SERVO CONTROL DEVICE

Abstract

Servo control device for a numerically controlled machine in which a target unit is driven by a servo motor to control a position or speed of the target unit according to a position command value or a speed command value issued by a host device, wherein the servo control device comprises: - a disturbance suppression system for estimating a disturbance value that is input into the target system and for inputting the estimated value back to a control input, wherein the disturbance suppression system comprises: - a disturbance estimation error calculator for calculating a disturbance estimation error signal based on a detected velocity of the target system and a system model velocity corresponding to a drive force command value for the servo motor, and - a noise suppression control configured by µ-design to improve noise suppression performance associated with a mid to low range of a frequency domain, while preventing the instruction following performance from changing in a high range of the frequency domain, with the noise estimation error signal being amplified in the noise suppression control and fed back to the drive force instruction value.

Description

TECHNICAL AREA

[0001] The present invention relates to a servo control device and in particular a speed control device and a position control device for controlling axes in a numerically controlled machine.

[0002] DE 60 2005 001 651 T2 discloses a device for damping vibrations. INVENTIONAL BACKGROUND

[0003] In general, servo control devices (such as speed control devices and position control devices) used to control the axes in numerically controlled machines must have good system stability (including good vibration suppression) and a high level of instruction sequencing performance and disturbance suppression performance. Fig. Figure 6 shows the general structure of a servo system in one of such servo control devices. According to Fig. In Figure 6, a servo control device forms a feedback system in which a controller C is used and a disturbance d is input into an input u of a target device 55 to generate an output y of the target device 55 based on a command value Y. In this servo system, the command-following performance is improved when a bandwidth (a range in which |T| ≈ 1) of a complementary sensitivity function T = C~P / (1 + C~P), which represents the transfer characteristic from Y → y, becomes wider, while the disturbance suppression performance for suppressing the disturbance d is improved when a value |S| in a mid- to low-frequency range of a sensitivity function S = 1 / (1 + C~P), which represents the transfer characteristic from d → u, becomes smaller.

[0004] Here, the instruction-following performance can be improved by implementing a feedforward feedback system (not shown). However, if the sensitivity function S is minimized in the mid to low frequency range to improve noise suppression performance, the bandwidth of the complementary sensitivity function T inevitably widens considerably. This often leads to problems such as the appearance of vibrations in a high-frequency band and a decrease in system stability. To avoid such problems, a prior art control method has been proposed to suppress only the noise in a smaller loop.A conventional disturbance suppression control method has been proposed, a known control method (hereinafter referred to as the disturbance observer method) in which a target plant of a lower-order model is estimated to estimate a disturbance d that is fed into the target plant and to provide feedback data to a control input to eliminate the disturbance d, so that the disturbance is suppressed.

[0005] Fig. Figure 7 is a block diagram illustrating an example of a conventional position control device employing the disturbance observer method. The target device 55 is driven by a drive motor (not shown), and a control input u of the target device 55 receives a signal obtained by adding the disturbance d to a drive force uc generated by a motor in an adder 55. An output v from the target device 55 represents a motor speed detected by a detector (not shown) or the like, and another output x from the target device 55 consists of a driver position and a motor position detected in a manner similar to that used by the detector.A speed control unit 100 is a speed control device for controlling a motor speed v exactly according to a speed command value Vc and is intended to amplify a speed deviation signal in a speed deviation amplifier Gv, which is obtained by subtracting the motor speed v from the speed command value Vc.

[0006] A disturbance monitor 53 internally stores a lower-order model of the target system 55 and outputs a disturbance estimate ^d, which is calculated from the motor-generated drive force uc and the motor speed v. A drive force command value u0, which is an output of the speed deviation amplifier Gv, is fed into a subtractor 52 as the motor-generated drive force uc, obtained by subtracting the disturbance estimate ^d from the drive force command value u0. To control a driver position x (or a motor position that indirectly indicates the driver position), exactly according to a position command value Xc, a position control device 101 subtracts the driver position x from the position command value Xc in a subtractor 50 to determine a position deviation and amplifies the position deviation in a position deviation amplifier Gp.An amplified output from the position control device 101 is passed on to the speed control unit 100 as the speed command value Vc.

[0007] In the conventional position control device, in which the previously described disturbance observer method is applied, the disturbance estimate ^d is obtained as a precise estimate of the disturbance d, as long as a transfer characteristic ~ P of the target system corresponds to a lower-order transfer characteristic P of the model stored in the disturbance observer, and the disturbance d can be eliminated by feedback of the exact disturbance estimate ^d. In this way, the disturbance suppression performance can be improved without affecting the instruction sequencing performance. However, since the transfer characteristic is generally ~If the input characteristic P of the target system does not correspond to the transmission characteristic P of the lower-order model, especially in a high-frequency band, the disturbance estimate ^d to be returned may contain data from an unintended state feedback loop. This often has a negative impact on command-following performance and leads to vibrations.

[0008] Based on Fig. Section 8 describes in more detail the control features associated with the conventional position control device constructed in this way. Fig. Figure 8 is a block diagram that illustrates in detail an embodiment of a speed control unit 100. A description of the components associated with Fig. The points explained in section 7 will not be repeated here. Here is the transmission characteristic. ~P of the target system is assumed to be a transmission characteristic from the control input u to the motor speed v and is defined as a double inertial system with a transmission pole ωp and a transmission zero ωz, expressed by equation (1) as follows. ~P=(s2+ωz2) / {I1s(s2+2ζωp⋅s+ωp2)}

[0009] Here, I1 represents a lateral drive moment of inertia, I2 represents a lateral load moment of inertia, and ζ represents a damping factor. Then, the lower-order model stored in the perturbation observer is defined as a simple inertial system, and the transfer characteristic P of the model is defined such that it represents the transfer characteristic ~ P of the target system in a low frequency range is met by equation (2) as follows. P=1 / (Is)=1 / {(I1+I2)s} Here, the equivalent moment of inertia is (I1 + I2).

[0010] The operation of outputting the disturbance estimate ^d, performed by the disturbance observer 53, can be expressed in the same way by an operation in which, in a subtractor 56, the motor-generated driving force uc is subtracted from the motor speed v, which is multiplied by an inverse transfer characteristic P-1 of the lower-order model, and the subtracted result is multiplied by a transfer characteristic K0 in a transfer function block 57. The transfer characteristic K0 in the transfer function block 57, which can be expressed using an observer gain ω0 as K0 = I · ω0, serves as the primary low-pass filter characteristic in equation (3) given below. PK0 / (1+PK0)=(ω0 / s) / {1+(ω0 / s)}=ω0 / (s+ω0)

[0011] Here are the parameters associated with the transfer characteristics. ~P are defined as I1 = 0.2 [kgm2], I2 = 0.4 [kgm2], ζ = 0.005, ωp = 628 [rad / s], and ωz = 364 [rad / s], and the velocity deviation amplifier Gv is suitably determined under the condition that the disturbance observer method is not applied (ω0 = 0). Frequency responses (Vc, d → v, v2) of the velocity control unit 100, obtained with the above-mentioned parameters and the velocity deviation amplifier Gv, are shown in Fig. Figure 9 shows that v2 represents the load-side velocity of the target plant 55, expressed in the dual inertial frame, and the velocity deviation amplifier Gv is configured by a commonly used proportional-integral compensator. In comparison, Figure 9 shows that v2 represents the load-side velocity of the target plant 55, expressed in the dual inertial frame, and the velocity deviation amplifier Gv is configured by a commonly used proportional-integral compensator. Fig. 10 the frequency responses (Vc, d → v, v2) obtained when the perturbation observer method is applied with observer gain ω0 chosen as ω0 = 300 [rad / s].

[0012] The interference suppression performance, represented by the frequency responses (d → v, v2), is improved in the mid to low range by applying the interference observer method. However, as shown from Fig. 10 becomes clearly impaired in a frequency band near the transmission zero ωz, in which there is a difference between the transmission characteristic ~ The application of the disturbance observer method maximizes the command following power, which is represented by the frequency response (Vc → v2), for which a gain is increased, leading to the occurrence of vibrations.

[0013] The following explains how the instruction sequencing performance is affected by the disturbance monitoring method. Since the transfer characteristic K0 in the transfer function block 57 in Fig. Since the maximum drop of only -20 dB / dec is shown in equation (3), the transfer function block 57 cannot clip any other growing components than the disturbance contained in an input to the transfer function block 57 in the frequency range in which a system fault, which is the difference between the transfer characteristics ~ P and P is enlarged.

[0014] To record the operation of the interference suppression system in Fig. 8. A noise suppression part is converted in the same way, assuming that u0 = 0. Fig. Figure 11 shows a result of this equivalent conversion. If the transfer characteristic ~If the target system P is expressed by a nominal characteristic of the transfer characteristic P and the system error, the problem of reducing the magnitude of equation (3) with respect to the system error can be replaced by a serious stability problem. This is used in JP H11-24708 A to designate the transfer characteristic K0 with H∞ control design (where K0 is an H∞ infinity control).

[0015] However, in JP H11-24708 A, a speed control unit is constructed as shown in Fig. Figure 12 illustrates how a response to the drive force command value u0 becomes equivalent to a response to the disturbance d in the disturbance suppression system. Therefore, when this disturbance suppression system is introduced, the command sequence characteristic in the mid to low range undergoes a change. To avoid this change, the command sequence performance should not be handled independently of the disturbance suppression performance, which would inevitably require a new design for the velocity deviation amplifier Gv.

[0016] As previously described, the conventional servo control device with its noise suppression system implemented in the smaller loop, while improving noise suppression performance in the mid- to low-range, can negatively impact the command sequence system, introducing vibrations or other adverse effects. Due to these effects, it was impossible to design the command sequence and noise suppression performance independently. Therefore, there is a need for a servo control device with a noise suppression system capable of improving only the noise suppression performance in the mid- to low-range without affecting the command sequence system. SUMMARY

[0017] The problem is solved by the features of the independent claim. The present invention achieves this goal by introducing a noise suppression system into a servo control device, in which a component contained in an output of a target system due to a noise estimation error is amplified in a noise suppression disturbance that has a frequency response corresponding to the magnitude of a system error, which is a difference in the transmission characteristic between the target system and the system model, and is fed back into the control input.

[0018] In the servo control device according to the present invention, a target plant model is used to calculate a disturbance estimation error signal. The disturbance suppression system is configured to amplify the disturbance estimation error signal to a greater extent in a medium to low frequency range, in which the plant error is small enough to detect the disturbance estimation error signal with a high degree of accuracy, and to amplify the disturbance estimation error signal to a lesser extent in a higher frequency range, in which the plant error is large enough to impair the accuracy of the detection of the disturbance estimation error signal, and to return the disturbance estimation error signal, which has been amplified in some way, to the control input.The interference suppression system designed in this way allows for an improvement in interference suppression performance exclusively in the mid to low frequency range, without affecting command-following performance across the entire frequency range. This has the advantageous effects of preventing vibrations caused by the formation of a smaller loop within the interference suppression system, and thus enabling interference suppression performance to be managed independently of command-following performance. Furthermore, since it prevents any changes to characteristics in the high frequency range, it ensures that the stability of the speed control and position control systems is not degraded. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The present invention is described below with reference to the accompanying drawings, in which the same reference numerals refer to the same parts in different views. Fig. Figure 1 is a block diagram illustrating the structure of a position control device according to an embodiment of the present invention; Fig. Figure 2 is a block diagram illustrating a method for designing a fault suppression control according to the present invention; Fig. Figure 3 is a diagram illustrating an example of frequency responses for each block in Fig. 2; Fig. Figure 4 is a diagram illustrating an example of frequency responses of the interference suppression control according to the present invention; Fig. Figure 5 is a diagram illustrating an example of frequency responses of a speed control unit in the position control device. Fig. 1; Fig. Figure 6 is a diagram illustrating a schematic setup of a typical servo control device; Fig. Figure 7 is a block diagram illustrating an example of a conventional position control device in which a disturbance observer method is used; Fig. Figure 8 is a block diagram that shows in detail an exemplary structure of a speed control unit in Fig. 7 shows; Fig. Figure 9 is a diagram showing an example of frequency responses in a conventional speed control unit where no perturbation observer method is used; Fig. Figure 10 is a diagram illustrating an example of frequency responses of a conventional speed control unit in which the perturbation observer method is applied; Fig. Figure 11 is a block diagram illustrating an arrangement obtained by equivalent conversion of a noise suppression unit, as described in Fig. 8 is shown and Fig. Figure 12 is a block diagram illustrating the structure of a conventional speed control device. DESCRIPTION OF THE EXECUTION FORMS

[0020] The present invention is described below with reference to an exemplary embodiment and the drawings. In the present invention, a noise suppression system is also formed in a smaller loop, through which data is fed back to a control input, as is conventionally the case. First, a query for forming the noise suppression system, which does not induce vibrations, is replaced, as described above, by a problem of finding a transfer characteristic K0, which makes it possible to reduce a feedback amount input to the control input in a frequency range in which a system fault, caused by a difference between transfer characteristics, occurs. ~P and P', which is represented, becomes larger. The solution to this problem can prevent a change in instruction sequencing performance in a high-frequency range due to the formation of the noise suppression system. Secondly, an equivalent block, which is in Fig. Figure 11, as described above, is used to consider the interference suppression performance. A query to increase the interference suppression performance in a medium to low frequency range, where the system error is small, is replaced by a problem to find the transfer characteristic K0, which minimizes a transfer characteristic of d → u in the medium to low frequency range. Fig. 11 makes it possible.

[0021] Fig. Figure 2 shows a block diagram obtained by incorporating the two problems, formulated as a robust control problem, into a generalized system G. Fig. In 2, the reference symbol P represents a transfer characteristic of a plant model created by modeling a target plant, and the reference symbol K represents a robust controller with a transfer characteristic corresponding to the transfer characteristic K0 described above. The robust controller K is the disturbance suppression controller according to the present invention. A method for solving the robust control problem to obtain the disturbance suppression controller K with the transfer characteristic that satisfies the above-mentioned requirements is described below. It is noted that M represents a transfer matrix of [w1, w2]T → [z1, z2]T and Mij represents a transfer function of elements in i rows and j columns.

[0022] If the system error is represented as a multiplication variable, a weighted transfer function Wm is used. Fig. 2 that of a phase system with a stable minimum, with a size of the system error which is represented as a multiplication variable of the frequency responses, and the function is determined such that it satisfies the relation expressed by the following equation (4). |Wm|>|1−~P / P|

[0023] Here, a transfer characteristic M22 for w2 → z2 can be expressed by equation (5) as follows. M22=−Wm⋅PK / (1+PK)

[0024] Since W m Since the value increases in the high range, the task of creating a transfer characteristic K that allows equation (5) to be minimized corresponds to the problem described above of preventing the control response power from changing in the high range.

[0025] In the following, a transfer function M11 for w1 → z1 is expressed by equation (6) as follows. M11=Wd⋅1 / (1+PK)

[0026] If the weighted transfer function Wd is chosen to correspond to the phase system with a stable minimum, which shows an increase in the mid to low range, the task of creating the transfer characteristic K that allows a minimization of equation (6) corresponds to the problem of improving the disturbance suppression performance in the mid to low range.

[0027] Since, as described above, the problem in which the two requirements described above are formulated is contained in the transfer matrix M, the disturbance suppression control K, which will be a general solution, is obtained by solving the robust control problem. Equation (7) described below represents a structure of the transfer matrix M that does not contain diagonal elements. [Expression 1] M=[Wd11+PK−WmK1+PKWdP1+PK−WmPK1+PK]

[0028] In a method for solving the robust control problem, it is undesirable that the non-diagonal elements in equation (7) do not constitute a constraint on the formation of the disturbance suppression control K, due to the fact that they are not included in the formulation of the conditions. Accordingly, according to the present invention, the robust control problem is solved using a µ-design such that the non-diagonal elements are multiplied by a scaling factor to minimize the design constraint.

[0029] Equation (8) given below shows the structure of the weighted transfer function W m , which is used in this embodiment and is such that it |1 - ~ P / P| covers the frequency responses. [Expression 2] Wm=(ks2+2ζbωbs+ωb2s2+2ζaωas+ωa2) Here, W is mA high-pass filter and a transfer characteristic of +80 dB / sec are chosen.

[0030] Similarly, equation (9) below shows the structure of the weighted transfer function W. d , which is designed to grow in the medium to low range. [Expression 3] Wd−s2+2ζfωfs+ωf2s2+2ζeωes+ωe2 Here, W is d The low-pass type and -40 dB / dec are chosen for the transmission characteristic.

[0031] Fig. Figure 3 shows the frequency gain functions of the transfer characteristics described above. ~ P and P, |1 - ~ P / P|, and the transfer functions Wm and Wd, which are determined using parameters (k, ζa, ζb, ζe, ζf, ωa, ωb, ωe, ωf) by a trial-and-error process for the target system 55 with the transfer characteristic ∼P, expressed by equation (1), (where I1 = 0.2 [kgm2], I2 = 0.4 [kgm2], ζ = 0.005, ωp = 628 [rad / s], and ωz = 364 [rad / s]), which was used to explain the prior art. It is noted that the transfer characteristic P of the plant model is also determined by defining the lower-order model with the transfer characteristic from equation (2), which was used in explaining the prior art.

[0032] The frequency response of the interference suppression control K, which is generated under the conditions described above, is in Fig. Figure 4 illustrates this. Since the transfer characteristic M11, defined by equation (6), uses a transfer characteristic of the weighted transfer function Wd of -40 dB / dec in the mid- to low-range, the H∞ norm of the transfer characteristic M11 is reduced when the interference suppression control K has a transfer characteristic of -20 dB / dec in the mid- to low-range. On the other hand, in the transfer characteristic M22, defined by equation (5), a transfer characteristic of the weighted transfer function Wm of +80 dB / dec is chosen, which corresponds to a range near the transmission zero ωz of the target system, and the H∞ norm of the transfer characteristic M22 is reduced when the interference suppression control K has a sharply falling curve in this range.

[0033] Fig. Figure 1 is a block diagram illustrating an example of a functional control device, which is the servo control device according to the present invention. Components that differ from the conventional position control device described above are explained below. This position control device 11 is internally equipped with a speed control unit 10, which forms the noise suppression system. The noise suppression controller K (designated by reference numeral 1) according to the present invention is contained in the smaller loop within this system. The system model has a transfer function P, which is modeled with respect to the target system. It receives the drive force command value u0 as input and outputs a model speed vm. A subtractor 3 subtracts the model speed vm from the motor speed v and outputs a subtraction result, which is the noise estimation error signal e described above.Thus, the subtractor 3 acts as the disturbance estimation error calculator for calculating the disturbance estimation error signal. The disturbance estimation error signal e is an input to the disturbance suppression controller K, from which the disturbance estimation value ^d is output according to the present invention. The output disturbance estimation value ^d is subtracted from the drive force command value u0 in a subtractor 4 and passed on as the drive force uc generated by the motor.

[0034] Here, the perturbation estimation error signal e is expressed by equation (10) as follows. [Expression 4] e=P˜{uo+(d−d^)}−Puo=P˜(d−d^)+(P˜−P)uo

[0035] From the definition of the disturbance estimation error (d - ^d), it follows that the disturbance estimation error signal e includes a disturbance estimation error with high accuracy in the medium to lower range where the system error is small, but a disturbance estimation error that is not obtained with high accuracy in the high range where the system error is large. Thus, the disturbance suppression system of the present invention has a disturbance structure in which the amount of feedback returned to the control input is increased, while the degree of accuracy of the detection of the disturbance estimation error is increased.

[0036] Thus, the transfer characteristic of u0, d → u in the disturbance suppression system is expressed by equation (11) as follows. [Expression 5] u=1+PK1+P˜Kuo+11+P˜Kd

[0037] Since the noise suppression control K is an amplifier that performs amplification at a higher gain rate in the mid- to low-range, where the system error is smaller, and at a lower gain rate in the high-range, where the system error is large, an improved noise suppression effect is obtained in the mid- to low-range, so that the influence of the noise d on the control input u is correspondingly attenuated. Equation (11) confirms that, because the control input u is obtained as u ≈ u0 when d = 0 across the entire range, the instruction sequence system is not affected by the noise suppression system.

[0038] Fig. Figure 5 shows frequency responses (Vc, d → v, v2) of the speed control unit 10, which is located in Fig. Figure 1 is shown. Here, the velocity deviation amplifier Gv is defined such that it has the same properties as the one used to achieve the frequency responses described above. Fig. 9 and Fig. 10. The frequency responses (Vc → v, v2), which represent the instruction sequencing power in Fig. The numbers 5 are similar to those that are described in the conventional way, as in Fig. Figure 9 is shown, provided that the perturbation observer method is not used. That is, in contrast to conventional progressions, as in Fig. 10 shown, and obtained through the disturbance observer method, causes the command sequencing performance according to Fig. 5. No vibrations caused by increased gain. Meanwhile, the frequency responses (d → v, v2), which represent the interference suppression performance, become clear from the following. The frequency response (d → v) in the mid to low range can be expressed by the following equation (12). [Expression 6] v=P˜(1+P˜K)+GvP˜(1+PK)d≒1GvPKd

[0039] Since the interference suppression control K has a transfer characteristic of -20 dB / dec in the mid to low range, the transfer characteristics of the frequency responses (d → v, v2) in the mid to low range are steeper than those of the conventional responses in Fig.10, which are obtained by the interference observer method. From equation (12) and the steeper transfer characteristic, it follows that the interference suppression performance is improved. Furthermore, in the middle to low range, where v ≈ v2, the frequency response (d → v2) is essentially identical to the frequency response (d → v).

[0040] As described above, the servo control device according to the present invention has a structure that includes the interference suppression system, which is capable of improving the interference suppression performance in the mid to low frequency range without affecting the command-following performance across the entire frequency range. Since this design prevents changes in the frequency response that can occur in the high frequency range due to the inclusion of the interference suppression system, it is possible to design the command-following performance and the interference suppression performance separately without affecting the stability of the speed control system and the position control system.

Claims

[1] A servo control device for a numerically controlled machine in which a target device is driven by a servo motor to control a position or speed of the target device according to a position command value or a speed command value output from a host device, the servo control device comprising: - a disturbance suppression system for estimating a disturbance value input to the target system and for inputting the estimated value back to a control input, the disturbance suppression system comprising: - a disturbance estimation error calculator for calculating a disturbance estimation error signal based on a detected speed of the target equipment and a equipment model speed corresponding to a driving force command value for the servo motor, and - a disturbance suppression controller configured by µ-design for improving a disturbance suppression performance associated with a middle to low range of a frequency domain while preventing the command following performance from changing in a high range of the frequency domain, wherein the disturbance estimation error signal is amplified in the disturbance suppression controller and fed back to the driving force command value. [2] A servo control device for a numerically controlled machine according to claim 1, wherein the disturbance suppression control performs amplification with a higher gain ratio as the frequency becomes lower. [3] A servo control device for a numerically controlled machine according to claim 1 or 2, wherein: if M22=−Wm.Pk / (1+PK) and M11=Wd.1 / (1+PK), where K represents a transfer characteristic of the disturbance suppression controller, P represents a transfer characteristic of a plant model, Wm represents a weighted transfer function indicating a plant fault, Wd represents a transfer characteristic from w1 → z1 (where w1 is a disturbance value, and z1 is an input to the target plant), the transfer characteristic K enabling minimization of M22 and M11 can be found in the µ-design.

Images