Drive control device, motor system, and drive control method

Abstract

The present invention comprises a position estimation means (6) having: a speed electromotive force calculation unit (61) that calculates a speed electromotive force (Es) from the current (Is) flowing through a permanent magnet synchronous motor, a voltage command (V* s) applied thereto, and the motor constant; a polyphase-to-two-phase converter (62) that converts an induced voltage (Es) into two-phase AC; a position calculator (63) that calculates a first estimated position (θ-hat) by performing an arctangent calculation on the induced voltage (Eαβ) converted into two-phase AC; and a pulsation remover (64) that calculates a second estimated position (θF-hat) by removing a periodic pulsation from the first estimated position (θ-hat), wherein the pulsation remover (64) is configured to remove the periodic pulsation so that the steady-state average error between the first estimated position (θ-hat) and the second estimated position (θF-hat) is zero.

Description

[Technical Field]

[0001] This application relates to a drive control device, a motor system, and a drive control method. [Background technology]

[0002] Linear synchronous motors (LSMs) have long been used in industries such as railways and factory automation (FA). Recently, in the FA industry, LSMs with the primary stator windings (coils) located on the ground and the secondary magnets located on the movable part have been attracting attention.

[0003] To control the thrust of an LSM, it is necessary to know the magnetic pole position of the movable element, just as with rotating machines. However, with LSMs, there is a challenge in that the cost of position sensors increases as the distance the device can travel increases. In response to this, a position sensorless control method has been disclosed for LSMs that detects the position using the induced voltage generated in a coil, also called the speed electromotive force, similar to that used in rotating machines (see, for example, Patent Document 1). [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2002-223587 (paragraph 0022, Figure 4, paragraphs 0033-0034, Figure 7) [Overview of the project] [Problems that the invention aims to solve]

[0005] However, recent LSMs have seen miniaturization of the movable elements, and as a result of this miniaturization, the effects of end effects have become relatively larger, leading to an increase in motors with large induced voltage distortions. When attempting to drive such motors with position sensorless control, speed estimation errors and position estimation errors occur due to the distortion of the induced voltage. These errors make it difficult to achieve highly responsive control.

[0006] Furthermore, in rotating machinery, permanent magnet synchronous motors with large induced voltage distortions similarly result in speed estimation errors and position estimation errors, leading to a decrease in the control performance of position sensorless control.

[0007] This invention discloses a technology for solving the above-mentioned problems, and aims to obtain a motor system that can be driven accurately without a position sensor by suppressing errors caused by distortion of induced voltage due to end effects in a permanent magnet synchronous motor. [Means for solving the problem]

[0008] The drive control device disclosed herein comprises a power supply device for controlling the current of a permanent magnet synchronous motor; an induced voltage calculation unit for calculating the induced voltage of the permanent magnet synchronous motor from the current flowing through the permanent magnet synchronous motor, the applied voltage, and the motor constants of the permanent magnet synchronous motor; a coordinate transformation means for converting the induced voltage into a two-phase alternating current; a position calculation unit for calculating a first estimated position of the magnetic poles of the permanent magnet synchronous motor by performing an inverse tangent calculation on the induced voltage converted to a two-phase alternating current; a pulsation remover for calculating a second estimated position by removing periodic pulsations caused by distortion of the induced voltage from the first estimated position; and a control calculation means for controlling the operation of the power supply device based on the second estimated position, wherein the pulsation remover removes the periodic pulsations such that the steady-state average error between the first estimated position and the second estimated position becomes zero.

[0009] The drive control method disclosed herein includes: an induced voltage calculation step of calculating an induced voltage of a permanent magnet synchronous motor from the current flowing through the permanent magnet synchronous motor, the applied voltage, and the motor constants of the permanent magnet synchronous motor; a coordinate transformation step of converting the induced voltage into a two-phase alternating current; a position calculation step of calculating a first estimated position of the magnetic poles of the permanent magnet synchronous motor by performing an inverse tangent calculation on the induced voltage converted to a two-phase alternating current; a pulsation removal step of calculating a second estimated position by removing periodic pulsations caused by distortion of the induced voltage from the first estimated position; and a control step of controlling the current flowing through the permanent magnet synchronous motor based on the second estimated position, wherein the pulsation removal step is characterized in that the periodic pulsations are removed such that the steady-state average error between the first estimated position and the second estimated position becomes zero. [Effects of the Invention]

[0010] According to the drive control device or drive control method disclosed herein, errors caused by distortion of induced voltage due to end effects in a permanent magnet synchronous motor can be suppressed, thereby enabling a motor system that operates without a position sensor. [Brief explanation of the drawing]

[0011] [Figure 1] This is a block diagram illustrating the configuration of a drive control device and a linear motor system using the drive control device according to Embodiment 1. [Figure 2] This is a block diagram illustrating the configuration of the coil current control means in the drive control device according to Embodiment 1. [Figure 3] This is a waveform diagram showing the speed electromotive force generated in each coil in the linear motor system according to Embodiment 1. [Figure 4] This is a waveform diagram obtained by performing a static two-phase conversion on the speed electromotive force generated in each coil in the linear motor system according to Embodiment 1. [Figure 5] This waveform diagram shows the result of calculating the arctangent of the speed electromotive force after static two-phase conversion in the linear motor system according to Embodiment 1. [Figure 6] This waveform diagram illustrates the position estimation error when performing an arctangent calculation due to the influence of end effects in the motor system according to Embodiment 1. [Figure 7] This is a block diagram illustrating the configuration of the movable element position estimation means in the drive control device according to Embodiment 1. [Figure 8] This is a block diagram illustrating the configuration of the position calculator for the movable element position estimation means in the drive control device according to Embodiment 1. [Figure 9] Figures 9A and 9B show the waveforms of the estimated position output when the comparative example filters are applied and the waveforms of the estimated position output when the filter of the present invention is applied, respectively, when the movable element is traveling at a constant speed. [Figure 10] This is a flowchart illustrating the operation of the drive control device and the drive control method according to Embodiment 1. [Figure 11] This is a block diagram showing an example of the hardware configuration of the part of the drive control device according to Embodiment 1 that performs calculation processing. [Figure 12] This is a block diagram illustrating the configuration of the movable element position estimation means in the drive control device according to Embodiment 2. [Figure 13] This is a block diagram showing an example in which the pulsation remover of the movable element position estimation means is configured as a type 3 in the drive control device according to Embodiment 2. [Figure 14] This is a block diagram illustrating the configuration of a motor system using a rotating machine with a drive control device and drive control method according to Embodiment 3. [Figure 15] Figures 15A to 15C show the waveforms of the regenerative electromotive force, the arctangent of the regenerative electromotive force, and the position estimation error, respectively, with the rotor position as the horizontal axis, in a three-phase rotating machine. [Figure 16] This is a block diagram illustrating the configuration of the rotor position estimation means in the drive control device according to Embodiment 3. [Modes for carrying out the invention]

[0012] Embodiment 1. Figures 1 to 10 illustrate the configuration and operation of a drive control device and a linear motor system using the drive control device according to Embodiment 1, as well as a drive control method. Figure 1 is a block diagram illustrating the configuration of the drive control device and the linear motor system including the linear motor that it controls, and Figure 2 is a block diagram illustrating the configuration of the coil current control means portion of the drive control device.

[0013] Figure 3 is a waveform diagram showing the speed electromotive force generated in each of the three adjacent coils of the stator when the movable element of the linear motor is driven at a constant speed in a linear motor system, with the speed electromotive force on the vertical axis and the movable element position on the horizontal axis. Figure 4 is a waveform diagram showing the speed electromotive force after static two-phase conversion processing of the speed electromotive force shown in Figure 3, and Figure 5 is a waveform diagram showing the result of arctangent calculation of the speed electromotive force after static two-phase conversion shown in Figure 4.

[0014] Figure 6 is a waveform diagram illustrating the position estimation error when performing an arctangent calculation due to the influence of end effects in a linear motor system, Figure 7 is a block diagram illustrating the configuration of the movable element position estimation means of the drive control device, and Figure 8 is a block diagram illustrating the configuration of the position calculator of the movable element position estimation means.

[0015] Furthermore, Figures 9A and 9B show the waveforms of the estimated position with time on the horizontal axis and the movable part position on the vertical axis when a first-order low-pass filter is applied as a filter to the position estimate value, which is calculated with an error due to the distortion of the induced voltage when the movable part is moving at a constant speed, and the waveforms of the estimated position with the same axis configuration when a filter is applied such that the steady-state error to the ramp response of the present invention becomes zero. And Figure 10 is a flowchart for explaining the operation of the drive control device and the drive control method.

[0016] The drive control device, drive control method, and motor system according to the embodiment of the present application will be described in detail below with appropriate reference to the drawings. In each figure, the same reference numerals indicate the same or corresponding parts.

[0017] The motor system according to Embodiment 1, as shown in Figure 1, is a linear motor system consisting of a non-three-phase LSM linear motor 100 and a drive control device 8 that drives and controls the linear motor 100. The system is characterized by estimating the position or speed of the movable element 1 of the linear motor 100 based on the induced voltage generated in the coil 3 of the stator 2, while removing the distortion of the induced voltage due to end effects. Before describing the characteristic features, the basic configuration and operation of the drive control of the non-three-phase LSM will be explained.

[0018] The linear motor 100 is a permanent magnet type synchronous motor in which three permanent magnets 11a to 11c (referred to as "permanent magnet 11" if not individually distinguished) are arranged in a straight line on the movable element 1 and magnetically coupled by a back yoke 10. On the other hand, the stator 2 consists of multiple coils 3a to 3i (referred to as "coil 3" if not individually distinguished) arranged to form the track for the movable element 1. Due to space limitations, only nine coils 3 are shown in the diagram, but in reality, hundreds or more coils 3 may be arranged. These coils 3 are magnetically coupled by a core back 4. Also, although only one movable element 1 is depicted in the diagram, in an actual linear motor 100, multiple movable elements 1 may exist on the track.

[0019] In the linear motor 100, the movable element 1 can be moved by controlling the current flowing through each coil 3. To control the current flowing through each coil 3, the drive control device 8 is provided with coil current control means 5a to 5i (referred to as "coil current control means 5" when not individually distinguished) corresponding to each of the multiple coils 3. In order for the multiple movable elements 1 to move in close proximity, it is necessary to individually control the current of each of the multiple coils 3, so each of the multiple coil current control means 5 individually controls the current of each coil 3 according to the current command determined by the control calculation means 7.

[0020] Each of the multiple coil current control means 5 can be configured, for example, as shown in Figure 2. In the figure, coil 3 is represented by an inductor symbol. A variable voltage source 52 is connected to coil 3. The variable voltage source 52 can be any circuit as long as it can independently control the current of each coil 3 by outputting an arbitrary voltage. The variable voltage source 52 can be configured as, for example, a single-phase inverter circuit. The current detection means 51 detects the current flowing through coil 3. The current control means 53 controls the current so that the current command matches the current flowing through the coil, and determines the voltage command. As a calculation method for current control, for example, PID control (proportional-integral-derivative control) is well known. The variable voltage source 52 operates according to the voltage command determined by the current control means 53, and the desired current flows through coil 3.

[0021] As explained in the background technology section, recent LSMs are characterized by "miniaturization of the movable elements" and "narrowing of the pitch between movable elements." To achieve this, a non-three-phase motor structure that cannot be driven by 120-degree energization is used. By removing the premise of "driving with 120-degree energization," the degree of freedom of the magnetic structure increases, making it easier to increase the thrust density.

[0022] Although this is merely one example, the linear motor 100 shown in Figure 1 is a type of LSM that is driven by a 135° current flow. The ratio of the lateral length of one movable element to the lateral length of one coil is 4:1, and three permanent magnets 11a to 11c are attached to the movable element 1. We will refer to this structure as a "3-pole 4-slot structure". If the lateral length of one magnetic pole is 180 degrees, then the electrical angle θ per coil in the 3-pole 4-slot structure is c As shown in equation (1), this is 135 degrees.

[0023]

number

[0024] By changing the combination of the length of the movable element 1 and the arrangement of the permanent magnets 11, various structures can be considered, such as a "5-pole 6-slot structure" and a "6-pole 8-slot structure". The phase difference of the current in each coil 3 changes depending on the number of pole slots, but the electrical angle θ of one coil remains the same regardless of the structure. c This can be calculated using a similar formula. In general rotating machines, from the perspective of cost performance, θ c Three-phase motors with a 120-degree rotational frequency (=120 degrees) are often used. However, in recent LSMs, flexibility of operation and added value such as cableless movable parts are becoming more important than cost performance, so non-three-phase motor structures that cannot be driven with 120-degree rotational frequency are being adopted.

[0025] As shown in Figure 1, the drive control device 8 can individually control the current of each coil 3, so thrust can be generated regardless of the combination of pole slots. However, to control the thrust of the LSM, it is necessary to know the magnetic pole position of the movable element, just as in the case of a rotating machine. Therefore, the magnetic pole position of the movable element 1 is measured by the movable element position estimation means 6. At the same time, the moving speed of the movable element 1 may also be calculated.

[0026] Of course, any known means, such as optical encoders, magnetic encoders, or video cameras, may be used to detect the movable element position and velocity. However, as mentioned above, this increases costs, so the movable element position is estimated from the voltage command and current of each coil 3.

[0027] In rotating machinery, the principle of position estimation generally utilizes the induced voltage, also known as the speed electromotive force, generated during rotation. As mentioned in the background information, this method has been applied to LSMs, so we will consider whether it can be applied to the linear motor 100 explained in Figure 1.

[0028] For example, when the movable element 1 is driven at a constant speed, the waveforms of the three adjacent coils 3d to 3f, centered on coil 3e, are as shown in Figure 3. In the figure, the three waveforms V3d, V3e, and V3f are the speed electromotive forces generated in coils 3d, 3e, and 3f, respectively, and the point where the center of coil 3e coincides with the center of the movable element 1 is considered to be zero.

[0029] The waveform V3e, shown by the solid line, is sinusoidal within the ±180-degree range. However, due to the non-uniformity of the magnetic properties at the movable element ends, distortion is observed in the waveform even within the ±180-degree range. This phenomenon, caused by the non-uniformity of the magnetic properties at the movable element ends, is generally known as the "end effect." As shown in Figure 1, LSMs with a small number of magnetic poles exhibit a larger end effect compared to LSMs with a large number of magnetic poles (for example, more magnetic poles than the number of movable elements).

[0030] When the waveform V3e, shown by the solid line, exceeds the ±180 degree range, it gradually attenuates and eventually becomes zero. The other two waveforms (waveforms V3d and V3f) have the same shape, but their phases are shifted by 135 degrees. Due to space limitations, the waveforms of the other coils 3 (coils 3a-3c, 3g-3i) are not shown here, but the speed electromotive forces of these coils 3 also have waveforms that are shifted by 135 degrees in the same way.

[0031] In rotating machinery control, it is common practice to use coordinate transformation to convert such multiphase AC waveforms into two-phase AC waveforms. However, in the case of non-three-phase LSMs that cannot be driven with 120-degree energization, it is not even certain whether they can be converted into two-phase AC waveforms. To the best of the inventors' knowledge, sensorless control technology and coordinate transformation technology for non-three-phase LSMs that cannot be driven with 120-degree energization are not publicly known, and none have yet become clear to those skilled in the art.

[0032] Therefore, we decided to perform position estimation using the following approach. Here, the rate electromotive force for 9 coils of coil 3 is given by dot θK. m9We will represent it with a vector. In this application, in order to convert this into a two-phase sine wave waveform of two orthogonal axes, we calculate it as shown in Equation (2) using the coordinate transformation matrix A9 shown in Equation (3).

[0033]

Number

[0034] When the motor structure is not of the 3-pole 4-slot type, the value of the electrical angle θ per coil c is different from 135 degrees. However, if a coordinate transformation equation similar to the above is established using the electrical angle θ per coil c corresponding to the structure, a two-phase sine wave waveform of two orthogonal axes can be obtained.

[0035] 0.5 In the case of a three-phase rotating machine, it is common to multiply the three-phase to two-phase coordinate transformation matrix or its inverse transformation matrix by a coefficient of (2 / 3).

[0036] Therefore, following this, it is also possible to multiply the above coordinate transformation matrix by some coefficient. However, in a non-three-phase LSM that cannot perform 120-degree conduction, unlike the case of a three-phase rotating machine, there is almost no merit in multiplying by a coefficient, so it is not necessary to multiply by a coefficient intentionally. Therefore, no special coefficient is multiplied in Equation (2). m7 When the speed electromotive force of 7 coils from coil 3b to 3h is represented by a vector called θ̇K

[0037]

Number

[0038] Figure 4 shows the waveform obtained by transforming the waveform shown in Figure 3 using static two-phase conversion, i.e., equations (4) and (5). This is the calculation result when coil 3e is centered, and if we consider only when the movable element 1 is within a range of ±180 degrees, although some distortion is observed, the two waveforms Va and Vb have almost the same amplitude, the phase difference is about 90 degrees, and a two-phase sinusoidal waveform is obtained.

[0039] Furthermore, it can be seen that when the distance between the central coil 3e and the movable element 1 exceeds ±180 degrees, the waveform of the recoil electromotive force after static two-phase conversion begins to attenuate. The waveform in this range cannot be said to be a two-phase sine wave. Increasing the number of coils 3 used in the calculation expands the range in which two-phase sine waves can be produced, but it also increases the amount of calculation required. In the case of this type of LSM, the number of coils can exceed several hundred, so it is best not to use too many coils 3 in the calculation.

[0040] In the drive control device 8 of this invention, the position of the movable element is estimated using the recoil electromotive force after the static two-phase conversion described above. Figure 5 shows the waveform Wr obtained by the four-quadrant arctangent (inverse tangent) calculation process for the recoil electromotive force after the static two-phase conversion shown in Figure 4. For reference, the waveform Wi obtained by the four-quadrant arctangent calculation process for an ideal two-phase sine wave with no end attenuation is also shown.

[0041] When the four-quadrant arctangent calculation is performed on an ideal two-phase sine wave, a sawtooth wave waveform (waveform Wi) is obtained. The waveform Wr obtained by the four-quadrant arctangent processing on the recoil electromotive force after static two-phase conversion also has a shape similar to a sawtooth wave, but as the position of the movable element 1 moves away from the origin, it deviates from the sawtooth wave waveform. Since the difference from the sawtooth wave waveform corresponds to the position estimation error, Figure 5 shows that position estimation by arctangent calculation is possible as long as the distance between the central coil 3e and the movable element 1 is within a range of approximately ±180 degrees. However, even when the distance between the central coil 3e and the movable element 1 is within a range of approximately ±180 degrees, a difference (position estimation error) between the arctangent calculation result (waveform Wr) and the sawtooth wave waveform (waveform Wi) occurs at a visually observable level.

[0042] It should be noted that this arctangent calculation only yields the direction of the recoil electromotive force vector. Even if the movable element position is the same, if the direction of movement of the movable element 1 is different, the direction of the recoil electromotive force vector will differ by 180 degrees. Therefore, if the movable element 1 can move in both directions, it is necessary to correct the movable element position according to the sign of the movable element velocity. However, since position sensorless control using recoil electromotive force is intended for operation in the medium to high speed range, the sign of the movable element velocity is known, so the correction is not particularly difficult.

[0043] In the LSM (linear motor 100) with the structure described in Figure 1, the number of magnetic poles is 3, and because it uses a movable element 1 with a small number of magnetic poles, it tends to be greatly affected by so-called end effects. End effects refer to the performance degradation caused by magnetic asymmetry at the ends of the movable element. Due to the influence of end effects, there is a large distortion in the speed electromotive force of this type of LSM. This causes an error in the estimation of the movable element position. Since this error is due to the motor structure, the arctangent calculation results in a position estimation error of a periodic pulsating waveform as shown in Figure 6.

[0044] While the movable element velocity can be calculated by differentiating the movable element position, if such pulsating components are superimposed on the estimated position information, then large pulsations will also be superimposed on the estimated velocity. Controlling the linear motor 100 using movable element position or velocity superimposed with pulsations will not yield a very good control response. Therefore, it is necessary to remove these pulsating components from the movable element position and velocity information.

[0045] Furthermore, there are cases where the voltage command or current changes suddenly, or where noise is superimposed on the detected current value. In such cases, the output of the position calculator 63 fluctuates, but the actual movable element position may not have changed yet. Therefore, it is necessary to prevent the estimated movable element position from changing too sensitively.

[0046] The simplest method for removing the high-frequency pulsation component from the estimated movable element position is to use a low-pass filter. However, a simple low-pass filter increases the DC component of the position estimation error. For example, when movable element 1 is moving at a constant speed, its position changes in a ramp-like manner. If this movable element position is input to a simple first-order low-pass filter, a steady-state error occurs in the filter's input and output. Since a DC error in the estimation of the movable element position prevents efficient thrust generation, filtering must be performed in a way that eliminates DC errors in the steady state.

[0047] Applying a low-pass filter with a low cutoff frequency relative to the estimated velocity, even if intended to eliminate pulsations, is not ideal. This is because applying a low-pass filter with a low cutoff frequency prevents rapid control of the movable element.

[0048] Thus, it is not easy to eliminate the influence of end effects from the position estimation results or velocity estimation results of a linear motor 100, which has large end effects. Therefore, in Embodiment 1, as a method that can eliminate the influence of end effects from the position estimation results and velocity estimation results with a simple configuration, a movable element position estimation means 6 was configured as shown in Figure 7.

[0049] The movable element position estimation means 6 includes a nearby coil selector 65 that selects a coil 3 from among several coils 3 that is close to the movable element 1 as a nearby coil. Specifically, it selects a coil 3 that is directly below the movable element 1 and within a range of less than the electrical angle of one coil 3 (135 degrees in this example) extending outward from both ends in the direction of travel. In Figure 1, the selected coils are 3c to 3f directly below the movable element 1 and coils 3b and 3g that are within a range of less than 135 degrees extending outward from both ends. The selected coils 3 are referred to as "nearby coils". The nearby coil selector 65 receives voltage command vectors V from each of the multiple coil current control means 5. * s and current vector I s Therefore, the voltage command V of the nearby coil * sn and current Isn Output only this.

[0050] Then, the voltage command V for the nearby coil is output from the nearby coil selector 65. * sn and current I sn From the recoil electromotive force E sn A speed electromotive force calculation unit 61 calculates the speed electromotive force E sn Converting the multiphase to twophase system to obtain the velocity electromotive force E in stationary Cartesian coordinates αβ It is equipped with a multiphase two-phase converter 62 that calculates the voltage command vector V. The reason for limiting the data output to the speed electromotive force calculation unit 61 is mainly to handle cases where multiple movable elements 1 are present on the track, and also to reduce the amount of calculation required. However, it is not always necessary to limit the data, and it is possible to omit the nearest coil selector 65. In that case, the speed electromotive force calculation unit 61 receives the voltage command vector V from each of the multiple coil current control means 5. * s and current vector I s This will result in the output being displayed as is.

[0051] Furthermore, the velocity electromotive force E in stationary Cartesian coordinates αβ The system includes a position calculator 63 that calculates the first estimated position hat θ by performing an arctangent operation. The position calculator 63 may also acquire additional information to determine the driving direction of the movable element 1. Here, the sign (dot θ) is the sign of the velocity command. * This is set to be input to the position calculator 63.

[0052] Then, the second estimated position hat θ is obtained by removing the aforementioned error from the first estimated position hat θ calculated by the position calculator 63. F The system is equipped with a pulsation remover 64 that calculates the estimated velocity dot-hat θ.

[0053] Speed ​​electromotive force E obtained by speed electromotive force calculation unit 61 s The method for calculating the induced voltage will be explained. The voltage equation for linear motor 100 is given by equation (6).

[0054]

number

[0055] Stator voltage vector V s Instead, the voltage command vector V * s If we decide to use this, the rate electromotive force vector E s This can be found using equation (7).

[0056]

number

[0057] Note that the rate electromotive force VectorE s The movable element magnetic flux vector φ m It is defined as the time derivative of the movable element flux vector φ. m Since is a function of the movable element position θ, and the movable element position θ is a function of time t, the recoil electromotive force vector E s It can also be written as equation (8).

[0058]

number

[0059] As mentioned above, when the linear motor 100 is driven (movable element 1 travels along the track provided by the stator 2), the recoil electromotive force of each coil 3 has a waveform as shown in Figure 3, and the recoil electromotive force can be roughly calculated using equation (7). Since the recoil electromotive force is a function of the movable element position θ, the movable element position θ can be estimated from this information. In this case, using the coordinate transformation equations shown in equations (2) to (5) above makes it easier to estimate the movable element position θ.

[0060] The position calculator 63 calculates the velocity electromotive force E in stationary Cartesian coordinates, for example, as shown in Figure 8. αβ and the speed command sign (dot θ) * The system includes an inverse tangent calculator 631 that estimates the movable part position by performing an inverse tangent calculation. It also includes an edge detection unit 632 that detects edges from the result of the inverse tangent calculation, an offset amount calculation unit 633 that calculates the offset amount, and an adder 634 that adds the offset amount calculated by the offset amount calculation unit 633 to the movable part position estimated by the inverse tangent calculator 631.

[0061] The inverse tangent unit 631 performs a four-quadrant inverse tangent calculation on the speed electromotive force, which has been coordinate-transformed to two-phase AC, to estimate the position of the movable element. When the movable element 1 moves in both forward and reverse directions, the calculation result needs to be corrected by 180 degrees according to the direction of movement of the movable element 1, so here the sign of the speed command sign(θ) * The sign (θ) of the velocity command is input to the arctangent unit 631, but if the movable element 1 moves in only one direction, the sign (θ) of the velocity command is input to the arctangent unit 631. * You do not need to enter ).

[0062] In linear motor systems requiring non-three-phase LSMs, hundreds or more coils 3 may be positioned on the ground side, resulting in a range of motion of the movable element 1 that can exceed tens of thousands of degrees in electrical angle terms. However, the output range of the inverse tangent calculator 631 is only ±180 degrees. If current were to be supplied to all coils 3, the phase of current supply could be determined solely by the output value of the inverse tangent calculator 631. However, supplying current to coils 3 located far from the movable element 1 is energy inefficient. Therefore, current is usually supplied only to coils 3 in the vicinity of the movable element 1. To achieve this, a mechanism is needed to determine the approximate position of the movable element 1.

[0063] The simplest way to roughly determine the position of the movable element 1 is to detect and count discontinuous changes (edges) in the result of the arctangent calculation. The edge detection unit 632 adds one count value when the result of the arctangent calculation changes abruptly from +180 degrees to -180 degrees. Conversely, it subtracts one count value when it changes abruptly from -180 degrees to +180 degrees. The offset amount calculation unit 633 outputs the approximate position (offset amount) of the movable element 1 by multiplying the value counted by the edge detection unit 632 by 360 degrees. This is added to the result of the arctangent calculation by the adder 634, and the resulting addition is used as the output (estimated position hat θ) of the position calculator 63.

[0064] There are various other methods for determining the offset amount. For example, the rate electromotive force vector E s One possible method for determining the offset amount is to square each element and examine their relative magnitudes. Another method involves determining the offset amount by obtaining a rough position using an inexpensive sensor such as a video camera, and then estimating the precise position using arctangent calculation.

[0065] By performing this process, the first estimated position hat θ is obtained. However, as explained in Figure 6, the first estimated position hat θ is superimposed with position estimation errors due to edge effects. Furthermore, when the movable element velocity is calculated based on the first estimated position hat θ, velocity estimation errors occur. Since position estimation errors or velocity estimation errors degrade the performance of position sensorless control, it is necessary to reduce both position estimation errors and velocity estimation errors as much as possible.

[0066] Therefore, in this application, a pulsation remover 64 is provided as a method that can accurately reduce the influence of edge effects from position estimation results and velocity estimation results with a simple configuration. The pulsation remover 64 removes the periodic error component from the first estimated position hat θ to obtain the second estimated position hat θ. FIt outputs the following. During this calculation, the estimated velocity dot-hat θ can be calculated, so it also functions as a velocity calculator that simultaneously outputs an estimated velocity signal. The pulsation remover 64 consists of an adder 641, a proportional-integral calculator 642, and an integrator 643.

[0067] The following explains why the pulsation remover 64 has this configuration. As mentioned above, the first estimated position hat θ is superimposed with pulsations due to edge effects, so some kind of filter is needed to remove them. Here, we consider filtering by performing a feedback operation on the difference between input and output. At this time, it is important not only to remove the pulsations due to edge effects, but also to make the steady-state difference between the input and output of the filter zero.

[0068] Here, we consider how to make the steady-state difference between the input and output of the filter zero when the movable element 1 is moving at a constant speed. Since the movable element position changes in a ramp-like manner at this time, we should adopt a controller structure that makes the steady-state error with respect to the ramp response zero. It is generally known as the "internal model principle" in classical control theory that a necessary condition for making the steady-state error with respect to the ramp response zero is to use a controller of type 2 or higher.

[0069] A Type 2 controller is a controller that includes two integrators connected in series. The pulsation remover 64 is equipped with a proportional-integral unit 642 and an integrator 643. Since the proportional-integral unit 642 has an integrator inside, two integrators are connected in series in this system. Therefore, the pulsation remover 64 is a Type 2 controller.

[0070] In the removal of pulsations caused by edge effects, the difference between using a general low-pass filter and using the pulsation remover 64 of this application will be explained as a comparative example. The movable element position (first estimated position hat θ) that becomes the input to each filter increases in a ramp-like manner, as shown in Figures 9A and 9B.

[0071] In contrast, the waveform hat θ when a first-order low-pass filter is appliedLPF As shown in Figure 9A, it increases in a ramp-like manner with a slight delay from the first estimated position hat θ. The error between the two waveforms (input and output) in the steady state is Δθ DC Therefore, with a first-order low-pass filter, Δθ will remain constant no matter how much time passes. DC Δθ will never be zero. DC The slope of the first estimated position hat θ, i.e., the estimated velocity dot hat θ, increases as it increases. Δθ DC Since thrust cannot be generated efficiently under these conditions, filtering with a first-order low-pass filter is not practical.

[0072] On the other hand, in the filter using the pulsation remover 64 of the present invention, according to the internal model principle, a first estimated position hat θ and a second estimated position hat θ which is the filter output thereto F Since it is guaranteed that they will eventually match, after a predetermined time has elapsed, Δθ will appear as shown in Figure 9B. DC This becomes zero. Therefore, using the pulsation remover 64 of this application, the problem of efficiency degradation in the high-speed region, which occurs with a first-order low-pass filter, does not occur.

[0073] Furthermore, in cases where the movable element speed changes at a constant rate, if the steady-state difference between the filter's input and output is to be reduced to zero, a Type 3 controller can be used. This will be discussed in Embodiment 2, but it is not particularly difficult to implement as it only requires adding a double integrator path to the pulsation remover 64 of this embodiment.

[0074] Regarding the ramp response, according to the internal model principle, the first estimated position and the second estimated position hat θ, which is the output of integrator 643, are given by the first estimated position and the second estimated position hat θ. F Since they consistently coincide, the input to integrator 643 can be considered as velocity information. From this, it can be said that this system can calculate velocity without performing differential operations.

[0075] The proportionality term in the proportional-integral operator 642 is necessary to manipulate the damping coefficient of the system. This will be explained in more detail. The transfer function of the proportional-integral operator 642 can be written as shown in equation (9).

[0076]

number

[0077] Then, in the pulsation remover 64, the second estimated position hat θ is obtained from the first estimated position hat θ. F The transfer function up to this point can be expressed as shown in equation (10).

[0078]

number

[0079] Furthermore, the transfer function from the first estimated position hat θ to the estimated velocity dot hat θ can be expressed as shown in equation (11).

[0080]

number

[0081] Furthermore, the following equation (Equation (12)) is well-known as the transfer function in quadratic normal form.

[0082]

number

[0083] By comparing the coefficients in equations (10) to (12), we can determine the attenuation coefficient ζ and natural angular frequency ω of the pulsation remover 64. nIt can be seen that this is determined by the following formulas (Equations (13) and (14)).

[0084]

number

[0085] Proportional gain K of proportional-integral unit 642 P When ω is zero, the damping coefficient is zero. It is known that when the damping coefficient is zero, the output signal undergoes sustained oscillation. Therefore, it is necessary to design the proportional gain so that the damping coefficient becomes the desired value. On the other hand, the natural angular frequency ω of the system n The integral gain K I This is determined by the integral gain K in the pulsation remover 64. I By adjusting this setting, you can change the frequency range at which the effects of edge effects are removed.

[0086] Adopting this configuration has the secondary effect of making it more resistant to sudden changes in voltage commands or current. It also becomes more resistant to current detection noise. If the voltage command or current changes suddenly, or if noise is introduced into the current detection, the first estimated position hat θ, which is the output of the position calculator 63, may fluctuate, but the actual movable part position may not have changed yet. Therefore, it is necessary to prevent the estimated movable part position from changing too sensitively. Second estimated position hat θ F Because it changes more slowly compared to the first estimated position hat θ, it is less susceptible to the effects of such disturbances.

[0087] The operation of the drive control device 8 according to Embodiment 1, that is, the flow of the drive control method, is summarized in the flowchart of Figure 10. First, the nearest coil is selected by the nearest coil selector 65 (step S100), and the voltage command vector V received from each of the multiple coil current control means 5 is selected. * s and current vector I s Among them, the voltage command V of the nearby coil * sn and current I snOnly this is output to the recurrent electromotive force calculation unit 61.

[0088] The speed electromotive force calculation unit 61 calculates the voltage command V of the nearby coil. * sn and current I sn From the recoil electromotive force E sn The (induced voltage) is calculated (step S110). The multiphase two-phase converter 62 receives the speed electromotive force E output from the speed electromotive force calculation unit 61. sn The velocity electromotive force E in stationary Cartesian coordinates is obtained by multiphase two-phase conversion (stationary two-phase sine wave conversion). αβ The calculation is performed (step S120).

[0089] The position calculator 63 receives the reciprocating electromotive force E in stationary Cartesian coordinates output from the multiphase two-phase converter 62. αβ The first estimated position hat θ is calculated by a four-quadrant inverse tangent operation (step S130) and output to the pulsation remover 64. The pulsation remover 64, which is a type 2 or higher controller, calculates the second estimated position hat θF by removing the pulsation error from the first estimated position hat θ (step S140) and outputs it. At the same time, the estimated velocity dot hat θ is also calculated and output.

[0090] In step S130, the edge detection unit 632 detects an edge from the result of the arctangent calculation and outputs a count value, increased or decreased according to the type of edge (positive or negative direction), to the offset amount calculation unit 633. The offset amount calculation unit 633 outputs the value obtained by multiplying the count value by 360 degrees as the offset amount to the adder 634, and the adder 634 outputs the movable element position, which is the result of adding the offset amount to the result of the arctangent calculation, as the estimated position hat θ.

[0091] Furthermore, the movable element position estimation means 6 or the control calculation means 7 that constitute the drive control device 8 may be configured by a single hardware 6H comprising a processor 6H1 and a storage device 6H2, as shown in Figure 11. Although not shown, the storage device 6H2 comprises a volatile storage device such as random access memory and a non-volatile auxiliary storage device such as flash memory. Alternatively, an auxiliary storage device such as a hard disk may be provided instead of flash memory. The processor 6H1 executes the program input from the storage device 6H2. In this case, the program is input from the auxiliary storage device to the processor 6H1 via the volatile storage device. The processor 6H1 may also output data such as calculation results to the volatile storage device of the storage device 6H2, or it may store the data in the auxiliary storage device via the volatile storage device.

[0092] As described above, by using the drive control device 8 or drive control method according to Embodiment 1, the influence of edge effects can be eliminated from the position estimation result or velocity estimation result with a simple configuration. High-performance control of LSM can also be realized at low cost using this position estimation system.

[0093] Incidentally, in rotating machinery, various position estimation methods or position sensorless control methods have been devised in addition to the method of performing an arctangent calculation on the recoil electromotive force. Such technologies may also be combined with the present invention.

[0094] Embodiment 2 In Embodiment 1, the case where the speed is constant was discussed, so the pulsation remover was a Type 2 controller. In Embodiment 2, an example of configuring the pulsation remover to handle the case where the acceleration is constant will be described. Figure 12 is a block diagram showing an example in which the pulsation remover of the movable element position estimation means is configured as a Type 3 in the drive control device according to Embodiment 2, and Figure 13 is a block diagram showing the configuration of the proportional-integral double integrator of the pulsation remover.

[0095] Except for the fact that the pulsation remover is configured as a Type 3, the configuration is the same as in Embodiment 1. Therefore, the explanation of the similar parts will be omitted, and Figures 1 to 6 and 8 to 11 used in Embodiment 1 will be referenced.

[0096] A constant acceleration is an example of a pattern where velocity gradually increases or decreases. In this case, if we want to make the steady-state difference between the filter's input and output zero, a Type 3 controller is required. Therefore, in this embodiment... 2 As shown in Figure 12, the pulsation remover 64 is equipped with a proportional-integral double integral calculator 644 instead of the proportional-integral calculator 642 shown in Embodiment 1.

[0097] This section explains the difference between the proportional-integral unit 642 and the proportional-integral double integrator 644. As shown in Figure 13, the proportional-integral double integrator 644 includes a proportional unit 6441, an integral unit 6442, a double integrator 6443, and an adder 6444. The proportional-integral double integrator 644 is equivalent to the proportional-integral unit 642 with the addition of the double integrator 6443. The double integrator 6443 outputs the result of integrating the input signal and then integrating that signal again. The output of the double integrator 6443 is multiplied by a coefficient as appropriate to stabilize the controller. The adder 6444 calculates the sum of the outputs of the proportional unit 6441, the integral unit 6442, and the double integrator 6443, and the result of this calculation is the output of the proportional-integral double integrator 644.

[0098] We will also explain the mathematical differences. The transfer function of the proportional-integral double integral operator 644 is expressed as shown in equation (15).

[0099]

number

[0100] Comparing equation (9) and equation (15), equation (15) contains the double integral term K. II / s 2 The following has been added. Embodiment 2Therefore, since this double integral term and integrator 643 are connected in series, the number of integrators connected in series in the pulsation remover 64 is 3. Thus, this is a type 3 controller. According to the internal model principle, it is known that a type 3 controller can make the steady-state error for a constant acceleration input zero. However, for this to be possible, the proportional gain K P , integral gain K I Double integral gain K II It is necessary to design it properly and ensure that the system is stable.

[0101] An example of a gain design method for the 644 proportional-integral double integral calculator is shown below. In the pulsation remover 64, the second estimated position hat θ is obtained from the first estimated position hat θ. F The transfer function up to this point can be expressed as shown in equation (16).

[0102]

number

[0103] In this system as well, the input to integrator 643 can be considered as the estimated velocity. Here, the transfer function from the first estimated position hat θ to the estimated velocity dot hat θ can be expressed as shown in equation (17).

[0104]

number

[0105] There are several possible polar configurations, but for example, a polar configuration with triple roots is commonly used in cubic systems, so it is advisable to use this. In the case of a polar configuration with triple roots, the denominator polynomial is as shown in equation (18).

[0106]

number

[0107] Here, -ω n is the pole of the system. To ensure the stability of the system, let ω n > 0. This system has three poles, and all three poles overlap at a single point on the real axis. The overlapping point is -ω n . Such a pole arrangement is called a "pole arrangement with triple roots."

[0108] Once the poles of the system are determined, the numerical values of each gain can be obtained by comparing the coefficients of Equation (18). The gain for which the poles of this system become triple roots is as shown in Equation (19).

[0109]

Number

[0110] In the case of the above-mentioned pole arrangement with triple roots, since the real-axis components of all poles are negative, the system is stable. If the system is to be stable, other pole arrangement methods other than triple roots can also be used. For example, when designing the gain with the pole arrangement of the third-order Butterworth, the gain can be determined as shown in Equation (20).

[0111]

Number

[0112] In the case of the above design, the poles are at the positions of -ω n , -ω n exp(±jπ / 3). Also in this case, since the real-axis components of all poles are negative, the system is stable.

[0113] The gain can be designed freely to some extent, but to stabilize this system, at least K P > 0, K I > 0 is required. Therefore, the proportional term or the single-integral term cannot be omitted from the proportional-integral double-integral calculator 644.

[0114] As described above, the pulsation remover 64, which has a properly designed proportional-integral double integrator 644, has the characteristic that no steady-state error occurs before and after the filter input / output, even when the acceleration is constant. Therefore, by using the movable element position estimation means 6 according to Embodiment 2, the influence of end effects can be eliminated from the position estimation result or velocity estimation result with a simple configuration. Using this position estimation system, high-performance control of the LSM (linear motor 100) can be realized at low cost.

[0115] Furthermore, by applying the method disclosed in this application, it is possible to handle cases where the change in the position or velocity of the movable element 1 is even steeper. For example, if it is desired to make the steady-state error zero in a case where the acceleration changes at a constant rate, a Type 4 controller should be adopted, and a proportional-integral-double-integral-triple-integral unit should be used. Similarly, it is also possible to consider Type 5 or higher controllers.

[0116] Embodiment 3. Although the drive control device or drive control method according to Embodiments 1 and 2 described above were applied to LSMs, they can also be applied to rotating machines. Embodiment 3 describes an example of application to a rotating machine.

[0117] Figure 14 is a block diagram illustrating the configuration of a drive control device and a motor system using a rotating machine according to Embodiment 3. Figures 15A to 15C show the waveform of the speed electromotive force with the rotor position as the horizontal axis, the waveform obtained by calculating the arctangent of the speed electromotive force, and the waveform of the position estimation error of the waveform obtained by calculating the arctangent, respectively, in a three-phase rotating machine. Figure 16 is a block diagram illustrating the configuration of the rotor position estimation means.

[0118] The motor system according to Embodiment 3, as shown in Figure 14, is a motor system consisting of a rotating machine 100R, which is a rotary permanent magnet synchronous motor, and a drive control device 8R that drives and controls the rotating machine 100R. The rotating machine 100R is a permanent magnet synchronous motor and consists of a rotor having magnetic poles of permanent magnets that are rotatably supported on the inner circumferential surface side of an annular stator (not shown).

[0119] The drive control device 8R is equipped with a variable voltage source 5R that applies voltage to the coils of the stator (not shown) of the rotating machine 100R, thereby enabling the rotation of the rotating machine 100R (more precisely, the rotor). It also includes a current detection means 51R that detects the current flowing through the rotating machine 100R (more precisely, the coils of the stator, not shown), and a rotor position estimation means 6R that estimates at least one of the rotor position and rotor speed.

[0120] Furthermore, the system includes a control means 7R that determines a voltage command to drive the rotating machine 100R to a variable voltage source 5R using information on the current flowing through the rotating machine 100R, the rotor speed of the rotating machine 100R, and the rotor position. Therefore, the current detection means 51R transmits the detected current value to the control means 7R and the rotor position estimation means 6R. The rotor position estimation means 6R calculates the speed electromotive force from the voltage command and current, and uses the speed electromotive force to estimate the rotor position and rotor speed.

[0121] Most 100R rotating machines are driven by a three-phase power supply, so the following explanation assumes a three-phase motor, but the motor does not necessarily have to be three-phase.

[0122] For example, the waveforms of the u-phase speed electromotive force Eu, the v-phase speed electromotive force Ev, and the w-phase speed electromotive force Ew in a three-phase rotating machine are distorted, as shown in Figure 15A. Although this is just one example, the speed electromotive force (waveforms Eu, Ev, Ew) has 5th and 7th harmonics superimposed on it. By converting this from three-phase to two-phase and performing the arctangent calculation as in Embodiment 1, a waveform WrR as shown in Figure 15B is obtained.

[0123] For comparison, a sawtooth wave waveform (dashed line: waveform WiR) is shown alongside the figure. However, as shown in Figure 15B, the arctangent calculation result does not match the sawtooth wave due to the distortion of the reticle electromotive force. Therefore, as shown in Figure 15C, periodic position estimation errors appear due to the 5th and 7th harmonics of the reticle electromotive force. In this case, the order of the pulsation in the position estimation error is 6th.

[0124] Thus, even in the rotating machine 100R, a position estimation error may occur due to distortion of the speed electromotive force. Although the rotational speed can be calculated by differentiating the waveform in FIG. 15B, there is a problem that a pulsation error also occurs in the rotational speed with this calculation method.

[0125] Therefore, in the third embodiment, as shown in FIG. 16, the rotor position estimation means 6R is configured. The rotor position estimation means 6R includes a speed electromotive force calculation unit 61R that calculates the speed electromotive force E * s from the voltage command V s and the current I s input from the control means 7R and the current detection means 51R respectively, and a three-phase to two-phase converter 62R that performs three-phase to two-phase conversion on the speed electromotive force E s to calculate the speed electromotive force E αβ in the stationary orthogonal coordinates.

[0126] Further, it includes a position calculator 63R that calculates the first estimated position hat θ by inverse tangent operation from the speed electromotive force E αβ in the stationary orthogonal coordinates, and a pulsation remover 64R that calculates the second estimated position hat θF from which the error has been removed from the first estimated position hat θ.

[0127] The speed electromotive force calculation unit 61R calculates the speed electromotive force E s on the three-phase coordinates based on the voltage equation. The three-phase to two-phase converter 62R performs coordinate conversion of the speed electromotive force E s on the three-phase coordinates to the speed electromotive force E αβ on the stationary orthogonal coordinates. Note that the order of these calculations may be reversed. That is, after performing coordinate conversion on the voltage command V * s and the current I s , the speed electromotive force on the stationary orthogonal coordinates may be calculated.

[0128] The position calculator 63R performs a four-quadrant inverse tangent operation using the speed electromotive force E αβ in the stationary orthogonal coordinates to estimate the first estimated position hat θ of the rotor. The position calculator 63R may acquire information for determining the driving direction of the rotor as additional information. Here, it is the sign of the speed command, sign(dot θ* The value ) is input to the position calculator 63R. This is necessary when the motor rotates in both forward and reverse directions. In applications where the motor rotates in only one direction, sign(dot θ) is used. * You do not need to enter ().

[0129] Similar to embodiments 1 and 2, embodiment 3 also includes position estimation errors caused by distortion of the recoil electromotive force in the first estimated position hat θ, and therefore it is necessary to remove these as much as possible. In embodiment 3, as with embodiments 1 and 2, a pulsation remover 64R is used as a method that can accurately reduce the effect of distortion of the recoil electromotive force from the position estimation result or the velocity estimation result with a simple configuration. Pulsation remover 64 R The operation is the same as in Embodiment 1. By adopting this configuration, good position estimation and velocity estimation become possible.

[0130] As described above, the position sensorless control using the drive control device 8 (including the drive control device 8R) or drive control method of the present application is applicable not only to the linear motor 100 but also to the rotating machine 100R, and is effective in reducing the cost of the drive device. Incidentally, in the rotating machine 100R, various position estimation methods or position sensorless control methods have been devised in addition to the method of performing an arctangent calculation on the speed electromotive force. Such technologies may also be combined with the configuration disclosed in the present application.

[0131] While this application describes various exemplary embodiments and examples, the various features, aspects, and functions described in one or more embodiments are not limited to the application of a particular embodiment, but are applicable individually or in various combinations to the embodiments. Accordingly, countless variations not illustrated are conceivable within the scope of the technology disclosed herein. For example, these include modifying, adding, or omitting at least one component, or even extracting at least one component and combining it with a component from another embodiment.

[0132] As described above, according to the drive control devices 8 and 8R of the present invention, a power supply device (coil current control means 5, variable voltage source 5R) controls the current of the permanent magnet synchronous motor (linear motor 100, rotating machine 100R), and the current (current vector I) that flows through the permanent magnet synchronous motor is controlled by a power supply device (coil current control means 5, variable voltage source 5 drive control devices 8 and 8R of the present invention control the current of the permanent magnet synchronous motor (linear motor 100, rotating machine 100R), and the current (current vector I s ) and the applied voltage (voltage command vector V * s ), and the motor constants of a permanent magnet synchronous motor (armature resistance R s The induced voltage (speed electromotive force E) of a permanent magnet synchronous motor is derived from the inductance matrix L, etc. s The induced voltage calculation unit (return electromotive force calculation unit 61) calculates the induced voltage (return electromotive force E) and the induced voltage (return electromotive force E s A coordinate transformation means (multiphase two-phase converter 62, three-phase two-phase converter 62R) that converts ) into two-phase AC, and the induced voltage E converted into two-phase AC αβ Position calculators 63, 63R calculate the first estimated position hat θ of the magnetic pole of the permanent magnet synchronous motor by performing an inverse tangent calculation, and the second estimated position hat θ obtained by removing the periodic pulsations caused by distortion of the induced voltage from the first estimated position hat θ. F Position estimation means 6, 6R having pulsation removers 64, 64R that calculate a second estimated position hat θ F Based on this, the power supply unit is equipped with control calculation means 7,7R for controlling the operation of the power supply unit, and the pulsation removers 64,64R are equipped with a first estimated position hat θ and a second estimated position hat θ F The system is configured to eliminate periodic pulsations so that the steady-state average error becomes zero. This allows the motor drive to be controlled based on the precise magnetic pole position, eliminating errors caused by distortion in the induced voltage, thus enabling a motor system that operates without position sensors.

[0133] The pulsation removers 64 and 64R have a first estimated position hat θ and a second estimated position hat θ F Subtractors 641 and 641R calculate the deviation, proportional-integral operators 642 and 642R perform proportional-integral operations on the deviation calculated by subtractors 641 and 641R, and integrate the values ​​obtained from the proportional-integral operations to obtain the second estimated position hat θ F If integrators 643 and 643R are provided to perform the calculation, errors caused by distortion of the induced voltage can be reliably eliminated.

[0134] If the position estimation means 6,6R outputs the output of the proportional-integral calculators 642,642R as the estimated velocity dot-hat θ, the velocity can be estimated without differential calculation, thereby obtaining the travel speed of the movable element 1 or the rotational speed of the rotor with minimal pulsation.

[0135] The position calculators 63 and 63R have an edge detection unit 632 that detects sawtooth wave edges obtained by arctangent calculation, and if configured to calculate a first estimated position hat θ from the sawtooth wave edges, the approximate position of the movable element 1 or the approximate rotational position of the rotor can be estimated. In particular, in LSMs, it is possible to narrow down the coils 3 to be energized and improve efficiency.

[0136] A permanent magnet synchronous motor is a linear synchronous motor (linear motor 100) composed of a movable element 1 and a stator 2 in which multiple coils 3 are arranged to form a path for the movable element 1. If the power supply is a current control device (coil current control means 5) that individually controls the current of each of the multiple coils 3, then position sensorless control suitable for modern LSMs can be achieved.

[0137] A linear synchronous motor (linear motor 100) can perform position sensorless control, which is suitable for the increasingly popular LSMs, as long as the phase difference between the induced voltages generated in two adjacent coils 3 is not 120 degrees.

[0138] Furthermore, the motor system of the present invention includes a permanent magnet synchronous motor (linear motor 100, rotating machine 100R) that is driven and controlled by the aforementioned drive control devices 8, 8R and power supply device (coil current control means 5, variable voltage source 5R), so that the position of the movable element 1 in a non-three-phase LSM can be estimated without a position sensor and driven efficiently.

[0139] Furthermore, according to the drive control method of the present invention, the current (current vector I) flowing through the permanent magnet synchronous motor (linear motor 100, rotating machine 100R) is controlled. s ) and the applied voltage (voltage command vector V* s ), and the motor constants of a permanent magnet synchronous motor (armature resistance R s The induced voltage (speed electromotive force E) of a permanent magnet synchronous motor is derived from the inductance matrix L, etc. s Step S110: Calculate the induced voltage (rate electromotive force E) s A coordinate transformation step (step S120) to convert ) into two-phase AC, and the induced voltage converted to two-phase AC. αβ A position calculation step (step S130) calculates the first estimated position hat θ of the magnetic pole of the permanent magnet synchronous motor by performing an inverse tangent calculation, and a second estimated position hat θ is obtained by removing the periodic pulsation caused by the distortion of the induced voltage from the first estimated position hat θ. F A pulsation removal step (step S140) calculates the second estimated position hat θ F Based on this, a control step is included in which the current flowing through the permanent magnet synchronous motor is controlled, and in the pulsation removal step (step S140), a first estimated position hat θ and a second estimated position hat θ are used. F The system is configured to eliminate periodic pulsations so that the steady-state average error becomes zero. This allows the motor drive to be controlled based on the precise magnetic pole position, eliminating errors caused by distortion in the induced voltage, thus enabling a motor system that operates without position sensors. [Explanation of Symbols]

[0140] 1: Movable element, 100: Linear motor (permanent magnet synchronous motor), 100R: Rotating machine (permanent magnet synchronous motor), 2: Stator, 3: Coil, 5: Coil current control means (power supply unit), 5R: Variable voltage source (power supply unit), 6: Movable element position estimation means (position estimation means), 6R: Rotor position estimation means (position estimation means), 61,61R: Speed ​​electromotive force calculation unit (induced voltage calculation unit), 62,62R: Multiphase two-phase converter, 63,63R: Position calculator, 631: Incandescent tangent calculator, 632: Edge detection unit, 633: Offset amount calculation unit, 64,64R: Pulsation remover, E s : Speed ​​electromotive force (induced voltage), E αβ :Speed ​​electromotive force (two-phase sine wave), I s : Current vector, Isn :Current, V * s : Voltage command vector, V * sn : Voltage command, Hat θ: First estimated position, Hat θ F :Second estimated position, dot hat θ:Estimated speed (driving speed).

Claims

1. A power supply unit that controls the current of a permanent magnet synchronous motor. A position estimation means comprising: an induced voltage calculation unit that calculates the induced voltage of the permanent magnet synchronous motor from the current flowing through the permanent magnet synchronous motor, the applied voltage, and the motor constants of the permanent magnet synchronous motor; a coordinate transformation means that converts the induced voltage into a two-phase alternating current; a position calculator that calculates a first estimated position of the magnetic poles of the permanent magnet synchronous motor by performing an inverse tangent calculation on the induced voltage converted to a two-phase alternating current; and a pulsation remover that calculates a second estimated position by removing periodic pulsations caused by distortion of the induced voltage from the first estimated position; and The system includes a control calculation means for controlling the operation of the power supply based on the second estimated position, The pulsation remover is characterized by removing the periodic pulsations such that the steady-state average error between the first estimated position and the second estimated position becomes zero.

2. The pulsation remover includes: A subtractor that calculates the deviation between the first estimated position and the second estimated position, A proportional-integral unit that performs a proportional-integral operation on the deviation calculated by the subtractor, and An integrator that calculates the second estimated position by integrating the value obtained in the proportional-integral operation, The drive control device according to claim 1, characterized in that it is provided.

3. The drive control device according to claim 2, characterized in that the position estimation means outputs the output of the proportional-integral calculator as the estimated speed.

4. The drive control device according to claim 1, wherein the position calculator has an edge detection unit that detects sawtooth wave edges obtained by the arctangent calculation, and calculates the first estimated position from the sawtooth wave edges.

5. The aforementioned permanent magnet synchronous motor is a linear synchronous motor comprising a movable element and a stator in which a plurality of coils are arranged to form a path for the movable element. The drive control device according to claim 1, characterized in that the power supply device is a current control device that individually controls the current of each of the plurality of coils.

6. The drive control device according to claim 5, characterized in that the linear synchronous motor has a phase difference of 120 degrees between the induced voltages generated in two adjacent coils.

7. A drive control device according to any one of claims 1 to 6, and The permanent magnet synchronous motor, which is driven and controlled by the power supply unit, A motor system characterized by having the following features.

8. An induced voltage calculation step for calculating the induced voltage of a permanent magnet synchronous motor from the current flowing through the permanent magnet synchronous motor, the applied voltage, and the motor constants of the permanent magnet synchronous motor. A coordinate transformation step to convert the induced voltage into a two-phase alternating current, A position calculation step of calculating the first estimated position of the magnetic pole of the permanent magnet synchronous motor by performing an inverse tangent calculation on the induced voltage converted to two-phase AC, A pulsation removal step of calculating a second estimated position by removing periodic pulsations caused by distortion of the induced voltage from the first estimated position, and The control step includes controlling the current flowing through the permanent magnet synchronous motor based on the second estimated position, The drive control method is characterized in that, in the pulsation removal step, the periodic pulsations are removed such that the steady-state average error between the first estimated position and the second estimated position becomes zero.

Images