control device

Abstract

This invention provides a control device that can track 2f disturbances caused by magnetic circuit imbalance and current detection system imbalance in real time and suppress 2f vibrations. [Solution] The control device 200 includes a speed control loop, an estimated q-axis current control loop, and a q-axis current control loop. The Nf component remover 106 removes the 2f component from the q-axis current based on the detected current value of the three-phase motor 104. The subtractor 101 subtracts the corrected q-axis current, from which the 2f component has been removed by the Nf component remover 106, from the second current command value. The gain compensator 102 maximizes the gain of the 2f component in the open-loop characteristics based on the output of the subtractor 101.

Description

Technical Field

[0001] The present disclosure relates to a control device for controlling a three-phase motor.

Background Art

[0002] Devices for controlling a three-phase motor are described in Patent Documents 1 and 2. In the devices described in Patent Documents 1 and 2, control is performed to suppress torque ripple of the three-phase motor.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0004] As causes of torque ripple that is twice the electrical angle of a three-phase motor, there are two: impedance imbalance between phases of the drive magnetic circuit of the three-phase motor and detection gain imbalance of the current detection system flowing through the coils of each phase. In this specification, the frequency that is twice the electrical angle of the three-phase motor is also denoted as 2f, and the torque ripple is also denoted as 2f vibration. The impedance imbalance is also denoted as magnetic circuit imbalance. The detection gain imbalance is also denoted as current detection system imbalance.

[0005] In the device described in Patent Document 1, measures are taken against 2f vibration caused by magnetic circuit imbalance. This measure is premised on grasping the imbalance amount, that is, the 2f disturbance, in advance. Therefore, it has not been possible to suppress 2f vibration in real time following changes in the imbalance amount.

[0006] The device described in Patent Document 2 incorporates measures to address 2f vibration caused by current detection system imbalance. However, this measure requires a learning operation during device installation. Therefore, it was not possible to suppress 2f vibration in real time by tracking changes in the imbalance.

[0007] This disclosure was made to solve the problems described above. The purpose of this disclosure is to provide a control device that can track 2f disturbances caused by magnetic circuit imbalance and current detection system imbalance in real time and suppress 2f vibrations. [Means for solving the problem]

[0008] The control device according to this disclosure comprises a speed control loop for making the speed of a three-phase motor follow a speed command value using negative feedback control, an estimated q-axis current control loop for making the estimated q-axis current based on the speed of the three-phase motor follow a first current command value which is the output of the speed control loop using negative feedback control, and a q-axis current control loop for making the q-axis current based on the detected value of the current of the three-phase motor follow a second current command value which is the output of the estimated q-axis current control loop using negative feedback control. The q-axis current control loop is based on the detected value of the current of the three-phase motor Based on the q-axis current and the electrical angle of the three-phase motor, the 2f component is estimated and extracted. An Nf component remover removes the 2f component from the q-axis current, a first subtractor subtracts the corrected q-axis current from which the 2f component has been removed by the Nf component remover from the second current command value, and the output of the first subtractor. Based on the electrical angle of the three-phase motor, the 2f component is estimated and extracted. It includes a gain compensator that maximizes the gain of the 2f component in the open-loop characteristics. 2f represents a frequency that is twice the electrical angle of a three-phase motor. [Effects of the Invention]

[0009] The control device according to this disclosure can track 2f disturbances caused by magnetic circuit imbalance and current detection system imbalance in real time and suppress 2f vibrations. [Brief explanation of the drawing]

[0010] [Figure 1] This figure shows an example of a drive system equipped with a control device according to Embodiment 1. [Figure 2] This is a diagram showing an example of a current controller. [Figure 3] This is a diagram showing an example of an Nf component open-loop gain infinity compensator. [Figure 4] This is a diagram showing an example of an Nf component detector. [Figure 5] This is a diagram showing an example of setting the upper limit value of the feedback gain. [Figure 6] This is a diagram showing an example of the estimated characteristics of an Nf component detector. [Figure 7] This is a diagram showing another example of an Nf component open-loop gain infinity compensator. [Figure 8] This is a diagram showing an example of an Nf component remover. [Figure 9] This is a diagram showing an example of an Nf component detector. [Figure 10] This is a diagram for explaining various characteristics of the q-axis current control loop. [Figure 11] This is a diagram showing an example of a q-axis current estimator. [Figure 12] This is a diagram for explaining various characteristics of the middle loop. [Figure 13] This is a diagram for explaining various characteristics of the speed control loop when various characteristics of the middle loop are set as shown in FIG. 12. ]> [Figure 14] An example of the waveform of the torque signal of the hoisting motor when a 2f disturbance signal is applied is shown. [Figure 15] An example of the waveform of the torque signal of the hoisting motor when a 2f disturbance signal is applied is shown. [Figure 16] An example of the waveform of the torque signal of the hoisting motor when a 2f disturbance signal is applied is shown. [Figure 17] This is a diagram showing an example of the hardware resources of the control device. [Figure 18] This is a diagram showing another example of the hardware resources of the control device.

Embodiments for Carrying Out the Invention

[0011] A detailed explanation follows, with reference to the drawings. Repetitive explanations will be simplified or omitted as appropriate. In each drawing, the same reference numerals indicate the same or corresponding parts.

[0012] Embodiment 1. Figure 1 shows an example of a drive system equipped with a control device 200 in Embodiment 1. The drive system comprises a three-phase motor 104 (not shown in Figure 1) and a control device 200. As an example, the three-phase motor 104 is a motor for driving an elevator car. As another example, the three-phase motor 104 is a motor used in electric power steering. The three-phase motor 104 may also be a motor used for other purposes, such as a motor used to power machine tools.

[0013] The control device 200 includes a dq-axis current control device and controls the three-phase motor 104. Figure 1 shows an example of the q-axis current control device, which is one of the functions of the control device 200. The control device 200 includes, for example, a current controller 100, a pulse generator 1, a time differentiator 2, a subtractor 3, a speed control compensator 4, a q-axis current estimator 5, a subtractor 6, and an estimated q-axis current control compensator 7. The current controller 100 has the function of controlling the d-axis current and q-axis current of the three-phase motor 104, i.e., a dq-axis current controller.

[0014] Figure 2 shows an example of the current controller 100. Figure 2 is a block diagram showing the details of the q-axis current control function, which is one of the dq-axis current control functions of the current controller 100. As shown in Figure 2, the q-axis current controller included in the current controller 100 comprises, for example, a subtractor 101, an Nf component open-loop gain infinity compensator 102, a current control compensator 103, a q-axis current detector 105, and an Nf component remover 106. Hereafter, the Nf component open-loop gain infinity compensator 102 will also be simply referred to as the gain compensator 102.

[0015] A pulse generator 1 is installed on a three-phase motor 104. The pulse generator 1 outputs the electrical angle θre of the three-phase motor 104. The electrical angle θre from the pulse generator 1 is input to a time differentiator 2. The time differentiator 2 calculates the speed of the three-phase motor 104 from the electrical angle θre. Specifically, the time differentiator 2 calculates and outputs the angular velocity ωre of the three-phase motor 104 from the input signal, the electrical angle θre.

[0016] The speed control loop shown in Figure 1 is a negative feedback control. This speed control loop is used to make the speed of the three-phase motor 104 follow the speed command value using negative feedback control. Specifically, the speed control loop makes the angular velocity ωre of the three-phase motor 104, which is the output from the time differentiator 2, follow the speed command value, which is the speed tracking target mv_ref.

[0017] Subtractor 3 subtracts the angular velocity ωre from the speed tracking target mv_ref, which is also an input signal, and outputs a speed error mv_er. The speed error mv_er from subtractor 3 is input to speed control compensator 4. The speed control loop is a loop that operates so that the speed error mv_er becomes 0. That is, speed control compensator 4 outputs a current command value iq_ref so that the input signal, the speed error mv_er, becomes 0. The current command value iq_ref is the first current command value, which is the output of the speed control loop. Speed ​​control compensator 4 is implemented, for example, by a PID controller.

[0018] The estimated q-axis current control loop shown in Figure 1 is a negative feedback control. This estimated q-axis current control loop is used to make the estimated q-axis current of the three-phase motor 104 follow the first current command value using negative feedback control. The estimated q-axis current is the q-axis current estimated based on the detected speed value of the three-phase motor. Specifically, the estimated q-axis current control loop makes the estimated q-axis current iqest of the three-phase motor 104, which is the output from the q-axis current estimator 5, follow the current command value iq_ref from the speed control compensator 4.

[0019] The q-axis current estimator 5 takes the angular velocity ωre from the time differentiator 2 as input and outputs the estimated q-axis current iqest. The subtractor 6 subtracts the estimated q-axis current iqest from the current command value iq_ref, which is also an input signal, and outputs the estimated q-axis current error er_iqest. The estimated q-axis current error er_iqest from the subtractor 6 is input to the estimated q-axis current control compensator 7. The estimated q-axis current control loop is a loop that operates so that the estimated q-axis current error er_iqest becomes 0. That is, the estimated q-axis current control compensator 7 outputs the current command value iq_ref2 so that the input signal, the estimated q-axis current error er_iqest, becomes 0. The current command value iq_ref2 is the second current command value, which is the output of the estimated q-axis current control loop. The estimated q-axis current control compensator 7 is implemented, for example, by a PID controller.

[0020] The overall function of the current controller 100 shown in Figure 2 is negative feedback control, which makes the detected q-axis current follow the current command value iq_ref2. In other words, the q-axis current control loop makes the q-axis current of the three-phase motor 104 follow the second current command value using negative feedback control. This q-axis current is calculated based on the detected current value of the three-phase motor.

[0021] Specifically, the q-axis current control loop includes a subtractor 101, a current control compensator 103, and a q-axis current detector 105, as well as a gain compensator 102 and an Nf component remover 106. The Nf component remover 106 is connected between the q-axis current detector 105 and the subtractor 101. The gain compensator 102 is connected between the subtractor 101 and the current control compensator 103.

[0022] The q-axis current detector 105 receives the values ​​of the currents flowing through each phase of the three-phase motor 104. These values ​​are, for example, values ​​detected by a current detector (not shown). Specifically, the q-axis current detector 105 receives the detected values ​​iu for the current flowing through the U phase, iv for the current flowing through the V phase, and iw for the current flowing through the W phase of the three-phase motor 104. The q-axis current detector 105 calculates and outputs the q-axis current iq of the three-phase motor 104 from the detected values ​​iu, iv, and iw of the currents flowing through each phase of the three-phase motor 104. In other words, the q-axis current iq, which is the output of the q-axis current detector 105, is a signal indicating the q-axis current value calculated based on the detected values ​​of the currents of each phase of the three-phase motor 104.

[0023] The Nf component remover 106 receives the q-axis current iq from the q-axis current detector 105 and the electrical angle θre from the pulse generator 1 as inputs. The Nf component remover 106 functions to remove the frequency component of the q-axis current iq that is N times the electrical angle θre, i.e., the Nf component. The Nf component includes the 2f component. Based on the input signals, the q-axis current iq and the electrical angle θre, the Nf component remover 106 calculates and outputs a removed signal, i.e., the corrected q-axis current iq_cor shown in Figure 2, which is obtained by removing the Nf component from the q-axis current iq.

[0024] The subtractor 101 receives the current command value iq_ref2 from the estimated q-axis current control compensator 7 and the corrected q-axis current iq_cor from the Nf component remover 106 as inputs. The subtractor 101 subtracts the corrected q-axis current iq_cor from the current command value iq_ref2, which is also an input signal, and outputs the error signal of the q-axis current control loop, i.e., the q-axis current error er_iq_in shown in Figure 2.

[0025] The gain compensator 102 receives the q-axis current error er_iq_in from the subtractor 101 and the electrical angle θre from the pulse generator 1 as inputs. Based on the output of the subtractor 101, the gain compensator 102 functions to infinitely increase the gain of the frequency component of the open-loop characteristics of the q-axis current control loop that is N times the electrical angle θre of the three-phase motor 104. Based on the input signals, the gain compensator 102 outputs a correction error signal.

[0026] The current control compensator 103 receives a correction error signal from the gain compensator 102. Based on the correction error signal from the gain compensator 102, the current control compensator 103 supplies power to control the rotation of the three-phase motor 104.

[0027] Figure 3 shows an example of an Nf component open-loop gain infinity compensator 102. The gain compensator 102 includes, for example, an adder 102a and an Nf component detector 102b. The Nf component detector 102b receives the electrical angle θre from the pulse generator 1 and the output from the adder 102a as inputs. The Nf component detector 102b estimates and extracts a frequency component (for example, the 2f component) that is a predetermined multiplier of the electrical angle θre from the output of the adder 102a, and outputs the extracted signal, i.e., the estimated Nf component nf_sens_q shown in Figure 2.

[0028] The adder 102a adds the output of the subtractor 101 and the output of the Nf component detector 102b. Specifically, the adder 102a receives the q-axis current error er_iq_in from the subtractor 101 and the estimated Nf component nf_sens_q from the Nf component detector 102b as inputs. The adder 102a adds the input signals, the q-axis current error er_iq_in and the estimated Nf component nf_sens_q, and outputs the result. The output from the adder 102a is input to the Nf component detector 102b and the current control compensator 103. Looking at the function of the gain compensator 102 shown in Figure 3, it is a loop that selects the Nf frequency component of the input signal, the q-axis current error er_iq_in, and provides positive feedback with a gain of 1.

[0029] Figure 4 shows an example of the Nf component detector 102b. In the example shown in Figure 4, the Nf component detector 102b comprises a magnification setter 102b1, a multiplier 102b2, a cosine arithmetic unit 102b3, a sine arithmetic unit 102b4, a subtractor 102b5, a multiplier 102b6, a magnified integrator 102b7, a multiplier 102b8, a multiplier 102b9, a magnified integrator 102b10, a multiplier 102b11, and an adder 102b12.

[0030] The multiplier 102b2 multiplies the electrical angle θre of the three-phase motor 104 by the multiplier N preset by the multiplier setter 102b1, and outputs the multiplied value Nθre. The multiplier N is preferably an integer. However, the multiplier N does not have to be an integer. The multiplier N may also be a fraction. The multiplied value Nθre from the multiplier 102b2 is input to the cosine arithmetic unit 102b3 and the sine arithmetic unit 102b4. The cosine arithmetic unit 102b3 calculates and outputs the cosine value cos(Nθre) of the multiplied value Nθre. The sine arithmetic unit 102b4 calculates and outputs the sine value sin(Nθre) of the multiplied value Nθre.

[0031] As shown in Figure 3, the input signal to the Nf component detector 102b is the output signal of the adder 102a. Therefore, the input signal input can be defined as the sum of the q-axis current error er_iq_in of the q-axis current control loop and the estimated Nf component nf_sens_q, which is the output signal of the Nf component detector 102b, as shown in Equation 1.

[0032]

number

[0033] The output signal from the Nf component detector 102b is expressed by the following equation.

[0034]

number

[0035] The subtractor 102b5 subtracts the output signal shown in Equation 2, i.e., the output of the Nf component detector 102b, from the input signal shown in Equation 1, i.e., the output of the adder 102a, and outputs the error er_sens_q. The multiplier 102b6 multiplies the output of the subtractor 102b5, i.e., the error er_sens_q, with the cosine value cos(Nθre), which is the output from the cosine arithmetic unit 102b3, and outputs the result of the multiplication. Similarly, the multiplier 102b9 multiplies the error er_sens_q from the subtractor 102b5 with the sine value sin(Nθre), which is the output from the sine arithmetic unit 102b4, and outputs the result of the multiplication.

[0036] The multiplier integrator 102b7 integrates the output from the multiplier 102b6 with a preset amplification factor. Specifically, the multiplier integrator 102b7 multiplies the output from the multiplier 102b6 by K, performs time integration, and outputs the result of the integration. Similarly, the multiplier integrator 102b10 integrates the output from the multiplier 102b9 with a preset amplification factor. Specifically, the multiplier integrator 102b10 multiplies the output from the multiplier 102b9 by K, performs time integration, and outputs the result of the integration. Note that K is a specific multiplier. Suitable values ​​for setting K will be described later.

[0037] Multiplier 102b8 multiplies the output from multiplier integrator 102b7 by the cosine value cos(Nθre) from cosine arithmetic unit 102b3 and outputs the result of the multiplication. Similarly, multiplier 102b11 multiplies the output from multiplier integrator 102b10 by the sine value sin(Nθre) from sine arithmetic unit 102b4 and outputs the result of the multiplication. Adder 102b12 adds the output from multiplier 102b8 and the output from multiplier 102b11, and the result of this addition is the output of Nf component detector 102b. That is, adder 102b12 outputs the estimated Nf component nf_sens_q.

[0038] The Nf component detector 102b, excluding the part that negatively feeds the output signal back to the input signal, has the function of estimating the frequency component of the electrical angle θre contained in the input signal if the value of K is 1. In the Nf component detector 102b shown in Figure 4, the function of convergeding the error between the actual value and the estimated value of the frequency component of the electrical angle θre contained in the input signal to 0 is realized by negatively feeding this estimated value back to the input signal. Therefore, it is desirable that the multiplier K in the multiplier integrator 102b7 and the multiplier K in the multiplier integrator 102b10 be set to a value greater than 1 in order to improve the responsiveness of convergence. In the example shown in this embodiment, the explanation will be given on the premise that the value of K is set to a value sufficiently greater than 1.

[0039] Next, referring to Figure 4, we will explain the signal estimation characteristics of frequency components at N times the electrical angle θre in the Nf component detector 102b.

[0040] The output of the Nf component detector 102b, i.e., the estimated Nf component nf_sens_q, is expressed by the following equation.

[0041]

number

[0042] The signal real_dis, which has frequency components that are N times the electrical angle θre to be detected, is defined by the following equation.

[0043]

number

[0044] Assuming that the q-axis current error er_iq_in of the q-axis current control loop is approximately equal to the signal real_dis, the estimation error er_sens_q in the Nf component detector 102b is expressed by the following equation.

[0045]

number

[0046] Furthermore, the first estimated coefficient, represented by the output of the multiplier integrator 102b7, is given by Equation 6, and the second estimated coefficient, represented by the output of the multiplier integrator 102b10, is given by Equation 7.

[0047]

number

[0048]

number

[0049] Therefore, each estimated coefficient has a double-angle term of Nθre. Also, if we ignore the double-angle term in Equation 6, we can obtain Equations 8 and 9. If we ignore the double-angle term in Equation 7, we can obtain Equations 10 and 11.

[0050]

number

[0051]

number

[0052]

number

[0053]

number

[0054] From equations 9 and 11, it can be seen that each estimated coefficient is essentially the detection characteristic of a first-order low-pass filter (LPF) with a cutoff frequency of 0.5K [rad / sec] relative to the true value, as shown in the following equation.

[0055]

number

[0056] Substituting equations 9, 11, and 12 into equation 3, we obtain the following equation.

[0057]

number

[0058] From the above, it can be seen that the LPF shown in Equation 12 is an indicator that determines the estimated bandwidth of the Nf component detector 102b. Furthermore, if the example shown in Figure 4 is adopted as the Nf component detector 102b, it can be seen that the Nf component detector 102b functions as a sensor that detects frequency components that are N times the electrical angle θre contained in the input signal. When the Nf component detector 102b functions as a sensor, its detectable frequency bandwidth is the cutoff frequency of the LPF. To widen the detectable bandwidth and improve responsiveness, the cutoff frequency should be increased, that is, the value of the feedback gain K should be increased.

[0059] However, since each estimated coefficient actually contains a double-angle term, we cannot simply increase the value of K arbitrarily. Specifically, if we increase the value of K too much, the double-angle component is amplified, and the above advantages are lost. The reason for this will be explained in detail below.

[0060] If the double-angle term in Equation 6 is not ignored, the actual estimated coefficient is expressed in Equation 14. Similarly, if the double-angle term in Equation 7 is not ignored, the actual estimated coefficient is expressed in Equation 15.

[0061]

number

[0062]

number

[0063] Equation 14, which formalizes the estimated coefficient of cosine, contains a term in the second right-hand side representing the estimation error related to the estimated coefficient of sine, which is unrelated to cosine. Similarly, in Equation 15, which formalizes the estimated coefficient of sine, contains a term in the second right-hand side representing the estimation error related to the estimated coefficient of cosine, which is unrelated to sine. The second right-hand side of both Equation 14 and Equation 15 is defined as a mutual interference term between the two estimated coefficients. This mutual interference term is a term we want to neutralize, and minimizing the function related to this mutual interference term is a key design point.

[0064] The function F1 relating to the second term on the right-hand side of Equation 14 is expressed in Equation 16. Similarly, the function F2 relating to the second term on the right-hand side of Equation 15 is expressed in Equation 17.

[0065]

number

[0066]

number

[0067] Equations 16 and 17 each specify a first-order low-pass filter (LPF). Furthermore, its cutoff frequency is expressed by a trigonometric function and therefore fluctuates periodically within the range of -1 to 1. The DC gain of this LPF is maximized when sin2Nθre in the numerator is at its maximum value of 1, as exemplified by the following equation.

[0068]

number

[0069] When sin2Nθre is 1, it represents the worst-case scenario, and the LPF characteristics in this case are those of a first-order LPF where both the estimated cosine and sine coefficients have a cutoff frequency of 0.5K [rad / sec]. This exhibits the same characteristics as the LPF that determines the estimated bandwidth of the single-frequency component detector shown in Equation 12.

[0070] Since the trigonometric function term to be disabled is a double-angle of the frequency component to be detected, if the cutoff frequency of the LPF shown in Equation 12 is lower than this double-angle frequency component, the said double-angle frequency component can be attenuated. In other words, if the feedback gain is set to a frequency below the double-angle frequency of the trigonometric function term, the unwanted double-angle component can be suppressed relative to the frequency component to be detected. This relationship is expressed by the following equation.

[0071]

number

[0072] Here, ω0 [rad / sec] is the frequency that we want to detect and suppress in the system. Furthermore, the following equation holds:

[0073]

number

[0074] If we want to suppress the N-fold angular component of the electrical angle, as in the example shown in this embodiment, the frequency ω0 becomes Nf [rad / sec], and the recommended upper limit of K becomes 2Nf [rad / sec]. This relationship is expressed by equations 21 to 23. Here, f is the electrical angular frequency.

[0075]

number

[0076]

number

[0077]

number

[0078] In actual operation, the three-phase motor 104 is used as a power source for a specific device. Therefore, it is expected that the rotational speed of the three-phase motor 104 will rise from 0 to a specific speed, and then return to 0 and end. Even in this case, the Nf component detector 102b can accurately detect the Nf component by following the rotational speed of the three-phase motor 104. Since Nf changes moment by moment depending on the rotational speed of the three-phase motor 104, it is desirable that the recommended upper limit of K be set so that the feedback gain K can be changed in accordance with the rotational speed of the three-phase motor 104.

[0079] Figure 5 shows an example of setting the upper limit of the feedback gain K. Figure 5 shows an example where the upper limit of the feedback gain K changes in synchronization with the speed pattern of the three-phase motor 104. The upper figure in Figure 5 shows the time-series waveform of the rotational speed of the three-phase motor 104. The lower figure in Figure 5 shows the time-series waveform of the upper limit of the feedback gain K. The horizontal axis of the lower figure coincides with the horizontal axis of the upper figure. Since Nf is proportional to the rotational speed of the three-phase motor 104 and the recommended upper limit of the feedback gain K is 2Nf [rad / sec], it is desirable that the upper limit of the feedback gain K be defined in a proportional relationship synchronized with the rotational speed, as shown in Figure 5. For example, if the upper limit of the rotational speed of the three-phase motor 104 is ωmax [rad / sec], then the upper limit of the feedback gain K will be 4ωmax [times].

[0080] Next, we will explain the setting value of the feedback gain K. Figure 6 shows an example of the estimated characteristics of the Nf component detector 102b. Specifically, Figure 6 is a frequency response diagram showing an example of the setting of the feedback gain K in the LPF shown in Equation 12.

[0081] As described above, the frequency characteristics shown in Equation 12 are those of a first-order LPF, with a cutoff frequency of 0.5K [rad / sec] and a DC gain of 0 [dB]. Based on Equation 22, the cutoff frequency of 0.5K [rad / sec] is set to a bandwidth lower than 2Nf [rad / sec]. Regarding the cutoff frequency, this setting is generally desirable. On the other hand, the lower limit of the cutoff frequency of 0.5K [rad / sec] is preferably set to the estimated disturbance frequency Nf [rad / sec], as shown in Figure 6. This is to further broaden the estimable bandwidth and enable fast response to the constantly changing frequency ω0 [rad / sec] to output accurate values ​​in real time. With this setting, the Nf component detector 102b can achieve both the characteristic of preventing unwanted information from being mixed in and the characteristic of fast and accurate estimation.

[0082] The above describes an example configuration of the Nf component open-loop gain infinity compensator 102 using Figures 3 and 4, and an example of parameter setting using Figures 5 and 6. Figure 7 shows another example of the Nf component open-loop gain infinity compensator 102. The configuration shown in Figure 7 can be realized by equivalently swapping the configurations shown in Figures 3 and 4.

[0083] The gain compensator 102 shown in the upper part of Figure 7 is simply a re-explanation of the gain compensator 102 shown in Figures 3 and 4. The single-frequency component detector 102b' shown in Figure 7 is identical to the portion enclosed by the dashed line indicated by reference numeral 102' in Figure 4. The lower part of Figure 7 can be obtained by equivalently swapping the upper part of Figure 7.

[0084] In the lower diagram of Figure 7, the gain compensator 102 comprises, for example, an adder 102c and a single-frequency component detector 102b', which was included in the Nf component detector 102b in Figure 4. The single-frequency component detector 102b' receives the electrical angle θre from the pulse generator 1 and the q-axis current error er_iq_in from the subtractor 101 as inputs. Similar to the example described above, the single-frequency component detector 102b' outputs the estimated Nf component nf_sens_q.

[0085] The adder 102c adds the q-axis current error er_iq_in and the estimated Nf component nf_sens_q, both of which are input signals, and outputs the result. In the lower part of Figure 7, the feedback loop within the gain compensator 102 shown in the upper part of Figure 7 appears to cancel each other out. However, the gain compensator 102 shown in the upper part of Figure 7 and the gain compensator 102 shown in the lower part of Figure 7 have identical dynamic characteristics and behave completely equivalently. Therefore, the gain compensator 102 shown in the lower part of Figure 7 may also be used.

[0086] Next, the Nf component remover 106 will be described. Figure 8 shows an example of the Nf component remover 106. The Nf component remover 106 includes, for example, an Nf component detector 106a and a subtractor 106b. The Nf component detector 106a receives the electrical angle θre from the pulse generator 1 and the q-axis current iq from the q-axis current detector 105 as input. The Nf component detector 106a estimates and extracts a frequency component (for example, the 2f component) that is a predetermined multiplier of the electrical angle θre from the q-axis current iq, and outputs the extracted signal, i.e., the estimated Nf component nf_sens_q shown in Figure 8.

[0087] The subtractor 106b receives the q-axis current iq from the q-axis current detector 105 and the estimated Nf component nf_sens_q from the Nf component detector 106a as inputs. The subtractor 106b subtracts the estimated Nf component nf_sens_q from the input signal q-axis current iq and outputs the result of the subtraction. That is, the corrected q-axis current iq_cor, which is the signal from which the Nf component of the q-axis current iq has been removed, is output from the subtractor 106b as the output of the Nf component remover 106.

[0088] Figure 9 shows an example of an Nf component detector 106a. In the example shown in Figure 9, the Nf component detector 106a comprises a magnification setter 106a1, a multiplier 106a2, a cosine arithmetic unit 106a3, a sine arithmetic unit 106a4, a subtractor 106a5, a multiplier 106a6, a magnified integrator 106a7, a multiplier 106a8, a multiplier 106a9, a magnified integrator 106a10, a multiplier 106a11, and an adder 106a12.

[0089] The Nf component detector 106a shown in Figure 9 is substantially identical to the Nf component detector 102b shown in Figure 4. For example, the magnification setter 106a1 has the same function as the magnification setter 102b1. The multiplier 106a2 has the same function as the multiplier 102b2. The functions indicated by symbols 106a3 to 106a12 are also the same as the corresponding functions of the Nf component detector 102b. The subtractor 106a5 subtracts the estimated Nf component nf_sens_q, which is the output of the Nf component detector 106a, from the q-axis current iq from the q-axis current detector 105, and outputs the result of the subtraction as the q-axis current error er_iq. Further detailed explanations of the Nf component detector 106a are omitted.

[0090] As described above, Figure 2 is a block diagram showing the details of the q-axis current control function to which the gain compensator 102 and Nf component remover 106 are applied, among the functions of the current controller 100 including the dq-axis current controller. By equipping the current controller 100 with the gain compensator 102 and Nf component remover 106, the Nf component of the open-loop gain characteristics of the q-axis current control is made infinite, and the 2f oscillation disturbance generated in the q-axis current detector 105 due to DCCT imbalance is removed. The various characteristics of the q-axis current control loop of the current controller 100 having such functions will be described in detail below.

[0091] Figure 10 illustrates various characteristics of the q-axis current control loop. Specifically, Figure 10(a) shows the closed-loop characteristics. Figure 10(b) shows the disturbance sensitivity characteristic C1 to vibration disturbances caused by magnetic circuit imbalance. Figure 10(c) shows the disturbance sensitivity characteristic C2 to vibration disturbances caused by current detection system imbalance.

[0092] As shown in Figure 10(a), the closed-loop characteristic is an LPF characteristic with a cutoff frequency of ωiq [rad / sec], which is the crossover frequency of the open-loop q-axis current control, and the DC gain is 0 [dB]. In addition, the closed-loop characteristic has a high peak gain in the 2f component, ω2f [rad / sec]. Therefore, if the input signal contains the ω2f component, the input signal will be amplified, and the output may saturate at the 2f component.

[0093] As shown in Figure 10(b), the disturbance sensitivity characteristic C1 of the q-axis current control loop is an HPF characteristic with a cutoff frequency of ωiq [rad / sec]. The current controller 100 includes a gain compensator 102 to suppress vibration disturbances caused by magnetic circuit imbalance, but its effect is canceled out by the Nf component remover 106. Therefore, the disturbance sensitivity characteristic C1 is the same as that of q-axis current control without the gain compensator 102 and Nf component remover 106. Consequently, although the gain compensator 102 is provided to suppress vibration disturbances caused by magnetic circuit imbalance, this suppression function by the gain compensator 102 is not performed. As a result, a 2f vibration component remains in the torque signal.

[0094] As shown in Figure 10(c), the disturbance sensitivity characteristic C2 of the q-axis current control loop is an LPF characteristic with a cutoff frequency of ωiq [rad / sec]. Since the sensitivity at frequency ω2f [rad / sec] is 0 [dB], there is no effect in suppressing disturbances. Therefore, the 2f oscillation component remains in the torque signal.

[0095] Thus, if both measures to counter vibration disturbances caused by magnetic circuit imbalance and measures to counter vibration disturbances caused by current detection system imbalance are introduced into the q-axis current control loop, their effects cancel each other out, resulting in no effective solution. The following describes a function to improve this problem.

[0096] As shown in Figure 10(a), the closed-loop characteristics of the q-axis current control have a high peak gain at frequency ω2f [rad / sec]. Therefore, if no other measures are taken, the q-axis current may become oscillating. This problem can be addressed by stabilizing the outer loop of the q-axis current control loop. That is, a control loop is provided outside the q-axis current control loop. The crossover frequency of the open loop of this control loop should be set to a value higher than frequency ω2f [rad / sec] and lower than the crossover frequency ωiq [rad / sec] of the open loop of the inner loop of the q-axis current control. If this frequency is ωiqest [rad / sec], the following equation holds.

[0097]

number

[0098] Furthermore, if the crossover frequency of the open-loop speed control loop is ωv [rad / sec], then the following equation holds:

[0099]

number

[0100] Thus, this control loop is an intermediate loop between the outer loop, the speed control loop, and the inner loop, the q-axis current control loop. Hereafter, this control loop will also be referred to as the middle loop. This middle loop is labeled as the estimated q-axis current control loop in Figure 1. The middle loop, or estimated q-axis current control loop, will be explained in detail below.

[0101] Figure 11 shows an example of the q-axis current estimator 5. Figure 11(a) shows a functional block diagram of the q-axis current estimator 5. Figure 11(b) shows the transfer function of each block shown in Figure 11(a). Figure 11(c) shows the transfer function shown in Figure 11(b) equivalently replaced into a single transfer function. In Figure 11, J is the inertia associated with the three-phase motor 104, Kτ is the force constant of the three-phase motor 104, ωs is the cutoff frequency of the band-limited LPF, and ζs is the attenuation coefficient of the band-limited LPF.

[0102] The function of the q-axis current estimator 5 is to estimate the estimated q-axis current iqest from the speed of the three-phase motor 104, i.e., the angular velocity ωre from the time differentiator 2. As shown in Figure 11(a), the q-axis current estimator 5 comprises a first functional unit 5a having the inverse characteristics of the transfer function from the q-axis current of the three-phase motor 104 to the motor speed, and a second functional unit consisting of a band-limiting LPF 5b. The transfer function from the q-axis current iq [A] of the three-phase motor 104 to the motor speed ωre [rad / sec] is expressed by the following equation.

[0103]

number

[0104] By rearranging equation 26, we can obtain the following equation.

[0105]

number

[0106] Equation 27 represents the q-axis current iq of a three-phase motor 104 when the motor speed ωre is input, and Js / Kτ represents the inverse characteristic. This inverse characteristic is a differential characteristic. Therefore, the right-hand side of Equation 27 will be an accurate value if the motor speed ωre is an ideal signal without noise. However, the motor speed ωre is a sensor signal and is superimposed with noise. For this reason, with respect to the right-hand side of Equation 27, the high-frequency noise is amplified due to the effect of the derivative of the inverse characteristic, and there is a risk that the signal will become saturated with noise.

[0107] To counter such noise, a band-limited LPF 5b is provided. The band-limited LPF 5b can be configured in any way, for example, a first-order LPF. Below, we will describe an example in which a second-order LPF with excellent high-frequency noise immunity is adopted as the band-limited LPF 5b, taking into consideration the S / N ratio of the motor speed ωre. The transfer function G of the band-limited LPF 5b is expressed by the following equation.

[0108]

number

[0109] The band-limited LPF5b is a typical second-order LPF with a DC gain of 0 [dB], a cutoff frequency of ωs [rad / sec], and an attenuation coefficient of ζs. The attenuation coefficient ζs is preferably set to 1, which represents a critical damping state without a peak at the cutoff frequency. However, the attenuation coefficient ζs can be any value close to 1. A suitable example where ζs = 1 will be described below.

[0110] The cutoff frequency ωs [rad / sec] is a parameter used to remove high-frequency noise contained in the motor speed ωre. Setting the cutoff frequency ωs to a low value increases the noise reduction effect, but it reduces the phase margin of the middle loop and impairs stability. To ensure this stability, it is preferable to set the cutoff frequency ωs higher than the open-loop crossover frequency ωiqest of the middle loop, as shown in the following equation.

[0111]

number

[0112] The following section describes an example where ωs = 2 × ωipest.

[0113] Figure 12 is a diagram illustrating the various characteristics of the middle loop. Figure 12 shows an example where the crossover frequency ωiqest is set to satisfy Equation 24. Specifically, Figure 12(a) shows the closed-loop characteristics. Figure 12(b) shows the disturbance sensitivity characteristic C1 to vibration disturbances caused by magnetic circuit imbalance. Figure 12(c) shows the disturbance sensitivity characteristic C2 to vibration disturbances caused by current detection system imbalance.

[0114] As shown in Figure 12(a), the closed-loop characteristic is an LPF characteristic with the middle loop's open-loop crossover frequency ωiqest [rad / sec] as the cutoff frequency, and the DC gain is 0 [dB]. In the closed-loop characteristic, there is a peak gain of several dB at the cutoff frequency. Also, the peak gain at frequency ω2f [rad / sec] shown in Figure 10(a) has disappeared, and there is a small gain fluctuation of several dB. Thus, the closed-loop characteristic shown in Figure 12(a) is an improved and more stable characteristic compared to the closed-loop characteristic shown in Figure 10(a).

[0115] As shown in Figure 12(b), the disturbance sensitivity characteristic C1 of the middle loop is an HPF characteristic with a cutoff frequency of ωiq [rad / sec]. The sensitivity at frequency ω2f [rad / sec] is significantly reduced in a notch filter-like manner only around this frequency. Therefore, the disturbance sensitivity characteristic C1 shown in Figure 12(b) can suppress 2f oscillation disturbances caused by magnetic circuit imbalance with significantly lower sensitivity compared to the disturbance sensitivity characteristic C1 shown in Figure 10(b).

[0116] As shown in Figure 12(c), the disturbance sensitivity characteristic C2 of the middle loop has a cutoff frequency of This is an HPF characteristic with ωiqest [rad / sec]. The sensitivity at frequency ω2r [rad / sec] is significantly reduced in a notch filter-like manner only around this frequency. Therefore, the disturbance sensitivity characteristic C2 shown in Figure 12(c) can suppress 2f oscillation disturbances caused by current detection system imbalance with significantly lower sensitivity compared to the disturbance sensitivity characteristic C2 shown in Figure 10(c).

[0117] Figure 13 illustrates the various characteristics of the speed control loop when the various characteristics of the middle loop are set as shown in Figure 12. Specifically, Figure 13(a) shows the closed-loop characteristics. Figure 13(b) shows the disturbance sensitivity characteristic C1 to vibration disturbances caused by magnetic circuit imbalance. Figure 13(c) shows the disturbance sensitivity characteristic C2 to vibration disturbances caused by current detection system imbalance. In Figure 13, the characteristics of the speed control loop in this disclosure are shown by solid lines, and the characteristics of a conventional speed control loop without a middle loop are shown by dashed lines.

[0118] As shown by the solid line in Figure 13(a), the closed-loop characteristic is an LPF characteristic with a cutoff frequency of the open-loop crossover frequency ωv [rad / sec] of the speed control loop, and the DC gain is 0 [dB]. Regarding the closed-loop characteristic, it is the same as the closed-loop characteristic of a conventional speed control loop up to the open-loop crossover frequency ωiqest [rad / sec] of the middle loop. On the other hand, at frequencies exceeding the crossover frequency ωiqest [rad / sec], the value shown by the solid line in Figure 13(a) becomes smaller than the value shown by the dashed line. However, since the gain value itself is small in this frequency band, it has almost no effect on the speed response to the speed command value.

[0119] Therefore, the behavior of speed control when a middle loop is introduced is substantially the same as the behavior of speed control when a middle loop is not introduced. In other words, even if a middle loop is introduced, the response of the outer loop, the speed control loop, to the speed command value is not substantially affected. Since the behavior in response to the speed command value does not change between the speed control loop shown in Figure 1 and the conventional speed control loop, there is no need to change the specifications of the product using the three-phase motor 104 from the conventional specifications.

[0120] As shown by the dashed line in Figure 13(b), in the disturbance sensitivity characteristic C1 of a conventional speed control loop, the gain is almost constant between frequencies ωv and ωiq [rad / sec]. Furthermore, at frequencies lower than ωv [rad / sec], the gain decreases as the frequency decreases, and at frequencies higher than ωiq [rad / sec], the gain decreases as the frequency increases.

[0121] On the other hand, in the disturbance sensitivity characteristic C1 shown by the solid line in Figure 13(b), the gain basically decreases as the frequency decreases below the frequency ωiqest [rad / sec]. That is, the sensitivity is further reduced in that frequency band. Also, the sensitivity at frequency ω2f [rad / sec] is significantly reduced only around this frequency, in a notch filter-like manner. Therefore, the disturbance sensitivity characteristic C1 shown by the solid line in Figure 13(b) can significantly reduce and suppress 2f oscillation disturbances caused by magnetic circuit imbalance compared to the disturbance sensitivity characteristic C1 shown by the dashed line.

[0122] As shown by the dashed line in Figure 13(c), the disturbance sensitivity characteristic C2 in a conventional speed control loop is an LPF characteristic with a cutoff frequency of frequency ωv [rad / sec].

[0123] On the other hand, in the disturbance sensitivity characteristic C2 shown by the solid line in Figure 13(c), the gain is almost constant between frequencies ωv and ωiqest [rad / sec]. Furthermore, at frequencies lower than ωv [rad / sec], the gain decreases as the frequency decreases. At frequencies higher than ωiqest [rad / sec], the gain decreases as the frequency increases. In the disturbance sensitivity characteristic C2 shown by the solid line in Figure 13(c), the disturbance sensitivity characteristic C2 is almost the same as that of a conventional speed control loop in the frequency band higher than ωiqest [rad / sec]. Moreover, compared to the disturbance sensitivity characteristic C2 shown by the dashed line, this disturbance sensitivity characteristic C2 is more sensitive in the frequency band lower than ωiqest [rad / sec].

[0124] Furthermore, in the disturbance sensitivity characteristic C2 shown by the solid line in Figure 13(c), the sensitivity at frequency ω2f [rad / sec] is significantly reduced in a notch filter-like manner only around this frequency. Therefore, the disturbance sensitivity characteristic C2 shown by the solid line in Figure 13(c) can suppress 2f oscillation disturbances caused by current detection system imbalances with significantly lower sensitivity compared to the disturbance sensitivity characteristic C2 shown by the dashed line.

[0125] Next, referring to Figures 14 to 16, we will explain an example in which the configuration shown in Figure 1, etc., is applied to an elevator. In this example, the three-phase motor 104 shown in Figure 1 is the hoisting motor of the elevator. The hoisting motor is the motor that drives the elevator car in which passengers ride. Figures 14 to 16 show examples of the torque signal waveform of the hoisting motor when a 2f disturbance signal is applied.

[0126] Specifically, in Figures 14 to 16, (a) shows an example of a signal applied as a drive current disturbance. The signal shown in (a) is an example of a 2f vibration disturbance caused by magnetic circuit imbalance. (b) shows an example of a signal applied as a detection current disturbance. The signal shown in (b) is an example of a 2f vibration disturbance caused by current detection system imbalance. (c) shows an example of the torque signal waveform of the hoisting motor when conventional speed control is performed. (d) shows an example of the torque signal waveform of the hoisting motor when the speed control shown in Figures 1 and 2 is performed. The horizontal axis in each figure is the time axis.

[0127] As shown in Figures 14(a) and (b) for signal waveforms, Figures 14(c) and (d) show the torque signal waveforms when there is magnetic circuit imbalance but no current detection system imbalance. The torque signal shown in Figure 14(c) is superimposed with 2f vibration. On the other hand, the torque signal shown in Figure 14(d) is not superimposed with 2f vibration, and its waveform is stable and exhibits a clean shape without vibration.

[0128] As shown in Figures 15(a) and (b) for signal waveforms, Figures 15(c) and (d) show the torque signal waveforms when there is no magnetic circuit imbalance but there is a current detection system imbalance. The torque signal shown in Figure 15(c) has 2f vibration superimposed on it. On the other hand, the torque signal shown in Figure 15(d) does not have 2f vibration superimposed on it, and its waveform is stable and has a clean shape without vibration.

[0129] As shown in Figures 16(a) and (b) for signal waveforms, Figures 16(c) and (d) show the torque signal waveforms when both magnetic circuit imbalance and current detection system imbalance are present. The torque signal shown in Figure 16(c) is superimposed with 2f vibration. On the other hand, the torque signal shown in Figure 16(d) is not superimposed with 2f vibration, and its waveform exhibits a stable and clean shape without vibration.

[0130] Thus, in the example shown in this embodiment, 2f oscillations appearing in the torque signal can be suppressed when there is either a magnetic circuit imbalance or a current detection system imbalance, or when both are present.

[0131] Furthermore, in the example shown in this embodiment, 2f vibrations appearing in the torque signal can be suppressed by tracking in real time the changes in amplitude and phase of 2f vibration disturbances caused by magnetic circuit imbalance and current detection system imbalance. For this reason, in the example shown in this embodiment, there is no need to limit the conditions for application or perform specific learning operations. In the example shown in this embodiment, 2f vibrations caused by magnetic circuit imbalance and current detection system imbalance can be effectively suppressed without performing learning operations, regardless of changes in the jerk and acceleration of the three-phase motor 104.

[0132] Figure 17 shows an example of the hardware resources of the control device 200. The control device 200 includes a processing circuit 10 as a hardware resource, which includes a processor 11 and memory 12. The processing circuit 10 may include multiple processors 11. The processing circuit 10 may also include multiple memory 12.

[0133] In this embodiment, the devices indicated by reference numerals 100 to 106 represent functions of the control device 200. The functions of the devices indicated by reference numerals 100 to 106 can be realized by software, firmware, or a combination of software and firmware, which are written as programs. The programs are stored in memory 12. The control device 200 realizes the functions of the parts indicated by reference numerals 100 to 106 by executing the programs stored in memory 12 using the processor 11 (computer).

[0134] The processor 11 is also called a CPU (Central Processing Unit), central processing unit, processing unit, arithmetic unit, microprocessor, microcomputer, or DSP. The memory 42 may include semiconductor memory, magnetic disks, flexible disks, optical disks, compact disks, minidiscs, or DVDs. Possible semiconductor memories include RAM, ROM, flash memory, EPROM, and EEPROM.

[0135] Figure 18 shows another example of the hardware resources of the control device 200. In the example shown in Figure 18, the control device 200 includes a processor 11, memory 12, and a processing circuit 10 including dedicated hardware 13. Figure 18 shows an example in which some of the functions of the control device 200 are realized by the dedicated hardware 13. All of the functions of the control device 200 may also be realized by the dedicated hardware 13. The dedicated hardware 13 can be a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC, an FPGA, or a combination thereof.

[0136] Examples of aspects that may be included in this disclosure are listed below as an addendum.

[0137] [Note 1] A speed control loop for controlling the speed of a three-phase motor to track a speed command value using negative feedback control, An estimated q-axis current control loop for causing the estimated q-axis current based on the speed of the three-phase motor to follow the first current command value, which is the output of the speed control loop, using negative feedback control, A q-axis current control loop is provided to make the q-axis current, based on the detected current value of the three-phase motor, follow the second current command value, which is the output of the estimated q-axis current control loop, using negative feedback control. Equipped with, The q-axis current control loop is, An Nf component remover that removes the 2f component from the q-axis current based on the detected current value of the three-phase motor, A first subtractor subtracts the corrected q-axis current from which the 2f component has been removed by the Nf component remover from the second current command value. Based on the output of the first subtractor, a gain compensator is provided that maximizes the gain of the 2f component in the open-loop characteristics to infinity. A control device equipped with the following features. [Note 2] The aforementioned gain compensator is First Nf component detector, A first adder that adds the output of the first Nf component detector and the output of the first subtractor, Equipped with, The control device described in Appendix 1, wherein the first Nf component detector estimates and extracts the 2f component from the output of the first adder based on the electrical angle of the three-phase motor. [Note 3] The first Nf component detector is A first multiplier that multiplies the electrical angle of the three-phase motor by a preset multiplier, A first cosine arithmetic unit that calculates the cosine value of the multiplication value by the first multiplier, A first sine arithmetic unit that calculates the sine value of the multiplication by the first multiplier, A second subtractor subtracts the output of the first Nf component detector from the output of the first adder, A second multiplier that multiplies the output of the second subtractor by the output of the first cosine arithmetic unit, A third multiplier that multiplies the output of the second subtractor by the output of the first sine arithmetic unit, A first multiplier integrator integrates the output of the second multiplier with a preset amplification factor, A second multiplier integrator integrates the output of the third multiplier with a preset amplification factor, A fourth multiplier that multiplies the output of the first multiplier integrator and the output of the first cosine arithmetic unit, A fifth multiplier that multiplies the output of the second multiplier integrator by the output of the first sine arithmetic unit, A second adder that adds the output of the fourth multiplier and the output of the fifth multiplier, and uses the result of this addition as the output of the first Nf component detector, The control device described in Appendix 2, which is equipped with the following: [Note 4] The aforementioned Nf component remover is A second Nf component detector estimates and extracts the 2f component from the q-axis current based on the detected current value of the three-phase motor, based on the electrical angle of the three-phase motor. A third subtractor subtracts the output of the second Nf component detector from the q-axis current based on the detected current of the three-phase motor, A control device as described in any one of the appendices 1 to 3, which is equipped with the following: [Note 5] The second Nf component detector is, A sixth multiplier that multiplies the electrical angle of the three-phase motor by a preset multiplier, A second cosine arithmetic unit that calculates the cosine value of the multiplication value by the sixth multiplier, A second sine arithmetic unit that calculates the sine value of the multiplication by the sixth multiplier, A fourth subtractor that subtracts the output of the second Nf component detector from the q-axis current based on the detected current of the three-phase motor, A seventh multiplier that multiplies the output of the fourth subtractor by the output of the second cosine arithmetic unit, An eighth multiplier that multiplies the output of the fourth subtractor by the output of the second sine arithmetic unit, A third multiplier integrator integrates the output of the seventh multiplier with a preset amplification factor, A fourth multiplier integrator integrates the output of the eighth multiplier with a preset amplification factor, A ninth multiplier that multiplies the output of the third multiplier integrator by the output of the second cosine arithmetic unit, A tenth multiplier that multiplies the output of the fourth multiplier integrator by the output of the second sine arithmetic unit, A third adder adds the output of the ninth multiplier and the output of the tenth multiplier, and uses the result of this addition as the output of the second Nf component detector. A control device as described in Appendix 4, equipped with the following features. [Note 6] The estimated q-axis current control loop includes a q-axis current estimator that estimates the estimated q-axis current based on the speed detection value of the three-phase motor. The q-axis current estimator is, A functional unit having the inverse characteristics of the transfer function from the q-axis current to the speed of the three-phase motor, Bandwidth limiting LPF and A control device as described in any one of Appendix 1 to Appendix 5, comprising the above. [Explanation of Symbols]

[0138] 1 Pulse generator, 2 Time differentiator, 3 Subtractor, 4 Speed ​​control compensator, 5 Q-axis current estimator, 6 Subtractor, 7 Estimated Q-axis current control compensator, 10 Processing circuit, 11 Processor, 12 Memory, 13 Dedicated hardware, 100 Current controller, 101 Subtractor, 102 Nf component open-loop gain infinity compensator (gain compensator), 102b' Single frequency component detector, 103 Current control compensator, 104 Three-phase motor, 105 Q-axis current detector, 106 Nf component remover, 102a Adder, 102b Nf component detector, 102b1 Multiplier setter, 102b2 Multiplier, 102b3 Cosine arithmetic unit, 102b4 Sine arithmetic unit, 102b5 Subtractor 102b6 Multiplier, 102b7 Multiplier-integrator, 102b8 Multiplier, 102b9 Multiplier, 102b10 Multiplier-integrator, 102b11 Multiplier, 102b12 Adder, 102c Adder, 106a Nf component detector, 106b Subtractor, 106a1 Multiplier setter, 106a2 Multiplier, 106a3 Cosine arithmetic unit, 106a4 Sine arithmetic unit, 106a5 Subtractor, 106a6 Multiplier, 106a7 Multiplier-integrator, 106a8 Multiplier, 106a9 Multiplier, 106a10 Multiplier-integrator, 106a11 Multiplier, 106a12 Adder, 200 Control device

Claims

1. A speed control loop for controlling the speed of a three-phase motor to track a speed command value using negative feedback control, An estimated q-axis current control loop for causing the estimated q-axis current based on the speed of the three-phase motor to follow the first current command value, which is the output of the speed control loop, using negative feedback control, A q-axis current control loop is provided to make the q-axis current, based on the detected current value of the three-phase motor, follow the second current command value, which is the output of the estimated q-axis current control loop, using negative feedback control. Equipped with, The q-axis current control loop is, An Nf component remover that estimates and extracts a 2f component based on the q-axis current, which is determined from the current of the three-phase motor, and the electrical angle of the three-phase motor, and removes the 2f component from the q-axis current. A first subtractor subtracts the corrected q-axis current from which the 2f component has been removed by the Nf component remover from the second current command value. A gain compensator that estimates and extracts the 2f component based on the output of the first subtractor and the electrical angle of the three-phase motor, and maximizes the gain of the 2f component in the open-loop characteristics to infinity, Equipped with, 2f is a control device that represents a frequency twice the electrical angle of the three-phase motor.

2. The aforementioned gain compensator is First Nf component detector, A first adder that adds the output of the first Nf component detector and the output of the first subtractor, Equipped with, The control device according to claim 1, wherein the first Nf component detector estimates and extracts the 2f component from the output of the first adder based on the electrical angle of the three-phase motor.

3. The first Nf component detector is A first multiplier that multiplies the electrical angle of the three-phase motor by a preset magnification factor, A first cosine arithmetic unit that calculates the cosine value of the multiplication value by the first multiplier, A first sine arithmetic unit that calculates the sine value of the multiplication by the first multiplier, A second subtractor subtracts the output of the first Nf component detector from the output of the first adder, A second multiplier that multiplies the output of the second subtractor by the output of the first cosine arithmetic unit, A third multiplier that multiplies the output of the second subtractor by the output of the first sine arithmetic unit, A first multiplier integrator integrates the output of the second multiplier with a preset amplification factor, A second multiplier integrator integrates the output of the third multiplier with a preset amplification factor, A fourth multiplier that multiplies the output of the first multiplier integrator and the output of the first cosine arithmetic unit, A fifth multiplier that multiplies the output of the second multiplier integrator and the output of the first sine arithmetic unit, A second adder adds the output of the fourth multiplier and the output of the fifth multiplier, and uses the result of this addition as the output of the first Nf component detector. The control device according to claim 2, comprising:

4. The aforementioned Nf component remover is A second Nf component detector estimates and extracts a 2f component from the q-axis current based on the detected current value of the three-phase motor, based on the electrical angle of the three-phase motor. A third subtractor subtracts the output of the second Nf component detector from the q-axis current based on the detected current of the three-phase motor, A control device according to any one of claims 1 to 3, comprising

5. The second Nf component detector is A sixth multiplier that multiplies the electrical angle of the three-phase motor by a preset magnification factor, A second cosine arithmetic unit calculates the cosine value of the multiplication by the sixth multiplier, A second sine arithmetic unit that calculates the sine value of the multiplication by the sixth multiplier, A fourth subtractor subtracts the output of the second Nf component detector from the q-axis current based on the detected current of the three-phase motor, A seventh multiplier that multiplies the output of the fourth subtractor by the output of the second cosine arithmetic unit, An eighth multiplier that multiplies the output of the fourth subtractor by the output of the second sine arithmetic unit, A third multiplier integrator integrates the output of the seventh multiplier with a preset amplification factor, A fourth multiplier integrator that integrates the output of the eighth multiplier with a preset amplification factor, A ninth multiplier that multiplies the output of the third multiplier integrator by the output of the second cosine arithmetic unit, A tenth multiplier that multiplies the output of the fourth multiplier integrator by the output of the second sine arithmetic unit, A third adder adds the output of the ninth multiplier and the output of the tenth multiplier, and the result of this addition is used as the output of the second Nf component detector. The control device according to claim 4, comprising:

6. The estimated q-axis current control loop includes a q-axis current estimator that estimates the estimated q-axis current based on the speed detection value of the three-phase motor. The q-axis current estimator is, A functional unit having the inverse characteristics of the transfer function from the q-axis current to the speed of the three-phase motor, Bandwidth limiting LPF and A control device according to any one of claims 1 to 3, comprising

Images