JP2018196297A - Ac motor controller - Google Patents

Abstract

To improve the transient accuracy of an estimated speed and improve responsiveness to speed control in a controller that performs sensorless speed control of an AC motor.SOLUTION: An AC motor controller 1 comprises: a state observation unit 17 for computing an estimated current and an estimated magnetic flux of an AC motor 4; a speed estimation unit 18 for outputting the estimated speed; an estimated slip frequency computation unit 19 for computing an estimated slip frequency on the basis of the estimated current, the estimated magnetic flux and a detected current so that the q-axis component of the estimated magnetic flux becomes 0; and a compensation unit 20 for generating a compensation signal ADD that compensates for disturbances generated in the state observation unit 18. The state observation unit 18 accepts input of the detected current and a voltage command and further the compensation signal ADD, an output frequency and the estimated slip frequency, and computes the estimated current and estimated magnetic flux of the AC motor 4.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a control device that controls an AC motor without using a speed sensor.

There are various methods for controlling the AC motor, and they are properly used depending on the application. However, when performing highly precise torque control and speed control, current control is performed as a minor loop of control.
Conventional speed control methods for induction motors include the following. The control device outputs a three-phase voltage command to the PWM inverter. In the control device, a torque current command is generated so that the detected speed follows the speed command. Further, the exciting current command is generated so that the magnetic flux calculated from the detected current follows the magnetic flux command. Further, the slip frequency and the output frequency are calculated, and the phase for coordinate conversion is obtained based on the output frequency. A voltage command on the two rotation axes is generated so that the detected current on the two rotation axes follows the torque current command and the excitation current command, and coordinate conversion is performed to generate a three-phase voltage command. In the conventional synchronous motor speed control method, the phase for coordinate conversion is set as the magnetic pole position of the synchronous motor (for example, see Non-Patent Document 1).

A conventional speed sensorless control method that performs speed control of such an AC motor without requiring speed detection is shown below.
A state observer called an observer, which is configured using a mathematical model of the induction motor, uses the voltage information applied to the induction motor to estimate the magnetic flux and current. In addition, the difference between the detected current of the induction motor and the estimated current is multiplied by a predetermined gain and fed back. Then, by controlling the feedback gain of the observer and adding a speed estimator for the induction motor, speed control is performed without detecting the rotation speed of the induction motor. The mathematical model of the induction motor incorporates the rotational speed information of the induction motor therein, and the observer is the same. Based on the idea of adaptive identification, if there is a difference between the estimated current output by the observer and the actual detected current (current estimation error), it is assumed that there is an error in the rotational speed information built in the observer, and based on the current estimation error. In this method, the estimated speed is obtained by correcting the rotational speed information built in the observer. If speed control is performed based on this estimated speed, the rotational speed of the induction motor can be controlled to a desired value (see, for example, Non-Patent Document 2).
Also in the conventional speed sensorless control method in a synchronous motor, speed estimation is performed using an observer (see, for example, Patent Document 1).

Japanese Patent No. 4672236

Practical design of AC servo system, General Electronic Publishing (1997), 122, 86 Hidehiko Sugimoto, et al. “Improvement of Stability of Speed Sensorless Vector Control of Induction Motors” 2001 IEEJ National Conference Proceedings, 4-109

In a conventional speed sensorless control method for an induction motor, the observer of the induction motor incorporates a slip frequency as a parameter in addition to the rotation speed information. Due to the difference in voltage-to-current transfer characteristics between the induction motor and the observer, there is a difference between the slip frequency and the estimated slip frequency obtained based on the estimated current and estimated magnetic flux output by the observer. There is. This difference becomes particularly noticeable as the current of the induction motor is changed transiently, and the current estimation error by the observer includes not only the error due to the rotation speed information but also the error due to the slip frequency. For this reason, the transient accuracy of the estimated speed is lowered, and the speed control response is deteriorated.
When increasing the speed control response of the induction motor, the gain of the speed controller is usually increased to increase the torque amplitude of the induction motor. However, as a result, the torque current of the induction motor fluctuates greatly, and the difference in slip frequency increases, so that the accuracy of the estimated speed decreases. For this reason, there has been a problem that improvement of the speed control response is limited.

  In a conventional speed sensorless control method in a synchronous motor, the frequency of the coordinate conversion phase is track-controlled so as to synchronize the two-axis rotational coordinate where the observer is constructed with the rotor magnetic flux of the synchronous motor. Due to the error due to the transient fluctuation of the frequency for the follow-up control at that time, there is a problem that the transient accuracy of the estimated speed is lowered and the improvement of the speed control response is limited as in the case of the induction motor.

  The present invention has been made to solve the above-described problems. In a control device that performs speed sensorless control of an AC motor, the transient accuracy of the estimated speed is improved and the speed control response is improved. The purpose is to plan.

  The control apparatus for an AC motor according to the present invention includes a command generation unit that generates a voltage command on the dq axis for driving the AC motor, a current detection unit that detects a current on the dq axis of the AC motor, and the above A state observation unit that inputs a detected current and the voltage command and outputs an estimated current and an estimated magnetic flux of the AC motor, and an estimated speed of the AC motor based on the estimated current, the estimated magnetic flux, and the detected current Based on the estimated current, the estimated magnetic flux, and the detected current, a first frequency calculating unit that calculates a first frequency so that a q-axis component of the estimated magnetic flux becomes 0, and Generated in the state observation unit due to the first frequency, and an output frequency calculation unit that calculates the output frequency that is the frequency of the dq axis rotation coordinate phase by adding the first frequency to the estimated speed of the AC motor And a compensation unit that generates a compensation signal for compensating for that disturbance. The state observing unit inputs the detection current, the voltage command, the compensation signal, the output frequency, and the first frequency, and calculates the estimated current and the estimated magnetic flux of the AC motor. Is.

  According to the control apparatus for an AC motor according to the present invention, the disturbance in the state observation unit can be compensated by the compensation signal, the influence of the disturbance is suppressed, the transient accuracy of the estimated speed is improved, and the speed control response is improved. Can be planned.

It is a figure which shows the structure of the control apparatus of the induction motor by Embodiment 1 of this invention. It is a figure which shows the structure which implement | achieved the function of each part in the control apparatus of the induction motor by Embodiment 1 of this invention using software. It is a figure which shows the structure of the control apparatus of the synchronous motor by Embodiment 2 of this invention.

Embodiment 1 FIG.
Hereinafter, a control device for an induction motor that performs speed sensorless control will be described as a control device for an AC motor according to Embodiment 1 of the present invention.
FIG. 1 is a diagram illustrating a configuration of a control device for an induction motor according to a first embodiment.
As shown in FIG. 1, the control device 1 controls the induction motor 4 by driving and controlling an inverter 2 having a switching element. Hereinafter, the subscript * indicates a command value, and the subscript # indicates an estimated value. The subscript s indicates the state quantity of the stator of the induction motor 4, and the subscript r indicates the state quantity of the rotor of the induction motor 4.
The control device 1 performs control arithmetic processing and outputs voltage commands Vuvw * (Vu *, Vv *, Vw *) to the inverter 2. In the inverter 2, PWM processing is performed, and the switching element is driven based on the PWM processing, and the output voltage in accordance with the voltage command Vuvw * is supplied to the induction motor 4. The current sensor 3 detects the current iuvw (iu, iv, iw) of the induction motor 4 and transmits a signal of the detected current to the control device 1.

The control device 1 includes coordinate conversion units 11a and 11b that perform coordinate conversion between three-phase stationary coordinates and dq axis rotation coordinates, a current control unit 12, a speed control unit 14, a magnetic flux control unit 15, and state observation. An observer 17 as a unit, a speed estimation unit 18, an estimated slip frequency calculation unit 19 as a first frequency calculation unit for calculating an estimated slip frequency ωs # as a first frequency, and an adder 13 as an output frequency calculation unit And an integrator 16 and a compensation unit 20. The current detection unit 21 includes the current sensor 3 and the coordinate conversion unit 11a, and the command generation unit 22 includes the current control unit 12 and the speed control unit 14.
The current sensor 3 may be provided outside the current detection unit. In this case, the current detection unit detects the current on the dq axis of the AC motor based on the detection current iuvw from the current sensor 3. Further, the induction motor control device 1 may include an inverter 2.

In the control apparatus 1, the process described below is performed to control the rotation speed of the induction motor 4 to a desired value.
The speed control unit 14 receives the speed command ωr * and the estimated speed ωr #, performs arithmetic processing so that they match, and outputs a torque current command iq *. The magnetic flux control unit 15 receives the magnetic flux command Φdr * and the estimated magnetic flux Φdr #, performs arithmetic processing so that they match, and outputs an excitation current command id *.
The detected current iuvw from the current sensor 3 is converted into a current idq (id, iq) on the dq axis rotation coordinates by the coordinate conversion unit 11 a and input to the current control unit 12. The current control unit 12 inputs the detection current idq (id, iq) on the dq axis, the torque current command iq *, and the excitation current command id *, performs arithmetic processing so that they match, Voltage command Vdq * (Vd *, Vq *) is output. The voltage command Vdq * is converted into a three-phase voltage command Vuvw * by the coordinate conversion unit 11 b and output to the inverter 2.

  The observer 17 outputs the estimated current idq # and the estimated magnetic flux Φdr # of the induction motor 4 based on the detected current idq and the voltage command Vdq *. The speed estimation unit 18 outputs an estimated speed ωr # based on the estimated current idq #, the estimated magnetic flux Φdr #, and the detected current idq. The estimated slip frequency calculation unit 19 calculates an estimated slip frequency ωs # based on the estimated current idq #, the estimated magnetic flux Φdr #, and the detected current idq. The adder 13 calculates the output frequency ω # by adding the estimated speed ωr # and the estimated slip frequency ωs #. The integrator 16 integrates the output frequency ω # to obtain a coordinate conversion phase θ and outputs it to the coordinate conversion units 11a and 11b. The compensation unit 20 generates a compensation signal ADD that compensates for disturbance generated in the observer 17 based on the estimated current idq #, the estimated magnetic flux Φdr #, the detected current idq, and the estimated slip frequency ωs #. This compensation signal ADD is additionally input to the observer 17 together with the detection current idq and the voltage command Vdq *. The observer 17 also receives the estimated slip frequency ωs # and the output frequency ω #.

Next, the process for estimating the rotational speed of the induction motor 4 will be described. In this embodiment, a compensation signal ADD that compensates for disturbance generated in the observer 17 is input to the observer 17. Here, the basic arithmetic processing will be described first.
Based on the arithmetic processing described in Non-Patent Document 2, a case where there is no compensation signal ADD will be described below. However, various methods have been studied for the calculation of the observer configuration and the estimated speed, and the method described below is merely an example, and this embodiment can be applied to those in which disturbance occurs inside the observer.

In the observer, the estimated current idqs # (ids #, iqs #) and the estimated magnetic flux Φdqr # (Φdr #, Φqr #) are taken as state variables, and the following equation (1) that is an observer state equation is calculated, Estimated current idqs # and estimated magnetic flux Φdqr # are output. Here, since a known method described in Non-Patent Document 2 is used, detailed description is omitted.
In Expression (1), an input signal to the observer is an input signal for a state variable to be integrated. That is, in Expression (1), the second term on the right side based on the voltage information and the third term on the right side for feeding back the current estimation error are input signals. Since the estimated current idqs #, which is the calculation result of Expression (1), is used for the calculation of the third term, the signals input from the outside to the observer are the detected current idqs (ids, iqs) and the voltage command Vds *.

Where Rs: primary resistance, Rr: secondary resistance, Ls: primary inductance, Lr: secondary inductance, M: mutual inductance, ωre #: estimated electrical angular frequency (number of pole pairs of estimated speed ωr #), g11 ˜g42: Observer feedback gain, ωs #: estimated slip frequency, ω #: output frequency (added value of estimated electrical angular frequency and estimated slip frequency).
Since A11 #, A12 #, and A22 # in the above formula (1) include ω #, ωre #, and ωs # inside, a subscript # indicating an estimated value is added.
The speed estimator 18 estimates the speed by a PI (proportional integration) operation shown in the following equation (2). However, Kp and Ki are speed estimation gains.

  The estimated slip frequency calculation unit 19 performs the calculation shown in the following equation (3). This is to control the q-axis estimated magnetic flux Φqr # to be zero, and aims to set the d-axis of the dq-axis rotational coordinate as the rotor magnetic flux of the induction motor 4.

The output frequency ω # obtained by adding the estimated speed ωr # and the estimated slip frequency ωs # is fed back and used as an observer parameter described in the above equation (1).
As described above, in the basic calculation process, the estimated speed ωr # is obtained by calculating the above formulas (1) to (3), and the estimated speed ωr # is fed back to the speed control unit 14. The sensorless speed control of the induction motor 4 can be realized.

  Next, the following formula (4) is a mathematical model of the induction motor 4 that rotates at the speed ωr. Where Φdqr (Φdr, Φqr): magnetic flux, Vds: d-axis voltage, ωre: electrical angular frequency (pole pair times the speed ωr), ωs: slip frequency, ω: output frequency (addition of electrical angular frequency and slip frequency) value). Symbols that are not particularly explained are the same as those in the above formula (1).

  As described above, a difference may occur between the slip frequency ωs and the estimated slip frequency ωs #, and the disturbance signal DI is generated in the observer due to the difference. As shown in the above equation (1), the integrated value of the estimated slip frequency ωs # and the state variables (estimated current idqs #, estimated magnetic flux Φdqr #) is in the integrand term on the right side. Similarly, in the mathematical model of the induction motor 4 shown in the above equation (4), the integrated value of the slip frequency ωs and the state variables (current idqs, magnetic flux Φdqr) is within the integrand term on the right side. For this reason, the disturbance signal DI resulting from the difference between the slip frequency ωs and the estimated slip frequency ωs # also takes the form of a product signal with the state variable of the observer.

Here, if a voltage according to the voltage command is applied to the induction motor 4 and there is no difference between the estimated speed ωr # and the speed ωr, the disturbance signal DI is expressed by the right side of the above equation (1) and the equation (4). It can be calculated by the difference from the right side and is expressed by the following equation (5).
When the estimated slip frequency ωs # is different from the slip frequency ωs, when the observer performs the calculation of the above equation (1), a disturbance signal DI is automatically generated inside the right side of the equation (1).

In this embodiment, in order to cancel the disturbance signal DI, the compensation unit 20 generates a compensation signal ADD obtained by inverting the polarity by applying a predetermined compensation feedback gain to the above equation (5) and adding it to the observer 17. input. The compensation signal ADD is expressed by the following equation (6). k1 to k4 are compensation feedback gains.
Further, the secondary magnetic flux Φdr and the slip frequency ωs of the induction motor 4 are obtained based on the detected current idqs. The secondary magnetic flux Φdr is calculated by the following equation (6a), and the slip frequency ωs is calculated by the following equation (6b).

  In the observer 17, the disturbance signal DI is compensated by adding the compensation signal ADD to the integrand calculation term on the right side of the above equation (1). That is, the observer 17 calculates the following equation (7), which is a state equation, and outputs the estimated current idqs # and the estimated magnetic flux Φdqr #.

  As shown in the above equation (6), the compensation signal ADD is a first signal formed by multiplying the estimated slip frequency ωs # calculated using the estimated current idqs # by the estimated current idqs # and the estimated magnetic flux Φdqr #. And a differential signal group based on a difference between each of the second signal group and a product obtained by multiplying the slip frequency ωs calculated using the detected current idqs by the detected current idqs and the magnetic flux Φdqr.

  As described above, in this embodiment, the observer 17 calculates the above equation (7), which is a state equation, and outputs the estimated current idqs # and the estimated magnetic flux Φdqr #, and the speed estimation unit 18 performs the above equation. (2) is calculated to calculate the estimated speed ωr #, and the estimated slip frequency calculating unit 19 calculates the above-described equation (3) to calculate the estimated slip frequency ωs #. The estimated speed ωr # is obtained by such calculation, and the estimated speed ωr # is fed back to the speed control unit 14 so that the sensorless speed control of the induction motor 4 can be realized.

  The observer 17 additionally inputs the compensation signal ADD as a disturbance signal DI generated internally due to the difference between the estimated slip frequency ωs # and the slip frequency ωs due to the estimated slip frequency ωs #. To compensate. For this reason, in the control apparatus 1 of the induction motor 4, the disturbance signal DI generated inside the observer 17 can be suppressed, and the accuracy of the estimated current idqs # and the estimated magnetic flux Φdqr # is improved. For this reason, the transient accuracy of the estimated speed ωr # is improved, and the speed control response is improved.

Note that the compensation signal ADD may be used only for the high-efficiency rows by setting the priority order among the first to fourth rows of the above equation (6), and the calculation load can be reduced. The priority order is determined based on the characteristics of the induction motor 4 itself and the configuration of the entire control system.
For example, in this embodiment in which the observer 17 and the speed estimation system are configured based on Non-Patent Document 2, the q-axis estimated magnetic flux Φqr # and the q-axis magnetic flux Φqr become zero by the action of the above equation (3). To be controlled. For this reason, the magnitude of the signal in the third row of the above equation (6) becomes very small and can be omitted. That is, the compensation signal ADD may be expressed by the following equation (8), and in this case, the state equation of the observer 17 is expressed by the following equation (8a).

By the way, generally, when operating an induction motor, control is performed so that the secondary magnetic flux is always constant regardless of the load torque and speed. This is to ensure the linearity of the torque current and the output torque of the induction motor, and accordingly, the d-axis current ids always flows, and the d-axis current ids is used for integration. For this reason, the signal of the second row related to the d-axis current ids in the compensation signal ADD is relatively larger than the signals of the other rows (first row, third row, and fourth row). The second row of the state equation of the observer 17 is a row having the q-axis current iqs as a state variable, and speed estimation is performed using the q-axis current estimation error as described in the above equation (2). Therefore, the sensitivity to the disturbance signal DI is higher than that of other rows.
For this reason, the signal in the second row of the compensation signal ADD contributes most to the suppression operation of the disturbance signal DI. That is, the compensation signal ADD may be expressed by the following equation (9). In this case, the state equation of the observer 17 is expressed by the following equation (9a).

  Furthermore, when the secondary magnetic flux is controlled so as to be constant and the operation conditions are limited so that the d-axis current ids is controlled to be substantially constant, the d-axis current ids and the d-axis estimated current ids # are treated as fixed values. Can do. At this time, the compensation signal ADD of the above equation (9) can be simplified as shown by the following equation (10). In this case, the state equation of the observer 17 is expressed by the following equation (10a). Here, both the secondary magnetic flux Φdr and the estimated magnetic flux Φdr # can be expressed by the product of the mutual inductance M and the fixed value of the d-axis current ids.

  As described above, when the priority order is set in the first to fourth rows of the above formula (6) and signals in some rows are omitted, it is compared with the case where all signals are used. Although the effect of suppressing the disturbance signal DI is somewhat inferior, the number of compensation feedback gains k1 to k4 can be reduced, and the design can be simplified.

In this embodiment, a four-dimensional observer having current and magnetic flux as state variables is taken as an example. However, if the observer has a structure having a product of a slip frequency and a state variable inside, the above-described principle of generating a disturbance signal is the same. Therefore, it can be applied regardless of the type of observer.
The estimated current, the estimated magnetic flux, and the estimated slip frequency include both a component that changes over time and a component that does not change over time. The same applies to the current, magnetic flux, and slip frequency of the induction motor. Integration is performed in the calculation of the compensation signal ADD, but the result of the integration processing includes a component that can be regarded as integration of a signal that changes over time and a signal that does not change over time. That is, the addition of the compensation signal ADD in the observer state equation can be regarded as equivalently manipulating the observer feedback gain. For this reason, it is necessary to design the compensation feedback gains k1 to k4 in the compensation signal ADD in consideration of the design of the existing observer feedback gain.

On the other hand, since the result of the integration process of the compensation signal ADD includes a component that can be regarded as a product of signals including components that change with time, it is different from a linear operation such as an observer feedback term. Therefore, in designing the compensation feedback gains k1 to k4 in the compensation signal ADD, it is necessary to consider this nonlinear calculation result.
If the compensation feedback gains k1 to k4 of the compensation signal ADD are simply set to 1, for example, to suppress the disturbance signal DI described in the above equation (5), the observer feedback caused by the observer feedback gain for the above-described reason. May affect the dynamic characteristics. Accordingly, the compensation feedback gains k1 to k4 are designed in consideration of the situation including the above-described linear calculation and nonlinear calculation in the calculation of the compensation signal ADD, and the disturbance signal DI is compensated through the design. Thereby, the effect of suppressing the influence of the disturbance signal DI and improving the speed control response in the speed sensorless control can be obtained while maintaining the characteristics of the original observer and ensuring the stability.

In addition, the control apparatus 1 shown in the said embodiment can also implement | achieve the function of each part using software, and is demonstrated below using FIG.
As shown in FIG. 2, the control device 1 includes a storage device 6 and a processor 7. As a result of signal processing of a program executed on the processor 7 and signal processing in a logic circuit provided on the processor 7. The function is realized. The processor 7 exchanges a read / write signal 8 with the storage device 6, reads the program from the storage device 6, and executes it. Further, the processor 7 writes information to be temporarily stored in the storage device 6 in the course of the processing, and reads from the storage device 6.
For example, the functions of the coordinate conversion units 11a and 11b, the current control unit 12, the speed control unit 14, the magnetic flux control unit 15, the integrator 16, the observer 17, the speed estimation unit 18, the estimated slip frequency calculation unit 19 and the compensation unit 20 are as follows. This can be realized by executing a program in the processor 7. Thus, even when the functions of the respective units are realized by the processing in the processor 7, the same effect can be obtained by the same operation as in the first embodiment.

Embodiment 2. FIG.
Next, a control device for a synchronous motor that performs speed sensorless control will be described as a control device for an AC motor according to Embodiment 2 of the present invention.
FIG. 3 is a diagram showing the configuration of the synchronous motor control device according to the second embodiment.
As shown in FIG. 3, the control device 1 controls the synchronous motor 5 by driving and controlling an inverter 2 having a switching element.
The control device 1 performs control arithmetic processing and outputs voltage commands Vuvw * (Vu *, Vv *, Vw *) to the inverter 2. In the inverter 2, PWM processing is performed, the switching element is driven based on the PWM processing, and the output voltage in accordance with the voltage command Vuvw * is supplied to the synchronous motor 5. The current sensor 3 detects the current iuvw (iu, iv, iw) of the synchronous motor 5 and transmits a signal of the detected current to the control device 1.

The control device 1 includes coordinate conversion units 11a and 11b that perform coordinate conversion between three-phase stationary coordinates and dq axis rotation coordinates, a current control unit 12, a speed control unit 14, and an observer 17A as a state observation unit. , A speed estimation unit 18, a tracking frequency calculation unit 19A as a first frequency calculation unit that calculates the tracking frequency ωt # as the first frequency, an adder 13, an integrator 16, and a compensation unit 20A. The current detection unit 21 includes the current sensor 3 and the coordinate conversion unit 11a, and the command generation unit 22 includes the current control unit 12 and the speed control unit 14.
The current sensor 3 may be provided outside the current detection unit. In this case, the current detection unit detects the current on the dq axis of the AC motor based on the detection current iuvw from the current sensor 3. Furthermore, the synchronous motor control device 1 may include an inverter 2.

In the control apparatus 1, the process described below is performed to control the rotational speed of the synchronous motor 5 to a desired value.
The speed control unit 14 receives the speed command ωr * and the estimated speed ωr #, performs arithmetic processing so that they match, and outputs a torque current command iq *. The excitation current command id * is normally set to zero, but may be set to a negative value when the output voltage of the inverter 2 is saturated. The detected current iuvw from the current sensor 3 is converted into a current idq (id, iq) on the dq axis rotation coordinates by the coordinate conversion unit 11 a and input to the current control unit 12. The current control unit 12 inputs the detection current idq (id, iq) on the dq axis, the torque current command iq *, and the excitation current command id *, performs arithmetic processing so that they match, Voltage command Vdq * (Vd *, Vq *) is output. The voltage command Vdq * is converted into a three-phase voltage command Vuvw * by the coordinate conversion unit 11 b and output to the inverter 2.

The observer 17A outputs the estimated current idq # and the estimated magnetic flux Φdr # of the synchronous motor 5 based on the detected current idq and the voltage command Vdq *. The speed estimation unit 18 outputs an estimated speed ωr # based on the estimated current idq #, the estimated magnetic flux Φdr #, and the detected current idq. The follow-up frequency calculating unit 19A calculates the follow-up frequency ωt # based on the estimated current idq #, the estimated magnetic flux Φdr #, and the detected current idq. The follow-up frequency ωt # is a correction frequency for synchronizing the d-axis of the dq-axis rotation coordinate with the rotor magnetic flux of the synchronous motor 5.
The adder 13 calculates the output frequency ω # by adding the estimated speed ωr # and the follow-up frequency ωt #. The integrator 16 integrates the output frequency ω # to obtain a coordinate conversion phase θ and outputs it to the coordinate conversion units 11a and 11b. The compensation unit 20A generates a compensation signal ADDa that compensates for the disturbance generated in the observer 17A based on the estimated current idq #, the estimated magnetic flux Φdr #, the detected current idq, and the tracking frequency ωt #. The compensation signal ADDa is additionally input to the observer 17A together with the detection current idq and the voltage command Vdq *. Note that the follower frequency ωt # and the output frequency ω # are also input to the observer 17A.

Next, the process for estimating the rotational speed of the synchronous motor 5 will be described. In this embodiment, a compensation signal ADDa for compensating for disturbance generated in the observer 17A is input to the observer 17A. Here, the basic arithmetic processing will be described first.
Based on the arithmetic processing described in Patent Document 1, a case where there is no compensation signal ADDa will be described below. However, various methods have been studied for the calculation of the observer configuration and the estimated speed, and the method described below is merely an example, and this embodiment can be applied to those in which disturbance occurs inside the observer.

In the observer, the estimated current idqs # (ids #, iqs #) and the estimated magnetic flux Φdqr # (Φdr #, Φqr #) are taken as state variables, and the following equation (11) that is an observer state equation is calculated, Estimated current idqs # and estimated magnetic flux Φdqr # are output. Here, since a known method is used, detailed description is omitted.
In Expression (11), an input signal to the observer is an input signal for a state variable to be integrated. That is, in Expression (11), the second term on the right side based on the voltage information and the third term on the right side for feeding back the current estimation error are input signals. Since the estimated current idqs #, which is the calculation result of Expression (11), is used for the calculation of the third term, the signals input from the outside to the observer are the detected current idqs (ids, iqs) and the voltage command Vds *.

Where Rs: resistance, L: inductance, ωre #: estimated electrical angular frequency (multiple times the estimated speed ωr #), g11 to g42: observer feedback gain, ω #: output frequency (estimated electrical angular frequency and tracking frequency) And addition value).
The speed estimation unit 18 estimates the speed by the PI calculation shown in the following formula (12). However, Kp and Ki are speed estimation gains.

  The follow-up frequency calculation unit 19A performs the calculation shown in the following formula (13). This calculation controls the q-axis estimated magnetic flux Φqr # to be zero, and the follow-up frequency ωt # is a coordinate conversion phase so that the d-axis of the dq-axis rotational coordinate is synchronized with the rotor magnetic flux of the synchronous motor 5. This is a frequency for correcting the frequency of θ (output frequency).

The output frequency ω # obtained by adding the estimated speed ωr # and the follow-up frequency ωt # is fed back and used as an observer parameter described in the above equation (11).
As described above, in the basic calculation process, the estimated speed ωr # is obtained by calculating the above equations (11) to (13), and the estimated speed ωr # is fed back to the speed control unit 14. The sensorless speed control of the synchronous motor 5 can be realized.

  Next, the following equation (14) is a mathematical model of the synchronous motor 5 that rotates at the speed ωr. However, Φdqr (Φdr, Φqr): magnetic flux, Vds: d-axis voltage, ωre: electrical angular frequency (pole pair times the speed ωr). Symbols that are not particularly explained are the same as in the above formula (11).

As shown in the above equation (14), the mathematical model of the synchronous motor 5 itself does not have a parameter of the frequency ωt that is the basis of the follow-up frequency ωt #. In the first embodiment, the disturbance signal DI is generated in the observer due to the difference between the slip frequency ωs and the estimated slip frequency ωs #. However, in the second embodiment, the follow-up frequency ωt # itself is generated. Disturbance occurs due to
The follow-up frequency ωt # is calculated so as to synchronize the d-axis of the dq-axis rotation coordinate where the observer shown in the above equation (11) is installed with the rotor magnetic flux of the synchronous motor 5. Since this follow-up frequency ωt # is calculated based on the estimation error of the current, when the torque current iqs of the synchronous motor 5 is transiently operated, the disturbance is caused by the difference in transfer characteristics between the observer and the synchronous motor. Occur.

  Here, when a voltage according to the voltage command is applied to the synchronous motor 5 and there is no difference between the estimated speed ωr # and the speed ωr, the disturbance signal DIa is expressed by the right side of the equation (11) and the equation (14). It can be calculated by the difference from the right side and is expressed by the following equation (15). The frequency corresponding to the frequency ωt is expressed as 0. When the observer performs the calculation of the above equation (11), the disturbance signal DIa is automatically generated inside the right side of the equation (11).

  In this embodiment, in order to cancel the disturbance signal DIa, the compensation unit 20A generates a compensation signal ADDa obtained by inverting the polarity by applying a predetermined compensation feedback gain to the above equation (15) and adding it to the observer 17A. input. The compensation signal ADDa is expressed by the following equation (16). k1 to k4 are compensation feedback gains. Further, Φdr # is a kind of electric motor constant that indicates the magnitude of the rotor magnetic flux, called an induced voltage constant, and is a value set in advance in the same manner as the resistance value and inductance value of the electric motor.

  In the observer 17A, the disturbance signal DIa is compensated by adding the compensation signal ADDa to the integral calculation term on the right side of the above equation (11). That is, the observer 17A calculates the following equation (17), which is a state equation, and outputs the estimated current idqs # and the estimated magnetic flux Φdqr #.

As shown in the above equation (16), the compensation signal ADDa is a first signal group composed of products obtained by multiplying the follow-up frequency ωt # calculated using the estimated current idqs # by the estimated current idqs # and the estimated magnetic flux Φdqr #. Consists of.
As described above, in this embodiment, the observer 17A calculates the above equation (17), which is a state equation, and outputs the estimated current idqs # and the estimated magnetic flux Φdqr #, and the speed estimation unit 18 performs the above equation. The estimated speed ωr # is calculated by calculating (12), and the follow-up frequency calculating unit 19A calculates the follow-up frequency ωt # by calculating the above equation (13). The estimated speed ωr # is obtained by such calculation, and the estimated speed ωr # is fed back to the speed control unit 14 to realize sensorless speed control of the synchronous motor 5.

  The observer 17A compensates the disturbance signal DIa generated internally due to the follow-up frequency ωt # by additionally inputting the compensation signal ADDa. For this reason, in the control device 1 of the synchronous motor 5, the disturbance signal DIa generated inside the observer 17A can be suppressed, and the accuracy of the estimated current idqs # and the estimated magnetic flux Φdqr # is improved. For this reason, the transient accuracy of the estimated speed ωr # is improved, and the speed control response is improved.

As in the first embodiment, the compensation signal ADDa may be used only for the high-efficiency rows by setting priorities among the first to fourth rows of the above equation (16). The load can be reduced. The priority order is determined based on the characteristics of the induction motor 4 itself and the configuration of the entire control system.
For example, in this embodiment in which the observer 17A and the speed estimation system are configured based on the above-mentioned Patent Document 1, the q-axis estimated magnetic flux Φqr # and the q-axis magnetic flux Φqr become zero by the action of the above equation (13). Be controlled. For this reason, the magnitude of the signal in the third row of the above equation (16) becomes very small and can be omitted. That is, the compensation signal ADDa may be expressed by the following equation (18). In this case, the state equation of the observer 17A is expressed by the following equation (18a).

  As described above, the compensation signal ADDa has the number of compensation feedback gains k1 to k4, although the suppression effect of the disturbance signal DI is somewhat inferior to the case of using all signals by omitting signals in some rows. Can be reduced, and the design can be simplified.

In this embodiment, a four-dimensional observer having current and magnetic flux as state variables is taken as an example. However, if the observer is constructed on the dq axis rotation coordinates, the d axis is synchronized with the rotor magnetic flux of the synchronous motor 5. Since the processing or similar processing is performed and a similar disturbance signal DIa is generated including a component corresponding to the follow-up frequency ωt #, it can be applied.
In addition, the compensation feedback gains k1 to k4 in the compensation signal ADDa are designed in consideration of the existing observer feedback gain and non-linear arithmetic processing, as in the first embodiment. Thus, the compensation feedback gains k1 to k4 are designed, and the disturbance signal DIa is compensated through the designed design. Thereby, while maintaining the characteristics of the original observer and ensuring stability, the effect of the disturbance signal DIa is suppressed, and the speed control response in the speed sensorless control is improved.

  Also in the above embodiment, the function of each unit may be realized by the processing in the processor 7 using the configuration shown in FIG.

  Also, within the scope of the present invention, the embodiments can be freely combined, or the embodiments can be appropriately modified and omitted.

Claims (6)

A command generator for generating a voltage command on the dq axis for driving the AC motor;
A current detector for detecting a current on the dq axis of the AC motor;
A state observation unit that inputs the detected current and the voltage command and outputs an estimated current and an estimated magnetic flux of the AC motor;
A speed estimation unit that outputs an estimated speed of the AC motor based on the estimated current, the estimated magnetic flux, and the detected current;
A first frequency calculating unit that calculates a first frequency based on the estimated current, the estimated magnetic flux, and the detected current so that a q-axis component of the estimated magnetic flux is zero;
An output frequency calculation unit that calculates the output frequency that is the frequency of the dq axis rotation coordinate phase by adding the first frequency to the estimated speed of the AC motor;
A compensation unit that generates a compensation signal that compensates for disturbance generated in the state observation unit due to the first frequency,
The state observing unit inputs the detection current, the voltage command, the compensation signal, the output frequency, and the first frequency, and calculates the estimated current and the estimated magnetic flux of the AC motor. ,
AC motor control device. The compensation unit generates the compensation signal using signals in a first signal group obtained by multiplying the estimated current and the estimated magnetic flux by the first frequency, respectively.
The control apparatus for an AC motor according to claim 1. The AC motor is an induction motor, the first frequency is an estimated slip frequency,
The compensation unit uses a slip frequency and a magnetic flux calculated based on the detection current, and a difference signal between the second signal group obtained by multiplying the slip frequency by the detection current and the magnetic flux, respectively, and the first signal group At least one of the groups is the compensation signal,
The control apparatus for an AC motor according to claim 2. The AC motor is a synchronous motor, and the first frequency is a follow-up frequency for synchronizing the d-axis with the rotor magnetic flux.
The control apparatus for an AC motor according to claim 2. The dq axis rotation coordinate phase of the AC motor calculated based on the output frequency is used for coordinate conversion between the dq axis rotation coordinate and the three-phase stationary coordinate in both the detection current and the voltage command.
The control apparatus for an AC motor according to any one of claims 1 to 4. The command generation unit includes a speed control unit that generates a current command so that the estimated speed follows the speed command, and a current control unit that generates the voltage command so that the detected current follows the current command. Prepared,
The control apparatus of the alternating current motor of any one of Claims 1-5.

Images