor by feeding back the corrected position signal obtained by correcting the position detection signal based on The difference between the maximum value and the minimum value of the current ripple is calculated as a current ripple amplitude by removing the DC component from the current ripple, and the similarity ratio candidate value that is the minimum value among the plurality of current ripple amplitudes is set as the similarity ratio. A correction data creation method characterized by the above. The correction data creation method according to claim 1 ,
The current command signal sampling is a correction data creation method that is performed for each rotation angle of the rotation shaft. The correction data creation method according to claim 2 ,
The sensor is a resolver;
In the first step, of the values for each rotation angle of the rotation shaft constituting the current ripple, a value where the rotation angle is zero is used as offset data, and the offset adjustment is performed by subtracting the offset data from the value for each rotation angle. A correction data creation method using completed data as the rotation command information. A position detection signal related to the rotation position of the rotary shaft is generated based on a signal output from the sensor corresponding to the rotation angle of the rotary shaft of the motor that rotates in accordance with the position command signal, and the position detection signal is corrected. A speed command signal is generated based on the corrected position signal obtained by correcting with data and the position command signal, and the motor driving speed and the speed command signal obtained by differentiating the corrected position signal are obtained. A correction data creating method for creating the correction data in a motor control device that feedback-controls the motor by a current command signal generated based on the current command signal,
Sampling obtained by sampling the current command signal while driving the motor with a driving pattern including a constant speed region in which the rotating shaft rotates at a constant speed while the rotating shaft rotates one or more times in the constant speed region. A first step of obtaining rotational command information from the data;
A second step of creating the correction data by multiplying the rotation command information by a similarity ratio ;
In the first step, the current command signal is obtained by sampling the current command signal while driving the motor with the forward drive pattern including the constant speed region while the rotating shaft rotates forward one or more times in the constant speed region. The forward rotation current ripple is obtained by removing the direct current component from the rotation sampling data, and the rotation axis is reversed by one rotation or more in the constant speed region while the motor is driven by the backward drive pattern including the constant speed region. A reverse rotation current ripple is obtained by removing a direct current component from reverse rotation sampling data obtained by sampling the current command signal during rotation, and the forward rotation current ripple and the reverse rotation current for each rotation angle of the rotation shaft. Obtain the average current ripple for each rotation angle obtained by calculating the average of ripples as the rotation command information,
In the second step, a plurality of different similarity ratio candidate values are prepared as candidates for the similarity ratio, and for each similarity ratio candidate value, the similarity ratio candidate value is obtained by multiplying the rotation command information. The corrected position signal obtained by correcting the position detection signal based on the obtained data is fed back to drive the motor back and forth, and the DC component is obtained from the sampling data obtained by sampling the current command signal during the forward drive. The positive rotation current ripple is obtained by removing the direct current component from the sampling data obtained by sampling the current command signal during the backward drive, and the reverse rotation current ripple is obtained, and the forward rotation is performed for each rotation angle of the rotary shaft. Of the average current ripple for each rotation angle obtained by calculating the average of the current ripple and the reverse rotation current ripple, the maximum value and the minimum value It was calculated as the current ripple amplitude, the correction data generation method a homothetic ratio candidate value having the minimum value among the plurality of current ripple amplitude and said <br/> be set as the similarity ratio. The correction data creation method according to claim 4 ,
In the second step, the similarity ratio-matched data in which the rotation angle is zero among the similarity ratio-matched data for each rotation angle of the rotating shaft is set as offset data, and the offset from the similarity ratio-matched data for each rotation angle A correction data creation method in which offset-adjusted data obtained by subtracting data is used as the correction data. The correction data creation method according to claim 1, 2, 4 or 5 ,
The sensor is a resolver;
The correction data creation method, wherein the constant speed is 15 [rps] to 100 [rps]. A position detection signal related to the rotation position of the rotary shaft is generated based on a signal output from the sensor corresponding to the rotation angle of the rotary shaft of the motor that rotates in accordance with the position command signal, and the position detection signal is corrected. A speed command signal is generated based on the corrected position signal obtained by correcting with data and the position command signal, and the motor driving speed and the speed command signal obtained by differentiating the corrected position signal are obtained. A motor control device that feedback-controls the motor by a current command signal generated based on the command,
Motor control device according to claim wherein the Ru a correction data creation unit that creates correction data <br/> it by the correction data generating method according to any one of claims 1 to 6.

Images