Control device for rotary machine

The control device for rotating machines addresses sensitivity issues in position sensorless control by using synchronized Cartesian coordinate systems to estimate rotor position and speed accurately, enhancing torque accuracy.

WO2026240194A1PCT designated stage Publication Date: 2026-11-19MITSUBISHI ELECTRIC CORP +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
PCT/JP2026/018420
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

Technical Problem

Existing position sensorless control methods for permanent magnet synchronous motors face challenges in accurately estimating rotor position due to high sensitivity to q-axis inductance, making it difficult to adjust sensitivity according to motor operating conditions, which affects torque accuracy.

Method used

A control device that estimates rotor position and speed by synchronizing a first Cartesian coordinate system with the rotating machine's d-axis and q-axis, and a second Cartesian coordinate system with different parameters, using a magnetic flux estimation unit to calculate phase differences, allowing for high accuracy in rotor position and speed estimation.

Benefits of technology

Enables accurate adjustment and optimization of rotor position and speed estimation by distributing sensitivity across multiple parameters, reducing errors and improving torque accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure JP2026018420_19112026_PF_FP_ABST
    Figure JP2026018420_19112026_PF_FP_ABST
Patent Text Reader

Abstract

In this control device (10, 10B, 10D) for a rotary machine, a magnetic flux estimation unit (6, 6A, 6D, 6E) for estimating the magnetic flux of a rotary machine estimates at least one set of a d-axis magnetic flux and a q-axis magnetic flux in a first orthogonal coordinate system of a d-axis and a q-axis, and estimates a plurality of sets of magnetic flux on each axis in a second orthogonal coordinate system with axes different from the d-axis and the q-axis on the basis of different parameters of the rotary machine. A position / speed estimation unit (7, 7A, 7B, 7C, 7D, 7E) synthesizes a set of the d-axis magnetic flux and the q-axis magnetic flux and a plurality of sets of magnetic fluxes in the second orthogonal coordinate system using a weighting coefficient, calculates a phase difference between the first orthogonal coordinate system and the second orthogonal coordinate system on the basis of the synthesized magnetic fluxes, and estimates the position and speed of the rotary machine.
Need to check novelty before this filing date? Find Prior Art

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. This requires information on the q-axis inductance Lq of a permanent magnet synchronous motor, but the sensitivity to the q-axis inductance Lq is very high among the parameters of a permanent magnet synchronous motor. In other words, 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 magnetic flux or voltage to be estimated, and thus an error is likely to occur in the estimated rotor position.

[0004] From a control system design perspective, concentrating sensitivity on only one parameter is undesirable in terms of robustness to characteristic variations of permanent magnet synchronous motors. Ideally, the sensitive parameters should be distributed as much as possible, while allowing the sensitivity of the parameters to be arbitrarily changed according to the permanent magnet synchronous motor used and its operating conditions.

[0005] 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 sensitivity of the q-axis inductance Lq can be reduced to a value between the d-axis inductance Ld and the q-axis inductance Lq.

[0006] Patent No. 4480696

[0007] However, in the configuration of Patent Document 1, although the sensitivity of the q-axis inductance Lq can be reduced to a value between the d-axis inductance Ld and the q-axis inductance Lq, the range of sensitivity is fixed, and it is not easy to set the sensitivity appropriately. In other words, it is difficult to arbitrarily adjust the sensitivity according to the operating point or motor. Furthermore, although it is assumed that control is performed on a coordinate that is phase-shifted from the d-axis or q-axis, it is not possible to detect that coordinate with a sensor or the like, so when torque accuracy is required, 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. That is to say, the conventional technology had the problem that it was not possible to arbitrarily change the sensitivity of the parameter according to the permanent magnet synchronous motor used and its operating conditions.

[0008] 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 by adjusting the sensitivity of multiple parameters of the rotating machine.

[0009] 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 at least one pair of d-axis magnetic flux and q-axis magnetic flux on a first Cartesian coordinate system having a d-axis and a q-axis that is synchronized with the rotation of the rotating machine, and estimates multiple pairs of magnetic flux for each axis on a second Cartesian coordinate system having axes different from the d-axis and q-axis based on different parameters of the rotating machine; and the position / velocity estimation unit synthesizes at least one pair of d-axis magnetic flux and q-axis magnetic flux and multiple pairs of magnetic flux for each axis on the second Cartesian coordinate system using a preset weight coefficient. Based on the combined magnetic flux, the phase difference between the first orthogonal coordinate and the second orthogonal coordinate is calculated, and the position and velocity of the rotor are estimated from the phase difference.

[0010] According to this invention, by adjusting the sensitivity of multiple parameters of the rotating machine, the position and speed of the rotor can be estimated with high accuracy, making it possible to obtain a control device for a rotating machine that is easy to adjust and optimize.

[0011] 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 diagram showing the magnetic flux vector estimated by the magnetic flux estimation unit according to Embodiment 1. This is a diagram showing the operation of the weight setter according to Embodiment 1. This is a functional block diagram showing the configuration of the position and velocity estimation unit of the control device for a rotating machine according to Embodiment 1. This is a diagram showing the magnetic flux vector used for estimating the magnetic flux phase in the position and velocity estimation unit according to Embodiment 1. This is a vector diagram explaining the operation of the magnetic flux combiner according to Embodiment 1. This is a diagram showing the sensitivity of the rotating machine parameters used for estimating the first magnetic flux according to Embodiment 1. This is a diagram showing the sensitivity of the rotating machine parameters used for estimating the second magnetic flux according to Embodiment 1. This is a diagram showing the sensitivity of the rotating machine parameters used for estimating the magnetic flux obtained by combining the first and second magnetic fluxes 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 diagram showing the magnetic flux vector estimated by the magnetic flux estimation unit according to Embodiment 2. This is a diagram showing the magnetic flux vector estimated by the magnetic flux estimation unit 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 vector diagram illustrating the operation of the magnetic flux combiner 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 control device for a rotating machine according to Embodiment 5. 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 5. 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 6.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 6. Variable L. *1 , L *2 This figure shows an example of the settings.

[0012] 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 these parts may be omitted to avoid redundancy.

[0013] 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.

[0014] 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 / speed estimation unit 7 that estimates the estimated position and estimated speed 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.

[0015] 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.

[0016] 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.

[0017] 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.

[0018] 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.

[0019] <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.

[0020] <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 first γδ-axis magnetic flux estimator 61 that estimates the magnetic flux of the γδ axis based on a voltage command generated by the voltage command generation unit 5 and an alternating current detected by the current detection unit 4, a dq-axis magnetic flux estimator 62 that estimates the magnetic flux of the dq axis based on an operation command, the magnetic flux of the rotor φm and the magnetic flux of the γδ axis estimated by the first γδ-axis magnetic flux estimator 61, and a second γδ-axis magnetic flux estimator 63 that estimates the magnetic flux of the γδ axis based on a voltage command generated by the voltage command generation unit 5 and an alternating current detected by the current detection unit 4. Here, the voltage command from the voltage command generation unit 5 is input to the first γδ-axis magnetic flux estimator 61 and the second γδ-axis magnetic flux estimator 63, but if there is a voltage sensor, the voltage from the voltage sensor may be input. Furthermore, the first γδ-axis magnetic flux estimator 61 estimates the magnetic flux in the γδ axis using the d-axis inductance Ld, and the second γδ-axis magnetic flux estimator 63 estimates the magnetic flux in the γδ axis using the q-axis inductance Lq. Details will be described later.

[0021] <Mathematical Model 1 of Rotating Machine 30> In order to explain the operation of the first γδ-axis magnetic flux estimator 61 and the dq-axis magnetic flux estimator 62, we will explain the mathematical model of the rotating machine 30 used in these two magnetic flux estimators. When IPMSM is assumed as the rotating machine 30, the voltage equation for the dq coordinate is expressed by the following equation (1).

[0022]

[0023] 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.

[0024] 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.

[0025]

[0026] Here, ed Ld , eq Ld are induced voltages when model transformation is performed, and take a form in which the magnetic flux generated by the stator current and the magnet magnetic flux are mixed. Since this induced voltage appears like the induced voltage generated in the rotor, it is referred to as a virtual induced voltage (virtual induced voltage). Specifically, since it is a virtual induced voltage by model transformation using the d-axis inductance Ld, it is referred to as a virtual induced voltage by the Ld model.

[0027] 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 expressed as being positioned at a phase delayed by θre from the d-axis, and the γ-axis of the estimated rotational coordinate is expressed as being positioned at a phase advanced by Δθre from the d-axis. Note that the αβ coordinate, which is a fixed coordinate formed by the orthogonal coordinates of 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 the orthogonal coordinate of 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 the orthogonal coordinate of the γ-axis and the δ-axis.

[0028]

[0029]

[0030] Here, vα, vβ are voltages on the fixed coordinate, that is, voltages on the αβ coordinate, and iα, iβ are currents on the fixed coordinate, eα Ld , eβ Ld is an induced voltage on the fixed coordinate when model transformation is performed. Further, vγ, vδ are voltages on the estimated rotational coordinate, that is, voltages on the γδ coordinate, iγ, iδ are currents on the fixed coordinate, eγ Ld , eδ Ld is an induced voltage on the estimated rotational coordinate when model transformation is performed. e Jθre , e -JΔθre means rotational coordinate transformation.

[0031] 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 LdDefining this as (Lq - Ld)iq, and rearranging equation (3) yields equation (5), and rearranging equation (4) yields equation (6), where I is the identity matrix.

[0032]

[0033]

[0034] In equations (5) and (6), φα Ld , φβ Ld φγ 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 fluxes (virtual rotor fluxes). More specifically, since they are virtual rotor fluxes resulting from model deformation using the d-axis inductance Ld, they are called virtual rotor fluxes based on the Ld model. Performing such deformation has the advantage that 2θre and 2Δθre do not occur before and after the coordinate transformation. Another feature is that the inductance included in the first term on the right-hand side of the voltage equation is only the d-axis inductance Ld.

[0035] <Relationship between γδ-axis magnetic flux and dq-axis magnetic flux> The first γδ-axis magnetic flux estimator 61 and the dq-axis magnetic flux estimator 62 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 LdThis 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. Ld This represents the coordinates. Also, θr is the magnetic flux vector φγ. Ld and magnetic flux vector φδ Ld This represents the phase.

[0036] <Operation of the first γδ axis magnetic flux estimator 61> The first γδ axis magnetic flux estimator 61 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.

[0037]

[0038] 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.

[0039] While an example using a same-dimensional observer as the state observer has been shown, this is not the only option. A minimum-dimensional observer, disturbance observer, sliding observer, and unknown input observer may be used instead of a same-dimensional observer. Alternatively, a filtering process including algebraic calculations and integration based on equation (6) may be used instead of a state observer. Furthermore, since the value of the d-axis inductance Ld changes depending on the operating conditions, its value may be changed according to operating conditions such as torque, rotational speed, current, and voltage using a function or map. In addition, estimated values ​​may be used.

[0040] <Operation of the dq-axis magnetic flux estimator 62> Next, the magnetic flux φd measured by the dq-axis magnetic flux estimator 62 Ld , φq Ld Let's explain the estimation method. First, the d-axis magnetic flux φd Ld This 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, since the rotor's magnetic flux φm also changes depending on the operating point and operating conditions, the value may be changed according to operating conditions such as torque, rotational speed, current, voltage, and temperature using a function or map. Alternatively, the estimated value may be used. Furthermore, the magnetic flux φγ Ld The amplitude can be used as a substitute.

[0041] 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.

[0042] 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 Ld Since only the amplitude can be determined, the q-axis magnetic flux φq Ld It is not possible to determine the direction (positive or negative). Therefore, for example, using the q-axis current command iq*, which is one of the operation commands, φq Ld =sign(iq*)√|φ all 2 -(φd Ld ) 2 By doing |, φq Ld An estimated value can be obtained. Here, sign(P) is the sign function, outputting 1 if P in parentheses is positive, -1 if it is negative, and 0 if it 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.

[0043] During position sensorless control, iq* does not necessarily equal iq, but if the position estimation error is within ±90 degrees, sign(iq*) and sign(iq) will coincide. Note that the q-axis magnetic flux is φq. Ld The example shown uses the q-axis current command iq* to determine the direction (positive or negative), but this is not the only option. Instead of the q-axis current command iq*, other operation commands, torque, current, power, rotational speed-related information, etc., can be used to determine the q-axis magnetic flux φq. Ld The sign of the current may be determined. In the case of current, for example, it may be an estimated value of the delta-axis current iδ or the q-axis current iq.

[0044] As described above, the dq-axis magnetic flux estimator 62 allows the magnetic flux φd Ld , φq Ld to be estimated. The output of the dq-axis magnetic flux estimator 62 is the estimated value of the magnetic flux φd Ld , φq Ld . Here, in the calculation of the estimated values of the magnetic flux φd Ld , φq Ld , since the q-axis inductance Lq is not used, it is found that the estimated values of the magnetic flux φd Ld , φq Ld have no sensitivity to the q-axis inductance Lq. Instead, it is found that they have sensitivity to the d-axis inductance Ld and the rotor magnetic flux φm.

[0045] <Mathematical Model 2 of Rotating Electric Machine 30> To explain the operation of the second γδ-axis magnetic flux estimator 63, a mathematical model of the rotating electric machine 30 will be described. When model transformation is performed to replace all the inductance terms in the first term on the right side of Equation (1), which is the dq-coordinate voltage equation assuming an IPMSM as the rotating electric machine 30, with the q-axis inductance Lq, the following Equation (8) is obtained.

[0046]

[0047] Here, ed Lq , eq Lq are induced voltages when the model transformation is performed, and take a form in which the magnetic flux generated by the stator current and the magnet magnetic flux are mixed. Since these induced voltages appear like induced voltages generated in the rotor, they are also referred to as virtual induced voltages (virtual induced voltages). Specifically, since these are virtual induced voltages obtained by model transformation using the q-axis inductance Lq, they are referred to as virtual induced voltages based on the Lq model.

[0048] When Equation (8) is transformed into the αβ coordinate which is a fixed coordinate, Equation (9) is obtained, and when Equation (8) is transformed into the γδ coordinate which is an estimated rotational coordinate, Equation (10) is obtained. As described above, the α-axis of the fixed coordinate is expressed as being positioned at a phase delayed by θre from the d-axis, and the γ-axis of the estimated rotational coordinate is expressed as being positioned at a phase advanced by Δθre from the d-axis.

[0049]

[0050]

[0051] Here, vα and vβ are voltages on a fixed coordinate system, i.e., voltages on an αβ coordinate system, iα and iβ are currents on a fixed coordinate system, and eα Lq , eβ Lq is the induced voltage in the fixed coordinate system when the model is deformed. Also, vγ and vδ are the voltages in the estimated rotating coordinate system, i.e., the voltage in the γδ coordinate system, iγ and iδ are the currents in the fixed coordinate system, and eγ Lq , eδ Lq This is the estimated induced voltage on the rotational coordinate system when the model is deformed. Jθre , e -JΔθre This signifies a rotational coordinate transformation.

[0052] The d-axis magnetic flux when the model is deformed is φd Lq = φm + (Ld - Lq)id, q-axis magnetic flux is φq Lq By defining = 0 and rearranging equation (9), we obtain equation (11), and by rearranging equation (10), we obtain equation (12).

[0053]

[0054]

[0055] In equations (11) and (12), φα Lq , φβ Lq φγ is the magnetic flux on fixed coordinates when the model is deformed. Lq ,φδ Lq 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 fluxes (virtual rotor fluxes). More specifically, since they are virtual rotor fluxes resulting from model deformation using the q-axis inductance Lq, they are called virtual rotor fluxes based on the Lq model. Performing such deformation has the advantage that 2θre and 2Δθre do not occur before and after the coordinate transformation. Another feature is that the inductance included in the first term on the right-hand side of the voltage equation is only the q-axis inductance Lq.

[0056] <Operation of the second γδ axis magnetic flux estimator 63> The second γδ axis magnetic flux estimator 63 of Embodiment 1 calculates the magnetic flux φγ based on the mathematical model of equation (12).Lq ,φδ Lq The magnetic flux φγ is estimated by the mathematical model of equation (12). Lq ,φδ Lq , magnetic flux φd Lq , φq Lq The relationship is explained using the vector diagram in Figure 5. Figure 5 shows the magnetic flux vector φγ in γδ coordinates and dq coordinates. Lq ,φδ Lq ,φd Lq This is a vector diagram showing φq. Lq Since = 0, φq Lq The magnetic flux vector φγ is not shown. Lq ,φδ Lq The dotted line represents the magnetic flux vector φd. Lq The dashed line indicates the magnetic flux vector φd. Lq It is pointed to the d-axis, and the magnetic flux vector is φγ Lq ,φδ Lq The superposition sum of the vectors is the magnetic flux vector φd Lq This matches.

[0057] The second γδ axis magnetic flux estimator 63 uses, for example, a state observer to determine the magnetic flux φγ Lq ,φδ Lq To estimate the following: For example, the state variables are estimated values ​​of current iγ and iδ and magnetic flux φγ. Lq ,φδ Lq Using the estimated values, we construct a same-dimensional observer as shown in equation (13) below.

[0058]

[0059] Observer gain is a parameter designed considering steady-state and transient performance to achieve the desired specifications. Since the calculation in equation (13) uses only the q-axis inductance Lq and not the d-axis inductance Ld, the magnetic flux φγ is unaffected by fluctuations and setting errors of the d-axis inductance Ld. Lq ,φδ Lq An estimate can be obtained.

[0060] Although an example using a same-dimensional observer as the state observer has been shown, this is not the only option. Similar to the first γδ-axis magnetic flux estimator 61, a minimum-dimensional observer, disturbance observer, sliding observer, and unknown input observer may be used instead of the same-dimensional observer. Alternatively, the state observer may be replaced by a filtering process including algebraic calculation and integration based on equation (12). Furthermore, since the value of the q-axis inductance Lq changes depending on the operating conditions, its value may be changed according to operating conditions such as torque, rotational speed, current, and voltage using a function or map. In addition, estimated values ​​may be used. Magnetic flux vector φd Lq It always points to the d-axis, and the magnetic flux vector φq Lq Since it is obvious that = 0, there is no need to estimate it, but the magnetic flux φd of the dq axis is estimated by the dq axis magnetic flux estimator 62. Ld , φq Ld Similarly to estimating the magnetic flux φd Lq , φq Lq It is also possible to estimate this.

[0061] <Operation of Weight Setter 64> Based on the operation command, the weight setter 64 adjusts the γδ-axis magnetic flux φγ estimated by the first γδ-axis magnetic flux estimator 61, which will be described later. Ld ,φδ Ld The estimated value and the dq-axis magnetic flux φd estimated by the dq-axis magnetic flux estimator 62 Ld , φq Ld The first magnetic flux estimated using and the γδ-axis magnetic flux φγ estimated by the second γδ-axis magnetic flux estimator 63 Lq ,φδ Lq Output a weighting coefficient for combining the second magnetic flux, which is based on the estimated value of the first flux.

[0062] Figure 6 shows the operation of the weight setter according to Embodiment 1, and is a diagram showing an example of the weight coefficient w when the operation command is the q-axis current command iq*. For example, as shown in Figure 6, the weight coefficient w is made smaller in the region where the q-axis current command iq* is small, and the weight coefficient w is made larger in the region where the q-axis current command iq* is large. This weight coefficient w is the γδ-axis magnetic flux φγ estimated by the first γδ-axis magnetic flux estimator 61. Ld ,φδ Ld The estimated value of the dq-axis magnetic flux φd, estimated by the dq-axis magnetic flux estimator 62.Ld , φq Ld , and the γδ-axis magnetic flux φγ estimated by the second γδ-axis magnetic flux estimator 63 Lq ,φδ Lq The parameter is set according to the sensitivity of the estimated value.

[0063] Figure 6 shows an example where the weighting coefficient w is determined based on the q-axis current command iq*, but it may also be the d-axis current command id* or other operation commands. Alternatively, it may be the voltage command applied to the rotating machine 30, the current flowing through the rotating machine 30, torque, rotational speed, etc. Furthermore, a filter or the like may be added to adjust the response to the input or output of the weighting setter 64 so that the weighting coefficient does not change abruptly during transient operation.

[0064] <Configuration of Position and Velocity Estimation Unit 7> Next, the configuration of the position and velocity estimation unit 7 will be described. Figure 7 is a functional block diagram showing the configuration of the position and velocity estimation unit 7 according to Embodiment 1. The position and velocity estimation unit 7 uses the γδ axis magnetic flux φγ estimated by the first γδ axis magnetic flux estimator 61. Ld ,φδ Ld The estimated value and the dq-axis magnetic flux φd estimated by the dq-axis magnetic flux estimator 62 Ld , φq Ld The first magnetic flux estimator 71 estimates the first magnetic flux using the estimated value, and the second γδ axis magnetic flux estimator 63 estimates the γδ axis magnetic flux φγ Lq ,φδ Lq The system includes a second magnetic flux estimator 72 that estimates a second magnetic flux based on the estimated value of the first magnetic flux, a magnetic flux combiner 73 that combines the first magnetic flux and the second magnetic flux, a magnetic flux phase calculator 74 that calculates a phase difference Δθre, which is the phase difference between the d-axis and the γ-axis or the q-axis and the δ-axis, from the combined magnetic flux combined by the magnetic flux combiner 73, and a position estimation error convergence calculator 75 that calculates the estimated position and estimated speed of the rotor from the phase difference Δθre.

[0065] <Operation of the First Magnetic Flux Estimator 71> The operation of the first magnetic flux estimator 71 will be explained. Figure 8 is a diagram showing the estimated values ​​of each magnetic flux on the γδ coordinate system, but from the γδ coordinate system, φγ Ld ,φδ Ld ,φd Ld , φq Ld Observing the vector of the estimated values, the estimated value of the phase difference Δθre is the magnetic flux vector φγLd ,φδ Ld 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 be performed as shown in equation (14).

[0066]

[0067] By using the estimated value of the phase difference Δθre calculated in equation (14), the magnetic flux vector φγ can be obtained from equation (15). Ld ,φδ Ld ,φd Ld , φq Ld From the estimated value, the magnetic flux vector φγ points to the d-axis. Ld ,φδ Ld It is possible to calculate an estimated value of [the value].

[0068]

[0069] Magnetic flux vector φγ pointing to the d-axis Ld ,φδ Ld By normalizing the estimated value using equation (16), the unit magnetic flux vector pointing to the d-axis can be obtained.

[0070]

[0071] <Operation of the second magnetic flux estimator 72> Next, the operation of the second magnetic flux estimator 72 will be explained. From Figure 5, the estimated value of the phase difference Δθre can be calculated as shown in equation (17). Using the calculated estimated value of the phase difference Δθre, the γδ axis magnetic flux φγ estimated by the second γδ axis magnetic flux estimator 63 can be calculated as shown in equation (18). Lq ,φδ Lq From the estimated values, the unit magnetic flux vector pointing to the d-axis can be determined.

[0072]

[0073]

[0074] <Operation of the Magnetic Flux Combiner 73> The magnetic flux combiner 73 combines the first magnetic flux and the second magnetic flux, both directed toward the d-axis, based on a weighting coefficient. Figure 9 is a vector diagram illustrating the operation of the magnetic flux combiner according to Embodiment 1. In Figure 9, the combined magnetic flux is calculated using the formula: Combined magnetic flux = (first magnetic flux) × w + (second magnetic flux) × (1 - w). However, this formula is not necessarily the only way to calculate it.

[0075] <Operation of the magnetic flux phase calculator 74> The magnetic flux phase calculator 74 uses equation (19) to estimate the phase difference Δθre based on the combined magnetic flux.

[0076]

[0077] However, in addition to the arctangent function, other trigonometric functions may be used, or functions and approximation formulas may be substituted. In this embodiment, the two magnetic fluxes are combined before calculating the phase in order to use the arctangent function only once, but a configuration in which the phases of the two magnetic fluxes are determined first and then combined may also be used.

[0078] <Operation of the Position Estimation Error Convergence Calculator 75> The position estimation error convergence calculator 75 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 control is performed so that the estimated value of the phase difference Δθre calculated by the magnetic flux phase calculator 74 becomes 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 may be added to or subtracted from the estimated position one control cycle ago.

[0079] The voltage command generation unit 5 uses the estimated position and estimated speed of the rotor estimated by the position estimation error convergence calculator 75 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.

[0080] <Effects of Embodiment 1> The effects of Embodiment 1 will be explained using the figures. Figure 10 shows the sensitivity of the parameters used to estimate the first magnetic flux with respect to torque: d-axis inductance Ld, q-axis inductance Lq, and rotor magnetic flux φm. The greater the sensitivity, the greater the error in estimating the magnetic flux and the greater the error in estimating the position when an error occurs between the parameters used in the first magnetic flux estimator 71 and the parameters of the rotating machine 30. When estimating the first magnetic flux, the d-axis inductance Ld and rotor magnetic flux φm are used, so the q-axis inductance Lq has no sensitivity regardless of torque. On the other hand, the sensitivity of the d-axis inductance Ld and rotor magnetic flux φm increases as the torque decreases, and decreases as the torque increases.

[0081] Figure 11 shows the sensitivity of the parameters used to estimate the second magnetic flux with respect to torque: d-axis inductance Ld, q-axis inductance Lq, and rotor magnetic flux φm. Similarly, higher sensitivity means that when there is an error between the parameters used in the second magnetic flux estimator 72 and the parameters of the rotating machine 30, a large error in magnetic flux estimation and a large error in position estimation will occur. When estimating the second magnetic flux, only the q-axis inductance Lq is used, so the d-axis inductance Ld and rotor magnetic flux φm have no sensitivity regardless of torque. On the other hand, the sensitivity of the q-axis inductance Lq decreases as the torque decreases and increases as the torque increases.

[0082] Thus, the sensitivity of the d-axis inductance Ld, the q-axis inductance Lq, and the rotor's magnetic flux φm has different characteristics depending on whether the first magnetic flux is being estimated or the second magnetic flux is being estimated. This embodiment utilizes these characteristics.

[0083] Figure 12 shows the sensitivity of the parameters used to estimate the combined magnetic flux obtained by combining the first and second magnetic fluxes: the d-axis inductance Ld, the q-axis inductance Lq, and the rotor magnetic flux φm, as solid lines. As shown in Figure 6, the weighting coefficient w is determined so that the weight of the second magnetic flux is increased in the region where the value of the q-axis current command iq* is small and the torque is small, and the weight of the first magnetic flux is increased in the region where the value of the q-axis current command iq* is large and the torque is large. In this way, the sensitivity of the parameters d-axis inductance Ld, q-axis inductance Lq, and rotor magnetic flux φm can be distributed among three, while always keeping the sensitivity of the parameters low.

[0084] Thus, by using a combined magnetic flux obtained by estimating the rotor position and speed using a first magnetic flux estimated with the d-axis inductance Ld and the rotor magnetic flux φm, and a second magnetic flux estimated using the q-axis inductance Lq, and by utilizing the fact that the first and second magnetic fluxes have different parameter sensitivities, the parameter sensitivity can be arbitrarily and appropriately changed according to the permanent magnet synchronous motor used and the operating conditions. Because the sensitivity can be changed, position sensorless control can be performed so that the parameter sensitivity is always small under all operating conditions, thereby improving the accuracy of rotor position estimation and making it less susceptible to the influence of characteristic fluctuations of the permanent magnet synchronous motor. Furthermore, since this control device 10 estimates the magnetic flux of the d-axis or q-axis and performs current vector control on the dq coordinate system, it is easy to adjust and optimize the parameters even in applications that require torque accuracy, such as automobiles and trains.

[0085] As described above, according to Embodiment 1, the control device for a rotating machine includes a power conversion unit that converts the voltage of the 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 and velocity estimation unit that estimates the position and velocity of the rotor of the rotating machine. The magnetic flux estimation unit estimates at least one pair of d-axis magnetic flux and q-axis magnetic flux on a first Cartesian coordinate system having a d-axis and a q-axis in synchronization with the rotation of the rotating machine, and estimates multiple pairs of magnetic flux for each axis on a second Cartesian coordinate system having axes different from the d-axis and q-axis, based on different parameters of the rotating machine. The position and velocity estimation unit is configured to synthesize at least one set of d-axis and q-axis magnetic fluxes with multiple sets of magnetic fluxes for each axis on the second Cartesian coordinate system using a preset weighting coefficient. Based on the synthesized magnetic fluxes, it calculates the phase difference between the first Cartesian coordinate system and the second Cartesian coordinate system, and estimates the rotor position and velocity from this phase difference. With this configuration, multiple sets of magnetic fluxes for each axis on the second Cartesian coordinate system are estimated based on different parameters of the rotating machine using three or more magnetic fluxes. By adjusting the sensitivity of the different parameters, the rotor position and velocity can be estimated with high accuracy, making it possible to obtain a rotating machine control device that is easy to adjust and optimize parameters for.

[0086] Furthermore, in Embodiment 1, as each axis on the second Cartesian coordinate system, axes on a Cartesian coordinate system having a γ axis and a δ axis synchronized with the rotation of the rotating machine are used. The set of magnetic flux on the γ axis and the magnetic flux on the δ axis is a plurality of sets, including a first magnetic flux set estimated using the voltage command generated by the voltage command generation unit, the current of the rotating machine, and the inductance of the rotating machine as parameters of the rotating machine, and a second magnetic flux set estimated using the voltage command generated by the voltage command generation unit, the current of the rotating machine, and an inductance different from the inductance of the rotating machine used in the first magnetic flux set as parameters of the rotating machine. Therefore, by adjusting the two magnetic flux sets using different inductance sensitivities, the rotor position and speed can be estimated with high accuracy, making it possible to obtain a rotating machine control device that is easy to adjust and optimize parameters for.

[0087] In this case, if the inductance of the rotating machine used to estimate the first magnetic flux set is the d-axis inductance, the set of d-axis magnetic flux and q-axis magnetic flux includes the d-axis magnetic flux estimated based on the rotor magnetic flux of the rotating machine and the q-axis magnetic flux estimated using the rotor magnetic flux of the rotating machine and the first magnetic flux set. The set of d-axis magnetic flux and q-axis magnetic flux is estimated so that the first γδ-axis magnetic flux, which is the combined magnetic flux from the first magnetic flux set, and the dq-axis magnetic flux, which is the combined magnetic flux from the set of d-axis magnetic flux and q-axis magnetic flux, coincide. The position and velocity estimation unit then combines the first γδ-axis magnetic flux and the second γδ-axis magnetic flux, which is the combined magnetic flux from the second magnetic flux set, using a weighting coefficient, and calculates the phase difference between the d-axis and q-axis and the γ-axis and δ-axis based on the combined magnetic flux to estimate the rotor's position and velocity. By controlling the sensitivity of the parameters used here—the d-axis inductance Ld, the q-axis inductance Lq, and the rotor's magnetic flux φm—to always be small, it becomes possible to accurately estimate the rotor's position and velocity.

[0088] In Embodiment 1, an example was shown in which there were two γ-axis and δ-axis magnetic flux sets and one d-axis and q-axis magnetic flux set with different parameter combinations, but this is not the only example. It is sufficient to have at least two γ-axis and δ-axis magnetic flux sets and at least one d-axis and q-axis magnetic flux set with different parameter combinations, and there may be three or more γ-axis and δ-axis magnetic flux sets and two or more d-axis and q-axis magnetic flux sets with different parameter combinations. For example, in addition to the first γδ-axis magnetic flux estimator 61 and dq-axis magnetic flux estimator 62 and the second γδ-axis magnetic flux estimator 63, by using third, fourth, and fifth γδ-axis magnetic flux estimators and second, third, and fourth dq-axis magnetic flux estimators with different parameter sensitivities, the sensitivity of each parameter can be finely adjusted according to the operating point.

[0089] 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 first γδ axis magnetic flux estimator 61, a dq axis magnetic flux estimator 62, a second γδ axis magnetic flux estimator 63, and a weight setter 64. In Embodiment 2, however, the first γδ axis magnetic flux estimator 61 and the second γδ axis magnetic flux estimator 63 are replaced with a first αβ axis magnetic flux estimator and a second αβ axis magnetic flux estimator, respectively. 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 and omitted here. The overall configuration of the control device for the rotating machine is the same as in Figure 1, but as described below, a controller 2A is provided instead of a controller 2, a voltage command generation unit 5A is provided instead of a voltage command generation unit 5, a magnetic flux estimation unit 6A is provided instead of a magnetic flux estimation unit 6, and a position / velocity estimation unit 7A is provided instead of a position / velocity estimation unit 7.

[0090] <Configuration and Operation of Magnetic Flux Estimation Unit 6A> Figure 13 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 13, the magnetic flux estimation unit 6A comprises a first αβ axis magnetic flux estimator 65A, a dq axis magnetic flux estimator 62A, a second αβ axis magnetic flux estimator 66A, and a weight setter 64A. The first αβ axis magnetic flux estimator 65A measures the magnetic flux φα on the fixed coordinates as described in equation (5). Ld , φβ Ld The following is estimated. The estimation method can be the same as the first γδ axis magnetic flux estimator 61 in Embodiment 1, for example, a method using a state observer.

[0091] Furthermore, the magnetic flux φα is expressed in the mathematical model of equation (5). Ld , φβ Ld ,φd Ld , φq Ld The relationship is shown in Figure 14. Figure 14 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 , φqLd 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 14, 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 , φβ Ld The coordinates and sign D are the magnetic flux vector φα. Ld This represents the coordinates. Also, θd is the magnetic flux vector φd. Ld and magnetic flux vector φq Ld This represents the phase.

[0092] Magnetic flux φd by dq-axis magnetic flux estimator 62A Ld , φq Ld The estimation method is the same as that of the dq-axis magnetic flux estimator 62 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.

[0093] 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.

[0094] The second αβ-axis magnetic flux estimator 66A estimates the magnetic flux φα based on the mathematical model of equation (11). Lq , φβ Lq Here, the magnetic flux φγ is expressed by the mathematical model of equation (11). Lq ,φδ Lq , magnetic flux φd Lq , φq Lq The relationship will be explained using the vector diagram in Figure 15. Figure 15 shows the magnetic flux vector φα in αβ coordinates and dq coordinates. Lq , φβ Lq ,φd Lq This is a vector diagram showing φq. Lq Since = 0, φq Lq The magnetic flux vector φα is not shown. Lq , φβ Lq The dashed line represents the magnetic flux vector φd. Lq The dashed line indicates the magnetic flux vector φd. Lq It is pointed to the d-axis, and the magnetic flux vector is φγ Lq ,φδ Lq The superposition sum of the vectors is the magnetic flux vector φd Lq This matches.

[0095] The second αβ axis magnetic flux estimator 66A determines the magnetic flux φα on the fixed coordinates in equation (11). Lq , φβ Lq The magnetic flux φα is estimated, and the estimation method can be the same as the second γδ axis magnetic flux estimator 63 of Embodiment 1, for example, a method using a state observer. For example, using a state observer, the magnetic flux φα Lq , φβ Lq To estimate the following: For example, the state variables are estimated values ​​of currents iα and iβ and magnetic flux φα. Lq , φβ Lq Alternatively, an estimate of the same-dimensional observer may be used to construct the same-dimensional observer.

[0096] The weight setter 64A, similar to the weight setter 64 in Embodiment 1, adjusts the αβ-axis magnetic flux φα estimated by the first αβ-axis magnetic flux estimator 65A (described later) based on the operation command. Ld , φβ Ld The estimated value and the dq-axis magnetic flux φd estimated by the dq-axis magnetic flux estimator 62A Ld , φq LdThe first magnetic flux estimated using and the αβ-axis magnetic flux φα estimated by the second αβ-axis magnetic flux estimator 66A Lq , φβ Lq Output a weighting coefficient wa for combining the second magnetic flux based on the estimated value.

[0097] <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 16 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 uses the αβ-axis magnetic flux φα estimated by the first αβ-axis magnetic flux estimator 65A. Ld , φβ Ld The estimated value and the dq-axis magnetic flux φd estimated by the dq-axis magnetic flux estimator 62A Ld , φq Ld The first magnetic flux estimator 71 estimates the first magnetic flux using the estimated value of and the αβ-axis magnetic flux φα estimated by the second αβ-axis magnetic flux estimator 66A. Lq , φβ Lq The system includes a second magnetic flux estimator 72A that estimates a second magnetic flux based on the estimated value of the first magnetic flux, a magnetic flux combiner 73A that combines the first magnetic flux and the second magnetic flux, a magnetic flux phase calculator 74A that calculates a phase difference θre, which is the phase difference between the d-axis and the α-axis or the q-axis and the β-axis, from the combined magnetic flux combined by the magnetic flux combiner 73A, and a speed estimator 76A that calculates the estimated speed of the rotor from the phase difference θre.

[0098] <Operation of the first magnetic flux estimator 71A> The phase difference θre in Figure 14 is the magnetic flux vector φα Ld , φβ Ld ,φd Ld , φq Ld Using the estimated values, the estimation can be performed using the following equation (20), similar to Embodiment 1.

[0099]

[0100] By using the estimated value of the phase difference θre calculated by equation (20), the magnetic flux vector φα pointing to the d-axis can be obtained, similar to Embodiment 1. Ld , φβ Ld We can obtain an estimate of this, and by normalizing it, we can obtain the unit magnetic flux vector φα pointing to the d-axis, as shown in equation (21). Ld , φβ LdAn estimated value can be obtained. This unit magnetic flux vector is called the first magnetic flux.

[0101]

[0102] <Operation of the second magnetic flux estimator 72A> Next, the operation of the second magnetic flux estimator 72A will be explained. From Figure 15, the estimated value of the phase difference θre can be calculated as shown in equation (22). Using the calculated estimated value of the phase difference θre, the αβ-axis magnetic flux φα estimated by the second αβ-axis magnetic flux estimator 66A can be calculated as shown in equation (23). Lq , φβ Lq From the estimated values, the unit magnetic flux vector pointing to the d-axis can be determined.

[0103]

[0104]

[0105] <Operation of the Magnetic Flux Combiner 73A> The magnetic flux combiner 73A combines the first magnetic flux and the second magnetic flux, both directed toward the d-axis, based on a weighting coefficient. Figure 17 is a vector diagram illustrating the operation of the magnetic flux combiner 73A according to Embodiment 2. In Figure 17, the combined magnetic flux is calculated using the formula: Combined magnetic flux = (first magnetic flux) × wa + (second magnetic flux) × (1 - wa). However, this formula is not necessarily the only way to calculate it.

[0106] <Operation of Magnetic Flux Phase Calculator 74A> The magnetic flux phase calculator 74 calculates an estimated value of the phase difference θre based on the combined magnetic flux using equation (24).

[0107]

[0108] However, you may use other trigonometric functions besides the arctangent function, or substitute functions and approximation formulas.

[0109] The velocity estimator 76A 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 magnetic flux phase calculator 74A.

[0110] <Effects of Embodiment 2> Embodiment 2 uses an αβ-axis magnetic flux estimator instead of a γδ-axis magnetic flux estimator to estimate the first and second magnetic fluxes in the same way as Embodiment 1. In Embodiment 2 as well, to estimate the rotor position and velocity, the first magnetic flux, estimated using the d-axis inductance Ld and rotor magnetic flux φm as parameters of the rotating machine, and the second magnetic flux, estimated using the q-axis inductance Lq, are combined to obtain a magnetic flux.

[0111] In other words, the sensitivity characteristics of the first magnetic flux parameter in Embodiment 2 are the same as those in Figure 10 of Embodiment 1, and the sensitivity characteristics of the second magnetic flux parameter in Embodiment 2 are the same as those in Figure 11 of Embodiment 1.

[0112] Therefore, similar to Embodiment 1, by utilizing the fact that the first magnetic flux and the second magnetic flux have different parameter sensitivities, the parameter sensitivity can be arbitrarily and appropriately changed according to the permanent magnet synchronous motor used and the operating conditions. Because the sensitivity can be changed, position sensorless control can be performed so that the parameter sensitivity is always small under all operating conditions, thereby improving the accuracy of rotor position estimation and making it less susceptible to the influence of characteristic fluctuations of the permanent magnet synchronous motor. Furthermore, since this control device 10 estimates the magnetic flux of the d axis or q axis and performs current vector control on the dq coordinate system, it is easy to adjust and optimize the parameters even in applications that require torque accuracy, such as automobiles and trains.

[0113] As described above, Embodiment 2 provides the same effects as Embodiment 1. Specifically, the control device for a rotating machine according to Embodiment 2 includes the same power conversion unit, current detection unit, voltage command generation unit, magnetic flux estimation unit, and position / velocity estimation unit as in Embodiment 1. The magnetic flux estimation unit estimates at least one pair of d-axis magnetic flux and q-axis magnetic flux on a first Cartesian coordinate system having a d-axis and a q-axis, synchronized with the rotation of the rotating machine, and estimates multiple pairs of α-axis magnetic flux and β-axis magnetic flux on a second Cartesian coordinate system having an α-axis and a β-axis, which are axes different from the d-axis and q-axis and coordinates based on the phase of the stator of the rotating machine, based on different parameters of the rotating machine. The position / velocity estimation unit is configured to synthesize at least one pair of d-axis magnetic flux and q-axis magnetic flux and multiple pairs of α-axis magnetic flux and β-axis magnetic flux using a preset weighting coefficient, calculate the phase difference between the d-axis and q-axis and the α-axis and β-axis based on the synthesized magnetic flux, and estimate the rotor position and velocity from this phase difference. This configuration uses three or more magnetic fluxes to estimate multiple sets of α-axis and β-axis magnetic fluxes based on different parameters of the rotating machine. By adjusting the sensitivity of these different parameters, the rotor position and speed can be estimated with high accuracy, making it possible to obtain a rotating machine control device that is easy to adjust and optimize.

[0114] In Embodiment 2, an example was shown in which there were two magnetic flux sets for the α and β axes and one magnetic flux set for the d and q axes using different parameter combinations, but this is not the only example. It is sufficient to have at least two magnetic flux sets for the α and β axes and at least one magnetic flux set for the d and q axes using different parameter combinations, and there may be three or more magnetic flux sets for the α and β axes and two or more magnetic flux sets for the d and q axes using different parameter combinations.

[0115] 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.

[0116] <Configuration of the Rotating Machine Control Device> Figure 18 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.

[0117] <Configuration and Operation of Voltage Estimation Unit 8B> Figure 19 is a functional block diagram showing an example of the configuration of the voltage estimation unit 8B. The voltage estimation unit 8B includes a first γδ axis voltage estimator 81B 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, a dq axis voltage estimator 82B 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 first γδ axis voltage estimator 81B, a second γδ axis voltage estimator 83B 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 weight setter 84B. Here, the voltage command from the voltage command generation unit 5B is input to the first γδ axis voltage estimator 81B and the second γδ axis voltage estimator 83B, but if there is a voltage sensor, the voltage from the voltage sensor may be input. Furthermore, the first γδ-axis voltage estimator 81B estimates the voltage along the γδ axis using the d-axis inductance Ld, and the second γδ-axis voltage estimator 83B estimates the voltage along the γδ axis using the q-axis inductance Lq.

[0118] The first γδ axis voltage estimator 81B 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 eγLd , eδ Ld The estimation of this is performed using methods such as the state observer described above.

[0119] The dq-axis voltage estimator 82B 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 ed 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.

[0120] The second γδ axis voltage estimator 83B is the induced voltage eγ on the γδ coordinate in equation (10). Lq , eδ Lq To estimate the voltage eγ Lq , eδ Lq The estimation of this is performed using methods such as the state observer described above.

[0121] Based on the operation command, the weight setter 84B adjusts the γδ axis voltage eγ estimated by the first γδ axis voltage estimator 81B, which will be described later. Ld , eδ Ld The estimated value and the dq-axis voltage ed estimated by the dq-axis voltage estimator 82B Ld , eq LdThe first voltage estimated using and the γδ axis voltage eγ estimated by the second γδ axis voltage estimator 83B Lq , eδ Lq Output a weighting coefficient wb for combining the second voltage based on the estimated value.

[0122] <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 20 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 receives the γδ axis voltage eγ estimated by the first γδ axis voltage estimator 81B. Ld , eδ Ld The estimated value and the dq-axis voltage ed estimated by the dq-axis voltage estimator 82B Ld , eq Ld The first voltage estimator 71B estimates the first voltage using the estimated value, and the γδ axis voltage eγ estimated by the second γδ axis voltage estimator 83B Lq , eδ Lq The system includes a second voltage estimator 72B that estimates a second voltage based on the estimated value of the first voltage, a voltage combiner 73B that combines the first voltage and the second voltage, a voltage phase calculator 74B that calculates a phase difference Δθre, which is the phase difference between the d axis and the γ axis or the q axis and the δ axis, from the combined voltage combined by the voltage combiner 73B, and a position estimation error convergence calculator 75B that calculates the estimated position and estimated speed of the rotor from the phase difference Δθre.

[0123] <Operation of the First Voltage Estimator 71B> The operation of the first voltage estimator 71B will now be explained. The first voltage estimator 71B calculates the phase difference Δθre using equation (25), similar to embodiments 1 and 2. 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.

[0124]

[0125] The estimated phase difference Δθre is calculated using the arctangent function, but this is not the only method. Other trigonometric functions, or substitutes with other functions and approximation formulas, may also be used.

[0126] By using the estimated value of the phase difference Δθre calculated by equation (25), the magnetic flux vector eγ that points to the d-axis is obtained, similar to embodiments 1 and 2. Ld , eδ Ld We can obtain an estimated value of this, and by normalizing it, we can obtain the unit voltage vector eγ pointing to the q-axis, as shown in equation (26). Ld , eδ Ld An estimated value can be obtained. This unit voltage vector is called the first voltage.

[0127]

[0128] <Operation of the second voltage estimator 72B> Next, the operation of the second voltage estimator 72B will be explained. By expanding the vector diagram showing the relationship between the γδ axis magnetic flux and the dq axis magnetic flux in Figure 5 into a vector diagram showing the relationship between the γδ axis voltage and the dq axis voltage, the estimated value of the phase difference Δθre can be calculated as shown in equation (27). Using the calculated estimated value of the phase difference θre, the αβ axis magnetic flux φα estimated by the second αβ axis magnetic flux estimator 66A can be calculated as shown in equation (28). Lq , φβ Lq From the estimated values, the unit magnetic flux vector pointing to the d-axis can be determined.

[0129]

[0130]

[0131] <Operation of Voltage Combiner 73B> The voltage combiner 73B combines a first voltage and a second voltage, both directed along the d-axis, based on a weighting coefficient wc. For example, the combined voltage can be calculated as: Combined voltage = (first voltage) × wc + (second voltage) × (1 - wc). However, this formula is not always the only way to calculate it.

[0132] <Operation of Voltage Phase Calculator 74B> The Voltage Phase Calculator 74B estimates the phase difference Δθre based on the combined voltage using equation (29).

[0133]

[0134] However, you may use other trigonometric functions besides the arctangent function, or substitute functions and approximation formulas.

[0135] <Operation of the Position Estimation Error Convergence Calculator 75B> The Position Estimation Error Convergence Calculator 75B is the same as the Position Estimation Error Convergence Calculator 75 shown in Embodiment 1. For example, it uses PLL processing to control the estimated value of the phase difference Δθre calculated by the voltage phase calculator 74B so that it becomes zero or close to zero, and calculates the estimated position and estimated speed of the rotor.

[0136] The voltage command generation unit 5B uses the estimated position and estimated speed of the rotor estimated by the position estimation error convergence calculator 75B 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.

[0137] <Effects of Embodiment 3> The effects of Embodiment 3 will now be explained. In the voltage estimation of Embodiment 3, information on the electrical angular velocity ωre can be added as a parameter. By setting the electrical angular velocity ωre information to cancel out errors caused by inductances such as the d-axis inductance Ld and the q-axis inductance Lq, the position estimation accuracy can be further improved.

[0138] 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 the 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 at least one pair of d-axis voltages and q-axis voltages on a first Cartesian coordinate system having d-axis and q-axis voltages synchronized with the rotation of the rotating machine, and estimates multiple pairs of voltages for each axis on a second Cartesian coordinate system having axes different from the d-axis and q-axis, based on different parameters of the rotating machine. The position and velocity estimation unit combines at least one pair of d-axis voltages and q-axis voltages with multiple pairs of voltages for each axis on the second Cartesian coordinate system using preset weight coefficients, calculates the phase difference between the first Cartesian coordinate system and the second Cartesian coordinate system based on the combined voltages, and estimates the rotor position and velocity from the phase difference. With this configuration, since multiple pairs of voltages for each axis on the second Cartesian coordinate system are estimated based on different parameters of the rotating machine using three or more voltages, the rotor position and velocity can be estimated with high accuracy by adjusting the sensitivity of the different parameters, making it possible to obtain a rotating machine control device that is easy to adjust and optimize parameters for.

[0139] Furthermore, in Embodiment 3, as the axes on the second Cartesian coordinate system, axes on the Cartesian coordinate system having a γ axis and a δ axis synchronized with the rotation of the rotating machine are used. The set of voltages for the γ axis and the δ axis is a plurality of sets, including a first voltage set estimated using the voltage command generated by the voltage command generation unit, the current of the rotating machine, and the inductance of the rotating machine as parameters of the rotating machine, and a second voltage set estimated using the voltage command generated by the voltage command generation unit, the current of the rotating machine, and an inductance different from the inductance of the rotating machine used in the first voltage set as parameters of the rotating machine. Therefore, by adjusting the plurality of voltage sets using different inductance sensitivities, the rotor position and speed can be estimated with high accuracy, making it possible to obtain a rotating machine control device that is easy to adjust and optimize parameters for.

[0140] In this case, if the inductance of the rotating machine used to estimate the first voltage set is the d-axis inductance, the set of d-axis voltage and q-axis voltage includes the d-axis voltage estimated based on the rotor magnetic flux of the rotating machine and the q-axis voltage estimated using the rotor magnetic flux of the rotating machine and the first voltage set. The set of d-axis voltage and q-axis voltage is estimated so that the first γδ-axis voltage, which is the combined voltage from the first voltage set, matches the dq-axis magnetic flux, which is the combined voltage from the set of d-axis voltage and q-axis magnetic flux. The position and velocity estimation unit then combines the first γδ-axis voltage and the second γδ-axis voltage, which is the combined voltage from the second voltage set, using a weighting coefficient. Based on the combined voltage, the unit calculates the phase difference between the d-axis and q-axis and the γ-axis and δ-axis to estimate the rotor's position and velocity. By controlling the sensitivity of the parameters used here, namely the d-axis inductance Ld, q-axis inductance Lq, and rotor magnetic flux φm, to always be small, it becomes possible to estimate the rotor's position and velocity with high accuracy.

[0141] In Embodiment 3, an example was shown in which there were two voltage sets for the γ-axis and δ-axis and one voltage set for the d-axis and q-axis, using different parameter combinations. However, the invention is not limited to this. It is sufficient to have at least two voltage sets for the γ-axis and δ-axis and at least one voltage set for the d-axis and q-axis using different parameter combinations. There may also be three or more voltage sets for the γ-axis and δ-axis and two or more voltage sets for the d-axis and q-axis, using different parameter combinations.

[0142] 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 first γδ axis voltage estimator 81B, a dq axis voltage estimator 82B, and a second γδ axis voltage estimator 83B. However, in Embodiment 4, an αβ axis voltage estimator is provided instead of the γδ axis voltage estimator. 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 18, but as described below, a controller 2C is provided instead of a controller 2B, a voltage command generation unit 5C is provided instead of a voltage command generation unit 5B, a voltage estimation unit 8C is provided instead of a voltage estimation unit 8B, and a position / velocity estimation unit 7C is provided instead of a position / velocity estimation unit 7B.

[0143] <Configuration and Operation of Voltage Estimation Unit 8C> Figure 21 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 21, the voltage estimation unit 8C includes a first αβ axis voltage estimator 85C that estimates the voltage of the αβ axis based on a voltage command generated by the voltage command generation unit 5C and the alternating current detected by the current detection unit 4, a dq axis voltage estimator 82C that estimates the voltage of the dq axis based on an operation command, the magnetic flux φm of the rotor and the voltage of the αβ axis estimated by the first αβ axis voltage estimator 85C, a second αβ axis voltage estimator 86C that estimates the voltage of the αβ axis based on a voltage command generated by the voltage command generation unit 5C and the alternating current detected by the current detection unit 4, and a weight setter 84C. Here, the voltage command from the voltage command generation unit 5C is input to the first αβ axis voltage estimator 85C and the second αβ axis voltage estimator 86C, but if there is a voltage sensor, the voltage from the voltage sensor may be input. Furthermore, the first αβ-axis voltage estimator 85C estimates the αβ-axis voltage using the d-axis inductance Ld, and the second αβ-axis voltage estimator 86C estimates the αβ-axis voltage using the q-axis inductance Lq.

[0144] The first αβ axis voltage estimator 85C is given by the induced voltage eα on the αβ coordinate in equation (3). Ld , eβ LdThe following is estimated. The voltage commands generated by the voltage command generation unit 5C used are voltages vα and vβ on the αβ coordinate system, and the AC currents detected by the current detection unit 4 are currents iγ and iδ on the γδ coordinate system. Voltage eα Ld , eβ Ld The estimation of this is performed using methods such as the state observer described above.

[0145] The dq-axis voltage estimator 82C calculates the induced voltage ed on the dq coordinate. Ld , eq Ld This is estimated in the same way as in Embodiment 3. That is, eq Ld =ωreφm, 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 Ld Since it is sufficient to determine the direction, instead of the q-axis current command iq*, information related to other operation commands, torque, other currents, power, and rotational speed is used to determine ed Ld The direction, i.e., the sign, may also be determined. Similar to Embodiment 4, the voltage estimation in Embodiment 3 also requires the electrical angular velocity ωre, and therefore is sensitive to, i.e., affected by, the rotational speed.

[0146] The second αβ axis voltage estimator 86C is given by the induced voltage eα on the αβ coordinate in equation (9). Lq , eβ Lq We estimate the voltage eα. Lq , eβ Lq The estimation of this is performed using methods such as the state observer described above.

[0147] Based on the operation command, the weight setter 84C adjusts the αβ axis voltage eα estimated by the first αβ axis voltage estimator 85C, which will be described later. Ld , eβ LdThe estimated value and the dq-axis voltage ed estimated by the dq-axis voltage estimator 82C Ld , eq Ld The first voltage estimated using and the αβ axis voltage eα estimated by the second αβ axis voltage estimator 86C Lq , eβ Lq Output a weighting coefficient wc for combining the second voltage based on the estimated value.

[0148] <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 22 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 receives the αβ axis voltage eα estimated by the first αβ axis voltage estimator 85C. Ld , eβ Ld The estimated value and the dq-axis voltage ed estimated by the dq-axis voltage estimator 82C Ld , eq Ld The first voltage estimator 71C estimates the first voltage using the estimated value of and the αβ axis voltage eα estimated by the second αβ axis voltage estimator 86C Lq , eβ Lq The system includes a second voltage estimator 72C that estimates a second voltage based on the estimated value of the first voltage, a voltage combiner 73C that combines the first voltage and the second voltage, a voltage phase calculator 74C that calculates an estimated value of the phase difference θre, which is the phase difference between the d axis and the α axis or the q axis and the β axis, from the combined voltage combined by the voltage combiner 73B, and a speed estimator 76C that calculates the estimated speed of the rotor from the estimated value of the phase difference θre.

[0149] <Operation of the First Voltage Estimator 71C> The operation of the first voltage estimator 71C will now be explained. The first voltage estimator 71C calculates the phase difference θre using equation (30), similar to embodiments 1 to 3. At this time, the vector diagram showing the relationship between the αβ axis magnetic flux and the dq axis magnetic flux in Figure 14 is expanded into a vector diagram showing the relationship between the αβ axis voltage and the dq axis voltage.

[0150]

[0151] The estimated phase difference θre is calculated using the arctangent function, but this is not the only method. Other trigonometric functions, or other functions and approximation formulas, may be used as substitutes.

[0152] By using the estimated value of the phase difference θre calculated by equation (30), the magnetic flux vector eα that points to the q-axis can be obtained, similar to embodiments 1 to 3. Ld , eβ Ld We can obtain an estimate of this, and by normalizing it, we can obtain the unit voltage vector eα pointing to the q-axis, as shown in equation (31). Ld , eβ Ld An estimated value can be obtained. This unit voltage vector is called the first voltage.

[0153]

[0154] <Operation of the second voltage estimator 72C> Next, the operation of the second voltage estimator 72C will be explained. By expanding the vector diagram showing the relationship between the αβ axis magnetic flux and the dq axis magnetic flux in Figure 15 into a vector diagram showing the relationship between the αβ axis voltage and the dq axis voltage, the estimated value of the phase difference θre can be calculated as shown in equation (32). Using the calculated estimated value of the phase difference θre, the αβ axis voltage eα estimated by the second αβ axis voltage estimator 86C can be calculated as shown in equation (33). Lq , eβ Lq From the estimated values, a unit voltage vector pointing to the d-axis can be determined.

[0155]

[0156]

[0157] <Operation of Voltage Combiner 73C> The voltage combiner 73C combines a first voltage and a second voltage, both directed along the d-axis, based on a weighting coefficient wc. For example, the combined voltage can be calculated as: Combined voltage = (first voltage) × wc + (second voltage) × (1 - wc). However, this formula is not always the only way to calculate it.

[0158] <Operation of Voltage Phase Calculator 74C> The Voltage Phase Calculator 74C estimates the phase difference θre based on the combined voltage using equation (34).

[0159]

[0160] However, you may use other trigonometric functions besides the arctangent function, or substitute functions and approximation formulas.

[0161] The speed estimator 76C 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 voltage phase calculator 74C. The estimated position and estimated speed of the rotor, estimated by the position and speed estimation unit 7C, are output to the voltage command generation unit 5C.

[0162] <Effects of Embodiment 4> The effects of Embodiment 4 will now be explained. In the voltage estimation of Embodiment 4, information on the electrical angular velocity ωre can be added as a parameter. By setting the electrical angular velocity ωre information to cancel out errors caused by inductances such as the d-axis inductance Ld and the q-axis inductance Lq, the position estimation accuracy can be further improved.

[0163] As described above, Embodiment 4 provides the same effects as Embodiment 3. Specifically, the control device for a rotating machine according to Embodiment 4 includes a power conversion unit, a current detection unit, a voltage command generation unit, a voltage estimation unit, and a position / velocity estimation unit, similar to those in Embodiment 3. The voltage estimation unit estimates at least one pair of d-axis voltages and q-axis voltages on a first Cartesian coordinate system having a d-axis and a q-axis, synchronized with the rotation of the rotating machine, and estimates multiple pairs of α-axis voltages and β-axis voltages on a second Cartesian coordinate system having α-axis and β-axis, which are axes different from the d-axis and q-axis and coordinates based on the phase of the stator of the rotating machine, based on different parameters of the rotating machine. The position / velocity estimation unit is configured to synthesize at least one pair of d-axis voltages and q-axis voltages and multiple pairs of α-axis voltages and β-axis voltages using a preset weight coefficient, calculate the phase difference between the d-axis and q-axis and the α-axis and β-axis based on the synthesized voltage, and estimate the rotor position and velocity from this phase difference. This configuration uses three or more voltages to estimate multiple sets of α-axis voltages and β-axis voltages based on different parameters of the rotating machine. By adjusting the sensitivity of the different parameters, the rotor position and speed can be estimated with high accuracy, making it possible to obtain a rotating machine control device that is easy to adjust and optimize.

[0164] In Embodiment 4, an example was shown in which there were two voltage sets for the α and β axes and one voltage set for the d and q axes using different parameter combinations, but this is not the only example. It is sufficient to have at least two voltage sets for the α and β axes and at least one voltage set for the d and q axes using different parameter combinations, and there may be three or more magnetic flux sets for the α and β axes and two or more voltage sets for the d and q axes using different parameter combinations.

[0165] Embodiment 5. Below, a control device for a rotating machine according to Embodiment 5 will be described with reference to the figures. Embodiments 1 and 2 described examples using a magnetic flux estimation unit, and Embodiments 3 and 4 described examples using a voltage estimation unit, but Embodiment 5 describes a configuration that combines both. In the following description, the differences from Embodiments 1 to 4 will be the main focus, and configurations and operations similar to those in Embodiments 1 to 4 will be referred to in those descriptions and omitted here.

[0166] <Configuration of the Rotating Machine Control Device> Figure 23 is a functional block diagram showing the configuration of the rotating machine control device according to Embodiment 5. The control device 10D includes a controller 2D, 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 2D, and a current detection unit 4 that detects the AC current output from the power conversion unit 3.

[0167] The controller 2D includes a voltage command generation unit 5D that generates a voltage command based on an operation command, a magnetic flux estimation unit 6D that estimates magnetic flux based on the voltage command generated by the voltage command generation unit 5D and the alternating current detected by the current detection unit 4, a voltage estimation unit 8D that estimates voltage based on the voltage command generated by the voltage command generation unit 5D and the alternating current detected by the current detection unit 4, and a position / speed estimation unit 7D that estimates the estimated position and estimated speed of the rotor of the rotating machine 30 based on the magnetic flux estimated by the magnetic flux estimation unit 6D and the voltage estimated by the voltage estimation unit 8D. The magnetic flux estimation unit 6D may be the magnetic flux estimation unit 6 of Embodiment 1 or the magnetic flux estimation unit 6A of Embodiment 2. The voltage estimation unit 8D may be the voltage estimation unit 8B of Embodiment 3 or the voltage estimation unit 8C of Embodiment 4.

[0168] <Configuration and Operation of Position and Velocity Estimation Unit 7D> Next, the configuration of the position and velocity estimation unit 7D will be described. Figure 24 is a functional block diagram showing the configuration of the position and velocity estimation unit 7D according to Embodiment 5. The position and velocity estimation unit 7D includes a first magnetic flux estimator 71Da, a second magnetic flux estimator 72Da, a first voltage estimator 71Db, a second voltage estimator 72Db, a magnetic flux / voltage combiner 73D, a phase calculator 74D, and a position estimation error convergence calculator 75D.

[0169] The first magnetic flux estimator 71Da estimates the first magnetic flux using the magnetic flux estimated by the magnetic flux estimation unit 6D, and the second magnetic flux estimator 72Da estimates the second magnetic flux using the magnetic flux estimated by the magnetic flux estimation unit 6D. For example, if the magnetic flux estimation unit 6D is the magnetic flux estimation unit 6 of Embodiment 1, the first magnetic flux estimator 71Da estimates the γδ axis magnetic flux φγ estimated by the first γδ axis magnetic flux estimator 61. Ld ,φδ Ld The estimated value and the dq-axis magnetic flux φd estimated by the dq-axis magnetic flux estimator 62 Ld , φq Ld The first magnetic flux estimator 71 estimates the first magnetic flux using the estimated value. The second magnetic flux estimator 72Da is the γδ-axis magnetic flux φγ estimated by the second γδ-axis magnetic flux estimator 63. Lq ,φδ Lq This corresponds to a second magnetic flux estimator 72 that estimates the second magnetic flux based on the estimated value.

[0170] Similarly, the first voltage estimator 71Db estimates the first voltage using the voltage estimated by the voltage estimation unit 8D, and the second voltage estimator 72Db estimates the second voltage using the voltage estimated by the voltage estimation unit 8D. For example, if the voltage estimation unit 8D is the voltage estimation unit 8B of Embodiment 3, the first voltage estimator 71Db estimates the γδ axis voltage eγ estimated by the first γδ axis voltage estimator 81B. Ld , eδ Ld The estimated value and the dq-axis voltage ed estimated by the dq-axis voltage estimator 82B Ld , eq Ld The first voltage estimator 71B corresponds to the first voltage estimator which estimates the first voltage using the estimated value. The second voltage estimator 72Db is the γδ axis voltage eγ estimated by the second γδ axis voltage estimator 83B. Lq , eδ LqThis corresponds to a second voltage estimator 72B that estimates a second voltage based on the estimated value. For example, the voltage estimation unit 8D may be the voltage estimation unit 8C of Embodiment 4, and configured as an αβ axis voltage estimator. That is, the voltage of an axis on a different Cartesian coordinate system than the magnetic flux estimator may be used.

[0171] <Operation of the Magnetic Flux / Voltage Combiner 73D> The operation of the magnetic flux / voltage combiner 73D will be explained below. The magnetic flux / voltage combiner 73D combines the first magnetic flux estimated by the first magnetic flux estimator 71Da, the second magnetic flux estimated by the second magnetic flux estimator 72Da, a weighting coefficient wda related to magnetic flux, the first voltage estimated by the first voltage estimator 71Db, the second voltage estimated by the second voltage estimator 72Db, and a weighting coefficient wdb related to voltage. Since voltage corresponds to the derivative of magnetic flux, for example, the combined magnetic flux can be normalized and the 90-degree vector advanced can be added together with the normalized combined voltage to obtain a combined voltage vector. Alternatively, the combined voltage can be normalized and the 90-degree vector can be delayed and then added together with the normalized combined magnetic flux. Alternatively, the calculation can be performed using weights instead of summation.

[0172] <Operation of Phase Calculator 74D> The phase calculator 74D estimates the phase difference Δθre using the arctangent function and the like.

[0173] <Operation of the Position Estimation Error Convergence Calculator 75D> The Position Estimation Error Convergence Calculator 75D calculates the estimated position and estimated velocity by configuring, for example, a PLL so that the estimated value of the phase difference Δθre becomes zero.

[0174] As described above, Embodiment 5 also achieves the same effects as Embodiments 1 to 4. That is, in the control device for a rotating machine, the magnetic flux estimation unit, which estimates the magnetic flux of the rotating machine, estimates at least one pair of d-axis magnetic flux and q-axis magnetic flux on a first Cartesian coordinate system having a d-axis and a q-axis, synchronized with the rotation of the rotating machine, and estimates multiple pairs of magnetic flux for each axis on a second Cartesian coordinate system having axes different from the d-axis and q-axis, based on different parameters of the rotating machine. The voltage estimation unit, which estimates the voltage of the rotating machine, estimates at least one pair of d-axis voltage and q-axis voltage on the first Cartesian coordinate system, and estimates multiple pairs of voltage for each axis on a second Cartesian coordinate system or a third Cartesian coordinate system having axes different from the d-axis and q-axis, based on different parameters of the rotating machine. The position and velocity estimation unit is configured to synthesize a set of magnetic flux from at least one d-axis and a set of magnetic flux from the q-axis, a set of magnetic flux from each axis on multiple second Cartesian coordinates, a set of voltage from at least one d-axis and a set of voltage from the q-axis, and a set of voltage from each axis on multiple second or third Cartesian coordinates, using a preset weighting coefficient. Based on the synthesized magnetic flux and voltage, it calculates the phase difference between the d-axis and q-axis and axes other than the d-axis and q-axis, and estimates the rotor's position and velocity from the phase difference. Since voltage corresponds to the derivative of magnetic flux, it is easy to synthesize the two, and by calculating the magnetic flux and voltage of multiple axes using different parameters, the rotor's position and velocity can be estimated with high accuracy. This makes it possible to obtain a control device for a rotating machine that is easy to adjust and optimize parameters for.

[0175] In the embodiments 1 to 4 described above, each has an effect by changing the way 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 a DC flow rate, 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.

[0176] While Figure 2 shows an example of the hardware configuration for controller 2 in Embodiment 1, it goes without saying that controllers 2A, 2B, 2C, and 2D in the other embodiments have a similar hardware configuration.

[0177] Embodiment 6. Below, a control device for a rotating machine according to Embodiment 6 will be described with reference to the figures. Embodiment 6 describes an example in which the model is modified using inductance values ​​different from those in Embodiments 1 to 5. In the following description, the differences from Embodiments 1 to 5 will be the main focus, and the same configuration and operation as in Embodiments 1 to 5 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 1, but as described below, a magnetic flux estimation unit 6E is provided instead of the magnetic flux estimation unit 6.

[0178] <Configuration and Operation of Magnetic Flux Estimation Unit 6E> Figure 25 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 6. In Figure 25, the magnetic flux estimation unit 6E includes a first γδ axis magnetic flux estimator 61E, a second γδ axis magnetic flux estimator 63E, a first dq axis magnetic flux estimator 62Ea, a second dq axis magnetic flux estimator 62Eb, and a weight setter 64E. In Embodiments 1 to 5, 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 and q-axis inductance Lq. In Embodiment 6, 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 equation (35) below.

[0179]

[0180] For example, the variable L in equation (35) * Substituting the d-axis inductance Ld into equation (2) yields equation (2), and substituting the q-axis inductance Lq yields equation (8). Converting equation (35) to the fixed coordinate system αβ coordinate system yields equation (36), and converting it to the estimated rotating coordinate system γδ coordinate system yields equation (37).

[0181]

[0182]

[0183] 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 (36), we obtain equation (38), and similarly, by rearranging equation (37), we obtain equation (39).

[0184]

[0185]

[0186] Note that variable L * The settings will be explained later.

[0187] <Operation of the first γδ-axis magnetic flux estimator 61E> The first γδ-axis magnetic flux estimator 61E receives a voltage command, current, and variable L. *1 The input is, for example, a state observer is used to determine the magnetic flux φγ L*1 ,φδ L*1 To estimate the variable L *1 is variable L * This is an arbitrary inductance input. For example, the state variables are estimated values ​​of current iγ and iδ and magnetic flux φγ. L*1 ,φδ L*1 Using the estimated values, construct a same-dimensional observer such as equation (7) or (13).

[0188] <Operation of the second γδ axis magnetic flux estimator 63E> The second γδ axis magnetic flux estimator 63E receives a voltage command, current, and variable L. *2 The input is, for example, a state observer is used to determine the magnetic flux φγ L*2 ,φδ L*2 To estimate the variable L *2 is variable L * An arbitrary inductance input to L *2 ≠L *1 For example, the state variables are estimated values ​​of currents iγ and iδ and magnetic flux φγ. L*2 ,φδ L*2 Using the estimated values, construct a same-dimensional observer such as equation (7) or (13).

[0189] <Operation of the first dq-axis magnetic flux estimator 62Ea> An operation command is input to the first dq-axis magnetic flux estimator 62Ea, and the d-axis magnetic flux φd L*1 = φm + (Ld - L *1 )id, q-axis magnetic flux φq L*1 = (Lq - L *1 ) Estimate iq. d-axis magnetic flux φd L*1 and q-axis magnetic flux φq L*1 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 *1 Using either of the following, d-axis magnetic flux φd L*1 , q-axis magnetic flux φq L*1 We just need to estimate that.

[0190] <Operation of the second dq-axis magnetic flux estimator 62Eb> An operation command is input to the second dq-axis magnetic flux estimator 62Eb, and the d-axis magnetic flux φd L*2 L*1=φm+(Ld-L *2 )id, q-axis magnetic flux φq L*2 = (Lq - L *2 ) Estimate iq. d-axis magnetic flux φd L*2 and q-axis magnetic flux φq L*2 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 *2 Using either of the following, d-axis magnetic flux φd L*2 , q-axis magnetic flux φq L*2 We just need to estimate that.

[0191] <Estimation of Position and Velocity> Figure 26 is a functional block diagram showing the configuration of the position and velocity estimation unit 7E according to Embodiment 6. The position and velocity estimation unit 7E receives the magnetic flux φγ output from the first γδ axis magnetic flux estimator 61E.L*1 ,φδ L*1 The estimated value and the magnetic flux φd output from the first dq-axis magnetic flux estimator 62Ea L*1 , φq L*1 The estimated values ​​are input to the first magnetic flux estimator 71E and the second γδ axis magnetic flux estimator 63E, and the output magnetic flux φγ is then used. L*2 ,φδ L*2 The estimated value and the magnetic flux φd output from the second dq-axis magnetic flux estimator 62Eb L*2 , φq L*2 It includes a second magnetic flux estimator 72E, a magnetic flux combiner 73E, a magnetic flux phase calculator 74E, and a position estimation error convergence calculator 75, to which the estimated value is input.

[0192] In this embodiment 6, the estimated magnetic flux φγ L*1 ,φδ L*1 ,φd L*1 , φq L*1 Using the relationship, we can estimate Δθre as follows:

[0193]

[0194]

[0195] Similar to embodiments 1 to 5, the relationship between equations (40) and (41) is used to convert the first magnetic flux estimator 71E and the second magnetic flux estimator 72E into a unit magnetic flux vector pointed to the d-axis, and the two magnetic fluxes are combined in the magnetic flux combiner 73E. Subsequently, the estimated position and estimated velocity are obtained by the magnetic flux phase calculator 74E and the position estimation error convergence calculator 75E.

[0196] <Variable L corresponding to inductance> * Settings > Variable L * Depending on the settings, the parameter sensitivity at the operating point will differ. For example, the variable L * If the sign of is set to be negative, then when an error occurs due to the q-axis inductance Lq, the errors that occur in the process of estimating the γδ-axis magnetic flux and the errors that occur in the process of estimating the dq-axis magnetic flux cancel each other out, thus making the sensitivity of the q-axis inductance Lq to position estimation as small as possible. Another example of setting is the variable L *Even if the sign 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.

[0197] Note that variable L * Although it was explained as a variable L* corresponding to inductance, values ​​other than inductance, such as the rotor magnetic 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.

[0198] In Figure 27, the variable L *1 The value and variable L *2 An example of setting the value of L is shown. With this setting, the magnetic flux estimation by the first magnetic flux estimator 71E and the first dq-axis magnetic flux estimator 62Ea tends to have a smaller sensitivity to errors in the q-axis inductance Lq, and the magnetic flux estimation by the second magnetic flux estimator 72E and the second dq-axis magnetic flux estimator 62Eb tends to have a smaller sensitivity to errors in the d-axis inductance Ld and rotor magnetic flux φm. As described above, the variable L has different parameter sensitivities. *1 and L *2 By setting the parameters and changing the weights to be combined according to the operating point, the rotor position and speed can be estimated with high accuracy. This makes it possible to obtain a control device for a rotating machine that is easy to adjust and optimize parameters for. Furthermore, compared to embodiments 1 to 5, the degree of freedom of the parameters that can be set as inductance is greater, allowing for fine adjustments according to the rotating machine used, its control specifications, and the operating point.

[0199] In this embodiment, two types of magnetic flux estimators are shown, but there may be three, four, or even more types. Furthermore, the estimator may be implemented in a fixed coordinate system instead of a rotating coordinate system, and voltage may be used instead of magnetic flux.

[0200] As described above, Embodiment 6 has the same effects as Embodiment 1. That is, in the control device for a rotating machine according to Embodiment 6, the magnetic flux estimation unit determines the set of d-axis magnetic flux and q-axis magnetic flux on a first Cartesian coordinate system having a d-axis and a q-axis, which is synchronized with the rotation of the rotating machine, as a variable L corresponding to inductance. *1 and L *2 The system is configured to estimate two sets of magnetic fluxes using the first Cartesian coordinate system, which has axes different from the d-axis and the q-axis, and to estimate two sets of γ-axis magnetic flux and δ-axis magnetic flux on the first Cartesian coordinate system, which has axes different from the d-axis and the q-axis, based on combinations that include different parameters such as inductance Ld, Lq, and magnetic flux. The position and velocity estimation unit combines the two sets of d-axis magnetic flux and q-axis magnetic flux and the two sets of γ-axis magnetic flux and δ-axis magnetic flux using a preset weighting coefficient, calculates the phase difference between the first Cartesian coordinate system and the second Cartesian coordinate system based on the combined magnetic flux, and estimates the rotor position and velocity from the phase difference. This makes it possible to obtain a control device for a rotating machine that is easy to adjust and optimize parameters, and to estimate the rotor position and velocity with high accuracy.

[0201] In this embodiment 6, when the magnetic flux estimation unit of embodiment 1 estimates 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 variable L corresponding to the inductance is used. * An example of estimating the position of a rotating machine using the variable L 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 use the following. In a configuration that includes both a configuration for estimating the magnetic flux of each axis and a configuration for estimating the voltage, as in Embodiment 5, the variable L * You may also use the variable L in embodiments 2 to 5. * It goes without saying that using the same method as in Embodiment 6 will produce the same effects.

[0202] 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.

[0203] 2, 2A, 2B, 2C, 2D: Controller, 3: Power conversion unit, 4: Current detection unit, 5, 5A, 5B, 5C, 5D: Voltage command generation unit, 6, 6A, 6D, 6E: Magnetic flux estimation unit, 61, 61E: First γδ axis magnetic flux estimator, 62, 62A: dq axis magnetic flux estimator, 62Ea: First dq axis magnetic flux estimator, 62Eb: Second dq axis magnetic flux estimator, 63, 63E: Second γδ axis magnetic flux estimator, 64, 64A, 64E, 84B, 84C: Weight setter, 65A: First αβ axis magnetic flux estimator, 66A: Second αβ axis magnetic flux estimator, 7, 7A, 7B, 7C, 7D, 7E: Position / velocity estimation unit, 71, 71A, 71Da, 71E: First magnetic flux estimator, 71B, 71C, 71Db: First voltage estimator, 72, 72A, 72Da, 72E: Second magnetic flux estimator, 72B, 72C, 72Db: Second voltage estimator, 73, 73A, 73E: Magnetic flux combiner, 73B, 73C: Voltage combiner, 73D: Magnetic flux / voltage combiner, 74, 74A, 74E: Magnetic flux phase calculator, 74B, 74C: Voltage phase calculator, 74D: Phase calculator, 75, 75B, 75D, 75E: Position estimation error convergence calculator, 76A, 76C: Velocity estimator, 8B, 8C, 8D: Voltage estimation unit, 81B: First γδ axis voltage estimator, 82B, 82C: dq axis voltage estimator, 83B: second γδ axis voltage estimator, 85C: first αβ axis voltage estimator, 86C: second αβ axis voltage estimator, 10, 10B, 10D: 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 at least one pair of d-axis magnetic flux and q-axis magnetic flux on a first Cartesian coordinate system having a d-axis and a q-axis that is synchronized with the rotation of the rotating machine, and estimates multiple pairs of magnetic flux for each axis on a second Cartesian coordinate system having axes different from the d-axis and q-axis based on different parameters of the rotating machine; and the position / velocity estimation unit synthesizes at least one pair of d-axis magnetic flux and q-axis magnetic flux and multiple pairs of magnetic flux for each axis on the second Cartesian coordinate system using a preset weight coefficient. A control device for a rotating machine, which calculates the phase difference between the first orthogonal coordinate and the second orthogonal coordinate based on the combined magnetic flux, and estimates the position and velocity of the rotor from the phase difference.

2. The control device for a rotating machine according to claim 1, wherein the set of magnetic fluxes for each axis on the second Cartesian coordinate system is a set of γ-axis magnetic flux and δ-axis magnetic flux for 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, and the set of γ-axis magnetic flux and δ-axis magnetic flux is a plurality of sets including: a first set of magnetic fluxes estimated using a voltage command generated by the voltage command generation unit, the current of the rotating machine, and the inductance of the rotating machine as parameters of the rotating machine; and a second set of magnetic fluxes estimated using a voltage command generated by the voltage command generation unit, the current of the rotating machine, and an inductance different from the inductance of the rotating machine used in the first set of magnetic fluxes as parameters of the rotating machine.

3. The control device for a rotating machine according to claim 2, wherein, in the case where the inductance of the rotating machine used to estimate the first magnetic flux set is the d-axis inductance, the set of the d-axis magnetic flux and the q-axis magnetic flux includes the set of the d-axis magnetic flux estimated based on the rotor magnetic flux of the rotating machine and the q-axis magnetic flux estimated using the rotor magnetic flux of the rotating machine and the first magnetic flux set.

4. The control device for a rotating machine according to claim 3, wherein the magnetic flux estimation unit estimates the d-axis magnetic flux and the q-axis magnetic flux set such that the first γδ-axis magnetic flux, which is the combined magnetic flux of the first magnetic flux set, and the dq-axis magnetic flux, which is the combined magnetic flux of the d-axis magnetic flux and the q-axis magnetic flux set coincide, and the position and velocity estimation unit combines the first γδ-axis magnetic flux and the second γδ-axis magnetic flux, which is the combined magnetic flux of the second magnetic flux set, using the weight coefficient, calculates the phase difference between the d-axis and the q-axis and the γ-axis and the δ-axis based on the combined magnetic flux, and estimates the position and velocity of the rotor from the phase difference.

5. The control device for a rotating machine according to claim 1, wherein the set of magnetic fluxes for each axis on the second Cartesian coordinate system is a set of α-axis magnetic flux and β-axis magnetic flux for 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, and the set of α-axis magnetic flux and β-axis magnetic flux is a plurality of sets including a third set of magnetic fluxes estimated using the voltage command generated by the voltage command generation unit, the current of the rotating machine, and the inductance of the rotating machine as parameters of the rotating machine, and a fourth set of magnetic fluxes estimated using the voltage command generated by the voltage command generation unit, the current of the rotating machine, and an inductance different from the inductance of the rotating machine used in the third set of magnetic fluxes as parameters of the rotating machine.

6. The control device for a rotating machine according to claim 5, wherein, in the case where the inductance of the rotating machine used to estimate the third magnetic flux set is the d-axis inductance, the set of the d-axis magnetic flux and the q-axis magnetic flux includes the set of the d-axis magnetic flux estimated based on the rotor magnetic flux of the rotating machine and the q-axis magnetic flux estimated using the rotor magnetic flux of the rotating machine and the third magnetic flux set.

7. The control device for a rotating machine according to claim 6, wherein the magnetic flux estimation unit estimates the d-axis magnetic flux and the q-axis magnetic flux set such that the first αβ-axis magnetic flux, which is the combined magnetic flux of the third magnetic flux set, and the dq-axis magnetic flux, which is the combined magnetic flux of the d-axis magnetic flux and the q-axis magnetic flux set coincide, and the position and velocity estimation unit combines the first αβ-axis magnetic flux and the second αβ-axis magnetic flux, which is the combined magnetic flux of the fourth magnetic flux set, using the weighting coefficient, calculates the phase difference between the d-axis and the q-axis and the α-axis and the β-axis based on the magnetic flux combined using the weighting coefficient, and estimates the position and velocity of the rotor from the phase difference.

8. The control device for a rotating machine according to any one of claims 1 to 7, wherein the magnetic flux estimation unit sets the weight coefficient based on any of the operation commands, which include a torque command, a rotational speed command, a current command, and a voltage command, the current flowing through the rotating machine, and the voltage applied to the rotating machine.

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 at least one pair of d-axis voltages and q-axis voltages on a first Cartesian coordinate system having a d-axis and a q-axis that is synchronized with the rotation of the rotating machine, and estimates multiple pairs of voltages for each axis on a second Cartesian coordinate system having axes different from the d-axis and the q-axis based on different parameters of the rotating machine; the position / velocity estimation unit synthesizes at least one pair of the d-axis voltages and q-axis voltages and multiple pairs of voltages for each axis on the second Cartesian coordinate system using a preset weight coefficient; and calculates the phase difference between the first Cartesian coordinate system and the second Cartesian coordinate system based on the synthesized 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 the set of voltages for each axis on the second Cartesian coordinate system is a set of γ-axis voltages and δ-axis voltages for axes on a Cartesian coordinate system having a γ-axis and a δ-axis that are synchronized with the rotation of the rotating machine, and the set of γ-axis voltages and δ-axis voltages is a plurality of sets including: a first set of voltages estimated using a voltage command generated by the voltage command generation unit, the current of the rotating machine, and the inductance of the rotating machine as parameters of the rotating machine; and a second set of voltages estimated using a voltage command generated by the voltage command generation unit, the current of the rotating machine, and an inductance different from the inductance of the rotating machine used in the first set of voltages as parameters of the rotating machine.

12. The control device for a rotating machine according to claim 11, wherein, in the case where the inductance of the rotating machine used to estimate the first voltage set is the d-axis inductance, the set of the d-axis voltage and the q-axis voltage includes the set of the d-axis voltage estimated based on the rotor magnetic flux of the rotating machine and the q-axis voltage estimated using the rotor magnetic flux of the rotating machine and the first voltage set.

13. The control device for a rotating machine according to claim 12, wherein the voltage estimation unit estimates the d-axis voltage and the q-axis voltage set such that the first γδ-axis voltage, which is a combined voltage of the first voltage set, matches the dq-axis voltage, which is a combined voltage of the d-axis voltage and the q-axis voltage set; the position and velocity estimation unit combines the first γδ-axis voltage and the second γδ-axis voltage, which is a combined voltage of the second voltage set, using the weight coefficient; calculates the phase difference between the d-axis and the q-axis and the γ-axis and the δ-axis based on the combined voltage; and estimates the position and velocity of the rotor from the phase difference.

14. The control device for a rotating machine according to claim 10, wherein the set of voltages for each axis on the second Cartesian coordinate system is a set of α-axis voltages and β-axis voltages for axes 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, and the set of α-axis voltages and β-axis voltages is a plurality of sets including a third set of voltages estimated using a voltage command generated by the voltage command generation unit, the current of the rotating machine, and the inductance of the rotating machine as parameters of the rotating machine, and a fourth set of voltages estimated using a voltage command generated by the voltage command generation unit, the current of the rotating machine, and an inductance different from the inductance of the rotating machine used in the third set of voltages as parameters of the rotating machine.

15. The control device for a rotating machine according to claim 14, wherein, in the case where the inductance of the rotating machine used to estimate the third voltage set is the d-axis inductance, the set of the d-axis voltage and the q-axis voltage includes the set of the d-axis voltage estimated based on the rotor magnetic flux of the rotating machine and the q-axis voltage estimated using the rotor voltage of the rotating machine and the third voltage set.

16. The control device for a rotating machine according to claim 15, wherein the voltage estimation unit estimates the d-axis voltage and the q-axis voltage set such that the first αβ-axis voltage, which is the combined magnetic flux of the third voltage set, matches the dq-axis voltage, which is the combined voltage of the d-axis voltage and the q-axis voltage set; the position and velocity estimation unit combines the first αβ-axis voltage and the second αβ-axis voltage, which is the combined voltage of the fourth voltage set, using the weight coefficient; calculates the phase difference between the d-axis and the q-axis and the α-axis and the β-axis based on the voltage combined using the weight coefficient; and estimates the position and velocity of the rotor from the phase difference.

17. The control device for a rotating machine according to any one of claims 10 to 16, wherein the voltage estimation unit sets the weight coefficient based on any of the operation commands, which include a torque command, a rotational speed command, a current command, and a voltage command, the current flowing through the rotating machine, and the voltage applied to the rotating machine.

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.

19. 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; 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 magnetic flux estimation unit estimates at least one pair of d-axis magnetic flux and q-axis magnetic flux on a first Cartesian coordinate system having a d-axis and a q-axis that is synchronized with the rotation of the rotating machine, and estimates multiple pairs of magnetic flux for each axis on a second Cartesian coordinate system having axes different from the d-axis and the q-axis based on different parameters of the rotating machine; and the voltage estimation unit A control device for a rotating machine, comprising: estimating at least one pair of d-axis voltage and q-axis voltage on the first Cartesian coordinate system, estimating multiple pairs of voltages for each axis on the second Cartesian coordinate system or the third Cartesian coordinate system having axes different from the d-axis and q-axis based on different parameters of the rotating machine; the position and velocity estimation unit synthesizes at least one pair of the d-axis magnetic flux and q-axis magnetic flux, multiple pairs of magnetic fluxes for each axis on the second Cartesian coordinate system, at least one pair of the d-axis voltage and q-axis voltage, and multiple pairs of voltages for each axis on the second Cartesian coordinate system or the third Cartesian coordinate system using a preset weighting coefficient; calculating the phase difference between the d-axis and q-axis and axes different from the d-axis and q-axis based on the combined magnetic flux and voltage using the weighting coefficient; and estimating the position and velocity of the rotor from the phase difference.