Control device for rotating machine
The control device estimates rotor position and speed by synchronizing multiple coordinate systems to suppress q-axis inductance fluctuations, ensuring accurate and stable position sensorless control for permanent magnet synchronous motors.
Patent Information
- Application Number
- PCT/JP2026/018419
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-15
- Filing Date
- 2026-05-15
- Publication Date
- 2026-11-19
AI Technical Summary
Conventional position sensorless control methods for permanent magnet synchronous motors are prone to instability due to fluctuations in the q-axis inductance Lq, leading to errors in rotor position estimation, which complicates parameter adjustment and optimization.
A control device that estimates rotor position and speed by synchronizing a first Cartesian coordinate system with the rotating machine's d- and q-axes, and a second system with different axes, calculating the phase difference between composite magnetic fluxes to suppress the influence of q-axis inductance fluctuations, allowing for accurate estimation and easy parameter adjustment.
Enables high-accuracy rotor position and speed estimation with improved stability and responsiveness, facilitating easy parameter adjustment and optimization, especially in applications requiring torque accuracy.
Smart Images

Figure JP2026018419_19112026_PF_FP_ABST
Abstract
Description
Control device for rotating machines
[0001] This disclosure relates to a control device for a rotating machine.
[0002] In rotating machines, torque is controlled by a current vector set based on the magnetic flux vector of the rotor. Among rotating machines, permanent magnet synchronous motors in particular require position sensors such as encoders, resolvers, or Hall elements to detect the rotor position because the rotor is equipped with permanent magnets. However, the installation of position sensors leads to increased costs and installation space. Against this backdrop, a position sensorless control method has been proposed that estimates the rotor position of a permanent magnet synchronous motor without using position sensors and controls the current vector based on that estimated rotor position.
[0003] Rotor position estimation can be achieved by estimating the phase of the rotor's magnetic flux or voltage. However, conventional approaches have the constraint that the rotor's magnetic flux or voltage must be directed along the d-axis or q-axis, making information on the q-axis inductance Lq of a permanent magnet synchronous motor indispensable. However, in rotor position estimation, the sensitivity to the q-axis inductance Lq is known to be very high among the parameters of a permanent magnet synchronous motor. If there is an error in the set value of the q-axis inductance Lq, an error is likely to occur in the phase of the estimated magnetic flux or voltage, and thus an error will occur in the estimated rotor position. Therefore, if the q-axis inductance Lq, which fluctuates while the rotating machine is running, is not set to an appropriate value, position sensorless control can easily become unstable, which is a challenge. Against this backdrop, there is a demand for a technology that can estimate the rotor's magnetic flux or voltage and estimate the rotor position with high accuracy even if the q-axis inductance Lq fluctuates.
[0004] In contrast, for example, Patent Document 1 discloses a method for estimating magnetic flux or voltage using a value between the d-axis inductance Ld and the q-axis inductance Lq instead of the q-axis inductance Lq. The phase of the estimated magnetic flux or voltage will be shifted from the d-axis or q-axis, but the method is characterized by controlling it on a coordinate system that is deliberately phase-shifted from the d-axis or q-axis. As a result, the influence of the setting error of the q-axis inductance Lq can be reduced to a value between the d-axis inductance Ld and the q-axis inductance Lq.
[0005] Patent No. 4480696
[0006] However, while the configuration in Patent Document 1 can reduce the influence of setting errors in the q-axis inductance Lq compared to conventional designs, it is not easy to set the value between the d-axis inductance Ld and the q-axis inductance Lq to an appropriate value. Furthermore, since it is based on the premise of control on a coordinate system that is phase-shifted from the d-axis or q-axis, the position of the rotor is not known. Because this phase-shifted coordinate system cannot be detected by sensors, etc., when torque accuracy is required, there is a problem in that it is difficult to use the value between the d-axis inductance Ld and the q-axis inductance Lq as a parameter and adjust and optimize it.
[0007] This application discloses a technology for solving the above-mentioned problems, and aims to provide a control device for a rotating machine that can estimate the rotor position and speed with high accuracy even when the q-axis inductance Lq fluctuates.
[0008] A control device for a rotating machine disclosed herein comprises: a power conversion unit that converts the voltage of a power supply and applies an AC voltage to the rotating machine; a current detection unit that detects the current flowing from the power conversion unit to the rotating machine; a voltage command generation unit that generates a voltage command to be given to the power conversion unit based on an operation command to the rotating machine; a magnetic flux estimation unit that estimates the magnetic flux of the rotating machine; and a position / velocity estimation unit that estimates the position and velocity of the rotor of the rotating machine, wherein the magnetic flux estimation unit estimates the d-axis magnetic flux and q-axis magnetic flux on a first Cartesian coordinate system having a d-axis and a q-axis that are synchronized with the rotation of the rotating machine, and the magnetic flux of each axis on a second Cartesian coordinate system having axes different from the d-axis and the q-axis, and estimates the d-axis magnetic flux and the q-axis magnetic flux such that the dq-axis magnetic flux, which is the composite magnetic flux of the d-axis magnetic flux and the q-axis magnetic flux, matches the first composite magnetic flux due to the magnetic flux of each axis on the second Cartesian coordinate system, and the position / velocity estimation unit The phase difference between the phase of the dq-axis magnetic flux and the phase of the first composite magnetic flux is calculated, and the position and velocity of the rotor are estimated from the phase difference.
[0009] According to this invention, even if the q-axis inductance Lq fluctuates, the rotor position and speed can be estimated with high accuracy, making it possible to obtain a control device for a rotating machine that allows for easy adjustment and optimization of parameters.
[0010] This is a functional block diagram showing the configuration of the control device for a rotating machine according to Embodiment 1. This is a configuration diagram showing an example of the controller hardware of the control device for a rotating machine according to Embodiment 1. This is a functional block diagram showing the configuration of the magnetic flux estimation unit of the control device for a rotating machine according to Embodiment 1. This is a diagram showing the magnetic flux vector estimated by the magnetic flux estimation unit according to Embodiment 1. This is a functional block diagram showing the configuration of the position and speed estimation unit of the control device for a rotating machine according to Embodiment 1. This is a diagram showing the magnetic flux vector used to estimate the magnetic flux phase in the position and speed estimation unit according to Embodiment 1. This is a functional block diagram showing the configuration of the magnetic flux estimation unit of the control device for a rotating machine according to Embodiment 2. This is a functional block diagram showing the configuration of the position and speed estimation unit of the control device for a rotating machine according to Embodiment 2. This is a diagram showing the magnetic flux vector used to estimate the magnetic flux phase in the position and speed estimation unit according to Embodiment 2. This is a functional block diagram showing the configuration of the control device for a rotating machine according to Embodiment 3. This is a functional block diagram showing the configuration of the voltage estimation unit of the control device for a rotating machine according to Embodiment 3. This is a functional block diagram showing the configuration of the position and speed estimation unit of the control device for a rotating machine according to Embodiment 3. This is a functional block diagram showing the configuration of the voltage estimation unit of the control device for a rotating machine according to Embodiment 4. This is a functional block diagram showing the configuration of the position and speed estimation unit of the control device for a rotating machine according to Embodiment 4. This is a functional block diagram showing the configuration of the magnetic flux estimation unit of the control device for a rotating machine according to Embodiment 5. Variable L * This figure shows an example of the settings.
[0011] The embodiments will be described below with reference to the figures. In each figure, the same reference numerals indicate the same or corresponding parts. Therefore, detailed explanations of them may be omitted to avoid redundancy.
[0012] Embodiment 1. <Configuration of the Rotating Machine Control Device> The rotating machine control device according to Embodiment 1 will be described below with reference to the figures. Figure 1 is a functional block diagram showing the configuration of the rotating machine control device according to Embodiment 1 (hereinafter referred to as the "control device"). The control device 10 includes a controller 2, a power conversion unit 3 that receives a DC voltage from a power supply 20 and applies an AC voltage to the rotating machine 30 based on a voltage command output from the controller 2, and a current detection unit 4 that detects the AC current output from the power conversion unit 3.
[0013] The controller 2 includes a voltage command generation unit 5 that generates voltage commands based on operation commands, a magnetic flux estimation unit 6 that estimates magnetic flux based on the voltage commands generated by the voltage command generation unit 5, the alternating current detected by the current detection unit 4, and the operation commands, and a position / velocity estimation unit 7 that estimates the estimated position and estimated velocity of the rotor of the rotating machine 30 based on the magnetic flux estimated by the magnetic flux estimation unit 6. The operation of each functional unit will be described later.
[0014] Figure 2 shows an example of the hardware configuration of the controller 2. The controller 2 consists of a processor 100 and a storage device 110. Although not shown, the storage device 110 includes a volatile storage device such as random access memory and a non-volatile auxiliary storage device such as flash memory. An auxiliary storage device such as a hard disk may be provided instead of flash memory. The processor 100 executes the functions of each functional unit of the controller 2 by executing a program input from the storage device 110. In this case, the program is input from the auxiliary storage device to the processor 100 via the volatile storage device. The processor 100 may also output data such as calculation results to the volatile storage device of the storage device 110, or it may store the data in the auxiliary storage device via the volatile storage device.
[0015] The power conversion unit 3 is, for example, an inverter equipped with switching elements configured in a full-bridge circuit, and is connected to the power supply 20 and the rotating machine 30 by wiring. It generates an AC voltage from the DC voltage of the power supply 20 based on a voltage command output from the controller 2, and outputs the AC voltage to the rotating machine 30. Naturally, when the rotational speed of the rotating machine 30 approaches zero, the AC current and AC voltage output from the power conversion unit 3 will include not only an AC component but also a DC component. The power conversion unit 3 may also be a converter that substitutes for the operation of an inverter.
[0016] The current detection unit 4 detects the current of each phase of the rotating machine 30. For example, if the rotating machine 30 consists of three phases, U-phase, V-phase, and W-phase, the currents iu, iv, and iw are detected. The current detection unit 4 does not need to detect all phases; it may detect only some phases, or it may substitute predetermined current command values for all phases. Alternatively, it may substitute with current estimates obtained by predicting or estimating the current. Furthermore, it may be substituted with a configuration that detects the current on the power supply side and calculates the current of each phase. Current estimation is not only used as a substitute for current sensors to reduce costs. If used as a substitute when the accuracy or response of the sensor in the current detection unit 4 is low, it has advantages such as improving current detection accuracy and being able to observe transiently changing currents.
[0017] The rotating machine 30 is assumed to be an IPMSM (Internal Permanent Magnet Synchronous Motor) or SPMSM (Surface Permanent Magnet Synchronous Motor), which are three-phase synchronous motors having permanent magnets, or a SynRM (Synchronous Reluctance Motor) that does not have permanent magnets. However, it may also be a motor other than three phase, such as a double three-phase winding motor or a five-phase motor, or an open-winding motor or a delta-connected motor, or a field-wound motor with field windings on the rotor, or an SRM (Switched Reluctance Motor). In this application, the dq axis is defined to produce a reverse salient pole, but in the case of SynRM, the dq axis may be defined to produce a forward salient pole. This embodiment is an example in which an IPMSM is used as the rotating machine 30.
[0018] <Configuration of each functional part of controller 2> The voltage command generation unit 5 generates a voltage command corresponding to the input operation command. Operation commands are, for example, torque, rotational speed, current, voltage, and position. When generating a voltage command, rotational coordinate transformation is generally used, and the rotational coordinate transformation uses the estimated position estimated by the position / speed estimation unit 7. Also, when generating a voltage command for the rotating machine 30, the rotational speed is required because the induced voltage caused by the rotational speed is calculated. The rotational speed used in calculating the induced voltage is the estimated speed estimated by the position / speed estimation unit 7. Note that in Figure 1, the converters that perform rotational coordinate transformation, 2-axis to 3-axis coordinate transformation, etc., which are equipped in the control device 10 are omitted.
[0019] <Configuration of Magnetic Flux Estimation Unit 6> Figure 3 shows an example of the configuration of the magnetic flux estimation unit 6. The magnetic flux estimation unit 6 includes a γδ-axis magnetic flux estimator 62 that estimates the magnetic flux of the γδ axis based on the voltage command generated by the voltage command generation unit 5 and the alternating current detected by the current detection unit 4, and a dq-axis magnetic flux estimator 64 that estimates the magnetic flux of the dq axis based on the operation command, the magnetic flux φm of the rotor, and the magnetic flux of the γδ axis estimated by the γδ-axis magnetic flux estimator 62. Here, the voltage command from the voltage command generation unit 5 is input to the γδ-axis magnetic flux estimator 62, but if there is a voltage sensor, the voltage from the voltage sensor may be input.
[0020] <Mathematical Model of Rotating Machine 30> In order to explain the operation of the γδ-axis magnetic flux estimator 62 and the dq-axis magnetic flux estimator 64, we will explain the mathematical model of the rotating machine 30 used in these two magnetic flux estimators. When IPMSM is assumed to be the rotating machine 30, the voltage equation for the dq coordinate is expressed by the following equation (1).
[0021]
[0022] Here, vd and vq are the voltages along the d and q axes, R is the winding resistance, s is the differential operator, Ld is the d-axis inductance, Lq is the q-axis inductance, ed and eq are the rotor induced voltages, ωre is the electrical angular velocity, and φm is the rotor magnetic flux. For example, in the case of IPMSMs and SPMSMs, φm represents the magnetic flux of the magnet, and in the case of field-wound motors, φm represents the magnetic flux of the field winding. When IPMSMs and SPMSMs are equipped with field windings, φm represents the combined magnetic flux of the magnet and the field winding.
[0023] By performing a model transformation in which all inductance terms in the first term on the right-hand side of equation (1) are replaced with d-axis inductance Ld, the following equation (2) is obtained.
[0024]
[0025] Here, ed Ld , eq Ld This is the induced voltage when the model is deformed, and it is a mixture of magnetic flux created by the stator current and magnetic flux from the magnet. This induced voltage looks like the induced voltage generated in the rotor, so it is called a virtual induced voltage (virtual induced voltage).
[0026] When Equation (2) is transformed into the αβ coordinate which is a fixed coordinate, Equation (3) is obtained, and when Equation (2) is transformed into the γδ coordinate which is an estimated rotational coordinate, Equation (4) is obtained. Note that the α-axis of the fixed coordinate is represented as being located at a phase delayed by θre from the d-axis, and the γ-axis of the estimated rotational coordinate is represented as being located at a phase advanced by Δθre from the d-axis. Note that the αβ coordinate of the fixed coordinate, which is an orthogonal coordinate formed by the α-axis and the β-axis, is obtained by performing two-phase conversion with the three phases of the stator of the rotating machine as a reference. With the αβ coordinate as a reference, the rotational coordinate synchronized with the rotation of the rotating machine is the dq coordinate which is an orthogonal coordinate formed by the d-axis and the q-axis. With the αβ coordinate as a reference, the estimated rotational coordinate on the rotating magnetic field, which is synchronized with the rotation of the rotating machine, is the γδ coordinate which is an orthogonal coordinate formed by the γ-axis and the δ-axis.
[0027]
[0028]
[0029] Here, vα and vβ are voltages on the fixed coordinate, that is, voltages on the αβ coordinate, iα and iβ are currents on the fixed coordinate, eα Ld , eβ Ld are induced voltages on the fixed coordinate when model transformation is performed. Further, vγ and vδ are voltages on the estimated rotational coordinate, that is, voltages on the γδ coordinate, iγ and iδ are currents on the fixed coordinate, eγ Ld , eδ Ld are induced voltages on the estimated rotational coordinate when model transformation is performed. e Jθre , e -JΔθre denotes rotational coordinate transformation.
[0030] The d-axis magnetic flux when model transformation is performed is defined as φd Ld =φm, and the q-axis magnetic flux is defined as φq Ld =(Lq-Ld)iq, then Equation (5) is obtained by modifying Equation (3), and Equation (6) is obtained by modifying Equation (4). Here, I is an identity matrix.
[0031]
[0032]
[0033] In Equations (5) and (6), φα Ld , φβ Ldφγ is the magnetic flux on fixed coordinates when the model is deformed. Ld ,φδ Ld This represents the estimated magnetic flux on the rotating coordinate system when the model is deformed. Since these magnetic fluxes also appear to be magnetic fluxes generated in the rotor, they are called virtual rotor magnetic fluxes (virtual rotor magnetic fluxes). Performing such a deformation has the advantage that 2θre and 2Δθre do not occur before and after the coordinate transformation. In addition, the inductance included in the first term on the right-hand side of the voltage equation is limited to only the d-axis inductance Ld, which contributes to the effectiveness of this embodiment.
[0034] <Relationship between γδ-axis magnetic flux and dq-axis magnetic flux> The γδ-axis magnetic flux estimator 62 and the dq-axis magnetic flux estimator 64 of Embodiment 1 are based on the mathematical model of equation (6) and represent the magnetic flux φγ Ld ,φδ Ld , magnetic flux φd Ld , φq Ld We estimate the magnetic flux φγ. Before explaining each estimator, let's look at the mathematical model of equation (6) Ld ,φδ Ld , magnetic flux φd Ld , φq Ld The relationship will be explained using the vector diagram in Figure 4. Figure 4 shows the magnetic flux vector φγ in γδ coordinates and dq coordinates. Ld ,φδ Ld ,φd Ld , φq Ld This is a vector diagram showing the magnetic flux vector φγ. Ld ,φδ Ld The dotted line represents the magnetic flux vector φd. Ld , φq Ld This is indicated by a dashed line. As mentioned above, the γ axis of the estimated rotation coordinate is located in a phase advanced by Δθre from the d axis. In Figure 4, the vector φ all φγ is the magnetic flux vector Ld ,φδ Ld The amplitude and magnetic flux vector φd Ld , φq Ld This represents the amplitude, where the sign O is the origin and the sign A is the magnetic flux vector φd. Ld The coordinates and sign B are the magnetic flux vector φq. Ld ,φδ Ld The coordinates and sign C are the magnetic flux vector φγ. LdThis represents the coordinates. Also, θr is the magnetic flux vector φγ. Ld and magnetic flux vector φδ Ld This represents the phase.
[0035] <Operation of the γδ axis magnetic flux estimator 62> The γδ axis magnetic flux estimator 62 uses, for example, a state observer to determine the magnetic flux φγ Ld ,φδ Ld To estimate the following: For example, the state variables are estimated values of current iγ and iδ and magnetic flux φγ. Ld ,φδ Ld Using the estimated values, we construct a same-dimensional observer as shown in equation (7) below.
[0036]
[0037] Observer gain is a parameter designed considering steady-state and transient performance to achieve the desired specifications. Since only the d-axis inductance Ld is used in the calculation of equation (7), and the q-axis inductance Lq is not used, the magnetic flux φγ is not affected by fluctuations and setting errors of the q-axis inductance Lq. Ld ,φδ Ld An estimate can be obtained.
[0038] While an example using a same-dimensional observer as the state observer was shown, this is not the only option. A minimum-dimensional observer, disturbance observer, sliding observer, and unknown input observer may be used instead of the same-dimensional observer. Furthermore, a filtering process involving algebraic calculations and integration based on equation (6) may be used as a substitute for the state observer. Although the d-axis inductance Ld does not change as much as the q-axis inductance Lq, it does change depending on the operating conditions; therefore, a function or map may be used to change the value according to operating conditions such as torque, rotational speed, current, and voltage. Additionally, estimated values may be used.
[0039] <Operation of the dq-axis magnetic flux estimator 64> Next, the magnetic flux φd measured by the dq-axis magnetic flux estimator 64 Ld , φq Ld Let's explain the estimation method. First, the d-axis magnetic flux φd LdThis cannot be directly estimated during position sensorless control. However, as mentioned above, φd Ld From = φm, it is clear that there exists a vector on the d-axis with a length of φm. In other words, if the magnetic flux φm of the rotor is measured in advance, then the magnetic flux φd Ld An estimated value can be obtained. Here, the rotor's magnetic flux φm also varies depending on the operating point, that is, it changes according to the operating conditions, so a function or map may be used to change the value according to operating conditions such as torque, rotational speed, current, voltage, and temperature. Alternatively, the estimated value may be used. Furthermore, the magnetic flux φγ Ld The amplitude can be used as a substitute.
[0040] Next, the q-axis magnetic flux φq Ld It cannot be directly estimated during position sensorless control. However, referring to Figure 4, the q-axis magnetic flux φq Ld It is clear that the phase lies on the q-axis, and the q-axis magnetic flux φq Ld The amplitude can also be calculated. In Figure 4, focusing on the right triangle OAB and the right triangle OCB, the q-axis magnetic flux φq Ld Amplitude |φq Ld Since | is the length of side AB, |φq Ld |=√|φ all 2 -(φd Ld ) 2 |=√|φ all 2 -φm 2 It can be found by |. Also, since the hypotenuse OB of the two right triangles OAB and OCB is common, φ all 2 = (φγ Ld ) 2 +(φδ Ld ) 2 That is the case.
[0041] However, the q-axis magnetic flux φq Ld The amplitude is √|φ on the q-axis. all 2 -(φd Ld ) 2 It can be seen that it is a vector of the form |, but its positive and negative directions change depending on the q-axis current. In the above equation, the q-axis magnetic flux is φq Ldcan only obtain the amplitude, the q-axis magnetic flux φq Ld cannot be judged for its direction (positive or negative). Therefore, for example, by using the q-axis current command iq*, which is one of the operation commands, φq Ld =sign(iq*)√|φ all 2 -(φd Ld ) 2 |, it is possible to obtain the estimated value of φq Ld . Here, sign(P) is a sign function, which outputs 1 if P in the parentheses is positive, -1 if P is negative, and 0 if P is zero. That is, sign(iq*) outputs 1 if the q-axis current command iq* is positive, -1 if it is negative, and 0 if it is zero.
[0042] During position-sensorless control, iq* does not necessarily equal iq, but when the position estimation error is within ±90 [deg], sign(iq*) and sign(iq) match. Note that although an example using the q-axis current command iq* to determine the direction (positive or negative) of the q-axis magnetic flux φq Ld is shown, the present invention is not limited to this. Instead of the q-axis current command iq*, other operation commands may be used, or information related to the torque, current, electric power, and rotational speed of the rotating machine may be used. In the case of current, for example, the estimated value of the δ-axis current iδ or the q-axis current iq may be used.
[0043] As described above, the dq-axis magnetic flux estimator 64 can estimate the magnetic fluxes φd Ld , φq Ld . The outputs of the dq-axis magnetic flux estimator 64 are the estimated values of the magnetic fluxes φd Ld , φq Ld . Here, since the q-axis inductance Lq is not used even in the calculation of the estimated values of the magnetic fluxes φd Ld , φq Ld , there is no influence from fluctuations and setting errors of the q-axis inductance Lq.
[0044] <Configuration of Position / Velocity Estimation Unit 7> Next, the configuration of the position / velocity estimation unit 7 will be described. Figure 5 is a functional block diagram showing the configuration of the position / velocity estimation unit 7 according to Embodiment 1. The position / velocity estimation unit 7 includes a γδ-axis magnetic flux phase estimator 72, a dq-axis magnetic flux phase estimator 74, an adder-subtractor 76, and a position estimation error convergence calculator 78. The γδ-axis magnetic flux phase estimator 72 and the dq-axis magnetic flux phase estimator 74 use the γδ-axis magnetic flux φγ estimated by the γδ-axis magnetic flux estimator 62 Ld , φδ Ld to calculate the phase of the γδ-axis magnetic flux from the estimated values thereof, and calculate the phase of the dq-axis magnetic flux from the estimated values of the dq-axis magnetic flux φd estimated by the dq-axis magnetic flux estimator 64 Ld , φq Ld respectively. For the calculated phase of each axis magnetic flux, the adder-subtractor 76 obtains an estimated value of the phase difference Δθre, the calculated estimated value of the phase difference Δθre is input to the position estimation error convergence calculator 78, and an estimated position and an estimated velocity are calculated. Details are described below.
[0045] The γδ-axis magnetic flux phase estimator 72 performs calculation using equation (8), obtains the angle ∠COB of the right triangle OCB in Figure 4, and estimates the phase θγ.
[0046]
[0047] The dq-axis magnetic flux phase estimator 74 performs calculation using equation (9), obtains the angle ∠AOB of the right triangle OAB in Figure 4, and estimates the phase θγ+Δθre.
[0048]
[0049] The dq-axis magnetic flux phase estimator 74 performs calculation using equation (9), obtains the angle ∠AOB of the right triangle OAB in Figure 4, and estimates the phase θγ+Δθre. By subtracting equation (8) from equation (9) with the adder-subtractor 76, the phase difference Δθre between the d-axis and the γ-axis can be estimated. That is, the position of the rotor can be estimated. On the other hand, Figure 6 is a diagram showing the estimated values of each magnetic flux on the γδ coordinate. When observing the vectors of the estimated values of φγ Ld , φδ Ld , φd Ld , φq Ld on the γδ coordinate, the phase difference Δθre is the magnetic flux vector φγ Ld , φδLd Vector sum φγδ Ld and magnetic flux vector φd Ld , φq Ld The vector sum φdq Ld It can be expressed using the cross product and dot product of . That is, it can also be calculated as in equation (10).
[0050]
[0051] By calculating as shown in equation (10), the number of arctangent calculations can be reduced to one. Furthermore, while equations (8) through (10) calculate the phase using the arctangent, this is not the only method. Other trigonometric functions may be used, or other functions and approximation formulas may be substituted.
[0052] The position estimation error convergence calculator 78 performs operations such as PLL (Phase Locked Loop) processing, which includes PI (Proportional Integral) control, P (Proportional) control, I (Integral) control, PID (Proportional Integral Derivative) control, and PII. 2 Using calculations such as control (control using a controller with two I controllers connected in parallel to a PI controller) and I-P (Integral-Proportional) control, the rotor is controlled so that the estimated value of the phase difference Δθre calculated by the adder / subtractor 76 is zero or close to zero, and the estimated position and speed of the rotor are calculated. When the estimated phase difference Δθre is zero, it means that the γδ axis and the dq axis coincide. However, it is not always necessary to use PLL processing, and a configuration in which the estimated phase difference Δθre is added to or subtracted from the estimated phase θre is also acceptable. For example, Δθre is added to or subtracted from the estimated position one control cycle ago.
[0053] The voltage command generation unit 5 uses the estimated position and estimated speed of the rotor estimated by the position estimation error convergence calculator 78 of the position and speed estimation unit 7 to generate a voltage command based on the input operation command and output it to the power conversion unit 3.
[0054] As described above, in the process of estimating the rotor position and speed from the γδ axis magnetic flux and dq axis magnetic flux, the q axis inductance Lq is not used at all, so position sensorless control can be performed without being affected by fluctuations and setting errors of the q axis inductance. Furthermore, since this control device 10 can estimate the magnetic flux of the d axis or q axis and perform current vector control on the dq coordinate, it is easy to adjust and optimize parameters even in applications that require torque accuracy, such as automobiles and trains. In other words, in this embodiment 1, the voltage equation in the dq axis is transformed into a voltage equation in the γδ axis, which is the estimated rotation coordinate, by focusing on the inductance, and the d axis magnetic flux and q axis magnetic flux are estimated so that the dq axis magnetic flux and γδ axis magnetic flux included in the voltage equation in the dq axis and the voltage equation in the γδ axis, respectively, match. With this configuration, not only the d axis and q axis but also the magnetic flux of other axes are handled, and the d axis magnetic flux and q axis magnetic flux are estimated so that the combined magnetic flux of both matches, so the influence of the q axis inductance Lq on the inductance can be suppressed. In this embodiment 1, the q-axis inductance Lq is not used in the process of estimating the rotor position and velocity from the γδ-axis magnetic flux and dq-axis magnetic flux.
[0055] As described above, according to Embodiment 1, the control device for a rotating machine includes a power conversion unit that converts the voltage of a power supply and applies an AC voltage to the rotating machine, a current detection unit that detects the current flowing from the power conversion unit to the rotating machine, a voltage command generation unit that generates a voltage command to be given to the power conversion unit based on an operation command to the rotating machine, a magnetic flux estimation unit that estimates the magnetic flux of the rotating machine, and a position / velocity estimation unit that estimates the position and velocity of the rotor of the rotating machine. The magnetic flux estimation unit estimates the d-axis magnetic flux and q-axis magnetic flux on a first Cartesian coordinate system which has a d-axis and a q-axis and is synchronized with the rotation of the rotating machine, and the magnetic flux of each axis on a second Cartesian coordinate system which has axes different from the d-axis and q-axis, and estimates the d-axis magnetic flux and q-axis magnetic flux such that the dq-axis magnetic flux, which is the composite magnetic flux of the d-axis magnetic flux and the q-axis magnetic flux, matches the first composite magnetic flux due to the magnetic flux of each axis on the second Cartesian coordinate system. The position and velocity estimation unit is configured to calculate the phase difference between the phase of the dq-axis magnetic flux and the first combined magnetic flux, and to estimate the rotor position and velocity from the phase difference. With this configuration, the influence of the q-axis inductance Lq is suppressed during the process of estimating the rotor position and velocity from the estimation of the combined magnetic flux of each axis, so that the rotor position and velocity can be estimated with high accuracy. Therefore, because the influence of fluctuations and setting errors of the q-axis inductance Lq is small, it is possible to realize control of a rotating machine using position sensorless control with excellent responsiveness and stability. In addition, since current vector control can be performed on the dq coordinate by estimating the dq-axis magnetic flux, parameter adjustment and optimization are easier, which also contributes to reducing the adjustment man-hours for the control device.
[0056] Furthermore, in Embodiment 1, each axis on the second Cartesian coordinate system is an axis on a Cartesian coordinate system having a γ axis and a δ axis that are synchronized with the rotation of the rotating machine. The magnetic flux estimation unit estimates the γ axis magnetic flux and δ axis magnetic flux on a Cartesian coordinate system having a γ axis and a δ axis, and estimates the d axis magnetic flux and q axis magnetic flux so that the γδ axis magnetic flux, which is the composite magnetic flux of the γ axis magnetic flux and the δ axis magnetic flux, matches the dq axis magnetic flux. The position and velocity estimation unit calculates the phase difference between the phase of the γδ axis magnetic flux and the phase of the dq axis magnetic flux, and estimates the rotor position and velocity from the phase difference. Moreover, the calculations related to the estimation of the γδ axis magnetic flux and dq axis magnetic flux in the magnetic flux estimation unit use inductance, but can be performed without using the q axis inductance Lq. Therefore, it is possible to estimate the rotor position and velocity with high accuracy. Consequently, since it is not affected by fluctuations and setting errors of the q axis inductance Lq, it is possible to control the rotating machine by position sensorless control with excellent responsiveness and stability. Furthermore, by estimating the dq-axis magnetic flux, current vector control can be performed on the dq coordinate system, making it easier to adjust and optimize parameters, thus contributing to a reduction in the adjustment man-hours required for the control device.
[0057] Embodiment 2. Below, a control device for a rotating machine according to Embodiment 2 will be described with reference to the figures. In Embodiment 1, an example was described in which the magnetic flux estimation unit 6 is equipped with a γδ axis magnetic flux estimator 62 and a dq axis magnetic flux estimator 64, but in Embodiment 2, an αβ axis magnetic flux estimator is provided instead of the γδ axis magnetic flux estimator 62. In the following description, the differences from Embodiment 1 will be the main focus, and the same configuration and operation as in Embodiment 1 will be referred to in that description and will not be described here. The overall configuration of the control device for the rotating machine is the same as in Figure 1, but as will be described below, a magnetic flux estimation unit 6A is provided instead of the magnetic flux estimation unit 6, and a position and velocity estimation unit 7A is provided instead of the position and velocity estimation unit 7.
[0058] <Configuration and Operation of Magnetic Flux Estimation Unit 6A> Figure 7 is a functional block diagram showing the configuration of the magnetic flux estimation unit of the control device for a rotating machine according to Embodiment 2. In Figure 7, the magnetic flux estimation unit 6A includes an αβ axis magnetic flux estimator 63A and a dq axis magnetic flux estimator 64A. The αβ axis magnetic flux estimator 63A measures the magnetic flux φα on the fixed coordinates described in equation (5). Ld , φβLd The following is estimated. The estimation method can be the same as that used for the γδ axis magnetic flux estimator 62 in Embodiment 1, for example, by using a state observer.
[0059] Magnetic flux φd by dq-axis magnetic flux estimator 64A Ld , φq Ld The estimation method is the same as that of the dq-axis magnetic flux estimator 64 in Embodiment 1. d-axis magnetic flux φd Ld and q-axis magnetic flux φq Ld φd cannot be directly estimated during position sensorless control. Ld = φm φq Ld =sign(iq*)√|φ all 2 ―φm 2 | =sign(iq*)√|{(φα Ld ) 2 +(φβ Ld ) 2}―φm 2 | as the d-axis magnetic flux φd Ld and q-axis magnetic flux φq Ld We estimate each of these.
[0060] However, sign(iq*) is the vector φq Ld Since it is sufficient to determine the direction, the current command value iq* does not necessarily have to be used for the determination. Instead of the current command value iq*, information related to torque, other currents, power, rotational speed, etc., can be used to determine the vector φq Ld You may also determine its orientation.
[0061] As described above, in the magnetic flux estimation unit 6A, the magnetic flux φα Ld , φβ Ld and magnetic flux φd Ld , φq Ld Since the calculation used to estimate this value does not use the q-axis inductance Lq, it is not affected by fluctuations in the q-axis inductance or setting errors.
[0062] <Configuration and Operation of Position and Velocity Estimation Unit 7A> Next, the configuration of the position and velocity estimation unit 7A will be described. Figure 8 is a functional block diagram showing the configuration of the position and velocity estimation unit 7A according to Embodiment 2. The position and velocity estimation unit 7A includes an αβ-axis magnetic flux phase estimator 73A, a dq-axis magnetic flux phase estimator 74A, an adder / subtractor 76A, and a velocity estimator 77A.
[0063] <Regarding the relationship between αβ-axis magnetic flux and dq-axis magnetic flux> Before describing the αβ-axis magnetic flux phase estimator 73A and the dq-axis magnetic flux phase estimator 74A of Embodiment 2, let's first discuss the magnetic flux φα Ld , φβ Ld , magnetic flux φd Ld , φq Ld The relationship will be explained using the vector diagram in Figure 9. Figure 9 shows the magnetic flux vector φα in αβ coordinates and dq coordinates. Ld , φβ Ld ,φd Ld , φq Ld This is a vector diagram showing the magnetic flux vector φα. Ld , φβ Ld The dashed line represents the magnetic flux vector φd. Ld , φq Ld The dashed line indicates this. As mentioned above, the α-axis of the fixed coordinate system is located at a phase lag of θre from the d-axis. In Figure 9, the vector φ all φα is the magnetic flux vector Ld , φβ Ld The amplitude and magnetic flux vector φd Ld , φq Ld This represents the amplitude, where the sign O is the origin and the sign A is the magnetic flux vector φd. Ld The coordinates and sign B are the magnetic flux vector φq. Ld ,φδ Ld The coordinates and sign D are the magnetic flux vector φαγ. Ld This represents the coordinates. Also, θd is the magnetic flux vector φd. Ld and magnetic flux vector φq Ld This represents the phase.
[0064] The αβ-axis magnetic flux phase estimator 73A calculates the angle ∠DOB of the right triangle ODB in Figure 9 using equation (11) and estimates the phase θd + θre.
[0065]
[0066] The dq-axis magnetic flux phase estimator 74A calculates the angle ∠AOB of the right triangle OAB in Figure 9 using equation (12) and estimates the phase θd.
[0067]
[0068] By subtracting equation (12) from equation (11) using the adder / subtractor 76A, the phase difference θre between the d-axis and the α-axis can be estimated. That is, the estimated position of the rotor can be determined. Also, similar to Embodiment 1, the phase difference θre can be calculated as shown in equation (13).
[0069]
[0070] As shown in equation (13), the number of arctangent calculations can be reduced to one. Furthermore, while equations (11) through (13) calculate the phase using the arctangent, this is not the only method. Other trigonometric functions may be used, or other functions and approximation formulas may be substituted.
[0071] The velocity estimator 77A is composed of filters including differential processing, and obtains the estimated velocity by differentiating the phase difference θre, which is the estimated position calculated by the adder / subtractor 76A.
[0072] Based on the above, since the q-axis inductance Lq is not used at all in the process of estimating the rotor position and speed from the αβ-axis magnetic flux and dq-axis magnetic flux, position sensorless control can be performed without being affected by fluctuations and setting errors of the q-axis inductance.
[0073] As described above, Embodiment 2 provides the same effects as Embodiment 1. Specifically, in the control device for a rotating machine according to Embodiment 2, each axis on the Cartesian coordinate system having axes different from the d-axis and q-axis is an axis on the Cartesian coordinate system having α-axis and β-axis, which are coordinates based on the phase of the stator of the rotating machine. The magnetic flux estimation unit estimates the α-axis magnetic flux and β-axis magnetic flux on the Cartesian coordinate system having α-axis and β-axis, and estimates the d-axis magnetic flux and q-axis magnetic flux such that the αβ-axis magnetic flux, which is the composite magnetic flux of the α-axis magnetic flux and β-axis magnetic flux, matches the dq-axis magnetic flux. The position and velocity estimation unit calculates the phase difference between the phase of the αβ-axis magnetic flux and the phase of the dq-axis magnetic flux, and is configured to estimate the rotor position and velocity from this phase difference. As a result, in the process of estimating the rotor position and velocity from the estimation of the αβ-axis magnetic flux and dq-axis magnetic flux, calculations can be performed without using the q-axis inductance Lq, making it possible to estimate the rotor position and velocity with high accuracy. Therefore, since it is not affected by fluctuations in the q-axis inductance Lq and setting errors, it is possible to control rotating machinery using position sensorless control with excellent responsiveness and stability. In addition, since current vector control can be performed on the dq coordinate by estimating the dq-axis magnetic flux, parameter adjustment and optimization are easier, which contributes to reducing the adjustment man-hours for the control device.
[0074] Embodiment 3. Below, a control device for a rotating machine according to Embodiment 3 will be described with reference to the figures. Embodiments 1 and 2 described examples that included a magnetic flux estimation unit, but Embodiment 3 is equipped with a voltage estimation unit instead of a magnetic flux estimation unit. In the following description, the differences from Embodiments 1 and 2 will be explained in detail, and the same configuration and operation as in Embodiments 1 and 2 will be referred to in those descriptions and omitted here.
[0075] <Configuration of the Rotating Machine Control Device> Figure 10 is a functional block diagram showing the configuration of the rotating machine control device according to Embodiment 3. The control device 10B includes a controller 2B, a power conversion unit 3 that receives a DC voltage from a power supply 20 and applies an AC voltage to the rotating machine 30 based on a voltage command output from the controller 2B, and a current detection unit 4 that detects the AC current output from the power conversion unit 3. The controller 2B includes a voltage command generation unit 5B that generates a voltage command based on an operation command, a voltage estimation unit 8B that estimates a voltage based on the voltage command generated by the voltage command generation unit 5B and the AC current detected by the current detection unit 4, and a position / speed estimation unit 7B that estimates the estimated position and estimated speed of the rotor of the rotating machine 30 based on the voltage estimated by the voltage estimation unit 8B.
[0076] <Configuration and Operation of Voltage Estimation Unit 8B> Figure 11 shows an example of the configuration of the voltage estimation unit 8B. The voltage estimation unit 8B includes a γδ axis voltage estimator 82B that estimates the voltage of the γδ axis based on the voltage command generated by the voltage command generation unit 5B and the alternating current detected by the current detection unit 4, and a dq axis voltage estimator 84B that estimates the voltage of the dq axis based on the operation command, the magnetic flux φm of the rotor, and the voltage of the γδ axis estimated by the γδ axis voltage estimator 82B. Here, the voltage command from the voltage command generation unit 5B is input to the γδ axis voltage estimator 82B, but if there is a voltage sensor, the voltage from the voltage sensor may be input.
[0077] The γδ axis voltage estimator 82B is the induced voltage eγ on the γδ coordinate in equation (4). Ld , eδ Ld The following is estimated. The voltage commands generated by the voltage command generation unit 5B used are voltages vγ and vδ on the γδ coordinate system, and the AC currents detected by the current detection unit 4 are currents iγ and iδ on the γδ coordinate system. Voltage eγ Ld , eδ Ld The estimation of this is performed using methods such as the state observer described above.
[0078] The dq-axis voltage estimator 84B calculates the induced voltage ed on the dq coordinate. Ld , eq Ld We estimate the following. From the relationship in equation (2), the induced voltage eq Ld eq is the product of the electric angular velocity ωre and the magnetic flux φm of the rotor.Ld Let = ωreφm. Similarly, ed Ld =-sign(iq*)√|e all 2 -(ωreφm) 2 | =-sign(iq*)√||{(eγ Ld ) 2 +(eδ Ld ) 2} -(ωreφm) 2 | as voltage ed Ld , eq Ld We estimate that, however, sign(iq*) is ed Ld Since it is sufficient to determine the direction, instead of the q-axis current command iq*, information related to torque, other currents, power, and rotational speed is used to determine ed Ld The direction may be determined. Unlike the magnetic flux estimation in embodiments 1 and 2, the voltage estimation in embodiment 3 requires the electrical angular velocity ωre, and therefore is sensitive to, i.e., affected by, the rotational speed.
[0079] <Configuration and Operation of Position and Velocity Estimation Unit 7B> Next, the configuration of the position and velocity estimation unit 7B will be described. Figure 12 is a functional block diagram showing the configuration of the position and velocity estimation unit 7B according to Embodiment 3. The position and velocity estimation unit 7B includes a γδ axis voltage phase estimator 72B, a dq axis voltage phase estimator 74B, an adder / subtractor 76B, and a position estimation error convergence calculator 78B.
[0080] The γδ axis voltage phase estimator 72B calculates and estimates the phase θγ using equation (14). At this time, the vector diagram showing the relationship between the γδ axis magnetic flux and the dq axis magnetic flux in Figure 4 is expanded into a vector diagram showing the relationship between the γδ axis voltage and the dq axis voltage.
[0081]
[0082] The dq-axis voltage phase estimator 74B calculates the phase θγ + Δθre using equation (15).
[0083]
[0084] By subtracting equation (14) from equation (15) using the adder / subtractor 76B, the phase difference Δθre between the d-axis and the γ-axis can be estimated. That is, the estimated position of the rotor can be determined. Also, similar to embodiments 1 and 2, the phase difference Δθre can be calculated as shown in equation (16).
[0085]
[0086] By calculating as shown in equation (16), the number of arctangent calculations can be reduced to one. Furthermore, while equations (14) through (16) calculate the phase using the arctangent, this is not the only method. Other trigonometric functions may be used, or other functions and approximation formulas may be substituted.
[0087] The position estimation error convergence calculator 78B is the same as the position estimation error convergence calculator 78 shown in Embodiment 1, and uses PLL processing, for example, to control the estimated value of the phase difference Δθre calculated by the adder / subtractor 76B so that it becomes zero or a value near zero, and calculates the estimated position and estimated velocity of the rotor.
[0088] The voltage command generation unit 5B uses the estimated position and estimated speed of the rotor estimated by the position estimation error convergence calculator 78B of the position and speed estimation unit 7B to generate a voltage command based on the input operation command and output it to the power conversion unit 3.
[0089] As described above, in the process of estimating the rotor position and speed from the γδ axis voltage and dq axis voltage, the q axis inductance Lq is not used at all, so position sensorless control can be performed without being affected by fluctuations and setting errors of the q axis inductance. Furthermore, since this control device 10B can estimate the voltage of the d axis or q axis and perform current vector control on the dq coordinate, it is easy to adjust and optimize parameters even in applications that require torque accuracy, such as automobiles and trains. In other words, in this embodiment 3, the voltage equation in the dq axis is transformed into a voltage equation in the γδ axis, which is the estimated rotation coordinate, by focusing on the inductance, and the d axis voltage and q axis voltage are estimated so that the dq axis voltage and γδ axis voltage included in the voltage equation in the dq axis and the voltage equation in the γδ axis, respectively, match. With this configuration, not only the d axis and q axis but also the voltages of other axes are handled, and the d axis voltage and q axis voltage are estimated so that the combined voltage of both matches, so the influence of the q axis inductance Lq on the inductance can be suppressed. In this third embodiment, the q-axis inductance Lq is not used in the process of estimating the rotor position and speed from the γδ-axis voltage and dq-axis voltage.
[0090] As described above, Embodiment 3 provides the same effects as Embodiment 1. Specifically, the control device for a rotating machine according to Embodiment 3 includes a power conversion unit that converts the voltage of a power supply and applies an AC voltage to the rotating machine, a current detection unit that detects the current flowing from the power conversion unit to the rotating machine, a voltage command generation unit that generates a voltage command to be given to the power conversion unit based on an operation command to the rotating machine, a voltage estimation unit that estimates the voltage of the rotating machine, and a position and speed estimation unit that estimates the position and speed of the rotor of the rotating machine. The voltage estimation unit estimates the d-axis voltage and q-axis voltage flux on a first Cartesian coordinate system having a d-axis and a q-axis that are synchronized with the rotation of the rotating machine, and the voltages of each axis on a second Cartesian coordinate system having axes different from the d-axis and q-axis, and estimates the d-axis voltage and q-axis voltage such that the dq-axis voltage, which is the combined voltage of the d-axis voltage and the q-axis voltage, matches the first combined voltage of the voltages of each axis on the second Cartesian coordinate system. The position and velocity estimation unit is configured to calculate the phase difference between the phase of the dq-axis voltage and the first combined voltage, and to estimate the rotor position and velocity from the phase difference. With this configuration, the influence of the q-axis inductance Lq is suppressed during the process of estimating the rotor position and velocity from the combined voltage of each axis, so that the rotor position and velocity can be estimated with high accuracy. Therefore, because the influence of fluctuations and setting errors of the q-axis inductance Lq is small, it is possible to control a rotating machine using position sensorless control with excellent responsiveness and stability. In addition, since current vector control can be performed on the dq coordinate by estimating the dq-axis voltage, parameter adjustment and optimization are easier, which contributes to reducing the adjustment man-hours for the control device.
[0091] Furthermore, in Embodiment 3, each axis on the second Cartesian coordinate system is an axis on a Cartesian coordinate system having a γ axis and a δ axis that are synchronized with the rotation of the rotating machine. The voltage estimation unit estimates the γ axis voltage and δ axis voltage on a Cartesian coordinate system having a γ axis and a δ axis, and estimates the d axis voltage and q axis voltage so that the γδ axis voltage, which is the combined magnetic flux of the γ axis voltage and the δ axis voltage, matches the dq axis voltage. The position and velocity estimation unit calculates the phase difference between the phase of the γδ axis voltage and the phase of the dq axis voltage, and estimates the rotor position and velocity from the phase difference. Moreover, the calculations related to the estimation of the γδ axis voltage and dq axis voltage in the voltage estimation unit use inductance, but can be performed without using the q axis inductance Lq. Therefore, it becomes possible to estimate the rotor position and velocity with high accuracy. Consequently, since it is not affected by fluctuations and setting errors of the q axis inductance Lq, it is possible to control the rotating machine by position sensorless control with excellent responsiveness and stability.
[0092] Embodiment 4. Below, a control device for a rotating machine according to Embodiment 4 will be described with reference to the figures. In Embodiment 3, an example was described in which the voltage estimation unit 8B was equipped with a γδ axis voltage estimator 82B and a dq axis voltage estimator 84B, but in Embodiment 4, an αβ axis voltage estimator is provided instead of the γδ axis voltage estimator 82B. In the following description, the differences from Embodiment 3 will be the main focus, and the same configuration and operation as in Embodiments 1 to 3 will be referred to in those descriptions and will not be described here. The overall configuration of the control device for the rotating machine is the same as in Figure 10, but as will be described below, a voltage estimation unit 8C is provided instead of the voltage estimation unit 8B, and a position and speed estimation unit 7C is provided instead of the position and speed estimation unit 7B.
[0093] <Configuration and Operation of Voltage Estimation Unit 8C> Figure 13 is a functional block diagram showing the configuration of the voltage estimation unit of the control device for a rotating machine according to Embodiment 4. In Figure 13, the voltage estimation unit 8C includes an αβ axis voltage estimator 83C that estimates the voltage of the αβ axis based on the voltage command generated by the voltage command generation unit 5B and the alternating current detected by the current detection unit 4, and a dq axis voltage estimator 84C that estimates the voltage of the dq axis based on the operation command, the magnetic flux φm of the rotor, and the voltage of the αβ axis estimated by the αβ axis voltage estimator 83C. The voltage command from the voltage command generation unit 5B is input to the αβ axis voltage estimator 83C, but if a voltage sensor is available, the voltage from the voltage sensor may be input.
[0094] The αβ axis voltage estimator 83C calculates the induced voltage eα on the fixed αβ coordinate system in equation (3). Ld , eβ Ld The following is estimated. The voltage command generated by the voltage command generation unit 5B used is the voltage vα, vβ on the αβ coordinate system, and the alternating current detected by the current detection unit 4 is the current iα, iβ on the αβ coordinate system. Voltage eα Ld , eβ Ld The estimation of this is performed using methods such as the state observer described above.
[0095] The dq-axis voltage estimator 84C calculates the induced voltage ed on the dq coordinate. Ld , eq Ld This is estimated in the same manner as in Embodiment 3. That is, the induced voltage ed Ld ωre is the product of the electrical angular velocity and the magnetic flux φm of the rotor, eq Ld Let = ωreφm. Similarly, ed Ld =-sign(iq*)√|e all 2 -(ωreφm) 2 | =-sign(iq*)√|{(eα Ld ) 2 +(eβ Ld ) 2} -(ωreφm) 2 | as voltage ed Ld , eq Ld We estimate that, however, sign(iq*) is ed LdSince it is sufficient to determine the direction, other operation commands may be used instead of the q-axis current command iq*, or information related to the torque of the rotating machine, other currents, power, rotational speed, etc., may be used to determine ed Ld You may also determine its orientation.
[0096] <Configuration and Operation of Position and Velocity Estimation Unit 7C> Next, the configuration of the position and velocity estimation unit 7C will be described. Figure 14 is a functional block diagram showing the configuration of the position and velocity estimation unit 7C according to Embodiment 4. The position and velocity estimation unit 7C includes an αβ axis voltage phase estimator 73C, a dq axis voltage phase estimator 74C, an adder / subtractor 76C, and a velocity estimator 77C.
[0097] The αβ-axis voltage phase estimator 73C calculates the phase θd + θre using equation (17). At this time, the vector diagram showing the relationship between the αβ-axis magnetic flux and the dq-axis magnetic flux in Figure 9 is expanded into a vector diagram showing the relationship between the αβ-axis voltage and the dq-axis voltage.
[0098]
[0099] The αβ-axis voltage phase estimator 73C estimates the phase θd using equation (18). That is, the estimated position of the rotor can be determined.
[0100]
[0101] By subtracting equation (18) from equation (17) using the adder / subtractor 76C, the phase difference θre between the d-axis and the α-axis can be estimated. That is, the estimated position of the rotor can be determined. In addition, similar to embodiments 1 to 3, the phase difference θre can also be calculated as shown in equation (19).
[0102]
[0103] The speed estimator 77C is composed of filters including differential processing, and obtains the estimated speed by differentiating the phase difference θre, which is the estimated position calculated by the adder / subtractor 76C. The estimated position and estimated speed of the rotor estimated by the position / speed estimation unit 7C are output to the voltage command generation unit 5B in Figure 10.
[0104] Based on the above, since the q-axis inductance Lq is not used at all in the process of estimating the rotor position and speed from the αβ-axis voltage and dq-axis voltage, position sensorless control can be performed without being affected by fluctuations and setting errors of the q-axis inductance.
[0105] As described above, Embodiment 4 has the same effects as Embodiment 3. Specifically, in the control device for a rotating machine according to Embodiment 4, each axis on the Cartesian coordinate system having axes different from the d-axis and q-axis is an axis on the Cartesian coordinate system having α-axis and β-axis, which are coordinates based on the phase of the stator of the rotating machine. The voltage estimation unit estimates the α-axis voltage and β-axis voltage on the Cartesian coordinate system having α-axis and β-axis, and estimates the d-axis voltage and q-axis voltage so that the αβ-axis voltage, which is the combined voltage of the α-axis voltage and β-axis voltage, matches the dq-axis voltage. The position and velocity estimation unit calculates the phase difference between the phase of the αβ-axis voltage and the phase of the dq-axis voltage, and is configured to estimate the rotor position and velocity from this phase difference. As a result, in the process of estimating the rotor position and velocity from the estimation of the αβ-axis voltage and dq-axis voltage, calculations can be performed without using the q-axis inductance Lq, making it possible to estimate the rotor position and velocity with high accuracy. Therefore, since it is not affected by fluctuations in the q-axis inductance Lq and setting errors, it is possible to control rotating machinery using position sensorless control with excellent responsiveness and stability. In addition, since current vector control can be performed on the dq coordinate by estimating the dq-axis magnetic flux, parameter adjustment and optimization are easier, which contributes to reducing the adjustment man-hours for the control device.
[0106] In the embodiments 1 to 4 described above, each has an effect depending on how the coordinate axes are defined. As in embodiments 1 and 3, when γδ coordinates are used to estimate the γδ axis magnetic flux or γδ axis voltage, the magnetic flux or voltage can be treated as DC flow, which has the effect of being applicable even to high-speed rotation. On the other hand, as in embodiments 2 and 4, when αβ coordinates are used to estimate the αβ axis magnetic flux or αβ axis voltage, θre can be directly determined instead of Δθre, which also has the effect of improving the response of position estimation and velocity estimation.
[0107] Embodiment 5. Embodiments 1 to 4 described examples of rotor position estimation and velocity estimation without using the q-axis inductance Lq by model deformation focusing on inductance. Embodiment 5 describes an example in which the q-axis inductance Lq is used during model deformation.
[0108] The control device for a rotating machine according to Embodiment 5 will be described below with reference to the figures. In Embodiment 1, an example was described in which the magnetic flux estimation unit 6 is equipped with a γδ-axis magnetic flux estimator 62 and a dq-axis magnetic flux estimator 64. However, in Embodiment 5, a γδ-axis magnetic flux estimator 62D and a dq-axis magnetic flux estimator 64D that operate differently from Embodiment 1 are provided. In the following description, the differences from Embodiment 1 will be the main focus, and the same configuration and operation as in Embodiment 1 will be referred to therein and omitted here. The overall configuration of the control device for the rotating machine is the same as in Figure 1, but as will be described below, a magnetic flux estimation unit 6D is provided instead of the magnetic flux estimation unit 6.
[0109] <Configuration and Operation of Magnetic Flux Estimation Unit 6D> Figure 15 is a functional block diagram showing the configuration of the magnetic flux estimation unit in the control device for a rotating machine according to Embodiment 5. In Figure 15, the magnetic flux estimation unit 6D includes a γδ-axis magnetic flux estimator 62D and a dq-axis magnetic flux estimator 64D. In Embodiments 1 to 4, a model modification was performed in which all inductance terms in the first term on the right-hand side of equation (1) were replaced with d-axis inductance Ld. In Embodiment 5, all inductance terms in the first term on the right-hand side of equation (1) are replaced with a variable L corresponding to inductance. * Using this, and performing model transformations, we obtain the following equation (20).
[0110]
[0111] For example, the variable L in equation (20) * Substituting the d-axis inductance Ld into equation (2) yields equation (2). Converting equation (20) to the fixed coordinate system αβ coordinate system yields equation (21), and converting it to the estimated rotating coordinate system γδ coordinate system yields equation (22).
[0112]
[0113]
[0114] Here, the d-axis magnetic flux when the model is deformed is φd L* = φm + (Ld - L * ) id, q-axis magnetic flux φq L* = (Lq - L * By defining iq as such and rearranging equation (21), we obtain equation (23), and similarly, by rearranging equation (22), we obtain equation (24).
[0115]
[0116]
[0117] Note that variable L * The settings will be explained later.
[0118] <Operation of the γδ-axis magnetic flux estimator 62D> The γδ-axis magnetic flux estimator 62D receives a voltage command, current, and variable L. * The input is, for example, a state observer is used to determine the magnetic flux φγ L* ,φδ L* To estimate the following: For example, the state variables are estimated values of current iγ and iδ and magnetic flux φγ. Ld ,φδ Ld Using the estimated values, we construct a same-dimensional observer as shown in equation (25) below.
[0119]
[0120] <Operation of the dq-axis magnetic flux estimator 64D> An operation command is input to the dq-axis magnetic flux estimator 64D, and the d-axis magnetic flux φd L* = φm + (Ld - L * )id, q-axis magnetic flux φq L* = (Lq - L * ) Estimate iq. d-axis magnetic flux φd L* and q-axis magnetic flux φq L* The magnetic flux of the rotor φm, the currents id and iq flowing through the rotating machine, the current commands id* and iq* which are one of the operating commands, and the inductances Ld, Lq, L * Using either of the following, d-axis magnetic flux φdL* , q-axis magnetic flux φq L* We just need to estimate that.
[0121] <Estimation of position and velocity> In Embodiment 1, the variable is magnetic flux φd. Ld , φq Ld , φγ Ld , φδ Ld In this embodiment 5, the magnetic flux φd L* , φq L* , φγ L* , φδ L* Although they differ, the estimated magnetic flux relationship in this embodiment 5 is consistent with the relationship diagram of the magnetic flux vectors of the dq axis and the γδ axis shown in Figure 4. Therefore, just as equations (9) to (11) were shown in embodiment 1, this embodiment 5 can be described as equations (26) to (28).
[0122]
[0123]
[0124]
[0125] In this embodiment 5, the position and velocity estimation unit is not shown, but it has the same configuration as the position and velocity estimation unit 7 shown in Figure 5 of embodiment 1. Therefore, by subtracting equation (26) from equation (27), the phase difference Δθre between the d axis and the γ axis can be estimated. That is, the estimated position of the rotor can be determined. Alternatively, the rotor position can be estimated using equation (28). Furthermore, the dq axis magnetic flux estimator 64D may be configured to directly obtain phase information that becomes equation (27) instead of magnetic flux.
[0126] <Variable L corresponding to inductance> * Settings > Variable L * Depending on the settings, the d-axis magnetic flux φd may be detected in the dq-axis magnetic flux estimator 64D. L* and the q-axis magnetic flux φq L* Since the q-axis inductance Lq is used in the estimation process, there is a risk that the rotor position error will become large due to the influence of the highly sensitive q-axis inductance Lq. For example, L *Setting Lq to = results in the widely known extended flux model, the ACTIVE FLUX model, where the influence of errors in the q-axis inductance Lq becomes large. Therefore, the variable L * For example, use the values shown in Figure 16.
[0127] Figure 16 shows examples of setting inductance command values, with seven examples shown. As in the seven examples from No. 1 to No. 7, the variable L is set using the d-axis inductance Ld, the q-axis inductance Lq, or both. * Set the sign of variable L to be negative. * If the sign of is set to be negative, then when an error occurs due to the q-axis inductance Lq, the error that occurs in the estimation process of equation (26) and the error that occurs in the estimation process of equation (27) cancel each other out, so the sensitivity of the q-axis inductance Lq to position estimation can be made as small as possible.
[0128] Another example of a setting is the variable L * Even if the sign of is positive, setting it to a value sufficiently larger than the original d-axis inductance Ld and q-axis inductance Lq will reduce the impact of errors, so it is acceptable to set it in that way. Note that the variable L * L is a variable that corresponds to inductance. * Although explained as such, values other than inductance, such as rotor flux, may also be used. However, depending on the setting of the coefficients multiplied by the d-axis inductance Ld and q-axis inductance Lq, the control may become oscillatory or the response may become slow, so the variable L * It is desirable to adjust this by using decimal values as well as integers. By doing so, the influence of setting errors in the q-axis inductance Lq can be minimized.
[0129] As described above, Embodiment 5 has the same effects as Embodiment 1. Specifically, the control device for a rotating machine according to Embodiment 5 is configured such that the magnetic flux estimation unit estimates the d-axis magnetic flux and q-axis magnetic flux on a first Cartesian coordinate system which has a d-axis and a q-axis that are synchronized with the rotation of the rotating machine, and the γ-axis and δ-axis magnetic flux on a second Cartesian coordinate system which has a γ-axis and a δ-axis that are different axes from the d-axis and q-axis and are synchronized with the rotation of the rotating machine, and estimates the d-axis magnetic flux and q-axis magnetic flux such that the dq-axis magnetic flux, which is the combined magnetic flux of the d-axis magnetic flux and the q-axis magnetic flux, and the γδ-axis magnetic flux, which is the combined magnetic flux of the γ-axis and δ-axis, coincides, and the position and velocity estimation unit is configured to calculate the phase difference between the phase of the dq-axis magnetic flux and the phase of the γδ-axis magnetic flux, and estimate the position and velocity of the rotor from the phase difference. During magnetic flux estimation, the inductance variable used was either a negative inductance value calculated using the d-axis inductance, the q-axis inductance, or both, or a positive inductance value greater than both the d-axis and q-axis inductances. This minimizes the influence of the q-axis inductance Lq during the process of estimating the rotor's position and velocity, enabling accurate estimation of the rotor's position and velocity. Consequently, because the system is unaffected by fluctuations and setting errors in the q-axis inductance Lq, it is possible to achieve highly responsive and stable control of a rotating machine using position sensorless control.
[0130] In this embodiment 5, when estimating the γ-axis magnetic flux and δ-axis magnetic flux on an orthogonal coordinate system having a γ-axis and a δ-axis that are synchronized with the rotation of the rotating machine, and the d-axis magnetic flux and q-axis magnetic flux on an orthogonal coordinate system having a d-axis and a q-axis that are synchronized with the rotation of the rotating machine, the inductance is set as the variable L. * An example of estimating the position of a rotating machine by replacing it with has been described, but it is not limited to this. As in Embodiment 2, when estimating the α-axis magnetic flux and β-axis magnetic flux on a Cartesian coordinate system having an α-axis and a β-axis based on the phase of the stator of the rotating machine, and the d-axis magnetic flux and q-axis magnetic flux on a Cartesian coordinate system having a d-axis and a q-axis synchronized with the rotation of the rotating machine, the variable L * You may also use this. Furthermore, in a configuration that estimates the voltage of each axis as in Embodiments 3 and 4, the variable L *You may also use the variable L in embodiments 2 to 4. * It goes without saying that using the same method as in Embodiment 5 will produce the same effects.
[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 art 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] 2, 2B: Controller, 3: Power conversion unit, 4: Current detection unit, 5, 5B: Voltage command generation unit, 6, 6A, 6D: Magnetic flux estimation unit, 62, 62D: γδ axis magnetic flux estimator, 63A: αβ axis magnetic flux estimator, 64, 64A, 64D: dq axis magnetic flux estimator, 7, 7A, 7B, 7C: Position / velocity estimation unit, 72: γδ axis magnetic flux phase estimator, 72B: γδ axis voltage phase estimator, 73A: αβ axis magnetic flux phase estimator, 73C: αβ axis voltage phase estimator, 74, 74A: dq axis magnetic flux phase estimator, 74B, 74C: dq axis voltage phase estimator, 76, 76A, 76B, 76C: Addition / subtraction unit, 77A, 77C: Velocity estimator, 78, 78B: Position estimation error convergence calculator, 8B, 8C: Voltage estimation unit, 82B: γδ axis voltage estimator, 83C: αβ axis voltage estimator, 84B, 84C: dq axis voltage estimator, 10, 10B: Control device, 20: Power supply, 30: Rotating machine, 100: Processor, 110: Memory device.
Claims
1. A control device for a rotating machine comprising: a power conversion unit that converts the voltage of a power supply and applies an AC voltage to the rotating machine; a current detection unit that detects the current flowing from the power conversion unit to the rotating machine; a voltage command generation unit that generates a voltage command to be given to the power conversion unit based on an operation command to the rotating machine; a magnetic flux estimation unit that estimates the magnetic flux of the rotating machine; and a position / velocity estimation unit that estimates the position and velocity of the rotor of the rotating machine, wherein the magnetic flux estimation unit estimates the d-axis magnetic flux and q-axis magnetic flux on a first orthogonal coordinate system having a d-axis and a q-axis that are synchronized with the rotation of the rotating machine, and the magnetic flux of each axis on a second orthogonal coordinate system having axes different from the d-axis and the q-axis, and estimates the d-axis magnetic flux and the q-axis magnetic flux such that the dq-axis magnetic flux, which is the composite magnetic flux of the d-axis magnetic flux and the q-axis magnetic flux, matches the first composite magnetic flux due to the magnetic flux of each axis on the second orthogonal coordinate system, and the position / velocity estimation unit, A control device for a rotating machine, which calculates the phase difference between the phase of the dq-axis magnetic flux and the phase of the first composite magnetic flux, and estimates the position and speed of the rotor from the phase difference.
2. Each axis on the second Cartesian coordinate system is an axis on a Cartesian coordinate system having a γ axis and a δ axis that are synchronized with the rotation of the rotating machine; the magnetic flux estimation unit estimates the γ axis magnetic flux and the δ axis magnetic flux on a Cartesian coordinate system having the γ axis and the δ axis, and estimates the d axis magnetic flux and the q axis magnetic flux such that the γδ axis magnetic flux, which is the composite magnetic flux of the γ axis magnetic flux and the δ axis magnetic flux, coincides with the dq axis magnetic flux; the position and velocity estimation unit calculates the phase difference between the phase of the γδ axis magnetic flux and the phase of the dq axis magnetic flux, and estimates the position and velocity of the rotor from the phase difference, the control device for a rotating machine according to claim 1.
3. The control device for a rotating machine according to claim 2, wherein the magnetic flux estimation unit estimates the γ-axis magnetic flux and the δ-axis magnetic flux based on the voltage command generated by the voltage command generation unit, the current flowing through the rotating machine detected by the current detection unit, and the inductance of the rotating machine.
4. The control device for a rotating machine according to claim 2 or 3, wherein the magnetic flux estimation unit estimates the d-axis magnetic flux based on the magnetic flux of the rotor, and estimates the q-axis magnetic flux based on the γ-axis magnetic flux, the δ-axis magnetic flux and the magnetic flux of the rotor, or estimates the d-axis magnetic flux and the q-axis magnetic flux based on the magnetic flux of the rotor, the current flowing through the rotating machine, a current command and / or inductance.
5. The control device for a rotating machine according to claim 1, wherein each axis on the second Cartesian coordinate system is an axis on a Cartesian coordinate system having an α-axis and a β-axis, which are coordinates based on the phase of the stator of the rotating machine, the magnetic flux estimation unit estimates the α-axis magnetic flux and the β-axis magnetic flux on a Cartesian coordinate system having the α-axis and the β-axis, and estimates the d-axis magnetic flux and the q-axis magnetic flux such that the αβ-axis magnetic flux, which is the composite magnetic flux of the α-axis magnetic flux and the β-axis magnetic flux, coincides with the dq-axis magnetic flux, and the position and velocity estimation unit calculates the phase difference between the phase of the αβ-axis magnetic flux and the phase of the dq-axis magnetic flux, and estimates the position and velocity of the rotor from the phase difference.
6. The control device for a rotating machine according to claim 5, wherein the magnetic flux estimation unit estimates the α-axis magnetic flux and the β-axis magnetic flux based on the voltage command generated by the voltage command generation unit, the current flowing through the rotating machine detected by the current detection unit, and the inductance of the rotating machine.
7. The control device for a rotating machine according to claim 5 or 6, wherein the magnetic flux estimation unit estimates the d-axis magnetic flux based on the magnetic flux of the rotor, and estimates the q-axis magnetic flux based on the α-axis magnetic flux, the β-axis magnetic flux and the magnetic flux of the rotor, or estimates the d-axis magnetic flux and the q-axis magnetic flux based on the magnetic flux of the rotor, the current flowing through the rotating machine, a current command and / or inductance.
8. The control device for a rotating machine according to claim 3 or 6, wherein the magnetic flux estimation unit estimates the magnetic flux using, as the inductance of the rotating machine, a negative inductance value calculated using the d-axis inductance, the d-axis inductance, or the q-axis inductance, or both, or a positive inductance value greater than the d-axis inductance and the q-axis inductance.
9. The control device for a rotating machine according to any one of claims 1 to 8, wherein the magnetic flux estimation unit estimates the sign of the q-axis magnetic flux using one of the following: an operation command to the rotating machine, a current flowing through the rotating machine, the power of the rotating machine, and the rotational speed of the rotating machine.
10. A control device for a rotating machine comprising: a power conversion unit that converts the voltage of a power supply and applies an AC voltage to the rotating machine; a current detection unit that detects the current flowing from the power conversion unit to the rotating machine; a voltage command generation unit that generates a voltage command to be given to the power conversion unit based on an operation command to the rotating machine; a voltage estimation unit that estimates the voltage of the rotating machine; and a position / velocity estimation unit that estimates the position and velocity of the rotor of the rotating machine, wherein the voltage estimation unit estimates the d-axis voltage and q-axis voltage on a first Cartesian coordinate system having a d-axis and a q-axis that are synchronized with the rotation of the rotating machine, and the voltages of each axis on a second Cartesian coordinate system having axes different from the d-axis and the q-axis, and estimates the d-axis voltage and the q-axis voltage such that the dq-axis voltage, which is the combined voltage of the d-axis voltage and the q-axis voltage, matches a first combined voltage of the voltages of each axis on the second Cartesian coordinate system, and the position / velocity estimation unit calculates the phase difference between the phase of the dq-axis voltage and the phase of the first combined voltage, A control device for a rotating machine that estimates the position and speed of the rotor from the phase difference.
11. The control device for a rotating machine according to claim 10, wherein each axis on the second Cartesian coordinate system is an axis on a Cartesian coordinate system having a γ axis and a δ axis that are synchronized with the rotation of the rotating machine, the voltage estimation unit estimates the γ axis voltage and the δ axis voltage on a Cartesian coordinate system having the γ axis and the δ axis, and estimates the d axis voltage and the q axis voltage such that the γδ axis voltage, which is the combined magnetic flux of the γ axis voltage and the δ axis voltage, matches the dq axis voltage, and the position and velocity estimation unit calculates the phase difference between the phase of the γδ axis voltage and the phase of the dq axis voltage, and estimates the position and velocity of the rotor from the phase difference.
12. The control device for a rotating machine according to claim 11, wherein the voltage estimation unit estimates the γ-axis voltage and the δ-axis voltage based on the voltage command generated by the voltage command generation unit, the current flowing through the rotating machine detected by the current detection unit, and the inductance of the rotating machine.
13. The control device for a rotating machine according to claim 11 or 12, wherein the voltage estimation unit estimates the d-axis voltage based on the magnetic flux of the rotor, and estimates the q-axis voltage based on the γ-axis voltage, the δ-axis voltage and the magnetic flux of the rotor, or estimates the d-axis voltage and the q-axis voltage based on any of the magnetic flux of the rotor, the current flowing through the rotating machine, a current command and / or inductance.
14. The control device for a rotating machine according to claim 10, wherein each axis on the second Cartesian coordinate system is an axis on a Cartesian coordinate system having an α-axis and a β-axis, which are coordinates based on the phase of the stator of the rotating machine, the voltage estimation unit estimates the α-axis voltage and the β-axis voltage on a Cartesian coordinate system having the α-axis and the β-axis, and estimates the d-axis voltage and the q-axis voltage such that the αβ-axis voltage, which is the combined voltage of the α-axis voltage and the β-axis voltage, matches the dq-axis voltage, and the position and velocity estimation unit calculates the phase difference between the phase of the αβ-axis voltage and the phase of the dq-axis voltage, and estimates the position and velocity of the rotor from the phase difference.
15. The control device for a rotating machine according to claim 14, wherein the voltage estimation unit estimates the α-axis voltage and the β-axis voltage based on the voltage command generated by the voltage command generation unit, the current flowing through the rotating machine detected by the current detection unit, and the inductance of the rotating machine.
16. The control device for a rotating machine according to claim 14 or 15, wherein the voltage estimation unit estimates the d-axis voltage based on the magnetic flux of the rotor, and estimates the q-axis voltage based on the α-axis voltage, the β-axis voltage and the magnetic flux of the rotor, or estimates the d-axis voltage and the q-axis voltage based on any of the magnetic flux of the rotor, the current flowing through the rotating machine, a current command and / or inductance.
17. The control device for a rotating machine according to claim 12 or 15, wherein the voltage estimation unit estimates the voltage using, as the inductance of the rotating machine, a negative inductance value calculated using the d-axis inductance, the d-axis inductance, or the q-axis inductance, or both, or a positive inductance value greater than the d-axis inductance and the q-axis inductance.
18. The control device for a rotating machine according to any one of claims 10 to 17, wherein the voltage estimation unit estimates the sign of the q-axis voltage using any one of the following: an operation command to the rotating machine, a current flowing through the rotating machine, the power of the rotating machine, and the rotational speed of the rotating machine.
