JP2013118817A - Motor control apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform motor drive control with high robustness without adjusting various parameters even in the case where both an inertia moment and an inertia mass or a multi-inertia model are unknown or may be varied in a drive mechanism.SOLUTION: In a current control system loop, a disturbance torque estimated by an inertia variation suppression section 3 using a first disturbance observer 31 is added to a first torque command and a second torque command is outputted. Furthermore, the first disturbance observer 31 estimating the disturbance torque in the inertia variation suppression section 3 includes a resonance suppression model 43 which outputs an acceleration/deceleration torque signal of an ideal motor suppressing influences caused by mechanical resonance of a drive mechanism 500 in accordance with a motor speed of the motor 200. The first disturbance observer 31 utilizes the acceleration/deceleration torque signal of the ideal motor to estimate the disturbance torque, such that the inertia variation suppression section 3 suppresses variation in inertia moment of the drive mechanism without vibration caused by mechanical resonance of the drive mechanism 500.

Description

  The disclosed embodiment relates to an electric motor control device that controls an electric motor.

  Conventionally, a schematic configuration of an AC motor control device, that is, an AC motor servo controller, includes a position control system feedback loop, a speed control system feedback loop, a PWM loop power conversion, and dq coordinates. It has a triple loop configuration including a current feedback loop including conversion. In the case of the current control system, since the torque of the motor is generated in proportion to the current, a torque command for the speed control system is input to the current control system.

  In such a control system, it is necessary to manually adjust various parameters such as a gain and a time constant in accordance with a specific driving mechanism whose configuration is known in advance. Further, depending on the configuration and use method of the drive mechanism, the moment of inertia ratio and the inertia model may fluctuate during the operation, and parameter adjustment needs to be performed so as to widely cope with such fluctuation. Conventionally, the extremely complicated and complicated parameter adjustment as described above has been a major obstacle to the use of the motor control system. On the other hand, in recent years, even when the inertia moment ratio between the rotor of the motor and the load mechanism is unknown or fluctuates, the response of the control system is made constant without identifying the inertia moment ratio by auto-tuning. "Disturbance suppression control" that can be maintained has been proposed and studied. In general, the fact that the fluctuation of the response characteristic of the control system with respect to the fluctuation of the parameter is small is expressed as “high robustness” or “the control system is robust”. This disturbance suppression control system estimates the disturbance torque from the first torque command and the actual detected speed of the motor, adds this to the first torque command, and uses it as a second torque command to be input to the current control system. Consists of fluctuation suppression unit.

  In Patent Document 1, in order to improve robustness even for a larger moment of inertia ratio, a first phase compensator that inputs an initial speed command and outputs a new speed command that advances the phase, A technique is provided in which a second phase compensation unit is provided that inputs a detected speed and a torque command and outputs a new speed command with advanced phase as a deviation target command.

Republished patent WO2005 / 093939

  On the other hand, the drive mechanism driven by the electric motor includes mechanical resonance characteristics. In some cases, the vibration generated by the mechanical resonance oscillates and the motor cannot be driven. The mechanical resonance characteristics can be modeled by a multi-inertia system modeled by multiple masses and springs. In the multi-inertia model, mechanical resonances of different frequencies occur. The prior art described above has room for improvement when driving a drive mechanism having vibration due to such mechanical resonance.

  In other words, there is a demand for an electric motor control device that can perform highly robust control with respect to moment of inertia fluctuations and mechanical resonance characteristics without requiring complicated parameter adjustment for an electric motor that drives a drive mechanism of any configuration. It was.

  The present invention has been made in view of such a problem, and even if any of the moment of inertia or inertial mass in the drive mechanism and the multi-inertia model can be unknown or fluctuated, the robustness can be achieved without adjusting various parameters. An object of the present invention is to provide an electric motor control device capable of controlling the driving of an electric motor having a high height.

  In order to solve the above-described problem, according to one aspect of the present invention, a position detection unit that detects a motor position of a motor that drives a driving mechanism in which any of an inertia moment or an inertia mass and an inertia model can be unknown or fluctuated; A speed calculation unit that inputs the motor position to calculate a first motor speed of the motor, a position control unit that inputs a difference between a position command and the motor position and outputs a first speed command, and the first A speed control unit that inputs a difference between the speed command and the second motor speed and outputs a first torque command or a first thrust command, and outputs the second motor speed advanced in phase with respect to the first motor speed. A disturbance observer having a first phase compensator and a resonance suppression model set to suppress mechanical resonance of the drive mechanism and capable of changing a gain change rate of a frequency characteristic of a rigid nominal model; An inertia fluctuation suppression unit that outputs a second torque command or a second thrust command based on a disturbance torque or a disturbance thrust estimated using a disturbance observer, the second torque command or the second thrust command is input, and the electric motor An electric motor control device is provided that includes a torque control unit or a thrust control unit that controls the torque or the thrust.

  According to the present invention, even when any of the inertia moment or inertia mass and the multi-inertia model in the drive mechanism is unknown or can be changed, it is possible to control the drive of the motor with high robustness without adjusting various parameters.

It is the block diagram which showed the structure of the electric motor control apparatus which concerns on embodiment in the transfer function format. It is a block diagram which shows the detailed structure of a 1st disturbance observer. It is the figure which showed the disturbance suppression control system loop with the feedback control system model. It is a block diagram which shows the detailed structure of a 1st phase compensation part. It is a figure showing the model figure of each inertia model of a rigid body model, 2 inertia models, and 3 inertia models, and a specific structural example. It is a frequency characteristic figure which shows an example of the relationship between the resonance suppression nominal model of a 1st disturbance observer, and the mechanical resonance of a drive mechanism. It is a figure showing the comparative example of the concept of the feedback control of disturbance suppression control. It is a figure showing the concept of the feedback control of the disturbance suppression control at the time of ideally deforming the resonance suppression nominal model. It is a Nyquist diagram when the disturbance suppression control of the embodiment is applied to a two-inertia model. FIG. 10 is an enlarged view of the vicinity of the reference coordinates in FIG. 9. It is a Nyquist diagram when the disturbance suppression control of the embodiment is applied to a three inertia model. It is an enlarged view of the reference | standard coordinate vicinity in FIG. It is a Nyquist diagram when the disturbance suppression control of the embodiment is applied to a four inertia model. It is an enlarged view of the reference | standard coordinate vicinity in FIG. It is a Nyquist diagram when the conventional disturbance suppression control is applied to the two-inertia model. FIG. 16 is an enlarged view near the reference coordinates in FIG. 15. It is a Nyquist diagram when the conventional disturbance suppression control is applied to the three inertia model. It is an enlarged view of the reference | standard coordinate vicinity in FIG. It is a Nyquist diagram in the case of applying the conventional disturbance suppression control to the four inertia model. FIG. 20 is an enlarged view near the reference coordinates in FIG. 19. It is a figure which shows the simulation result of a conventional comparative example with a 2 inertia model. It is a figure which shows the simulation result of embodiment by a 2 inertia model. It is a figure which shows the simulation result of the conventional comparative example with a three inertia model. It is a figure which shows the simulation result of embodiment by a 3 inertia model. It is a figure which shows the simulation result of a conventional comparative example with a four inertia model. It is a figure which shows the simulation result of embodiment by 4 inertial models. It is a figure which shows the simulation result of a conventional comparative example by 0 times of inertia moment ratio. It is a figure which shows the simulation result of embodiment by 0 times of inertia moment ratio. It is a figure which shows the simulation result of a conventional comparative example by 30 times of inertia moment ratio. It is a figure which shows the simulation result of embodiment by an inertia moment ratio 30 times. It is a figure which shows the simulation result of standard P-PI control by the inertia moment ratio of 30 times. It is a figure which shows the simulation result of embodiment by an inertia moment ratio 30 times. It is the block diagram which showed the structure of the motor control apparatus of the 1st modification which provided the 2nd phase compensation part in the transfer function format. It is a block diagram which shows the detailed structure of a 2nd phase compensation part. It is a figure which shows the simulation result of a 1st modification in 2 inertia models. It is a figure which shows the simulation result of a 1st modification in 3 inertia models. It is a figure which shows the simulation result of a 1st modification in 4 inertia models. It is a figure which shows the simulation result of a conventional comparative example by a 2 times inertia moment ratio. It is a figure which shows the simulation result of a 1st modification by 2 times of inertia moment ratio. It is a figure which shows the simulation result of a prior art comparative example by 3.5 times of inertia moment ratio. It is a figure which shows the simulation result of a 1st modification by 3.5 times of inertia moment ratio. It is a figure which shows the simulation result of a conventional comparative example by 5 times of inertia moment ratio. It is a figure which shows the simulation result of a 1st modification by 5 times of inertia moment ratio. It is a block diagram which shows the detailed structure of the 2nd disturbance observer with which a 2nd modification is provided. It is the figure which showed the simulation result of the 2nd modification in 2 inertia models. It is the figure which showed the simulation result of the 2nd modification in 3 inertia models. It is the figure which showed the simulation result of the 2nd modification in 4 inertia models. It is a figure which shows the simulation result of a 1st modification by 0 times of inertia moment ratio. It is a figure which shows the simulation result of a 2nd modification by 0 times of inertia moment ratio. It is a figure which shows the simulation result of a 1st modification by 30 times of inertia moment ratios. It is a figure which shows the simulation result of a 2nd modification by the inertia moment ratio of 30 times. It is a figure which shows the simulation result of a 1st modification with an inertia moment ratio of 35 times. It is a figure which shows the simulation result of a 2nd modification by 35 times of inertia moment ratio.

  Hereinafter, an embodiment will be described with reference to the drawings.

  The motor control device of the present embodiment shown in FIG. 1 controls the rotational position of the motor (hereinafter referred to as “motor position”) based on a position command input from a host control device (not shown). The illustrations and explanations below are all explained in transfer function format. In the example of this embodiment, a case where the electric motor is a rotary motor will be described.

<Schematic configuration of electric control device>
In FIG. 1, an electric motor control device 100 according to the present embodiment includes a position control unit 1, a speed control unit 2, an inertia variation suppression unit 3, a first current control unit 4, a position detection unit 5, and a speed calculation unit. 6 and a first phase compensator 7.

  The position control unit 1 reduces the position deviation based on the position deviation between the position command input through the subtracter 8 and the motor position of the motor 200 detected by the position detection unit 5 described later. The speed command is output to.

  The speed control unit 2 calculates a speed deviation between the speed command input from the position control unit 1 and the second motor speed phase-compensated from the first phase compensation unit 7 described later, which is input via the subtractor 9. Based on this, a first torque command is output so as to reduce this speed deviation.

  The inertia fluctuation suppression unit 3 estimates the disturbance torque based on the rotation speed of the electric motor 200 (hereinafter referred to as the first electric motor speed) output from the speed calculation unit 6 described later and the second torque command from the speed control unit 2. A first disturbance observer 31 for calculating the output, a low-pass filter 32 for removing harmonic noise of the output, a gain multiplier 33 for multiplying the output by a predetermined gain and outputting a correction amount of the estimated disturbance torque, And an adder 34 that adds the correction amount of the estimated disturbance torque to the first torque command and outputs it as the second torque command.

  The first current control unit 4 outputs a drive current by, for example, PWM control to the electric motor 200 based on the second torque command from the inertia variation suppressing unit 3.

  The electric motor 200 generates torque by the drive current from the first current control unit 4 and drives the load machine 400 via the drive unit 300 such as a speed reducer. In the present embodiment, the entire machine including the drive unit 300 and the load machine 400 that are driven by the electric motor 200 is referred to as a drive mechanism 500.

  The position detection unit 5 is configured by a rotary encoder mechanically coupled to the rotor of the electric motor 200, for example. The position detector 5 detects a motor position corresponding to the rotational position of the motor 200.

  The speed calculation unit 6 calculates the first motor speed of the electric motor 200 based on the change in the electric motor position of the electric motor 200 detected by the position detection unit 5. Specifically, a differentiator may be used for the speed calculation unit 6.

  The first phase compensation unit 7 advances the phase with respect to the first motor speed based on the first motor speed of the electric motor 200 output from the speed calculation unit 6 and the first torque command output from the speed control unit 2. A new second motor speed compensated for such phase is output.

  The electric motor control apparatus 100 of the present embodiment having the above configuration has a triple loop configuration of a position control system feedback loop, a speed control system feedback loop, and a current control system feedback loop. That is, after a position command is input from a host controller (not shown), the control signal and the detection signal are output in the order of the position control unit 1, the speed control unit 2, the first current control unit 4, the electric motor 200, and the position detection unit 5. A feedback loop of a position control system (hereinafter referred to as a position control system loop) that transmits the feedback is provided. In addition, the control signal and the detection signal are transmitted in the order of the speed control unit 2, the inertia variation suppression unit 3, the first current control unit 4, the electric motor 200, the position detection unit 5, the speed calculation unit 6, and the first phase compensation unit 7. A feedback loop of a speed control system that feeds back (hereinafter referred to as a speed control system loop) is also provided.

  Of these, in the speed control system loop, the inertia variation suppression unit 3 adjusts the torque command so as to reduce the influence of disturbance applied to the electric motor 200 via the drive mechanism 500. Thus, robustness in response characteristics of the electric motor 200 can be ensured even when the moment of inertia ratio of the drive mechanism 500 is unknown or fluctuates.

<Configuration of the first disturbance observer>
In FIG. 2, the first disturbance observer 31 includes a current loop model 41, a resonance suppression model 43, and equivalent low-pass filters 42a and 42b (abbreviated as equivalent LPF in the drawing). In the figure, a low-pass filter 32 that removes harmonic noise from the output of the first disturbance observer 31 and a gain that outputs a correction amount of the estimated disturbance torque by multiplying the output of the low-pass filter 32 by a predetermined gain. A multiplier 33 is also shown.

Here, each positioning of the component which the said 1st disturbance observer 31 has is demonstrated, referring FIG. FIG. 3 shows a disturbance including the inertia fluctuation suppression unit 3 (including the first disturbance observer 31 in FIG. 2), the first current control unit 4, the electric motor 200, and the drive mechanism 500 (the drive unit 300 and the load 400) in FIG. It is the figure which showed the feedback loop of the suppression control system using the block diagram. The control target P (s) (corresponding to the electric motor 200 and the drive mechanism 500) is represented by the product of the mechanical resonance system R (s) and the mechanical rigid system P _r (s). Among these, the mechanical rigid system P _r (s) is _expressed as P _r (s) = 1 / (J _m + J _L ) s when the motor inertia moment J _m , the load inertia moment J _L , and the integral element 1 / s are used. Become.

Further, the first disturbance observer 31 in the present embodiment includes a current loop model P _im (s), a resonance suppression model P _n ^-1 (s), and equivalent low-pass filters 42a and 42b (equivalent LPFs in FIG. 2). , Omitted in FIG. The current loop model P _im (s) is designed to be a model equivalent to the first current control unit 4 (P _i (s)) in FIG. 1, and P _i (s) = P _im ( s).

In general, when designing a disturbance observer, after defining a nominal model (normative model), an inverse system of the nominal model is determined, and the disturbance observer is provided with an inverse system of the nominal model. The inverse system, for example, the system ^{G (s), G (s} ) · G -1 (s) = 1 satisfies the ^G -1 (s) corresponds to the inverse system. The nominal model (normative model) is defined by a model (an equation such as a transfer function) that idealizes a control target that is a drive mechanism including an electric motor. In this embodiment, emphasizing that the nominal model to be controlled is as simple as possible, the machine rigid system corresponding to the machine rigid body part and the motor is defined as 1 / Js.

In this embodiment, a resonance suppression nominal model P _n (s) is proposed as a specific model of the control target nominal model. The resonance suppression nominal model P _n (s) is not particularly illustrated, but in the example of this embodiment, the frequency characteristic controller (1 + T ₁ s) / (1 + T ₂ s) and the mechanical rigid system model 1 / J to be controlled are included. _The resonance suppression nominal model is described in detail later. The machine rigid system model 1 / J ₀ s to be controlled does not necessarily match the machine rigid system to be controlled. In the mechanical rigid system model 1 / J ₀ s to be controlled, for example, the inertia moment J ₀ may be set to the inertia moment J _m of the motor alone (J ₀ = J _m ).

A resonance model P _n ⁻¹ (s) included in the first disturbance observer 31 of the present embodiment is an inverse system of the resonance suppression nominal model P _n (s), and is an inverse system (1 + T) of the frequency characteristic controller. ₂ s) / (1 + T ₁ s) and the inverse system J ₀ s of the mechanical rigid body model to be controlled are integrated.

本実施形態の外乱抑圧制御系ループは、上記（２）式の閉ループ伝達関数Ｇ_{Ｃｌｏｓｅ}を適切に設定することで、多様な多慣性モデルを取り得る機械共振要素に対し広い周波数帯域で抑圧制御が可能となる。この外乱抑圧制御系の低周波帯域での特性を説明するために、（２）式でｓ＝０を代入すると（Ｋ_ｆ＝１とした）、Ｇ_ＬＯＯＰ（ｓ）＝１／Ｊ_ｍｓとなり、負荷慣性モーメントＪ_Ｌの影響が除去されて電動機単体の特性に制御できることがわかる（低周波帯域であるから、ｇ（０）＝１、Ｐ_ｉ（０）＝１、Ｒ（０）＝１となる）。なお、この外乱抑圧制御系の高周波帯域での特性は後述する。 The loop transfer function G _Loop of the feedback control system (disturbance suppression control system loop) in FIG. 3 is expressed by the following equation (1), and the entire closed loop from the torque command r to the motor speed v with the disturbance suppression control system as a closed loop. The transfer function G _Close is expressed by the following equation (2).

The disturbance suppression control system loop of the present embodiment can perform suppression control in a wide frequency band with respect to a mechanical resonance element that can take various multi-inertia models by appropriately setting the closed-loop transfer function G _Close in the above equation (2). It becomes possible. In order to explain the characteristics of the disturbance suppression control system in the low frequency band, if s = 0 is substituted in equation (2) (K _f = 1), G _LOOP (s) = 1 / J _m s. It can be seen that the influence of the load inertia moment J _L is removed and the characteristics of the motor can be controlled (because it is a low frequency band, g (0) = 1, P _i (0) = 1, R (0) = 1. Becomes). The characteristics of the disturbance suppression control system in the high frequency band will be described later.

  Here, the configuration of a general disturbance observer will be described. The disturbance observer is designed with a state observer that incorporates a disturbance model. When the disturbance is assumed to be constant over time, it is called a zero-order disturbance observer. To reduce the order of the observer, the minimum-dimensional state observer is used. Use. Also, the disturbance observer described by the state equation can be expressed by an inverse system of a nominal model and an observer low-pass filter (hereinafter referred to as an equivalent low-pass filter). In the 0th-order disturbance observer, the equivalent low-pass filter is a first-order low-pass filter, and the band of this filter is the frequency band of the disturbance observer.

  The equivalent low-pass filter is a component of a disturbance observer, and the shape of the filter is defined. For example, the first-order disturbance observer is a low-pass filter of the denominator second-order and numerator first-order, and its parameters are not arbitrary but are determined by the design method of the disturbance observer. In disturbance suppression control using an Nth-order disturbance observer, the higher the order, the higher the disturbance suppression performance. However, when the disturbance observer is widened to improve the disturbance suppression performance, the control stability of the disturbance suppression control system is affected by the dynamic characteristics (mainly low-pass characteristics and control time delay) of the first current controller. There is a case where high frequency noise is generated from the electric motor due to a decrease and a peak in the frequency characteristics of the high frequency band. In addition, since the drive machine driven by the electric motor has mechanical resonance characteristics, the disturbance suppression control system may easily be affected by the mechanical resonance due to the dynamic characteristics of the first current control unit. In order to solve such a problem, the first disturbance observer 31 in the present embodiment includes a current loop model 41, a resonance suppression model 43, and equivalent low-pass filters 42a and 42b.

  In designing a disturbance observer used for disturbance suppression control, a torque command and a motor speed are used as input signals. Actually, a current control unit exists between the torque command and the motor speed. Since the response frequency of the current control is sufficiently higher than the response frequency of the speed control system, a disturbance observer can be designed with the transfer function of the current control unit set to “1” (a disturbance observer with the lowest order possible in terms of software implementation). Is required). However, when trying to compensate for the operation of the drive mechanism when the response frequency of the speed loop is several tens of Hz and the load inertia moment relative to the motor inertia moment is about 30 times, the disturbance suppression control system is required to have high response characteristics. The In this case, since it is necessary to accurately estimate the operation of the drive mechanism up to the high frequency band, it is necessary to add to the control target a model of the current control unit that has been ignored in the design of the disturbance observer.

  The current loop model 41 has a function of preventing the gain of mechanical resonance from increasing due to the influence of the current control loop. That is, the torque signal generated by the motor is output by adding the delay of the current control loop to the second torque command input as shown in FIG.

The resonance suppression model 43 includes an inverse system 51 of a frequency characteristic controller that controls the frequency characteristic of the electric motor 200 with respect to the first motor speed, and an output of the inverse system 51 of the frequency characteristic controller (load of the motor driven). And the controlled object inverse model 52 that multiplies the moment of inertia J ₀ of the rotor of the electric motor 200 by the acceleration s and the operator s together (substantially performs a differential operation). It has a function of obtaining an acceleration / deceleration torque signal (torque component used for acceleration / deceleration by an ideal motor assumed in a nominal model) of the motor.

  As a function of the entire first disturbance observer 31, the acceleration / deceleration torque signal of the ideal motor obtained by the resonance suppression model 43 based on the first motor speed of the motor 200 and passed through the equivalent low-pass filter 42b is obtained as the current loop model 41. The subtractor 44 subtracts the torque from the torque command that has passed through the equivalent low-pass filter 42a to estimate the disturbance torque. Then, the estimated disturbance torque is appropriately adjusted and output by the low-pass filter 32 for removing the harmonic noise set by the time constant T and the gain multiplier 33 having a predetermined gain. The equivalent low-pass filters 42 a and 42 b function as filters that determine the frequency band of the first disturbance observer 31.

The resonance suppression model 43 includes an inverse system 51 of the frequency characteristic controller, a multiplier that multiplies the moment of inertia J ₀ , and a differentiator that realizes the operator s, as originally illustrated. Since the differentiator realized by this operator s becomes a leading element, it cannot be mounted on a digital circuit and software as it is. In this case, by combining the resonance suppression model 43 and the equivalent low-pass filters 42a and 42b, the order of the denominator and the order of the numerator can be made the same. For example, a secondary filter may be used as the equivalent low-pass filters 42a and 42b. The inverse system 51 of the frequency characteristic controller will be described in detail later.

<Configuration of first phase compensation unit>
In FIG. 4, the first phase compensation unit 7 includes a controlled object model 61, a low-pass filter 62, and a controller 63. The first phase compensator 7 inputs the first torque command to the control target model 61, passes the output through the low-pass filter 62, generates a phase compensation signal, and uses the subtractor 64 to generate the phase compensation signal. The deviation subtracted from the one motor speed is passed through the controller 63, and this is fed back to the adder 65 that is inputting the first torque command. Among these signals, the first phase compensator 7 outputs the second motor speed as a new speed feedback signal using the signal output from the control target model 61. The low-pass filter 62 corresponds to the first low-pass filter described in each claim.

  Then, as shown in FIG. 1, by inputting a deviation between the speed command output from the position control unit 1 and the second motor speed output from the first phase compensation unit 7 to the speed control unit 2, The speed control system loop at is configured. Since the first phase compensation unit 7 functions as a phase compensation speed observer that estimates the motor speed of the motor 200 and advances the phase, the phase margin of the position control system loop can be improved by advancing the phase in the speed control system loop. . For this reason, the response is stabilized, and the inertia variation range can be expanded.

  However, when the speed loop gain (not shown) in the speed controller 2 is increased, the phase to be compensated for by the first phase compensator 7 is shifted, so that vibration occurs. Therefore, it is necessary to readjust the phase compensated by the first phase compensation unit 7 in accordance with the speed loop gain in the speed control unit 2.

  For example, in the present embodiment in which phase compensation is performed by the speed observer shown in FIG. 4, the amount of phase to be compensated can be set by adjusting the time constant of the low-pass filter 62 of the speed observer. Therefore, by simulation or actual machine test, values that optimize the time constant of the low-pass filter 62 corresponding to changes in the speed loop gain in the speed control unit 2 are sequentially plotted on a graph (not shown in particular). An approximate expression for interpolating a plot of measured data is obtained.

For example, when the disturbance observer as shown in FIG. 2 is used as in the present embodiment, the low-pass filter time constant T _LPF of the speed observer is a speed loop gain K such that T _LPF = a × K _v + b. _An approximate expression can be determined by a linear function of _v . In addition, what can be approximated by a quadratic function of T _LPF = a × K _v ² + b × K _v + c can also be used. Using these approximate expression expressed by a polynomial of the speed loop gain K _v and independent variables, for example, when changing the speed loop gain K _v in the speed control section 2, a speed observer by the function first The time constant T _LPF of the low-pass filter 62 of the one-phase compensation unit 7 can be appropriately changed, and the phase compensation value can be readjusted automatically.

  The most characteristic feature of the present embodiment having the above-described configuration is that a resonance suppression nominal model is set for the drive mechanism 500 driven by the electric motor 200, and the resonance suppression model 43 (resonance suppression nominal model of the first disturbance observer 31). The frequency characteristic controller reverse system 51 provided in the reverse system) controls the frequency characteristics so as to suppress the occurrence of vibration due to mechanical resonance in the electric motor 200 and the drive mechanism 500.

  First, a multi-inertia model expressing the mechanical resonance characteristics of the drive mechanism 500 will be described with reference to FIG.

<Multi inertia model and its frequency characteristics>
The multi-inertia model in the present embodiment is a classification that is abstracted by the number of inertia elements provided in the drive mechanism 500 and a connection configuration using elastic elements between them, and various inertia models depending on the configuration of the drive mechanism 500. Exists. For example, in the case of the electric motor 201 of the rotary motor shown in the figure, when only the rigid rotating plate 501 having the mass m ₁ is fixed as the driving mechanism 500 that is connected first from the output shaft, the inertia of the driving mechanism 500 The model is a rigid model. In this rigid body model, the drive mechanism 500 is configured by only one inertia element, and does not include an elastic element.

For example, when a weight 503 having a mass m ₂ is suspended by a band member 502 wound around the outer periphery of the rigid rotating plate 501, the drive mechanism 500 becomes a two-inertia model. That is, the rigid rotating plate 501 (m ₁ ) and the weight 503 (m ₂ ) each serve as an inertia element, and the band member 502 (k ₁ ) that is an elastic element is connected between them to constitute a two-inertia model. .

Further, for example, the drive mechanism 500 in which the ball screw 505 is connected to the output shaft of the rotary motor 201 via the coupling 504 and the feed base 506 is directly moved by the ball screw 505 is a three-inertia model. That is, the coupling input side member 504a (m ₁ ), the ball screw 505 (m ₂ ) including the coupling output side member 504b, and the feed base 506 (m ₃ ) are inertial elements, respectively. The connecting portion (k ₁ ) in the ring 504 and the rolling mechanism (k ₂ ) between the ball screw 50 and the feed base 506 each function as an elastic element to constitute a three-inertia model.

In the present embodiment, only models having a configuration in which inertia elements and elastic elements are connected in series in an alternating arrangement are handled, and an “n inertia model” means a model having n inertia elements. In addition, as shown in the figure, the inertia moment ratio in the present embodiment is the ratio of the inertia moment J _L of the entire drive mechanism 500 to the inertia moment J ₀ of the rotor including the output shaft of the electric motor 200 J _L / It refers to J _0. Depending on the configuration and use form of the drive mechanism 500, the inertia model and the inertia moment ratio of the drive function may vary. For example, the case where the moving target object which has an inertia element and an elastic element is mounted on the feed stand of 3 inertia model shown in figure is considered.

  As shown in FIG. 6, in the gain characteristic of the Bode diagram, the n inertia model draws a characteristic curve having n vertices and n bottom points. Each vertex is a resonance point, and each bottom point is an anti-resonance point. In the case of the rigid body model, a characteristic curve is drawn in which the gain gradually decreases without having a vertex (resonance point) and a bottom point (antiresonance point). Further, the gain characteristic becomes lower as the inertia moment ratio is larger (the graph moves downward).

  In the phase characteristic, the n inertia model draws a characteristic curve having n−1 phase advance regions. The phase characteristic advances in the frequency band between the anti-resonance point and the resonance point in the gain characteristic. In the case of the rigid body model, the phase of −90 ° is maintained in all frequency bands at any moment of inertia ratio.

  In both the gain characteristic and the phase characteristic, the characteristic curve changes even in the same n inertia model depending on the setting of the mass of each inertial element and the elastic coefficient of each elastic element. However, the characteristics of the above-mentioned curves are commonly changed. Absent.

  In contrast to the characteristics of each inertial model as described above, the resonance suppression nominal model in the present embodiment has characteristics of a rigid body model (that is, a state in which no elastic element is provided) having a zero moment of inertia ratio as shown in the figure. Nearly, the gain characteristic is set to be slightly higher in the middle frequency band and the high frequency band, and the phase characteristic is set to draw a characteristic curve in which the phase slightly advances in the central band.

  In the present embodiment, a resonance suppression nominal model having such characteristics is arbitrarily set, and the first disturbance observer includes a resonance suppression model 43 that is an inverse system of the set resonance suppression nominal model, so that the drive mechanism Generation of vibration due to mechanical resonance in 500 can be suppressed. Hereinafter, the principle of suppressing the mechanical resonance will be described.

<Principle of machine resonance suppression in this embodiment>
First, as described above, in the design of the conventional disturbance observer, the nominal model is assumed to be a rigid model. In this design method, the nominal model is a rigid model, so that the order of the operator s in the disturbance observer is minimized and the implementation is facilitated. However, the actual drive mechanism 500 is often composed of a multi-inertia model having two or more inertias, and there is a problem in that the nominal model is likely to oscillate due to actual mechanical resonance characteristics when the nominal model is designed as a rigid model. As a countermeasure against this problem, by adding a notch filter or a low-pass filter in the loop of the disturbance suppression control system including the inertia fluctuation suppressing unit 3, the mechanical resonance characteristic is lowered and the disturbance suppression control system is stabilized. It is possible. However, in this case, the robustness against the change in the moment of inertia ratio is reduced due to the decrease in the disturbance suppression performance due to the phase delay due to the notch filter and the gain decrease due to the low-pass filter.

  Therefore, as a solution to reduce the effect of actual mechanical resonance without using a notch filter or low-pass filter, in this embodiment, returning to the origin and why there is a mechanical resonance characteristic, why does the disturbance suppression control system oscillate? As a result of detailed theoretical examination by analytical calculation of the control system and computer simulation, a new fact shown in FIG. 7 was found.

<Fact 1> Disturbance suppression control is a technique for performing control so that the characteristics of the controlled object coincide with the nominal model. In the present embodiment, the control parameters are set so that the gain characteristic (the dashed line) of the control target is lower than the nominal model (solid line) as in the low frequency band of FIG. In this case, the disturbance suppression control works to raise the gain by positive feedback, but since the circuit transfer characteristic G _Loop shown in the equation (1) is designed to be less than 1, it is stable even with positive feedback. It is. Therefore, even if the gain characteristics and phase characteristics of the controlled object fluctuate in a complicated manner at a lower part than the nominal model, it was found that the circuit transfer function G _Loop shown in the equation (1) is always stable because it is less than 1. .

<Fact 2> On the other hand, when the apex portion of the mechanical resonance is above the nominal model as in the high frequency band of FIG. 7, the disturbance suppression control operates to reduce the gain characteristic by negative feedback. . In the curve above the nominal model, the round transfer characteristic G _Loop shown in the equation (1) is 1 or more. In the high frequency band, in addition to the fact that the phase delay of the control system is originally large, the phase greatly changes due to the mechanical resonance characteristics. Therefore, the control system is operated at an appropriate frequency at which the loop transfer characteristic G _Loop shown in the equation (1) is 1 or more. It becomes unstable and oscillates. This is the reason why the disturbance suppression control system oscillates due to the mechanical resonance characteristics of the controlled object.

It can be seen that in order to stably operate the disturbance suppression control system with respect to the controlled object including the mechanical resonance characteristics, it is sufficient to satisfy the above-mentioned <Fact 1> condition in all frequency bands. This is a new concept that is different from the conventional method of simply band-limiting with a filter. In the present embodiment, as a technique for realizing this, a resonance suppression nominal model obtained by modifying the nominal model is set as shown in FIG. That is, by using a resonance suppression nominal model in which the slope of the nominal model is changed at the frequencies f ₁ and f ₂ , the mechanical resonance characteristics can be made lower than that of the resonance suppression nominal model. By appropriately selecting the frequencies f ₁ and f ₂ so as to sandwich the top of the mechanical resonance characteristic, the control system can always be stabilized even if the mechanical resonance characteristic fluctuates. In addition, this method only modifies the nominal model, and no excessive delay occurs in the control system, so that the robust performance of the disturbance suppression control system can be maintained. In order for the resonance suppression nominal model of this embodiment to have such a gain characteristic, the resonance suppression nominal model includes a frequency characteristic controller (1 + T ₁ s) / (1 + T ₂ s) and a mechanical rigid system model 1 to be controlled. / J ₀ s (for example, the moment of inertia may be the moment of inertia J _m of the motor alone (J ₀ = J _m )).

Further, in the resonance suppression model, the inverse system (1 + T ₂ s) / (1 + T ₁ s) of the frequency characteristic controller and the inverse system of the mechanical rigid system model to be controlled and the inverse system J ₀ s of the mechanical rigid system Is configured to accumulate.

  However, the above method is limited to an ideal case that can be solved only by changing the gain characteristic of the resonance suppression nominal model. In the actual control target, no matter how the resonance suppression nominal model is designed, the top of some mechanical resonance characteristics exceeds the resonance suppression nominal model in a narrow band as shown in the gain characteristic diagram of FIG. May end up.

On the other hand, as described above, the condition where mechanical resonance occurs is a case where both the gain characteristic and the phase characteristic have a low margin at the same time, that is, (1) in the disturbance suppression control system loop including the inertia fluctuation suppressing unit 3. This is a case where the oscillation oscillates because the phase of the loop transfer characteristic G _Loop shown in the equation reaches −180 ° and the gain exceeds 1 and becomes large. Therefore, it is difficult to design a resonance suppression nominal model so that there is a margin in all frequency bands for either one of the characteristics. However, by providing as much margin as possible for both gain and phase as much as possible, it is possible to design to ensure a sufficient margin on the other side even if one margin is locally reduced in some frequency bands.

<Resonance suppression nominal model design example and its stability>
Therefore, the inverse system 51 of the frequency characteristic controller provided in the resonance suppression model 43 of the present embodiment has a phase of −180 degrees in the low frequency band with the round transfer characteristic G _Loop in the disturbance suppression control system loop expressed by the above equation (1). In the frequency region where the gain is less than 1 and the gain is 1 or more in the high frequency band, the phase is controlled to be other than −180 degrees (the phase is larger or smaller than −180 °). Specific examples of inverse system 51 of the frequency characteristic controller in this embodiment, it is configured to be expressed by a transfer function represented as shown in FIG. 2 and _{(1 + T 2 s) /} (1 + T 1 s), Of these, the two time constants T ₁ and T ₂ may be appropriately set in advance assuming all the drive mechanisms 500 that can be controlled.

For example, in the case of FIG. 8 described above, the two time constants T ₁ and T ₂ of the inverse system 51 ((1 + T ₂ s) / (1 + T ₁ s)) of the frequency characteristic controller are T ₁ = 1 / 2πf ₁ , T ₂ = 1 / 2πf ₂ is set. Further, by setting a wide width between f ₁ and f ₂ , it is possible to cope to some extent even if the mechanical resonance frequency changes, and it is possible to deal with various multi-inertia model controlled objects. In this way, the time constants T ₁ and T ₂ possessed by the inverse system 51 of the frequency characteristic controller have a plurality of frequency setting values f ₁ , f that change the gain change rate of the frequency characteristic based on a rigid nominal model. it is to be determined by f _2. If the gain characteristic of the resonance suppression nominal model is too high than the gain characteristic to be controlled, the round transfer characteristic G _Loop may become unstable. Therefore, an appropriate height difference (feedback amount) in any frequency band. The gain characteristics are set so that

  The frequency characteristic of the resonance suppression nominal model provided with this frequency characteristic controller is as shown in FIG. In FIG. 6, although the tops of some mechanical resonance characteristics exceed the resonance suppression nominal model in the gain characteristic diagram and the gain margin is impaired, the phase characteristics in the corresponding band have the resonance suppression nominal model. The phase characteristics have greatly advanced, and compensation is made by the amount of phase margin. By estimating the disturbance torque using the resonance suppression model 43 provided with the inverse system 51 of the frequency characteristic controller having such a design, a comprehensive gain margin and phase margin can be obtained for any multi-inertia model. It is possible to secure and suppress the occurrence of mechanical resonance.

Such determination of the stability of the disturbance suppression control system loop provided in the present embodiment is shown by the Nyquist diagrams in FIGS. 9, 11, and 13 show the stability of the closed-loop transfer function G _Close shown in the above equation (2) in polar coordinates on the complex plane of the frequency characteristics of the open-loop transfer function shown in the above equation (1). FIGS. 10, 12, and 14 are enlarged views of the vicinity of the reference coordinates (X = −1, Y = 0; −1 + j0) in FIGS. 9, 11, and 13, respectively. As shown. In the case of FIGS. 9 and 10 applied to the two-inertia model, the vector locus intersects the negative real axis at the coordinates (X = −0.997, Y = 0), and the angular frequency ω increases. The reference coordinates are located on the left side when viewed from the direction of travel of the trajectory. This shows that the feedback control system of the closed loop transfer function G _Close applied to the two inertia model is stable. Similarly, in each of FIGS. 11 and 12 applied to the three-inertia model and FIGS. 13 and 14 applied to the four-inertia model, the control loop of the disturbance suppression control system is stable. These feedback control systems depicting the Nyquist diagrams of FIGS. 9 to 14 are all the same system set to the same control parameter, that is, any of the 2 to 4 inertia models without requiring individual adjustment. It can be seen that the stability can be secured.

  15 to 20 are Nyquist diagrams of the conventional disturbance suppression control system. 15 and 16 to which the two-inertia model is applied, and FIGS. 19 and 20 to which the four-inertia model is applied are stable. However, in the case of FIGS. 17 and 18 to which the three-inertia model is applied, the vector trajectory is drawn so as to surround the reference coordinate. The reference coordinates are located on the right side. Thus, it can be seen that the conventional disturbance suppression control loop applied to the three inertia model becomes unstable and oscillates.

  As described above, the first disturbance observer 31 of this embodiment and the resonance suppression model 43 provided in the first disturbance observer 31 have a new configuration that is not conventionally provided, and have a function of suppressing generation of vibration caused by mechanical resonance.

<Demonstration data of disturbance suppression performance of this embodiment>
Hereinafter, the effectiveness of the motor control device 100 of the present embodiment will be described with reference to specific step response data of FIGS. Note that “Position Error” in the figure is a deviation between a position command from a host controller (not shown) and the motor position from the position detector 5, and is a signal directly input to the position controller 1 ( See A in FIG. 1 above; In addition, “Speed” in the figure is a speed component (position command differential value) of a position command input by the motor control device 100 and a first motor speed output by the speed calculation unit 6 (the same applies hereinafter). Further, “Torque” in the figure is the second torque command immediately before being output from the inertia fluctuation suppressing unit 3 and input to the first current control unit 4 (see B in FIG. 1 above; the same applies hereinafter). Comparison between FIG. 21 of the conventional comparative example when the two inertia model is applied and FIG. 22 of the present embodiment is not so conspicuous, but a comparison between FIG. 23 and FIG. 24 when the other three inertia models are applied, 4 In the comparison between FIG. 25 and FIG. 26 in the case where the inertia model is applied, the effect of suppressing fluctuations according to the present embodiment is remarkably exhibited in the motor speed and the torque command, that is, the robustness against mechanical resonance is improved. I can confirm.

  Next, FIG. 27 to FIG. 30 are diagrams comparing simulation results of the conventional comparative example and this embodiment according to the moment of inertia ratio, and all show cases where a rigid body model is applied. As can be seen from these, it is confirmed that the torque command fluctuation suppressing effect according to this embodiment is exhibited regardless of the magnitude of the moment of inertia ratio, that is, the robustness with respect to the moment of inertia ratio is also maintained in this embodiment. it can.

  FIG. 31 shows an ideal waveform for performing standard P-PI control when a rigid body model having a large moment of inertia ratio is applied. FIG. 32 shows a simulation result of the present embodiment under the same conditions. It is the figure (it is the same as the said FIG. 30) represented. Comparison between the two figures indicates that the waveform surrounded by the broken-line circle is close to the ideal waveform, and thus it can be confirmed that this embodiment is suitable for application to standard P-PI control.

<Effect of this embodiment>
As described above, according to the motor control device 100 of the present embodiment, a triple loop configuration of a position control system loop, a speed control system loop, and a current control system loop is provided. Among these, in the speed control system loop, the disturbance torque of the electric motor 200 estimated by the inertia variation suppressing unit 3 using the first disturbance observer 31 is added to the first torque command, and the second torque command is output. Thereby, robustness in the control characteristics of the electric motor 200 can be ensured even when the moment of inertia ratio of the drive mechanism 500 is unknown or fluctuates.

  In addition, the first phase compensator 7 provided in the speed control system loop functions as a phase compensation speed observer that estimates the speed and advances the phase, thereby advancing the phase of the speed control system and increasing the phase margin of the position control system. Can be improved, and the response characteristics of the inertia fluctuation suppressing unit 3 are improved. That is, the robustness with respect to the inertia moment ratio of the drive mechanism 500 is improved.

  Furthermore, in the present embodiment, the first disturbance observer 31 that estimates the disturbance torque in the inertia variation suppressing unit 3 suppresses the influence of the mechanical resonance of the drive mechanism 500 corresponding to the first motor speed of the motor 200. The resonance suppression model 43 that outputs the acceleration / deceleration torque signal is provided. The first disturbance observer 31 estimates the disturbance torque using the acceleration / deceleration torque signal of the ideal motor, and as a result, the inertia variation suppressing unit 3 generates vibration due to the mechanical resonance of the drive mechanism 500. And suppresses fluctuations in the moment of inertia of the drive mechanism. That is, robustness in the control characteristics of the electric motor 200 can be ensured even when the inertia model of the drive mechanism 500 is unknown or fluctuates.

  As described above, according to the present embodiment, even when both the moment of inertia and the multi-inertia model in the drive mechanism 500 are unknown or can be changed, it is possible to control the drive of the electric motor 200 with high robustness without adjusting various parameters. It becomes.

  As a result, it is possible to maintain the robust performance with the drive mechanism 500 including mechanical resonance characteristics, which has been impossible in the past. In addition, the mechanical resonance low tolerance (ease of oscillation), which was a weak point of the conventional disturbance suppression control system, can be greatly improved.

  Since the resonance suppression model 43 can be fixed by an appropriate design, a complicated control object having mechanical resonance characteristics can be operated without adjustment with a fixed control system and fixed parameters. The conventional motor control device can be operated only after adjusting the control gain individually (including manual and automatic) according to the mechanical characteristics of the drive mechanism 500. In other words, the conventional motor control device was high in performance but difficult to use, but in this embodiment, the use of the motor control device was made easier and the application range was expanded, and the setup time of the mechanical device was greatly increased. Since it can be shortened, it is very useful in using the motor control device.

In this embodiment, the resonance suppression model 43 is an inverse system of the resonance suppression nominal model proposed in the present invention. This resonance suppression nominal model is a frequency characteristic control having a plurality of time constants T ₁ and T ₂ determined by a plurality of frequency setting values f ₁ and f ₂ that can change the gain change rate of the frequency characteristic of the rigid body nominal model. (Not shown, (1 + T ₁ s) / (1 + T ₂ s)). By appropriately setting a plurality of time constants of the frequency characteristic controller, it is possible to design the resonance suppression nominal model by deforming it to the extent that resonance suppression is possible using the mechanical resonance characteristics of the control target as a prototype. As a result, the resonance suppression nominal model and the resonance suppression model 43 can be designed to appropriately correspond to the control targets (the electric motor 200 and the drive mechanism 500) that can be assumed.

In this embodiment, in particular, the plurality of time constants T ₁ and T ₂ indicate that the resonance suppression nominal model has a gain margin on the frequency characteristics in all frequency bands with respect to the mechanical resonance characteristics of the electric motor 200 and the drive mechanism 500. And at least one of the phase margin is set so as to be maintained.

By appropriately setting a plurality of time constants T ₁ and T ₂ determined by the frequency setting values f ₁ and f ₂ , a resonance suppression nominal model in which the slope of the rigid nominal model is changed is obtained from the mechanical resonance characteristics of the controlled object. Can also be made high in general. Furthermore, by appropriately setting a plurality of time constants T ₁ and T ₂ , the gain characteristics can be handled even when the top of some mechanical resonance characteristics exceeds the resonance suppression nominal model and the gain margin is impaired. With respect to the phase characteristics in the band, the phase characteristics of the resonance suppression nominal model greatly advance and can be compensated for by the phase margin. In this manner, the frequency characteristic controller can be designed so that the resonance suppression nominal model can control at least one of the gain margin and the phase margin on the frequency characteristic with respect to the mechanical resonance characteristic to be controlled. And by estimating the disturbance torque using the resonance suppression model, which is the inverse system of the resonance suppression nominal model equipped with this frequency characteristic controller, the overall gain margin and phase for any multi-inertia model The machine resonance can be suppressed by securing a margin.

In this embodiment, in particular, the resonance suppression model 43 in the first disturbance observer 31 is a round-trip transmission in a disturbance suppression control system loop including the inertia variation suppression unit 3, the first current control unit 4, the electric motor 200, and the drive mechanism 500. Characteristic G _Loop (Equation (1) above) is controlled so that the phase can be maintained at -180 degrees and the gain is less than 1 in the low frequency band, and the phase can be maintained at other than -180 degrees in the frequency domain where the gain is 1 or more in the high frequency band. A frequency characteristic controller inverse system 51 (1 + T ₂ s) / (1 + T ₁ s) having a plurality of time constants T ₁ and T ₂ is set. This inverse system 51 of the frequency characteristic controller controls the frequency characteristic with respect to the motor speed of the electric motor 200. Then, by converting the output of the inverse system 51 of the frequency characteristic controller, which is a signal corresponding to speed, into a signal corresponding to torque by multiplication of the moment of inertia J ₀ and differentiation of the operator s, the resonance suppression model 43 is mechanically The acceleration / deceleration torque signal of the ideal motor that suppresses the influence of resonance can be output.

Since the moment of inertia J ₀ is the moment of inertia of the rotor of the motor 200, it can be regarded as a disturbance element loads the entire drive mechanism 500. Thereby, parameter adjustment (adjustment of a time constant, an attenuation coefficient, various gains, etc.) performed with respect to the motor control apparatus 100 sets each parameter at the time of factory shipment corresponding to the various electric motors 200 applied as a control object. Therefore, parameter adjustment is not necessary for the drive mechanism 500 driven by the electric motor 200. A frequency characteristic controller (not shown) which is a component of the resonance suppression nominal model has a plurality of time constants T ₁ and T ₂ determined by a plurality of changeable frequency setting values f ₁ and f _2. Since it is defined by a function of (1 + T ₁ s) / (1 + T ₂ s), the disturbance suppression control system using the resonance suppression model that is the inverse system of the resonance suppression nominal model is suitable for any multi-inertia model. In addition, the overall gain margin and phase margin can be ensured and fluctuations in the moment of inertia of the drive mechanism can be suppressed without causing mechanical resonance.

  In this embodiment, in particular, the first disturbance observer 31 outputs the acceleration / deceleration torque signal of the ideal motor that suppresses the influence of the mechanical resonance obtained by the resonance suppression model 43 based on the first motor speed of the motor 200 as the second torque. Subtract from the command to estimate the disturbance torque.

  In this embodiment, in particular, the current loop model 41 in the first disturbance observer 31 processes the second torque command in the same manner as the first current control unit 4, so that the ideal motor addition obtained by the resonance suppression model 43 is obtained. Since the phase difference between the deceleration torque signal (torque signal calculated from the motor speed) and the torque signal (torque signal as a command) output from the current loop model can be reduced, the disturbance torque can be accurately estimated, and the inertia variation suppression unit 3 The response characteristics of the are improved. As a result, the response frequency in the speed control loop can be improved to a high frequency band, and a larger moment of inertia ratio can be handled. That is, the robustness with respect to the inertia moment ratio of the drive mechanism 500 can be improved.

  The present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the spirit and technical idea of the present invention. Hereinafter, such modifications will be described in order.

(1) When the second phase compensation unit as a phase advance filter is provided In the above embodiment, the phase margin of the position control system is improved by advancing the phase of the speed control system by providing the first phase compensation unit 7. However, the phase margin of the position control system can be further improved by providing the second phase compensator that advances the phase by a specific frequency band.

  In FIG. 33 corresponding to FIG. 1 of the above embodiment, in the motor control device 100A of the first modified example, a second phase compensation unit 11 is newly provided between the position control unit 1 and the subtracter 9. That is, the second phase compensation unit 11 inputs the first speed command output from the position control unit 1, advances the phase by a specific frequency band, and outputs the second speed command to the subtracter 9.

  For example, as shown in FIG. 34, the second phase compensation unit 11 is provided with a low-pass filter 71 and a high-pass filter 72, and outputs the respective outputs added by an adder 73. What is necessary is just a control block using.

Phase advance transfer of the filter function G (s) is represented by equation (3), the relationship of constant T _L and the high-pass constant T _H when the filter when the low-pass _{filter, 1 / T H <1 /} T Each time constant is adjusted to satisfy the condition of _L , that is, the relationship between the cutoff frequency ω ₁ (= 1 / T _L ) of the low-pass filter and the cutoff frequency ω ₂ (= 1 / T _H ) of the high-pass filter. The phase delay can be improved by adjusting the cut-off frequencies so that ω ₂ <ω ₁ . If the conditions are opposite, a phase delay filter is obtained and the phase advance effect is lost.
G (s) = (1 + sT _H ) / (1 + sT _L )
_{= 1 / (1 + sT L} ) + sT H / (1 + sT L)
= Ω _L / (ω _L + s) + T _H ω _L s / (ω _L + s) (3)
Also, the configuration of the phase advance filter can be written as in the following equation (4).
G (s) = ω _L / (ω _L + s) + T _L ω _H s / (ω _H + s) (4)
At this time, by setting the cutoff frequency of the low-pass filter 71 higher than the cutoff frequency of the high-pass filter 72, a phase advance filter is obtained.

  The first speed command from the position control unit 1 is input to the low-pass filter 71 and the high-pass filter 72, and their outputs are added by the adder 73 and output as a second speed command. The phase does not advance with the low-pass filter 71 alone, and the phase advances to a low frequency band that is not necessary with the high-pass filter 72 alone. The high-pass filter 72 is originally compensated by the speed loop gain or speed integration. The gain in the low frequency band will be reduced. However, by configuring the second phase compensator 11 as illustrated, it functions as a phase advance filter that advances the phase by a desired specific frequency band. The low-pass filter 71 corresponds to the second low-pass filter described in each claim.

  Further, when the speed loop gain (not shown) in the speed control unit 2 is increased, the phase to be compensated for by the second phase compensation unit 11 is shifted, so that vibration is generated. On the other hand, in the second phase compensation unit 11, the cutoff frequencies of the low-pass filter 71 and the high-pass filter 72 may be changed according to the speed loop gain in the speed control unit 2. Therefore, by simulation or actual machine test, the values at which the cutoff frequencies of the low-pass filter 71 and the high-pass filter 72 are optimal corresponding to changes in the speed loop gain in the speed control unit 2 are sequentially plotted on the graph (in particular, (Not shown), an approximate expression for interpolating the plot of the actual measurement data is obtained.

For example, when using a disturbance observer, such as shown in Figure 2 as in this modified example, the cutoff frequency of the high-pass filter 72 is previously fixed, the cutoff frequency f _c of the low-pass filter 62, f _c = An approximate expression can be determined by a quadratic function of the speed loop gain K _v such as a × K _v ² + b × K _v + c. In addition, what can be approximated by a linear function of f _c = a × K _v + b can also be used. Alternatively, it is acceptable to fix the cut-off frequency of the low pass filter 62, can be approximated the cutoff frequency by a quadratic function or a linear function of the cutoff frequency f _c is similarly speed loop gain K _V for the high-pass filter 72. Using these approximate expression expressed by a polynomial of the speed loop gain K _v and independent variables, for example, when changing the speed loop gain K _v in the speed control section 2, a speed observer by the function first The cutoff frequency of the low-pass filter 71 or the high-pass filter 72 of the two-phase compensator 11 can be changed appropriately, and the phase compensation value can be readjusted automatically. That is, the cutoff frequency of the low-pass filter 71 or the high-pass filter 72 of the second phase compensation unit 11 is appropriately changed using an approximate expression expressed by a polynomial having the velocity loop gain _Kv as an independent variable. Thus, the phase compensation value can be readjusted automatically.

  FIGS. 35, 36, and 37 are data diagrams corresponding to FIGS. 22, 24, and 26, respectively, in the above-described embodiment. As can be seen from these, FIG. 21, FIG. 23, and FIG. In comparison, the variation suppressing effect is remarkably exhibited also in this modification, that is, it can be confirmed that the robustness against mechanical resonance is improved in any inertia model.

  FIG. 38 to FIG. 43 are diagrams comparing the experimental results of the conventional comparative example and the present modification for each moment of inertia ratio, and show the case where the position command is continuously changed to perform the positioning operation. As can be seen from these, in the conventional comparative examples of FIGS. 38, 40, and 42, the torque command is likely to vibrate as the inertia moment ratio increases, whereas this modification of FIGS. 39, 41, and 43 is used. However, regardless of the magnitude of the moment of inertia ratio, fluctuations in the torque command are suppressed and the vibration is very stable. That is, in this modification, it can be confirmed that the robustness with respect to the inertia moment ratio is high.

  As described above, according to the motor control device 100A of this modification, the phase margin of the position control system is improved by the second phase compensation unit 11 provided in the position control system loop functioning as a phase advance filter. Thus, the response characteristic of the inertia variation suppressing unit 3 is improved. That is, the robustness with respect to the inertia moment ratio of the drive mechanism 500 can be improved.

  In the above embodiment, the current loop model 41 provided in the first disturbance observer 31 reduces the delay of the current loop with respect to the torque command, but the present invention is not limited to this. In addition, although not particularly illustrated, the response characteristic of the inertia variation suppressing unit 3 may be improved by providing a phase compensation current observer inside the current control unit to reduce the delay of the current loop.

(2) When a current loop inverse model is provided in the disturbance observer In FIG. 44, the difference between the second disturbance observer 31A and the first disturbance observer 31 in FIG. 2 is that the current loop model 41 (broken line frame portion) is not provided. Instead, a current loop inverse model 41A is provided on the output side of the resonance suppression model 43. Here, the current loop inverse model 41A is an inverse system of the current loop model 41. For example, when the transfer function denominator and the numerator are reversed, the order of the numerator operator s becomes higher than the order of the denominator. Cannot be realized. In order to deal with this for the convenience of mounting, although not particularly shown, the low-pass filter 32 outside the second disturbance observer 31A is moved to two paths on the input side of the internal subtractor 44, respectively. May be multiplied by the current loop inverse model 41A to reduce the numerator order of the transfer function of the inverse model. The current loop inverse model 41A advances the phase of the ideal motor acceleration / deceleration torque signal (torque signal calculated from the motor speed) output from the resonance suppression model 43, thereby determining the acceleration / deceleration of the ideal motor obtained by the resonance suppression model 43. Since the phase difference between the torque signal (torque signal calculated from the motor speed) and the second torque command (torque signal as a command) can be reduced, the disturbance torque can be estimated accurately.

  Here, as described above, when a high response characteristic is required for the disturbance suppression control system, it is necessary to add the model of the current control unit that has been ignored in the design of the disturbance observer to the control target. However, in this case, the order of the disturbance observer becomes extremely large, which is not practical in software implementation. Even if you can design a disturbance observer that incorporates the current control unit model with luxury hardware resources, the disturbance observer's control bandwidth is narrower than the low-order disturbance observer, so the robustness against fluctuations in the moment of inertia is reduced. As a result, the purpose of increasing robustness cannot be achieved.

In order to solve this point, it is necessary to improve the delay of the current control unit. Therefore, in this modification, the current loop model 41 is provided with a function of phase advance compensation. In contrast to the actual current control system P _i (s), the current loop inverse model 41A is ideally set to P _i ⁻¹ (s), but the actual current control system P _i (s) is complicated. Therefore, (1 + T ₄ s) / (1 + T ₃ s) is assumed as a simple phase advance compensator. Since it is forward compensation, the time constant is selected so that T ₃ <T ₄ . As a result, the delay of the actual current control system P _i (s) is reduced, and the transfer function approximates an ideal form of “1”. Therefore, the transfer function of the current loop inverse model 41A in the disturbance observer is set to “1 (× phase Since the error with the actual current control system P _i (s) is small even when approximated by “advance compensation”, the disturbance observer can be designed in a wide band, and the robust performance of the disturbance suppression control system can be improved.

となる。また、上記（２）式に相当する閉ループ伝達関数Ｇ_{Ｃｌｏｓｅ}は、

となる。 Therefore, the round transfer function G _Loop corresponding to the above equation (1) in this case is

It becomes. Further, the closed loop transfer function G _Close corresponding to the above equation (2) is

It becomes.

  45, 46, and 47 are data diagrams corresponding to FIGS. 22, 23, and 26, respectively, in the above-described embodiment. As can be seen from these data, FIG. 21, FIG. 23, and FIG. In comparison, the variation suppressing effect is remarkably exhibited also in this modification, that is, it can be confirmed that the robustness against mechanical resonance is improved in any inertia model. In the present modification, both the first phase compensation unit 7 and the second phase compensation unit 11 are provided as in the first modification.

  48 to 53 are diagrams comparing the simulation results of the first modification and the present modification according to the moment of inertia ratio. As can be seen from these figures, as the moment of inertia ratio increases, the present modification of FIGS. 49, 51, and 53 is more suitable for the electric motor 200 than the first modification of FIGS. 48, 50, and 52. It can be confirmed that fluctuations in the motor speed and the torque command can be suppressed (particularly, the broken line circle).

  As described above, according to the motor control device 100B of this modification, phase advance compensation is performed by the current loop inverse model 41A. As a result, the delay of the actual current control system is reduced and the transfer function approximates to an ideal form. Therefore, even if the transfer function of the current loop in the second disturbance observer 31A is approximated by “1”, the actual current control system The error becomes smaller. For this reason, the second disturbance observer 31A can be designed in a wide band, and the response characteristics of the inertia variation suppressing unit 3 are improved. That is, the robustness with respect to the inertia moment ratio of the drive mechanism 500 can be further improved.

  In the above-described embodiment and each modification, the case where the electric motor 200 is a rotary motor has been described. However, the present invention is not limited to this, and can be applied to a direct acting motor. In that case, replace the term related to torque with thrust, the term related to moment of inertia with inertial mass, the rotary encoder with linear scale, the accumulated rotation position with accumulated movement position, and the rotor with mover. Good.

  In addition to those already described above, the methods according to the above-described embodiments and modifications may be used in appropriate combination.

  In addition, although not illustrated one by one, the present invention is implemented with various modifications within a range not departing from the gist thereof.

DESCRIPTION OF SYMBOLS 1 Position control part 2 Speed control part 3 Inertia fluctuation suppression part 4 1st current control part (torque control part, thrust control part)
4A Second current control unit (torque control unit, thrust control unit)
DESCRIPTION OF SYMBOLS 5 Position detection part 6 Speed calculation part 7 1st phase compensation part 11 2nd phase compensation part 31 1st disturbance observer 31A 2nd disturbance observer 32 Low-pass filter 33 Gain 41 Current loop model 41A Current loop inverse model 43 Resonance suppression model 51 Inverse system of frequency characteristic controller 52 Inverse model to be controlled 62 Low-pass filter (first low-pass filter)
71 Low-pass filter (second low-pass filter)
DESCRIPTION OF SYMBOLS 100,100A Electric motor control apparatus 200 Electric motor 201 Rotary motor 500 Drive mechanism

Claims (9)

A position detector for detecting the motor position of the motor that drives the drive mechanism;
A speed calculator that inputs the motor position and calculates a first motor speed of the motor;
A position control unit for inputting a difference between a position command and the motor position and outputting a first speed command;
A speed control unit for inputting a difference between the first speed command and the second motor speed and outputting a first torque command or a first thrust command;
A first phase compensator for outputting the second motor speed advanced in phase with respect to the first motor speed;
A disturbance observer having a resonance suppression model that is set to suppress mechanical resonance of the drive mechanism and that can change a gain change rate of a frequency characteristic of a rigid-body nominal model, and a disturbance torque estimated using the disturbance observer or An inertia fluctuation suppressing unit that outputs a second torque command or a second thrust command based on a disturbance thrust;
A motor control device comprising: a torque control unit or a thrust control unit that inputs the second torque command or the second thrust command and controls torque or thrust of the motor. The inertia variation suppression unit
2. The motor control device according to claim 1, wherein the disturbance torque or the disturbance thrust is added to the first torque command or the first thrust command to output the second torque command or the second thrust command. . The resonance suppression model in the disturbance observer is an inverse system of a resonance suppression nominal model,
The resonance suppression nominal model is
3. The motor control device according to claim 1, further comprising a frequency characteristic controller having a plurality of time constants determined by a plurality of frequency setting values. The plurality of time constants of the frequency characteristic controller are:
The resonance suppression nominal model is set to control to maintain at least one of a gain margin and a phase margin on the frequency characteristics in all frequency bands with respect to the mechanical resonance characteristics of the electric motor and the drive mechanism. The electric motor control device according to claim 3. The resonance suppression model in the disturbance observer is
The loop transfer characteristic in the disturbance suppression control system loop including the inertia variation suppression unit, the torque control unit or the thrust control unit, the electric motor, and the drive mechanism has a phase of −180 degrees and a gain of less than 1 in the low frequency band. In addition, the frequency characteristic controller has an inverse system in which the plurality of time constants are set so as to control the phase so that the phase can be maintained at other than −180 degrees in a frequency region of a gain of 1 or more in a high frequency band,
5. The motor control device according to claim 3, wherein the value obtained by passing the first motor speed through the inverse system of the frequency characteristic controller is multiplied by an inertia moment or an inertial mass of the motor, and further differentiated. . The disturbance observer is
The second torque command or the second thrust command;
Subtracting the acceleration / deceleration torque signal of the ideal motor or the acceleration / deceleration thrust signal of the ideal motor calculated using the resonance suppression model based on the first motor speed, and outputting the disturbance torque or the disturbance thrust The electric motor control device according to any one of claims 1 to 5, wherein The disturbance observer is
A current loop model that simulates a current control loop in the torque control unit or the thrust control unit;
A value obtained by passing the second torque command or the second thrust command through the current loop model;
Subtracting the acceleration / deceleration torque signal of the ideal motor or the acceleration / deceleration thrust signal of the ideal motor calculated using the resonance suppression model based on the first motor speed, and outputting the disturbance torque or the disturbance thrust The electric motor control device according to any one of claims 1 to 5, wherein The disturbance observer is
A current loop inverse model that is an inverse model of a current control loop in the torque control unit or the thrust control unit;
The second torque command or the second thrust command;
The disturbance is obtained by subtracting the ideal motor acceleration / deceleration torque signal calculated using the resonance suppression model based on the first motor speed or the value obtained by passing the ideal motor acceleration / deceleration thrust signal through the current loop inverse model. The motor control device according to any one of claims 1 to 5, wherein torque or the disturbance thrust is output. The motor controller is
A second low-pass filter in which a cut-off frequency calculated by a polynomial having a speed loop gain in the speed control unit as an independent variable is set, and the first speed command is input and based on the second low-pass filter A second phase compensator for outputting a second speed command having advanced phase;
The speed control unit inputs a difference between the second speed command and the second motor speed instead of the first speed command, and outputs the first torque command or the first thrust command. The motor control device according to any one of claims 1 to 8.

Images