JPWO2010024194A1 - Electric motor pulsation suppression device - Google Patents

Abstract

[PROBLEMS] To reliably and easily learn a pulsation compensation signal while performing stable control while preventing overcompensation by a torque pulsation compensation signal or the like, unstable vibration, and interference. A controller 5 performs a Fourier transform on a shaft torque detection value of an electric motor 1 driven by an inverter 7, detects a torque pulsation component with two Fourier coefficients, and generates a compensation current for suppressing the torque pulsation by a PID learning controller. Learning is performed, and torque pulsation is suppressed by superimposing torque pulsation compensation current on d and q axis current command values id * and iq * in vector control. The controller 5 constitutes a learning control system for suppressing torque pulsation by detecting two Fourier coefficients on a complex vector plane. Also, the controller uses the complex vector plane to represent the frequency characteristics of the system to be controlled or the system identification result as a complex vector, and based on the interference term of two Fourier coefficients from the pulsation compensation current and the motor drive system. , Deactivate the real and imaginary parts of the complex vector.

Description

  The present invention relates to an apparatus for suppressing torque pulsation or the like generated due to the structure of an electric motor in an electric motor torque control apparatus.

  Since the electric motor has structural magnetic flux distortion and cogging torque, it generates torque pulsation that contributes to vibration and noise according to the rotation. In addition, when a multi-inertia system is configured between the motor and the load, the mechanical resonance point and the torque pulsation frequency component coincide with each other, generating an excessive shaft torsion torque, which adversely affects operating characteristics and damages the system. There is a danger of.

  As countermeasures for these, there are a method for suppressing shaft torsional resonance by feedback control and a method for reducing torque pulsation itself as a basis of resonance by feedback control. However, in the resonance suppression method using feedback, resonance suppression becomes difficult when the resonance frequency and the response frequency of the inverter are close, or when the sampling time of the control system cannot be sufficiently shortened.

  In order to solve these problems, a compensation signal that cancels the torque ripple of the IPM motor output is added to the d, q-axis current command or torque command of the IPM motor by feedforward control.

  On the other hand, torque pulsation generated in an electric motor is complicatedly associated with various factors such as magnetic imperfection of a motor structure, response / current error of an inverter driving the motor, and characteristics of a mechanical system. However, when focusing on the pulsation cycle of the motor, it has been found that it occurs mainly depending on the rotor position, and a method for generating a compensation signal repeatedly considering this as a periodic disturbance has been conventionally considered. .

  For example, there is a method in which a detected value of torque pulsation is Fourier transformed and the Fourier coefficient is controlled by a PI controller (see, for example, Patent Document 1). When the torque pulsation detection value is Fourier-transformed, two Fourier coefficients of sine and cosine can be derived. Therefore, if PI control is performed so that both of these coefficients become zero, torque pulsation of the frequency component having the coefficient is obtained. Can be suppressed.

  A block configuration of this method is shown in FIG. 17, and a tension control device that obtains a current command Is in the PI control unit according to the deviation between the tension command Ts and the detection signal Tf and controls the current of the motor M by the current control unit ACR. , The sine term coefficient and cosine term coefficient of the pulsation included in the tension detected by the tension sensor are calculated, the deviation between these and the command value “0” is obtained by PI control, and these are converted into a sine wave signal and a cosine wave signal. This is converted and added as a sine wave and cosine wave correction value of the current command.

Japanese Patent Publication No. 8-17585, which is a Japanese patent publication

  In the above-mentioned Patent Document 1, the torque pulsation of the frequency component can be suppressed by performing PI control so that the two Fourier coefficients are both zero. However, Patent Document 1 does not propose a method for adjusting a PI control parameter, and is limited to application to a simple system that does not require strict parameter adjustment. For example, if a torque pulsation suppressing device in which the PI control parameter is not adjusted is applied to a multi-inertia system, overcompensation is caused at the mechanical resonance point, or unstable vibration is caused.

  On the other hand, among the electric motors, the embedded magnet type synchronous motor is a high-efficiency electric motor that can effectively utilize not only the torque generated by the permanent magnet but also the reluctance torque using magnetic anisotropy. On the other hand, from the viewpoint of torque pulsation, both magnet torque pulsation and reluctance torque pulsation occur in a composite manner. In this regard, Patent Document 1 describes only superimposing the pulsation compensation current on the current command value, and touches on the problem that the reluctance torque of the embedded magnet synchronous motor and the pulsation of the magnet torque interfere with each other due to the compensation current. Not. Also, it is not specified that vector control is performed with the dq axes subjected to rotational coordinate conversion. Therefore, in the case of an embedded magnet synchronous motor, torque pulsation cannot always be optimally suppressed, and other frequency components may be adversely affected.

  An object of the present invention is to provide a pulsation compensation signal in a pulsation suppression device that detects a torque pulsation component or the like in the form of a Fourier coefficient and obtains a pulsation compensation signal by controlling a PI controller that sets the magnitude of the Fourier coefficient to “0”. It is an object of the present invention to provide a pulsation suppression device for an electric motor that can reliably and easily learn a pulsation compensation signal while performing stable control while preventing overcompensation, unstable vibration, and interference due to the above.

  In order to solve the above-described problems, the present invention performs Fourier transform on the torque of an electric motor driven by an inverter, detects a pulsation component of an arbitrary frequency in the form of two Fourier coefficients, and pulsation by detecting these two Fourier coefficients. The suppression learning control system is configured with a complex vector plane so as to obtain the pulsation compensation current, and has the following configuration.

(1) A pulsation detection value of an electric motor driven by an inverter is Fourier transformed, a pulsation component of an arbitrary frequency is detected in the form of two Fourier coefficients, and learning control is performed by a learning controller to suppress the pulsation, and vector control is performed. In the pulsation suppressing device for an electric motor that suppresses the pulsation by superimposing the learned pulsation compensation current on both the d-axis current command value or the q-axis current command value or the d-axis q-axis current command value of the rotating coordinate system in
The pulsation suppression learning control system based on the detection of the two Fourier coefficients is configured by a complex vector plane, and has means for obtaining the pulsation compensation current.

  (2) The means for obtaining the pulsation compensation current is expressed as a complex vector using the frequency characteristic of the system to be controlled or the system identification result on the complex vector plane, and two Fouriers are obtained from the pulsation compensation current and the motor drive system. A decoupling control element for decoupling the real part and the imaginary part of the complex vector based on the interference term of the coefficient is provided.

  (3) The means comprises a PID controller as a learning controller, obtains a compensation current from the output of the PID controller through the non-interacting control element, and sets the gain of the PID controller to a desired closed-loop response. It is characterized by having learning control means arranged in a pole.

  (4) A low frequency pass filter having an initial value of 0 is inserted in the output stage of the learning controller, and is poled so as to obtain a desired closed-loop response. Means is provided for preventing the occurrence of overshoot.

  (5) Insert a low-frequency pass filter with an initial value of 0 that functions only at the start of pulsation suppression into the feedback signal path to the learning controller, and after the start of pulsation suppression and after the time constant of the low-frequency pass filter has elapsed Means is provided for switching to a route not passing through the filter.

  (6) The gain of the learning controller is provided with an automatic control gain adjusting means for adjusting and automatically adjusting the gain in accordance with the rotational speed of the electric motor.

  (7) The Fourier transform cycle of the learning controller is changed in accordance with the rotational speed of the electric motor.

  (8) A means for setting the response frequency of the learning controller in conformity with a Fourier transform period is provided.

  (9) The learning controller may be applied to different frequency components, and parallelized so that a means for simultaneously suppressing torque pulsations of a plurality of frequency components is provided.

  (10) The pulsation component is a shaft torque detection value of the electric motor, and suppresses pulsation of the shaft torque of the electric motor.

  (11) The pulsation component is a pulsation component of the frame of the electric motor, and suppresses the pulsation of the frame of the electric motor.

  (12) The pulsation component is a pulsation component of a rotational speed detection value or a rotational position detection value of an electric motor, and suppresses the pulsation of the speed of the electric motor or the pulsation of the rotational position.

  (13) The pulsating component is a pulsating component of a current of the electric motor, and suppresses the pulsating of the electric current of the electric motor.

(14) The amplitude M and phase φ of the torque ripple compensation current i _qc ^* learned in the steady operation state (steady torque command / steady rotation speed) designated when the torque ripple suppression control is executed by the controller are used as the torque command T _ref ^* And an amplitude / phase compensation table for generating a two-dimensional amplitude table and a phase table with the rotational speed ω as a variable,
The amplitude M and phase φ at that time are read out from the amplitude / phase compensation table from the torque command T _ref ^* and the rotational speed ω in the operating state of the motor, and the amplitude M and phase φ and the rotational phase θ of the motor at that time are read out. A compensation current generation unit that generates a table compensation current i _qct ^* = M · sin (nθ + φ) using
Torque pulsation compensation for detecting the pulsation component of the same frequency as the table compensation current i _qct ^{* in} the form of two Fourier coefficients and making this pulsation component zero by Fourier transforming the pulsation detection value T _det in the motor operating state _Torque ripple suppression control means for _obtaining current i _qcc ^* ;
Compensation current synthesis means for synthesizing the torque pulsation compensation current i _qcc ^* and the table compensation current i _qct ^* and superimposing them on the d and q axis current command values is provided.

(15) The amplitude M and the phase φ of the torque ripple compensation current i _qc ^* learned in the steady operation state (steady torque command / steady rotation speed) designated when the torque ripple suppression control is executed by the controller are used as the torque command T _ref ^* And an amplitude / phase compensation table for generating a two-dimensional amplitude table and a phase table with the rotational speed ω as a variable,
The pulsation detection value T _det in the operating state of the motor is Fourier-transformed, the pulsation component having the same frequency as the table compensation current i _qct ^* is detected in the form of two Fourier coefficients, and the compensation current amplitude is set to zero. Torque ripple suppression control means for obtaining a command value M ^* and a compensation current phase command value φ ^* ;
The amplitude M and phase φ at that time are read out from the amplitude / phase compensation table from the torque command T _ref ^* and the rotational speed ω in the operating state of the motor, and the amplitude M and phase φ and the compensation current amplitude command value are read out. A synthesis means for synthesizing M ^* and the compensation current phase command value φ ^* , respectively, to generate a synthesis compensation current amplitude value M ′ and a synthesis compensation current phase value φ ′;
Compensation current generation for generating a table compensation current i _qc ^* = M ′ · sin (nθ + φ ′) using the combined compensation current amplitude value M ′, the combined compensation current phase value φ ′ and the rotational phase θ of the motor at that time And a section.

(16) Compensation current cosine Fourier coefficient table value A and compensation current sine Fourier coefficient table of torque ripple frequency components learned in a steady operation state (steady torque command / steady rotation speed) designated when torque ripple suppression control is executed by the controller A Fourier coefficient compensation table for generating the value B as a two-dimensional table value with the torque command T _ref ^* and the rotational speed ω as variables;
Two compensation current cosine Fourier coefficients A ^* and a compensation current sine for Fourier transforming the pulsation detection value T _det in the operating state of the motor to make the pulsation component of the same frequency as the table values A and B of the Fourier coefficient compensation table zero. Torque ripple suppression control means for obtaining a Fourier coefficient B ^* ;
The Fourier coefficient table values A and B at that time are read from the Fourier coefficient compensation table from the torque command T _ref ^* and the rotational speed ω in the operating state of the electric motor, and these Fourier coefficient table values A and B, and the compensation current A synthesis means for synthesizing the cosine Fourier coefficient A ^* and the compensation current sine Fourier coefficient B ^* to generate a compensation current cosine Fourier coefficient synthesis value A ′ and a compensation current sine Fourier coefficient synthesis value B ′;
Using the Fourier coefficient composite value A ′, the compensation current sine Fourier coefficient composite value B ′, and the rotational phase θ of the motor at that time, the amplitude M = √ (A ′ ² + B ′ ² ) and the phase φ = tan ⁻¹ ( B ′ / A ′) table compensation current i _qc ^* = M · sin (nθ + φ).

  (17) The compensation table, the synthesizing means, the torque ripple suppression control means, and the compensation current generating section individually generate the amplitude M and the phase φ or the Fourier coefficient with a plurality of order components, and generate the compensation current by combining these orders. A means for generating is provided.

The block diagram of the torque control apparatus of the electric motor which concerns on this invention.

The block block diagram of the learning control system which concerns on this invention.

1 is a schematic diagram of a system identification device in Embodiment 1. FIG.

The characteristic figure of a torque pulsation compensation current and an identification system.

The block diagram of non-interacting control.

FIG. 6 is a configuration diagram of non-interference control in the second embodiment.

The block diagram of the control system after non-interference implementation is realized.

The control block diagram after non-interference.

FIG. 10 is a block diagram of non-interference control in the fourth embodiment.

The block diagram of a primary low-pass filter.

FIG. 6 is an IP control waveform diagram according to the third and fourth embodiments.

FIG. 10 is a configuration diagram of non-interference control in the sixth embodiment.

FIG. 10 is a configuration diagram of non-interference control in the seventh embodiment.

FIG. 10 is a configuration diagram of non-interference control in the eighth embodiment.

The block diagram of the non-interacting control in Embodiment 9. FIG.

FIG. 18 is a configuration diagram of non-interference control in the tenth embodiment.

The block diagram of the tension control apparatus with a torque pulsation suppression block.

The block diagram of the torque control apparatus of the electric motor in Embodiment 11. FIG.

The block diagram of the torque control apparatus of the electric motor in Embodiment 12. FIG.

The block diagram of the torque control apparatus of the electric motor in Embodiment 13. FIG.

The block diagram of the torque control apparatus of the electric motor in Embodiment 14. FIG.

(Basic configuration)
Before describing embodiments of the present invention, the basic configuration of a pulsation suppressing device according to the present invention will be described.

FIG. 1 shows a configuration of a torque control device for an electric motor according to the present invention. An electric motor 1 that is a generation source of torque pulsation and some load device 2 are coupled by a shaft 3, and the shaft torque is measured by a torque meter 4 and input to a controller 5. Moreover, the rotor position information of the electric motor is input using a rotational position sensor 6 such as a rotary encoder. The controller 5 is equipped with torque pulsation suppression means, and gives the inverter 7 a command value obtained by adding a torque pulsation compensation current to the current command value generated based on the torque command value (or speed command value). In the example of FIG. 1, in consideration of the current vector control by the inverter 7, the d-axis and q-axis current command values i _d ^* and i _q ^* on the rotation coordinates (orthogonal dq axes) synchronized with the rotation of the motor are expressed as follows. Giving.

  The controller 5 detects the torque pulsation from the torque detection value by the shaft torque meter 4, but this form is only an example, frame vibration detection by an acceleration sensor installed in the frame, rotation speed / position fluctuation detection by an encoder or the like. Alternatively, it can be replaced with current pulsation detection by a current sensor to suppress motor shaft pulsation, motor frame pulsation, motor speed pulsation or rotational position pulsation, and motor current pulsation. is there.

  In the apparatus configuration of FIG. 1, it is known that torque pulsation periodically occurs according to the rotor position due to the structure of the motor. The torque pulsation of the three-phase motor is mainly a frequency 6 × n times the electrical rotation fundamental frequency (n is a positive integer: hereinafter, the 6-fold component is expressed as 6f, the 12-fold component is expressed as 12f, etc.) It is also known that the component becomes large. In addition, 1f and 2f torque pulsations may appear due to incompleteness of the inverter. Further, in a system in which axial torsional resonance can occur as shown in FIG. 1, a pulsation component close to the mechanical resonance point appears after amplification.

In the present invention, the torque pulsation in which these various frequency components are mixed is extracted based on Fourier transform. That is, the Fourier transform for each arbitrary order is performed on the detected value T _det of the shaft torque meter 4 in FIG. 1 using the motor rotor phase θ. Here, the Fourier series is defined by the following equation. ω represents an electrical angular frequency, and _Tr represents one cycle of a pulsating component that is arbitrarily defined.

ただし、

However,

It is noted that the Fourier coefficients T _An and T _Bn expressed by Equation 2 above output a direct current value if the motor rotation speed and the pulsation component that is a periodic disturbance are constant in an arbitrary period. If some control is given and both of these coefficients are zero and steady, it means that the torque pulsation of the frequency component is suppressed.

  Therefore, in the present invention, a PI control system with a target value of 0 is considered, and a method of learning control is proposed so that the Fourier coefficients of torque pulsation components are both 0. This method is similar to FIG. 17 in Patent Document 1 described above, but in the present invention, as shown in FIG. 2, the method of Patent Document 1 is extended to a vector control system, and the divergence prevention processing of a multi-inertia system is performed. Further, the torque pulsation is suppressed with high accuracy by compensating the torque pulsation in a learning manner based on the actual inverter drive device and the shaft torque feedback signal.

With reference to FIG. The sine / cosine wave generator 11 sets the rotor phase angle θ of the motor and the order n of the pulsating component to be suppressed, and generates a sine / cosine wave having a reference frequency. In the Fourier transform unit 12, the sine / cosine wave and an arbitrary Fourier transform period _Tr are set, and the torque detection value T _det is Fourier transformed based on Equation 2 to calculate n-order Fourier coefficients T _An and T _Bn . To do. Here, it is also possible to replace the integration process of the Fourier transform of Equation 2 with a low frequency pass filter (low pass filter) having an appropriate time constant.

Next, the two PI controllers 13A and 13B are controlled so that the Fourier coefficients of these pulsation components become the target value 0 (that is, the pulsation is suppressed). The configuration method of the PI controller is arbitrary, and outputs I _An and I _Bn for controlling the Fourier coefficient of torque pulsation to 0 are directly used as the Fourier coefficients of the compensation current signal, and multiplied by the same sine / cosine wave as the Fourier transform unit. Multipliers 14A and 14B respectively perform inverse transformation. That is, it is restored as a compensation current signal having a time waveform. The restored torque pulsation compensation current is superimposed on the d-axis or q-axis current command value of the vector control inverter 16. FIG. 2 shows an example in which only the q-axis current command value is superimposed. Here, the d-axis and q-axis current command values before superposition are defined as I _do and I _qo , respectively, and these are converted from the torque command values to the respective axis current command values by arbitrary distribution by the command value conversion unit 15. Use things. For example, the command value conversion unit 15 includes a command value conversion table that realizes maximum torque control. The inverter 16 is a general motor driving device that realizes current vector control in the d-axis and the q-axis obtained by rotating coordinate conversion from three phases, and drives the motor load 17.

  Here, when the mechanical system having the device configuration as shown in FIG. 1 is regarded as a two-inertia system including the electric motor 1 and the load device 2, the detected shaft torque has a resonance point depending on the mechanical characteristics, and an arbitrary operation The gain and phase characteristics change suddenly at the point. Therefore, if feedback learning control is performed while the mechanical characteristics are unknown, there may be a case where the pulsation is compensated in the divergence direction instead of simply being stable.

Therefore, in order to feedback the shaft torque detection value and generate and generate an optimal torque pulsation compensation signal, pulsation compensation is performed using the time until the response of the PI control output (Fourier coefficient of compensation current) settles to a steady value as the learning time. The Fourier coefficients I _An and I _Bn of the current are recorded in the memory 19 through a low frequency pass filter (LPF) 18. Once learned, the selectors 20A and 20B are switched to read values from the memory 19, and torque pulsation is suppressed by feedforward control. The RMS evaluation & gain adjuster 21 evaluates the amplitude value of the pulsation component using the root mean square of the Fourier coefficient of the torque pulsation detection value: RMS (Formula 3), discriminates the divergence / convergence condition, and determines the learning time is doing.

  Also, the RMS evaluation & gain adjuster 21 is a combination of the polarities of the two proportional gains of the PI controllers 13A and 13B when the RMS evaluation value is in the direction of divergence and the torque pulsation learning control system becomes unstable. Thus, a search is made and adjusted for a combination in the convergence direction, thereby responding to fluctuations in the gain / phase characteristics of the multi-inertia system.

  Regarding the gain adjustment of this PI controller, it is a simple adjustment method that gradually increases or decreases the adjustment amount after seeing the convergence of the torque pulsation evaluation value, but the adjustment time and learning time become longer There is. In particular, when the initial value is set small, the learning delay becomes remarkable, and it is difficult to determine the convergence. Conversely, if the initial value is too large, it becomes divergent and oscillating, and in the worst case, the search for the convergence condition itself may be difficult.

  In the present invention, an appropriate value for the initial nominal value of the gain of the PI controller is obtained for learning the torque pulsation compensation signal, and a learning speed and stability adapted to the response of the torque control system are obtained. Hereinafter, embodiments will be described.

(Embodiment 1)
In this embodiment, the basic control block configuration is the same as that in FIG. 2, and a Fourier coefficient learning control system is used.

In the present embodiment, the process starts by grasping in advance the frequency characteristics of the motor drive device. There are various methods for grasping frequency characteristics and system identification. However, since these are known general techniques, the present embodiment is not limited to specific methods. For example, as shown in FIG. 3, in an arbitrary steady operation state of a system in which the inverter 16 and the motor load 17 are operated with a current command generated by the torque command value conversion unit 15, the system identification device 22 generates a white noise signal as a torque. A pulsation compensation current i _qc is superimposed on the torque current command value (FIG. 3 is superimposed on the q-axis current, but it may be the d-axis or both). System input / output data is obtained using the detected shaft torque value T _det at that time as a system output signal. The frequency response is spectrum-analyzed from the input / output time series data, and the frequency transfer function from the torque pulsation compensation current command value to the detected shaft torque value is obtained.

  Although it can be formulated based on the system identification theory, in the present embodiment, the transfer function is expressed by a set of complex vectors as in the following Expression 4. The elements of the set correspond to complex vectors indicating gain / phase characteristics at each frequency.

ただし、Ｐ_SYS＾：同定システム、ｍ：周波数成分の分割要素番号、Ｐ_Am：ｍ番要素の同定システム実数部、Ｐ_Bm：ｍ番要素の同定システム虚数部

Where P _SYS ^: identification system, m: frequency component division element number, P _Am : m-th element identification system real part, P _Bm : m-th element identification system imaginary part

  Here, m will be described. For example, if the frequency component necessary for identification is DC to 5 kHz and the resolution is 1 Hz, the number of elements is 5001. Therefore, as m = {0, 1, 2, 3,..., 4999, 5000}, a spectrum analysis result or a system identification result is applied to each element (frequency component). Since the gain / phase characteristic of the system for each frequency component can be expressed by a complex vector on a complex plane, it can be expressed as Equation 4.

  By the way, torque pulsation of an arbitrary frequency component can be extracted by using Fourier transform. As shown in Equation 2, since the two Fourier coefficients are orthogonal, if the complex vector of Equation 4 above is used focusing on the system state of an arbitrary frequency component, torque pulsation suppression learning using the Fourier coefficient is performed. The control system can be constructed in a complex vector space.

  Equation 5 and FIG. 4 below show the torque pulsation compensation current and the characteristics of the identification system in a complex vector space. Here, the identification system of Formula 4 and the phase reference of the Fourier coefficient and the definition of the delay / leading direction are unified.

  In this way, when the system characteristics are expressed by the Fourier coefficient learning control system after Fourier transform, as shown in FIG. 4, the system has a term in which two Fourier coefficient control systems (real part system and imaginary part system) interfere with each other. I understand that I have it. Therefore, in the present embodiment, non-interference control between the real part and the imaginary part of the complex vector is realized. FIG. 5 is an example of the configuration, and a non-interacting control element 24 is inserted in the previous stage of the actual system 23.

When the response when extracting the frequency component by Fourier transform is expressed by the transfer function G _FT , the Fourier coefficients T _An (real part: cos component) and T _Bn (imaginary part: sin component) of the torque pulsation of the arbitrary frequency component are It is shown by Formula 6.

Since the real part and the imaginary part of the pulsation compensation current interfere as shown in the above mathematical formula 6, the non-interacting control element 24 shown in FIG. 5 is added so as to cancel the influence of the interference term. The non-interacting control element 24 replaces Equation 6 as Equation 7 below, where the real part T _An of the torque pulsation is the real part I _An of the pulsation compensation current and the imaginary part T _Bn is the real part I of the pulsation compensation current. It becomes possible to separate by the relational expression of _Bn .

  According to Equation 7, the real part control system and the imaginary part control system are independent and have exactly the same control configuration.

  Therefore, if this embodiment is applied, not only can the adverse effects due to the interference of the compensation currents of the two Fourier coefficient learning control systems be removed, the vibrational response can be reduced, but also the control system can be simplified. It becomes possible.

(Embodiment 2)
According to the first embodiment, since the torque pulsation Fourier coefficient after Fourier transform and the compensation current can be made non-interfering in the complex vector space, a controller is constructed using this. The controller uses general PID control as in the conventional system shown in FIG. Although there are various types of PID control, in the present embodiment, description will be made using proportional precedence type IP control as an example. The configuration method of the PID controller is arbitrary, and can be similarly configured in other cases.

The overall configuration of this embodiment is shown in FIG. In the figure, “s” is a Laplace operator. FIG. 6 is a control configuration diagram focusing only on the torque pulsation component, and a transfer function G for interpolating the non-interacting control element 24 to the IP controller 26 and extracting frequency components by the Fourier transform units 25A and 25B. _The non-interfering output through the _FT is used as a feedback signal.

In FIG. 6, target values T _An ^* and T _Bn ^* of two Fourier coefficients are normally 0, and are set so as to suppress pulsation of the frequency components. The detected values T _An and T _Bn of the Fourier coefficient are detected through the real system and the Fourier transform response as shown in Equation 7. The target value and the detected value are input to a proportional leading IP controller as shown in FIG. The proportional gain at this time K _p, an integral gain and K _i. The output of the IP controller is output as the Fourier coefficient of the pulsation compensation current, but the influence of the interference term is removed using the non-interacting control element of the first embodiment. I _qc obtained by restoring the torque pulsation compensation current into a complex vector is superimposed on the q-axis current command value with a time-axis waveform, and the motor load is driven via the vector control inverter. At this time, since the torque pulsation component is detected as a periodic disturbance, the pulsation component included in the shaft torque T _sn is extracted by Fourier transform and separated into a real part T _An and an imaginary part _TBn in the complex vector space. Feedback to the IP control system. With the above control, a Fourier coefficient learning control system can be constructed in a complex vector space.

Here, when the control system after realizing non-interference is simplified and expressed by the above-described Expression 7, two Fourier coefficient IP learning control systems can be independently expressed as shown in FIG. Based on this control system, the closed-loop system transfer functions from the command values T _An ^* and T _Bn ^* to the detected values T _An and T _Bn are calculated as shown in Equation 8 below. However, in this embodiment, as an example, the Fourier transform response transfer function G _FT is approximated by a simple first-order lag system expressed by the following Equation 9.

ｗ_fは、フーリエ変換応答周波数［ｒａｄ／ｓ］

w _f is the Fourier transform response frequency [rad / s]

  If the proportional gain and the integral gain are determined in consideration of the closed-loop system response of Equation 8, the learning control system can be designed with an arbitrary pole arrangement. Hereinafter, an example in which Formula 8 is a pole arrangement that matches a binomial coefficient standard system model will be given. There are other pole placement methods such as the Butterworth standard system, which can be calculated in the same way.

  Since Equation 8 is a quadratic system, model matching is performed with a secondary binomial coefficient standard system shown in Equation 10 below.

ｗ_cは、学習制御応答周波数［ｒａｄ／ｓ］であり、フーリエ変換応答周波数ｗ_fを超えない条件下で所望の値を任意に設定する。

w _c is a learning control response frequency [rad / s], and a desired value is arbitrarily set under conditions that do not exceed the Fourier transform response frequency w _f .

The coefficients 8 and 10 are compared with each other to obtain the proportional gain K _p and the integral gain K _i as follows.

As described above, if P _Am , P _Bm, Fourier transform response frequency w _f , and desired learning control response frequency w _c previously obtained by system identification are substituted into Equation 11, the control gain can be uniquely determined. Therefore, torque pulsation suppression control is started using this as the nominal initial value of the learning control system.

  The learning control method may be fine-adjusted while determining the convergence of the torque pulsation evaluation value of Equation 3, but if the system identification result is obtained to a certain degree of accuracy, the initial value is maintained with little adjustment. Sufficient suppression is possible.

  Therefore, according to the present embodiment, the response of the non-interfering closed loop system can be approximated to a desired pole arrangement based on the system identification result, so that the control gain adjustment time of the learning control system is unnecessary or significantly reduced. can do.

(Embodiment 3)
In the proportional leading I-P learning control according to the second embodiment, an inverse response may temporarily occur at the timing of starting torque pulsation suppression, resulting in an excessive shaft torque. Although it is a transient behavior, there is no problem if time passes. However, an instantaneous excessive torque may cause a malfunction in, for example, failure determination or operation stop processing.

  That is, in the configurations of FIGS. 6 and 7, since the proportional preceding path of the IP control is directly connected to the control output to form a compensation current, the Fourier coefficient (DC value) of the detected torque pulsation value immediately before the suppression is started. Has some value via a proportional gain. Therefore, the compensation current value changes stepwise at the moment when the suppression control is started. Although the latter-stage non-interacting control is also related, depending on the controller characteristics at that time, the above-mentioned inverse response problem occurs in a stepwise manner, causing a temporary excessive torque.

  Therefore, in the present embodiment, a low frequency band pass filter (low pass filter) is inserted in the output stage of the IP learning control system. The initial value of the integrator used for the low-pass filter is set to 0, and is always reset before the start of suppression to maintain the initial value. As a result, a low-pass filter with an initial value of 0 enters the proportional preceding path, and the stepwise reverse response that causes excessive torque is suppressed at the suppression start timing.

FIG. 8 shows a control configuration diagram after the non-interference of the present embodiment. As can be seen by comparing FIG. 7 and FIG. 8, a first-order low-pass filter that is a transfer function of G _LPF1 shown by the following Expression 12 is added to the _IP control output stage.

ｗ_Lは、ローパスフィルタのカットオフ周波数［ｒａｄ／ｓ］

w _L is the cut-off frequency of the low-pass filter [rad / s]

  The closed-loop transfer function at this time is expressed by Equation 13 below.

  When this transfer function is approximated to a third-order binomial coefficient standard system expressed by the following Equation 14, the proportional gain, the integral gain, and the cutoff frequency of the low-pass filter are expressed by Equation 15.

ｗ_cは、学習制御応答周波数［ｒａｄ／ｓ］であり、フーリエ変換応答周波数ｗ_fを超えない条件下で所望の値を任意に設定する。ただし、ｗ_Lは正の周波数でなければならないので、ｗ_c＞ｗ_f／３でなければならない。

w _c is a learning control response frequency [rad / s], and a desired value is arbitrarily set under conditions that do not exceed the Fourier transform response frequency w _f . However, since w _L must be a positive frequency, w _c > w _f / 3.

  From the above, also in this embodiment, the proportional gain, the integral gain, and the low-pass filter cutoff frequency can be varied adaptively. Moreover, according to the present embodiment, it is possible to adaptively determine the learning control parameter based on Expression 15, and it is possible to improve the response at the start of torque pulsation suppression.

(Embodiment 4)
In the third embodiment, a low-pass filter is inserted into the output of the IP controller, and the closed-loop characteristic of the third-order system is redesigned. The present embodiment provides a technique for improving the reverse response at the start of suppression while maintaining the secondary closed-loop characteristics of Equations 8 and 10 without increasing the order.

FIG. 9 is a control configuration diagram of the present embodiment. G _{LPF2 in} FIG. 9 is a low-pass filter that functions only at the start of torque pulsation suppression, and after starting the torque pulsation suppression control, switches to a path that does not go through the low-pass filter.

  FIG. 10 shows a configuration example when a first-order low-pass filter is used. Has a switching mechanism synchronized with the suppression start signal. Before the start of suppression, the Fourier coefficient (DC value) of the torque pulsation detection value is input to the low-pass filter, so that the low-pass filter output and the torque pulsation detection value are almost the same in the steady state, and the deviation obtained by adding or subtracting them. Is 0. This deviation 0 becomes the initial value of the output at the start of suppression.

Next, when torque pulsation suppression is started, the switch is switched and the input of the low-pass filter becomes zero. Since the integrator of the low-pass filter uses the torque pulsation value immediately before the start of suppression as an initial value, the value gradually approaches 0 according to the response frequency w _L characteristic. As a result, the deviation output at the rear stage gradually approaches the value of the torque pulsation detection value from the initial value 0. Eventually, the low-pass filter output becomes 0, and the torque pulsation detection value becomes a path that directly leads to the output.

  If the configurations of FIGS. 9 and 10 are used, an excessive torque due to a reverse response can be suppressed by giving a low-pass filter characteristic only at the start of torque pulsation suppression without increasing the order of the control system.

  The improvement effect by applying said Embodiment 3, 4 is shown in FIG. (A) of the same figure shows each part waveform by normal IP control, (b) shows each part waveform of IP control with a filter which becomes the method of Embodiment 3,4. From the top, the waveforms are a shaft torque detection value, a Fourier coefficient extracted value of torque pulsation (for example, a 12th order pulsation component), a compensation current command value (12th order compensation component), and an amplitude evaluation value of torque pulsation (Formula 3).

  It can be seen that with the normal IP control method, the shaft torque pulsation component increases instantaneously at the start of suppression. In addition, the compensation current also causes reverse response and overshoot. On the other hand, in the method using the IP control with a filter, the shaft torque pulsation is smoothly suppressed, and the compensation current does not generate a reverse response or an overshoot.

(Embodiment 5)
In Embodiments 2 to 4, assuming that the motor drive system is in a steady operation condition, the nominal initial values of the proportional gain K _p and the integral gain K _i are given based on Equation 11, but in this embodiment, feedback learning is performed. Even during the control, it is always variable based on Formula 11 or Formula 15. That is, control gain automatic adjustment adapted to fluctuations in the rotational speed of the motor is performed.

This embodiment will be specifically described. Among the parameters of the above formula 11, P _Am and P _Bm obtained by system identification are components depending on the frequency. That is, it is necessary to read one element (one-dimensional complex vector) of the corresponding frequency component from the system P _sys (Equation 4) recorded as a set of complex vectors representing the system response according to the rotation speed and the compensation order. There is. If the rotational speed fluctuates, for example, the system gain and phase characteristics fluctuate greatly in the vicinity of the mechanical resonance point.

Therefore, in this embodiment, P _Am and P _Bm in Expression 11 are changed according to the rotation speed. Similarly, P _Am and P _Bm in Expression 15 can be changed according to the rotation speed.

  According to the present embodiment, the system identification result is read out in accordance with the speed fluctuation, so that even if the system characteristics fluctuate greatly near the resonance point or the like, the error of the pole arrangement of the closed loop system and the non-interacting control is reduced, Torque pulsation suppression control can be maintained.

(Embodiment 6)
The Fourier transform can be calculated with a fixed period, but in the case of an electric motor, periodic torque pulsation synchronized with the rotation is generated. Therefore, it is desirable to change the period of the Fourier transform in accordance with the rotational speed. .

Therefore, in this embodiment, in addition to the fifth embodiment, the Fourier transform cycle is also adapted to the rotation speed. Along with this, the Fourier transform response frequency w _f approximated by Equation 9 is also changed in accordance with the rotational speed. Further, since the learning control response frequency w _c cannot in principle give a response performance higher than the Fourier transform, w _c is similarly adapted to the rotation speed.

The block configuration of this embodiment shown in FIG. 12, a proportional gain K _p in accordance with the rotational speed by the automatic gain adjusting unit 27 a integral gain K _i is automatically adjusted, P _Am in accordance with the rotational speed by the speed automatic adjusting section 28 , P _Bm is automatically adjusted.

  According to the present embodiment, since the Fourier transform cycle can be changed according to the rotation speed, the torque pulsation extraction performance can be improved at each speed, and the learning control parameter expressed by Expression 11 also has a desired closed-loop response. Can be adjusted automatically.

(Embodiment 7)
The above embodiments have been suppression control means for any one pulsation frequency component. However, if the above configuration is parallelized, a plurality of torque pulsation frequency components can be suppressed. In this case, the complex vector table of the system identification result can be used together, but the Fourier transform process and the adaptive gain automatic adjustment process are separately prepared for each frequency component.

FIG. 13 shows a configuration diagram of this embodiment. 29 _{1 to} 29 _n are control elements for suppressing pulsation components according to the first to _nth orders of the Fourier series, and the compensation current signals i _{qc1 to} i _qcn are superposed on each other by the adder 30 to _generate torque pulsation compensation signals. _Let i _qc .
According to this embodiment, a plurality of torque pulsation frequency components can be suppressed simultaneously.

(Embodiment 8)
In the embodiments described above, the torque pulsation compensation method has been proposed based on the configuration that feeds back the shaft torque, but it is also possible to change the object to be fed back.

  In the present embodiment, as shown in FIG. 14, the vibration by the acceleration sensor 8 is detected and fed back. The control method is the same as in the above-described embodiments.

  According to this embodiment, vibration and torque pulsation of the motor frame can be suppressed.

(Embodiment 9)
In the present embodiment, fluctuations in the phase / speed information provided from the rotational position sensor are detected, and control is performed to suppress the fluctuations. The control contents are basically the same as those in the above-described embodiments.

  The device configuration is as shown in FIG.

  According to this embodiment, speed fluctuations and rotational phase fluctuations can be suppressed.

(Embodiment 10)
In the present embodiment, the current pulsation caused by the torque pulsation is suppressed by detecting the current of the three-phase line supplied with power from the inverter 7 to the electric motor 1 to be compensated for torque pulsation.

  The apparatus configuration is as shown in FIG. 16, and a current sensor 9 is newly added, and the controller 5 performs compensation current processing using this detected current value.

  According to the present embodiment, inverter current pulsation can be suppressed using the inverter current detection value. Furthermore, torque pulsation can also be suppressed using the estimated torque value converted from the inverter current.

(Embodiment 11)
In the torque ripple suppression control according to the previous embodiments, the system transfer function of the frequency component synchronized with the torque ripple frequency is expressed by a one-dimensional complex vector, and the real part and the imaginary part of the torque ripple of the arbitrary frequency component are extracted by Fourier transform or the like. By applying the cosine and sine Fourier coefficients to the real and imaginary parts of the complex vector, a feedback suppression control system is constructed.

  Further, the proportional / integral gain is determined by the partial model matching method so that the closed-loop characteristic of the proportional leading IP control system matches the pole arrangement of an arbitrary standard system reference model. In addition, they automatically adapt parameters using system identification results and rotational speed information to facilitate implementation in a multi-inertia resonance system.

  In addition, the amplitude and phase of the compensation current signal when the suppression is completed at an arbitrary steady operating point (steady torque / steady rotational speed) are stored and implemented at a plurality of operating points to obtain a two-dimensional torque / rotational speed. It can also be implemented as a table. In this case, torque / rotational speed information is input to the table, and the compensation current is reproduced from the read compensation current amplitude / phase data, and is controlled by feedforward.

  In the table compensation method that suppresses torque ripple in a feed-forward manner using the compensation current table described above, a learned compensation table that has good responsiveness to changes in rotational speed and torque is read out to compensate for a previously learned result. Since it is used, there is no problem even if the load connected to the motor is replaced, there is no need for an axial torque detector, and there is an advantage that the calculation load can be reduced because only a table is basically mounted.

  However, when the characteristics of the electric motor, the inverter, etc. change with time, the conditions may differ from the learned conditions, which may reduce the compensation accuracy.

  The eleventh embodiment and the following twelfth to fourteenth to fourteenth embodiments propose a control device that can ensure compensation accuracy even when the characteristics of an electric motor, an inverter, and the like change over time in torque ripple suppression by the above-described table compensation method. .

  FIG. 18 shows a block configuration of the torque control apparatus for an electric motor according to the present embodiment. In the inverter control of the current vector control system, on-line compensation is used together with table compensation to suppress torque ripple.

In FIG. 18, the current vector control unit 31 of the inverter 7 is synchronized with the motor rotation coordinates by the coordinate conversion unit 33 from the motor drive currents i _u , i _v , i _w detected by the current sensor 32 and the rotor rotation angle θ of the motor 1. The motor current is controlled by comparison with the d and q axis current detection values converted into the current of the dq axis orthogonal rotation coordinate system. The rotor rotation angle θ is obtained from the encoder waveform abz by the rotational position sensor 6 together with the speed ω by the speed / phase detector 34.

The torque / id, iq conversion unit 35 (command value conversion unit 15 in FIG. 2) calculates the d-axis and q-axis currents of the rotation dq coordinate system in vector control from the torque command value T _ref from the controller 5 and the motor rotation speed ω. Converted to command values i _d ^* (I _do ), i _q0 ^* (I _qo ), and among these current command values, superimpose torque pulsation compensation current i _qcm ^* on q-axis current command value i _q0 ^*. Use command value.

Here, the compensation table 36 and the compensation current generator 37 correspond to the low-pass filter 18 and the memory 19 shown in FIG. 2, and the Fourier coefficient of torque pulsation is obtained at an arbitrary steady operating point (steady torque / steady rotational speed). Instead of the compensation current signal of the memory 19 for storing the outputs I _An and I _Bn for controlling to 0, the data is stored as table data of amplitude M and phase φ, and a compensation current is generated from these data.

Similar to the memory 19 of FIG. 2, the amplitude / phase compensation table 36 records the amplitude M and the phase φ corresponding to the torque ripple compensation current outputs I _An and I _Bn learned by the controller 5. This operation is similarly recorded at a plurality of steady operation points to generate a two-dimensional amplitude table and phase table using the torque command T _ref ^* and the rotational speed ω as variables. Information between data is interpolated by linear interpolation or the like.

The compensation current generation unit 37 reads the current amplitude M and phase φ from the operating state (torque command / rotation speed) from the amplitude / phase compensation table 36, and the read amplitude M and phase φ are the rotational phase at that time. Table compensation current i _qct ^* = M · sin (nθ + φ) is generated using θ (n is the compensation order). The generated table compensation current i _qct ^* is combined with the online compensation current _iqc ^* from the controller 5 to obtain a q-axis current command value i _qcm ^* .

As in FIG. 1, the controller 5 is equipped with torque ripple suppression control means. As this torque ripple suppression control means, for example, as shown in FIG. 2, a sine / cosine wave having a reference frequency (order n) and phase θ. Sine / cosine wave generator 11, setting of these sine / cosine waves and an arbitrary Fourier transform period _Tr , and a Fourier transform of the axial torque detection value T _det based on Equation 2, A Fourier transform unit 12 for calculating the Fourier coefficients T _An and T _Bn , two PI controllers 13A and 13B for controlling the Fourier coefficients of these pulsating components so as to become the target value 0 (that is, suppressing pulsation), Multipliers 14A and 14B that obtain these outputs I _An and I _Bn as the Fourier coefficients of the compensation current signal as torque pulsation compensation current i _qcc ^* are provided. Then, the obtained torque pulsation compensation current i _qcc ^* is combined with the table compensation current _iqct ^* from the compensation current generation unit 37 to form a compensation signal for the q-axis current command value i _q0 ^{* of} the vector control inverter 7.

Therefore, even when the compensation current i _qct ^* of the torque ripple learned by the compensation table 36 and the compensation current generator 37 is shifted due to a change in ambient temperature or a change in physical properties over time, the online compensation by the controller 5 is performed. Thus, it is possible to prevent the accuracy of the table compensation current from being lowered by the feedback loop.

The table compensation current i _qct ^* generates the compensation current by instantaneously obtaining the amplitude / phase from the learned compensation current table 36, so that a suitable compensation current can be quickly obtained even when a speed change or torque change occurs. Can be output.

Embodiment 12
FIG. 19 shows a block configuration of the torque control apparatus for an electric motor according to the present embodiment. Although the basic configuration and operation are the same as those in FIG. 18, the error of the amplitude M and phase φ of the compensation current generated in the compensation table 36 is shown. An online compensation feedback loop is constructed from the controller 5 to correct.

In FIG. 19, M ^* is a compensation current amplitude command value (on-line compensation), φ ^* is a compensation current phase command value (on-line compensation), and the synthesizers 38 and 39 use these as the compensation current amplitude M · phase φ. The compensation current amplitude value (composite value) M ′ and the compensation current phase value (composite value) φ ′ to the compensation current generator 37 are combined.

In the eleventh embodiment, the table compensation current i _qct ^* and the online compensation current i _qc ^* are combined after the generation of the compensation current. However, in the present embodiment, they are combined in the state of amplitude and phase. That is, the compensation table amplitude value M and the compensation table phase value φ are combined with the online compensation amplitude value M ^* and the online compensation phase value φ ^* , respectively, and the combined compensation current amplitude value M ′ and the combined compensation current phase value φ ′. Is generated. The compensation current generator 37 generates a final compensation current i _qc ^* with i _qc ^* = M ′ · sin (nθ + φ ′). (N is the compensation order)

  According to the present embodiment, the table error can be grasped and corrected with an intuitively easy-to-understand state quantity such as amplitude and phase, and the compensation current generator 37 can be gathered in one place.

(Embodiment 13)
FIG. 20 shows a block configuration of the torque control apparatus for an electric motor according to the present embodiment. Instead of the amplitude M and phase φ of the compensation table 36 of the twelfth embodiment (FIG. 19), the memory 19 and the combiners 14A and 14B in FIG. Similarly to the above, the compensation current is obtained as a cosine Fourier coefficient and a sine Fourier coefficient.

The compensation table 40 generates a compensation current cosine Fourier coefficient table value A and a compensation current sine Fourier coefficient table value B according to the torque command T _ref ^* and the rotational speed ω, and the controller 5 calculates the online compensation current cosine Fourier coefficient A ^* . The online compensation current sine Fourier coefficient B ^* is generated, and the synthesizers 38 and 39 synthesize these to obtain a compensation current cosine Fourier coefficient synthesis value A ′ and a compensation current sine Fourier coefficient synthesis value B ′.

  The controller 5 equipped with torque ripple suppression control means basically uses a technique for extracting and suppressing the torque ripple frequency component by Fourier transform (or a similar technique). In FIG. 2, there are two Fourier coefficients (sine Fourier coefficient and cosine Fourier coefficient), and the amplitude value and phase value of the compensation current are calculated using these.

  In the configuration of FIG. 2, the PID controllers 13A and 13B are controlled so that the sine / cosine Fourier coefficients of the torque ripple frequency component detection value are both zero, and the output value of the PID controller generated as a result is compensated current cosine Fourier. The coefficient A and the compensation current sine Fourier coefficient B are used. From these, the amplitude M and the phase φ are obtained as follows.

M = √ (A ² + B ² ), φ = tan ⁻¹ (B / A)

  In the eleventh and twelfth embodiments, the compensation table is generated or the online compensation is performed after obtaining the amplitude M and the phase φ. While the compensation current characteristic is easy to understand intuitively, the above-described arithmetic conversion is required. It was.

Therefore, in the present embodiment, as shown in FIG. 20, a compensation table is generated in the state of two Fourier coefficients A and B before being converted into amplitude and phase, and online compensation is also synthesized in the state of the Fourier coefficient. The compensation current i _qc ^* is generated at the final stage. As a result, it is possible to save the labor of the above-described calculation for converting into amplitude / phase and simplify the compensation process.

(Embodiment 14)
In the present embodiment, there is provided a method of parallelizing the configurations of any of the embodiments 11 to 13 and simultaneously suppressing the torque ripple of the multi-order component.

  For example, FIG. 21 shows a case where the configuration of the eleventh embodiment is parallelized. (Since this is the same for the other embodiments, description thereof is omitted.) FIG. 21 shows that the two compensation components 36A and 36B individually generate amplitude M and phase φ with two order components. The two compensation current generators 37A and 37B generate compensation currents for two orders, and simultaneously suppress torque ripple for these two orders. Of course, it is possible to extend the case where three or more orders are simultaneously suppressed.

  As described above, according to the present invention, the torque of an electric motor driven by an inverter is Fourier-transformed, a pulsation component of an arbitrary frequency is detected in the form of two Fourier coefficients, and pulsation suppression by detecting these two Fourier coefficients is performed. Since the learning control system is configured with a complex vector plane and the pulsation compensation current is obtained, the torque pulsation compensation signal, etc. Learning is sure and easy.

  Further, on the basis of the “table compensation method”, on-line compensation is also used to generate a compensation current that is suppressed in a feedforward manner, so that compensation accuracy can be ensured even when characteristics of the motor, the inverter, and the like change with time.

[0003]
It does not address the problem that pulsations interfere with compensation current. Also, it is not specified that vector control is performed with the dq axes subjected to rotational coordinate conversion. Therefore, in the case of an embedded magnet synchronous motor, torque pulsation cannot always be optimally suppressed, and other frequency components may be adversely affected.
[0011]
An object of the present invention is to provide a pulsation compensation signal in a pulsation suppression device that detects a torque pulsation component or the like in the form of a Fourier coefficient and obtains a pulsation compensation signal by controlling a PI controller that sets the magnitude of the Fourier coefficient to “0”. It is an object of the present invention to provide a pulsation suppression device for an electric motor that can reliably and easily learn a pulsation compensation signal while performing stable control while preventing overcompensation, unstable vibration, and interference due to the above.
[0012]
In order to solve the above-described problems, the present invention performs Fourier transform on the torque of an electric motor driven by an inverter, detects a pulsation component of an arbitrary frequency in the form of two Fourier coefficients, and pulsation by detecting these two Fourier coefficients. The suppression learning control system is configured with a complex vector plane so as to obtain the pulsation compensation current, and has the following configuration.
[0013]
(1) A pulsation detection value of an electric motor driven by an inverter is Fourier transformed, a pulsation component of an arbitrary frequency is detected in the form of two Fourier coefficients, and learning control is performed by a learning controller to suppress the pulsation, and vector control is performed. In the pulsation suppressing device for an electric motor that suppresses the pulsation by superimposing the learned pulsation compensation current on both the d-axis current command value or the q-axis current command value or the d-axis q-axis current command value of the rotating coordinate system in
A learning control system for suppressing pulsation by detecting the two Fourier coefficients is configured with a complex vector plane, and includes means for obtaining the pulsation compensation current,
The means for obtaining the pulsation compensation current is expressed by a complex vector using the frequency characteristic of the control target system on the complex vector plane or the system identification result, and interference between two Fourier coefficients from the pulsation compensation current and the motor drive system. A decoupling control element for decoupling the real part and the imaginary part of the complex vector based on the term is provided.
[0014]

Claims (17)

The pulsation detection value of the motor driven by the inverter is Fourier transformed, the pulsation component of an arbitrary frequency is detected in the form of two Fourier coefficients, and learning control is performed by the learning controller so as to suppress the pulsation. In the pulsation suppressing device for an electric motor that suppresses the pulsation by superimposing the learned pulsation compensation current on both the d-axis current command value, the q-axis current command value, or the d-axis q-axis current command value of the system,
A pulsation suppressing device for an electric motor comprising: a learning control system for suppressing pulsation by detecting the two Fourier coefficients, comprising a complex vector plane; and means for obtaining the pulsation compensation current.   The means for obtaining the pulsation compensation current is expressed by a complex vector using the frequency characteristic of the control target system on the complex vector plane or the system identification result, and interference between two Fourier coefficients from the pulsation compensation current and the motor drive system. The pulsation suppressing device for an electric motor according to claim 1, further comprising a non-interacting control element that makes the real part and the imaginary part of the complex vector non-interact based on the terms.   The means comprises a learning controller comprising a PID controller, and a compensation current is obtained from the output of the PID controller through the non-interacting control element, and the gain of the PID controller is arranged so as to have a desired closed-loop response. The pulsation suppressing device for an electric motor according to claim 1, further comprising learning control means for performing the learning control.   A low-frequency pass filter with an initial value of 0 is inserted in the output stage of the learning controller, and is poled to obtain a desired closed-loop response, and an inverse response or overshoot occurs when the learning control system starts to suppress pulsation. The pulsation suppressing device for an electric motor according to any one of claims 1 to 3, further comprising means for preventing the generation.   A low-frequency pass filter with an initial value of 0 that functions only at the start of pulsation suppression is inserted into the feedback signal path to the learning controller, and the filter is inserted after the start of pulsation suppression and after the time constant of the low-frequency pass filter has elapsed. The pulsation suppressing device for an electric motor according to any one of claims 1 to 3, further comprising means for switching to a route that does not pass.   The pulsation of an electric motor according to any one of claims 1 to 5, further comprising a control gain automatic adjustment unit that adjusts and automatically adjusts the gain of the learning controller in accordance with the rotational speed of the electric motor. Suppression device.   The pulsation suppressing device for an electric motor according to any one of claims 1 to 6, wherein a Fourier transform period of the learning controller is changed in accordance with a rotational speed of the electric motor.   The pulsation suppressing device for an electric motor according to claim 7, further comprising means for setting a response frequency of the learning controller in conformity with a Fourier transform period.   9. The learning controller according to any one of claims 1 to 8, further comprising means for applying the learning controller in different frequency components, parallelizing them, and simultaneously suppressing torque pulsation of a plurality of frequency components. Motor pulsation suppression device.   The pulsation component of the motor according to claim 1, wherein the pulsation component is a detected value of an axial torque of the electric motor and suppresses the pulsation of the axial torque of the electric motor.   11. The motor pulsation suppressing device according to claim 10, wherein the pulsation component is a pulsation component of a frame of an electric motor, and suppresses pulsation of the frame of the electric motor.   11. The pulsation control of an electric motor according to claim 10, wherein the pulsation component is a pulsation component of a rotation speed detection value of the electric motor or a rotation position detection value, and suppresses the pulsation of the speed of the electric motor or the pulsation of the rotation position. apparatus.   The pulsation control device for an electric motor according to claim 10, wherein the pulsation component is a pulsation component of the electric current of the electric motor, and suppresses the pulsation of the electric current of the electric motor. The amplitude M and phase φ of the torque ripple compensation current i _qc ^* learned in the steady operation state (steady torque command / steady rotation speed) designated when the torque ripple suppression control is executed by the controller are used as the torque command T _ref ^* and the rotation speed. An amplitude / phase compensation table for generating a two-dimensional amplitude table and a phase table with ω as a variable;
The amplitude M and phase φ at that time are read out from the amplitude / phase compensation table from the torque command T _ref ^* and the rotational speed ω in the operating state of the motor, and the amplitude M and phase φ and the rotational phase θ of the motor at that time are read out. A compensation current generation unit that generates a table compensation current i _qct ^* = M · sin (nθ + φ) using
Torque pulsation compensation for detecting the pulsation component of the same frequency as the table compensation current i _qct ^{* in} the form of two Fourier coefficients and making this pulsation component zero by Fourier transforming the pulsation detection value T _det in the motor operating state _Torque ripple suppression control means for _obtaining current i _qcc ^* ;
_{14. The} compensation current synthesis means for synthesizing the torque pulsation compensation current i _qcc ^* and the table compensation current i _qct ^* and superimposing them on the d and q axis current command values. The pulsation control device for an electric motor according to claim 1. The amplitude M and phase φ of the torque ripple compensation current i _qc ^* learned in the steady operation state (steady torque command / steady rotation speed) designated when the torque ripple suppression control is executed by the controller are used as the torque command T _ref ^* and the rotation speed. An amplitude / phase compensation table for generating a two-dimensional amplitude table and a phase table with ω as a variable;
The pulsation detection value T _det in the operating state of the motor is Fourier-transformed, the pulsation component having the same frequency as the table compensation current i _qct ^* is detected in the form of two Fourier coefficients, and the compensation current amplitude is set to zero. Torque ripple suppression control means for obtaining a command value command value M ^* and a compensation current phase command value φ ^* ;
The amplitude M and phase φ at that time are read out from the amplitude / phase compensation table from the torque command T _ref ^* and the rotational speed ω in the operating state of the motor, and the amplitude M and phase φ and the compensation current amplitude command value are read out. Combining means for synthesizing the command value M ^* and the compensation current phase command value φ ^* to generate a combined compensation current amplitude value M ′ and a combined compensation current phase value φ ′;
Compensation current generation for generating a table compensation current i _qc ^* = M ′ · sin (nθ + φ ′) using the combined compensation current amplitude value M ′, the combined compensation current phase value φ ′ and the rotational phase θ of the motor at that time The pulsation control device for an electric motor according to any one of claims 1 to 13, wherein the pulsation control device for an electric motor is provided. The compensation current cosine Fourier coefficient table value A and the compensation current sine Fourier coefficient table value B of the torque ripple frequency component learned in the steady operation state (steady torque command / steady rotation speed) designated when the torque ripple suppression control is executed by the controller. A Fourier coefficient compensation table generated as a two-dimensional table value using the torque command T _ref ^* and the rotational speed ω as variables;
Two compensation current cosine Fourier coefficients A ^* and a compensation current sine for Fourier transforming the pulsation detection value T _det in the operating state of the motor to make the pulsation component of the same frequency as the table values A and B of the Fourier coefficient compensation table zero. Torque ripple suppression control means for obtaining a Fourier coefficient B ^* ;
The Fourier coefficient table values A and B at that time are read from the Fourier coefficient compensation table from the torque command T _ref ^* and the rotational speed ω in the operating state of the electric motor, and these Fourier coefficient table values A and B, and the compensation current A synthesis means for synthesizing the cosine Fourier coefficient A ^* and the compensation current sine Fourier coefficient B ^* to generate a compensation current cosine Fourier coefficient synthesis value A ′ and a compensation current sine Fourier coefficient synthesis value B ′;
Using the Fourier coefficient composite value A ′, the compensation current sine Fourier coefficient composite value B ′, and the rotational phase θ of the motor at that time, the amplitude M = √ (A ′ ² + B ′ ² ) and the phase φ = tan ⁻¹ ( The electric motor according to claim 1, further comprising: a compensation current generation unit configured to generate a table compensation current i _qc ^* = M · sin (nθ + φ) of B ′ / A ′). Pulsation control device.   The compensation table, the synthesis unit, the torque ripple suppression control unit, and the compensation current generation unit individually generate amplitude M and phase φ with a plurality of order components or generate a Fourier coefficient, and generate a compensation current by combining these orders. The pulsation control device for an electric motor according to any one of claims 14 to 16, further comprising:

Images