Technical Field
[0001] The present invention relates to a technique for controlling a synchronous motor comprising a field and an armature.
[0002] More particularly, the present invention relates to a technique for controlling a rotary electric motor on the basis of a so-called primary magnetic flux which is a synthesis of a field flux that the field generates and a magnetic flux of an armature reaction generated by an armature current flowing in the armature.
Background Art
[0003] Conventionally, various controls of a rotary electric motor on the basis of a primary magnetic flux, i.e., a so-called primary magnetic flux controls have been proposed. Briefly speaking, the primary magnetic flux control is a technique for stably controlling the rotary electric motor by controlling the primary magnetic flux of the rotary electric motor in accordance with a command value thereof.
[0004] It is assumed, for example, that a phase of a field flux Λ0 is employed at a d axis in rotating coordinate system, a phase of a primary magnetic flux λ1 is employed at a δ axis in another rotating coordinate system, and a phase difference of the δ axis with respect to the d axis is a load angle ϕ. Herein, however, a γ axis is employed at a 90-degree leading phase with respect to the δ axis. Further, a δc axis and a γc axis are defined as control axes in the rotating coordinate system which is employed in the primary magnetic flux control. The δc axis and the γc axis are corresponding to the δ axis and the γ axis, respectively, and a phase difference of the δc axis with respect to the d axis is assumed as ϕc.
[0005] In this case, a command value of the primary magnetic flux λ1 (hereinafter, referred to as a "primary magnetic flux command value") has a δc-axis component Λδ*, and a γc-axis component is zero. Therefore, when the primary magnetic flux λ1 is equal to the primary magnetic flux command value, the δc-axis component λ1δc of the primary magnetic flux λ1 is equal to the δc-axis component Λδ*, the phase difference ϕc is equal to the load angle ϕ, and the δc axis is coincident with the δ axis.
[0006] The δc-axis component λ1δc and the γc-axis component λ1γc of the primary magnetic flux λ1 vary with a change of the primary magnetic flux command value, a variation in a load, an influence of control disturbance, or/and the like. For example, the change of the primary magnetic flux command value and the variation in the load invites a transient change of the primary magnetic flux λ1, and the control disturbance invites a variation in the γc axis/δc axis. As states where the control disturbance occurs, for example, a state where a voltage applied to the rotary electric motor is different from a voltage command due to an influence of a time delay, an on-loss, and dead time, and a state where there is a deviation between a device constant of the rotary electric motor and that assumed by a control system. Therefore, a deviation arises between the primary magnetic flux λ1 and the primary magnetic flux command value, and accordingly a deviation also arises between the load angle ϕ and the phase difference ϕc.
[0007] In the primary magnetic flux control, when there is a deviation between the primary magnetic flux λ1 and the primary magnetic flux command value, a control, for example, of a voltage command value to be corrected is performed so that the δc-axis component λ1δc of the primary magnetic flux λ1 may be made equal to the δc-axis component Λδ* of the primary magnetic flux command value and the γc-axis component λ1γc of the primary magnetic flux λ1 may become zero. The phase difference ϕc is thereby coincident with the load angle ϕ.
[0008] In such a primary magnetic flux control, control is made with a torque of the rotary electric motor being made in direct proportion to a γc-axis component of an armature current, not depending on a rotation angular velocity thereof.
[0009] Among the following prior-art documents, in Non-Patent Document 6, the γc axis and the δc axis are exchanged and employed, as compared with those in the other prior-art documents.
Prior-Art Documents
Patent Documents
[0010]
[Patent Document 1] Japanese Patent No. 3672761
[Patent Document 2] Japanese Patent Application Laid Open Gazette No. 4-91693
[Patent Document 3] US 2008/0191656 A1
Non-Patent Documents
[0011]
[Non-Patent Document 1] Hotta, Asano, and Tsunehiro, "Method of controlling Position Sensorless DC brushless motor", 1988 Tokai-Section Joint Conference of the Institutes of Electrical and Related Engineers, p. 161
[Non-Patent Document 2] Kaku and Tsunehiro, "A Novel Technique for a DC Brushless Motor Having No Position-Sensors ", 1990 Tokai-Section Joint Conference of the Institutes of Electrical and Engineers, p. 172
[Non-Patent Document 3] Kaku, Yamamura, and Tsunehiro, "A Novel Technique for a DC Brushless Motor Having No Position-Sensors", IEEJ Transaction on Industry Applications, 1991, Volume 111, No. 8, pp. 639 to 644
[Non-Patent Document 4] Urita, Tsukamoto, and Tsunehiro, "Constant estimation method for synchronous machines with the primary magnetic flux controlled", 1998 Tokai-Section Joint Conference of the Institutes of Electrical Engineers, p. 101
[Non-Patent Document 5] Urita, Yamamura, and Tsunehiro, "On General Purpose Inverter for Synchronous Motor Drive", IEEJ Transaction on Industry Applications, 1999, Volume 119, No. 5, pp. 707 to 712
[Non-Patent Document 6] Yabe and Sakanobe, "A Sensor-less Drive of IPM Motor with Over-modulation PWM", The papers of Joint Technical Meeting on Rotating Machinery, IEE Japan, 2001 (159), pp. 7 to 12
Summary of Invention
Problems to be Solved by the Invention
[0012] In Patent Document 1, a feedback is achieved by using a deviation in a δ-axis component, not by using a γ-axis component of the armature current. Further, in Non-Patent Document 3, assumed is a range in which the load angle ϕ can equally approximate a sine value sinϕ thereof.
[0013] Still further, in Patent Document 3 there is described a synchronous machine control apparatus which can produce a torque level as close as possible to a desired torque command level while considering limitations on an output current of a power conversion unit.
[0014] In any one of the above documents, however, except Non-Patent Document 6, with respect to an inductance of the armature winding, a d-axis component thereof and a γ-axis component which is 90-degree phase advance therewith are handled isotropically, and the technique cannot be applied to a rotary electric motor having so-called saliency such as an interior permanent magnet rotary electric motor.
[0015] Further, the amount of feedbacks employed in any one of the prior-art documents does not include any information of the load angle ϕ. For example, a δ-axis current and a γ-axis current are employed in Patent Document 2 and Non-Patent Documents 1, 2, 3 and 6 and the δ-axis current is employed in Patent Document 1 and Non-Patent Documents 4 and 5, respectively, for the amount of feedbacks. For this reason, in an area where the load angle ϕ is large, the primary magnetic flux cannot be controlled to a desired value. When a large torque is outputted, the load angle ϕ also becomes large. Therefore, in the conventional primary magnetic flux control, it is hard to appropriately perform a stable drive or a high-efficient drive in the area where the torque is large.
[0016] In order to solve the above problem, it is an object of the present invention to provide a technique to apply the primary magnetic flux control to the rotary electric motor even having saliency by performing a feedback based on the deviation of the primary magnetic flux. It is another object of the present invention to provide a primary magnetic flux control in which a drive can be performed at a stable and high-efficient operating point even in an area where an output torque is large.
Means for Solving the Problems
[0017] An electric motor controller according to the present invention is a device for controlling a primary magnetic flux ([λ1]) on a rotary electric motor including an armature having an armature winding and a rotor which is a field rotating relatively to the armature, the primary magnetic flux being a synthesis of a field flux (Λ0) that the field generates and a magnetic flux (λa: id · Ld, iq · Lq) of an armature reaction generated by an armature current ([I]) flowing in the armature.
[0018] A first aspect of the electric motor controller according to the present invention includes a first coordinate conversion unit (101) that converts the armature current into a first current ([i]) in a rotating coordinate system (δc-γc) having a predetermined phase (
φ c) with respect to rotation of the rotor, a first calculation unit (102) that obtains a first term ([F]) as a sum of an inductive voltage (ω* • [Λ1*]) by a primary magnetic flux command value ([Λ1*]) which is a command value of the primary magnetic flux and a voltage drop ({R}[i]) by the first current on the basis of a voltage equation at a time when the rotary electric motor is in a steady state, a second calculation unit (103A, 103B) that obtains a sum of a second term ([B]) obtained by performing an operation expressed by a non-zero matrix ({K}) on a deviation ([ΔΛ]) of the primary magnetic flux with respect to the primary magnetic flux command value and the first term as a first voltage command value ([v*]) which is a command value of a voltage to be applied to the rotary electric motor in the rotating coordinate system, and a second coordinate conversion unit (104) that coordinate-converts the first voltage command value into a second voltage command value ([V*]) which is a command value of the voltage to be applied to the rotary electric motor in another coordinate system.
[0019] A second aspect of the electric motor controller according to the present invention is the first aspect thereof in which the second calculation unit (103A) employs an estimation value ([λ1^]) of the primary magnetic flux as the primary magnetic flux.
[0020] A third aspect of the electric motor controller according to the present invention, which is the second aspect thereof, further includes a primary magnetic flux estimation unit (105) that obtains the estimation value ([λ1^]) of the primary magnetic flux from the predetermined phase (ϕc), a first component (Lq) orthogonal to the field flux of an inductance of the armature winding, a second component (Ld) in phase with the field flux of the inductance, the first current, and the field flux (Λ0).
[0021] A fourth aspect of the electric motor controller according to the present invention, which is the second or third aspect thereof, further includes a primary magnetic flux command correction unit (107) that corrects the primary magnetic flux command value ([Λ1*]) to output a primary magnetic flux command correction value ([Λ1**]) by using the predetermined phase (ϕc), a first component (Lq) orthogonal to the field flux of an inductance of the armature winding, a second component (Ld) in phase with the field flux of the inductance, the first current, the field flux (Λ0), and the estimation value ([λ1^]) of the primary magnetic flux. The second calculation unit (103B) employs the primary magnetic flux command correction value as the primary magnetic flux command value.
[0022] An estimation value of the predetermined phase may be employed as the predetermined phase. For example, the predetermined phase (ϕc) is obtained from the first voltage command value ([v*]), a resistance value ({R}) of the armature winding, the first component (Lq), a rotation angular velocity (ω*) of the rotor, and the first current ([i]).
Effects of the Invention
[0023] In the electric motor controller of the first aspect according to the present invention, since the second term obtained on the basis of the deviation of the primary magnetic flux functions as a feedback for the first voltage command value, the second term has information of a load angle, and even when the deviation between the predetermined phase and the load angle is large, it becomes easier to perform the primary magnetic flux control while correcting this. Further, the primary magnetic flux control does not depend on whether or not there is saliency.
[0024] In the electric motor controller of the second aspect according to the present invention, it is not necessary to perform a direct detection of the primary magnetic flux.
[0025] In the electric motor controller of the third aspect according to the present invention, it is possible to perform the primary magnetic flux control while correcting the deviation of the load angle regardless of whether or not there is saliency.
[0026] In the electric motor controller of the fourth aspect according to the present invention, it is possible to achieve accuracy on the same level with the second or third aspect, regardless of a method of detecting or estimating the primary magnetic flux.
[0027] These and other objects, features, aspects and advantages of the present invention will become more apparent from the following detailed description of the present invention when taken in conjunction with the accompanying drawings.
Brief Description of Drawings
[0028]
[Fig. 1] A vector diagram illustrating a first preferred embodiment.
[Fig. 2] A vector diagram illustrating the first preferred embodiment.
[Fig. 3] A vector diagram illustrating the first preferred embodiment.
[Fig. 4] A block diagram illustrating the first preferred embodiment.
[Fig. 5] A block diagram illustrating the first preferred embodiment.
[Fig. 6] A block diagram illustrating a second preferred embodiment.
[Fig. 7] A vector diagram illustrating the second preferred embodiment.
[Fig. 8] A block diagram illustrating the second preferred embodiment.
[Fig. 9] A block diagram illustrating a third preferred embodiment.
[Fig. 10] A block diagram illustrating the third preferred embodiment.
[Fig. 11] A vector diagram illustrating a variation.
Description of Embodiments
[0029] In the following embodiments, description will be made with a three-phase interior permanent magnet rotary electric motor taken an example. It is obvious that a rotary electric motor of polyphase, other than three-phase, as well as a rotary electric motor other than an interior permanent magnet type is also applicable.
<The First Preferred Embodiment>
[0030] Figs. 1 and 2 are vector diagrams each illustrating a primary magnetic flux control.
[0031] In the primary magnetic flux control, set is a δc-γc coordinate system which is in phase advance with a d-q coordinate system (a d axis in phase with a field flux Λ0, a q axis is in 90-degree phase advance with the d axis) with reference to a phase of the field flux Λ0 (i.e. with the rotation of a rotor), by a phase difference ϕc. Then, a voltage to be applied to the rotary electric motor (a γc-axis component and a δc-axis component thereof are assumed to be vγc and vδc, respectively) is adjusted so that a δc axis may be coincident with a δ axis, which is in phase with the primary magnetic flux.
[0032] First, Fig. 1 shows a case where the phase difference ϕc is coincident with the load angle ϕ. As shown in Fig. 1, a magnetic flux λa of an armature reaction is a synthesis of a magnetic flux Lq • iq in a q-axis positive direction and a magnetic flux Ld • id in a d-axis negative direction. Then, the primary magnetic flux is a synthesis of the magnetic flux λa and the field flux AO and takes a positive value Λδ (coincident with a command value Λδ* thereof) in the δ axis (coincident with the δc axis in Fig. 1).
[0033] An inductive voltage ω • Λδ (= ω • Λδ*) by the primary magnetic flux appears on a γc axis (herein, coincident with a γ axis). Further, for explanatory convenience, when it is grasped that the inductive voltage ω • Λ0 in a case where the armature reaction is disregarded (in other words, it is assumed that the magnetic flux λa = 0) is an inductive voltage by the field flux, the inductive voltage ω • Λ0 appears on the q axis.
[0034] Therefore, the inductive voltage by the armature reaction is represented as a synthesis of a voltage ω • Lq • iq in the d-axis negative direction and a voltage ω • Ld • id in a q-axis negative direction.
[0035] Introducing a resistance value R of an armature winding, a voltage drop by an armature current appears as a voltage R • iδc on the δc axis and as a voltage R • iγc on the γc axis.
[0036] Therefore, assuming that the γc-axis component and the δc-axis component of the voltage to be applied to the rotary electric motor are a voltage vγc and a voltage vδc, respectively, when the primary magnetic flux is coincident with the primary magnetic flux command value, vγc - R • iγc = ω • Λδ*, vδc = R • iδc are true, as shown in Fig. 1.
[0037] Now, the δc-axis component λ1δc and/or the γc-axis component λ1γc of the primary magnetic flux λ1 vary with a variation in the load, an influence of control disturbance, or/and the like. Therefore, as shown in Fig. 2, a deviation arises between the phase difference ϕc and the load angle ϕ. From the definition of the δ axis, since the primary magnetic has no γ-axis component, the primary magnetic flux which actually occurs is also referred to as a primary magnetic flux Λδ.
[0038] In the δc-γc rotating coordinate system on which the primary magnetic flux control is performed, a control is made so that the δc-axis component λ1δc of the primary magnetic flux Λδ may be coincident with (a δc-axis component Λδ* of) the primary magnetic flux command value and the γc-axis component λ1γc of the primary magnetic flux Λδ may be coincident with (a γc-axis component Λγ* = 0 of) the primary magnetic flux command value.
[0039] In order to make the δc-axis component λ1δc coincident with the δc-axis component Λδ* of the primary magnetic flux command value, the inductive voltage ω • Λδ* on the γc axis needs to appear. Also in consideration of the voltage drop in the armature winding, it is necessary to set the voltage command value as a sum of the inductive voltage ω • Λδ* and the voltage drop. Herein, the sum is represented as a feedforward term [F] = [Fγ Fδ]
^{t} (the former component represents the γc-axis component and the latter component represents the δc-axis component: the superscript after blackets "t" represents a transpose of a matrix: the same applies to the following unless otherwise indicated). Eqs. (1) and (2) are derived from a voltage equation of a rotary electric motor, where a differential operator p is introduced.
[Mathematical 1]
[0040] In Eq. (1), it can be grasped that a matrix {R} is a tensor indicating a resistance of the armature winding, and as shown in Eq. (2), the matrix {R} has the same component R both on the δc axis and the γc axis and off-diagonal components are zeros. Further, a current vector [i] = [iγc iδc]
^{t} indicating a current flowing the armature winding is introduced. The first term on the right side of each of Eqs. (1) and (2) represents a voltage drop {R}[i]. The third term of Eq. (2) is a transient term and can be disregarded. This is because an influence of the transient term can be also handled as the deviation between the load angle ϕ and the phase difference ϕc, as described above.
[0041] Further, assuming herein that both the δc axis and the δ axis rotate with respect to the d axis at an angular velocity ω which is equal to a command value ω* of the angular velocity, ω = ω*. By appropriately performing the primary magnetic flux control, ω = ω* is true.
[0042] When the phase difference ϕc is equal to the load angle ϕ, since the δc-axis component Λδ* of the primary magnetic flux command value [Λ1*] = [0 Λδ*]
^{t} in the δc-γc rotating coordinate system is coincident with the primary magnetic flux Λδ, the feedforward term [F] is the voltage command value [v*] for the rotary electric motor (also see Fig. 1).
[0043] When the γ axis is not coincident with the γc axis, however, the phase (ϕc - ϕ) is not resolved only by employing only the feedforward term [F] as the voltage command value. In the primary magnetic flux control, since no control is made on the basis of the deviation with respect to the voltage command value [v*] of the voltage [v] to be applied to the rotary electric motor, a voltage deviation [ve] = [v] - [v*] arises. The phase difference (ϕc - ϕ) thereby remains. Therefore, in order to resolve the phase difference (ϕc - ϕ) (in order to make ϕc = ϕ), as the voltage command value [v*] = [vγc* vδc*]
^{t} to be determined with respect to the primary magnetic flux Λδ, a vector represented at a position obtained by rotationally transfer the feedforward term [F] represented at a position J1 to be in phase advance by the phase difference (ϕc - ϕ) (in the counterclockwise direction in Fig. 2) should be employed. This is because only with the feedforward term [F], a voltage [v] arises at a position in phase lag with the voltage command value [v*] by the phase difference (ϕc - ϕ).
[0044] The matrix operation of such rotational transfer of the vector, however, cannot be performed. This is because the load angle ϕ which actually arises is not known.
[0045] As is clear from Fig. 2, the difference between the positions J1 and J2 is caused by the difference [ΔΛ] = [0-λ1γc Λδ*-λ1δc]
^{t} between the magnetic flux in the primary magnetic flux control, which takes the primary magnetic flux command value Λδ* in the δc-γc rotating coordinate system (on the δc axis), and the primary magnetic flux Λδ which actually arises in a δ-γ rotating coordinate system (on the δ axis). The difference is grasped, by the substances thereof, as the deviation of the primary magnetic flux with respect to the primary magnetic flux command value.
[0046] Therefore, by calculating the voltage command value [v*] with the sum of the feedback term [B] = [Bγ Bδ]
^{t} and the feedforward term [F] (see Fig. 3), in spite of the presence of the voltage deviation [ve], it is possible to reduce the difference between the feedforward term [F] and the voltage [v]. Now, the feedback term [B], however, can be obtained by Eqs. (3) and (4).
[Mathematical 2]
[0047] At least one of components Kγγ, Kγδ, Kδγ, and Kδδ in a matrix {K} for performing an arithmetic operation on the deviation [ΔΛ] of the magnetic flux is not zero. In other words, the matrix {K} is a non-zero matrix.
[0048] The feedforward term [F] functions as a feedforward based on the armature current and the feedback term [B] functions as a feedback based on the deviation of the magnetic flux.
[0049] When both the two elements of a column vector [Kγγ Kδγ]
^{t} forming the matrix {K} are not zero, for example, the γc-axis component (-λ1γc) of the deviation of the magnetic flux can be fed back to the voltage command value [v*] with respect to both the γc axis and the δc axis. Alternatively, when both the two elements of a column vector [Kγδ Kδδ]
^{t} are not zero, the δc-axis component (Λδ*-λ1δc) of the deviation of the magnetic flux can be fed back to the voltage command value [v*] with respect to both the γc axis and the δc axis.
[0050] Further, when both the column vectors [Kγγ Kδγ]
^{t} and [Kγδ Kδδ]
^{t} are non-zero vectors, the magnetic flux component of both axes can be fed back, and it is therefore possible to improve stability and responsibility of the control system.
[0051] Since the feedback term [B] functions as the feedback based on the deviation [ΔΛ] with respect to the voltage command value, if the phase difference ϕc deviates with respect to the load angle ϕ, it becomes easier to perform the primary magnetic flux control by correcting this. In conformity with Fig. 3, the deviation [ΔΛ] decreases, the phase difference ϕc approximates the load angle ϕ, and the γc axis approximates the γ axis. Then, when such a feedback proceeds and the γc axis becomes coincident with the γ axis, λ1γc = 0 and λ1δc = Λδ*, and the state shown in Fig. 1 is achieved. In other words, Fig. 3 is a vector diagram showing a condition while the phase difference ϕc approximates the load angle ϕ.
[0052] As is clear from Eq. (4), the voltage command value can be determined in consideration of the feedback based on the deviation [ΔΛ] of the primary magnetic flux. The matrix {K} functioning as a feedback gain may or may not have a diagonal component or an off-diagonal component only if the matrix {K} is a non-zero matrix. Further, each component may include an integral element.
[0053] On the basis of the above idea, Fig. 4 is a block diagram showing a configuration of an electric motor controller 1 in accordance with the present embodiment and its peripheral devices.
[0054] A rotary electric motor 3 is a three-phase electric motor, and comprises a not-shown armature and a rotor which is a field. As a technical common sense, the armature has an armature winding and the rotor rotates relatively to the armature. The field comprises, for example, a magnet which generates a field flux. Herein, description will be made on a case where an interior permanent magnet type is adopted.
[0055] A voltage supply source 2 comprises, for example, a voltage control type inverter and a control unit thereof, and applies a three-phase voltage to the rotary electric motor 3 on the basis of a three-phase voltage command value [V*] = [Vu* Vv* Vw*]
^{t}. A three-phase current [I] = [Iu Iv Iw]
^{t} thereby flows in the rotary electric motor 3. The components which the voltage command value [V*] and the three-phase current [I] have are, for example, described as a U-phase component, a V-phase component, and a W-phase component in this order.
[0056] The electric motor controller 1 is a device for controlling the primary magnetic flux [λ1] and the rotation velocity (hereinafter, illustrated as the rotation angular velocity) on the rotary electric motor 3. The primary magnetic flux [λ1] is a synthesis of the field flux Λ0 that a field magnet generates and the magnetic flux λa (see the components of Fig. 1, id • Ld, iq • Lq in Fig. 1) of the armature reaction generated by the armature current (this is also the three-phase current [I]) flowing in the armature. The magnitude of the primary magnetic flux [λ1] is a component Λδ on the actual δ axis, and is represented as [λ1] = [λ1γc λ1δc]
^{t} in the δc-γc rotating coordinate system. In the present embodiment, the primary magnetic flux [λ1] is handled as an observable value or an already-estimated value.
[0057] The electric motor controller 1 comprises a first coordinate conversion unit 101, a first calculation unit 102, a second calculation unit, 103A, a second coordinate conversion unit 104, and an integrator 106.
[0058] The first coordinate conversion unit 101 converts the three-phase current [I] into a current [i] in the δc-γc rotating coordinate system where the primary magnetic flux control is performed.
[0059] The first calculation unit 102 obtains the feedforward term [F]. The second calculation unit 103A obtains the voltage command value [v*] in the δc-γc rotating coordinate system as a sum of the feedforward term [F] and the feedback term [B].
[0060] The second coordinate conversion unit 104 performs a coordinate conversion of the voltage command value [v*] into a voltage command value [V*] of a voltage to be applied to the rotary electric motor 3 in another coordinate system. This "another coordinate system" may be, for example, a d-q rotating coordinate system, an α-β fixed coordinate system (for example, the α-axis is set in phase with the U phase), or a uvw fixed coordinate system, or a polar coordinate system. Which one of the coordinate systems is employed as "another coordinate system" depends on which control the voltage supply source 2 performs. For example, when the voltage command value [V*] is set in the d-q rotating coordinate system, [V*] = [Vd* Vq*]
^{t} (where the former component is the d-axis component and the latter component is the q-axis component).
[0061] The integrator 106 calculates a phase θ of the δc axis with respect to the α axis on the basis of the rotation angular velocity ω. On the basis of the phase θ, the first coordinate conversion unit 101 and the second coordinate conversion unit 104 can perform the above coordinate conversion. The rotation angular velocity ω is obtained as an output of a subtracter 109. The rotation angular velocity ω is obtained by subtracting a Km times multipled value, in a constant multiplier unit 108, of the γc-axis component iγc of the current [i] that has been removed its DC component in a high pass filter 110, from a command value ω* of the rotation angular velocity in a subtracter 109. When the primary magnetic flux control is appropriately performed, ω = ω* as described above.
[0062] Fig. 5 is a block diagram showing configurations of the first calculation unit 102 and the second calculation unit 103A. In Fig. 5, a reference sign "x" surrounded by a circle represents a multiplier, a reference sign "+" surrounded by a circle represents an adder, and a circle to which reference signs "+-" are attached represents a subtracter. Since the resistance value R, the primary magnetic flux command value Λδ* = 0 on the γc axis, feedback gains Kγγ, Kγδ, Kδγ, and Kδδ are already known, these can be set in the first calculation unit 102 and the second calculation unit 103A.
<The Second Embodiment>
[0063] The present embodiment shows a technique in which the electric motor controller 1 obtains an estimation value [λ1^] of the primary magnetic flux [λ1].
[0064] As shown in Fig. 6, the configuration of the electric motor controller 1 of the present embodiment further comprises a primary magnetic flux estimation unit 105 in that of the electric motor controller 1 of the first embodiment. As the primary magnetic flux [λ1], the second calculation unit 103A employs the estimation value [λ1^] thereof.
[0065] In general, the phase of the field flux Λ0 is employed on the d axis, and a q axis which is in 90-degree phase advance therewith is assumed. When such a d-q rotating coordinate system rotates at the angular velocity ω, introducing a d-axis voltage vd which is a d-axis component of the voltage to be applied to the rotary electric motor, a q-axis voltage vq which is a q-axis component of the voltage to be applied to the rotary electric motor, a d-axis inductance Ld which is a d-axis component of the inductance of the armature winding, a q-axis inductance Lq which is a q-axis component of the inductance of the armature winding, and the differential operator p, Eq. (5) is held.
[Mathematical 3]
[0066] The above equation is expressed in a ξ-η rotating coordinate system having a ξ axis rotating while maintaining the phase difference ψ with respect to the d axis and a η axis in 90-degree phase advance with the ξ axis, the following Eqs. (6), (7), and (8) are held. Note that introduced are a ξ-axis component iξ of the armature current, a η-axis component iη of the armature current, a ξ-axis component vξ and a η-axis component vη of the voltage to be applied to the rotary electric motor, and a ξ-axis component λξ and a η-axis component λη of the primary magnetic flux. Herein, it is not assumed that the primary magnetic flux control is performed.
[Mathematical 4]
[0067] The first term on the right side of Eq. (7) is a magnetic flux (armature reaction) generated by the armature current flow, and the second term is a magnetic flux contributing to the field flux Λ0.
[0068] Since the Eqs. (6), (7), and (8) are held regardless of the phase difference ψ, if the phase difference ψ is replaced with the phase difference ϕc, in other words, the ξ-η rotating coordinate system is replaced with the δc-γc rotating coordinate system, the meanings of Eqs. (6), (7), and (8) are not changed. Since a phase of the actual primary magnetic flux Λδ having the load angle ϕ with respect to the d axis is taken on the δ axis, with the above replacement, the value λξ in Eq. (7) represents the δc-axis component λ1δc of the primary magnetic flux Λδ and the value λη represents the γc-axis component λ1γc of the primary magnetic flux Λδ. The vector diagram at that time is shown in Fig. 7.
[0069] Therefore, from the phase difference ϕc, the d-axis inductance Ld, the q-axis inductance Lq, the armature currents iγc and iδc, and the field flux Λ0, the estimation value of the primary magnetic flux [λ1], [λ1^] = [λ1γc^ λ1δc^]
^{t} is obtained by Eqs. (9) and (10).
[Mathematical 5]
[0070] Herein, introduced is a field flux vector [Λ0] = [-Λ0 • sinϕc Λ0 • cosϕc]
^{t}, representing the field flux Λ0 in the δc-γc rotating coordinate system.
[0071] Further, it can be grasped that a matrix {L} in Eq. (9) is a coefficient of the current vector [iγc iδc]
^{t} of the first term on the right side in Eq. (10) and a tensor in which the inductance of the armature winding s expressed in the δc-γc rotating coordinate system. When the rotary electric motor has no saliency, since Ld = Lq, as is clear from Fig. 10, the off-diagonal component of the matrix {L} is zero. In other words, Eq. (10) can be employed in the rotary electric motor having saliency.
[0072] It can be grasped that the first term on the right side of each of Eqs. (9) and (10) is the magnetic flux caused by the armature reaction.
[0073] Further, the phase difference ϕc can employ an estimated value on the basis of Eq. (11). In this case, the used voltages vγc and vδc may employ the already-obtained voltage command values vγc* and vδc* to be used for estimation of a new phase difference ϕc.
[0074] Fig. 8 is a block diagram showing a configuration of the primary magnetic flux estimation unit 105. The primary magnetic flux estimation unit 105 comprises a delay unit 105a, a load angle estimating unit 105b, an armature reaction estimation unit 105c, a field flux vector generation unit 105d, and an adder 105e.
[0075] The armature reaction estimation unit 105c inputs thereto the phase difference ϕc, the d-axis inductance Ld, the q-axis inductance Lq, and the armature currents iγc and iδc, and calculates the first term on the right side of Eq. (10). Fig. 8 uses the expression {L}[i] of the first term on the right side of Eq. (9), and that two values of the γc-axis component and the δc-axis component are outputted is indicated by two slashes.
[0076] The field flux vector generation unit 105d inputs thereto the field flux Λ0 and calculates the second term on the right side of Eq. (10). Fig. 8 uses the expression [Λ0] of the second term on the right side of Eq. (9), and that two values of the γc-axis component and the δc-axis component are outputted is indicated by two slashes.
[0077] The adder 105e performs addition in the two components, the γc-axis component and the δc-axis component, to thereby achieve addition of the first term and the second term on the right side in each of Eqs. (9) and (10), and outputs the estimation value [λ1^] of the primary magnetic flux.
[0078] In order to estimate the phase difference ϕc, used are the voltage command values vγc* and vδc* obtained by the second calculation unit 103A at the immediately preceding control timing. In other words, the delay unit 105a delays the voltage command values vγc* and vδc* obtained by the second calculation unit 103A and the load angle estimating unit 105b calculates the phase difference ϕc in accordance with Eq. (11) at the immediately following control timing. Further, instead of employing the voltage command values vγc* and vδc* obtained at the immediately preceding control timing, the voltage command values vγc* and vδc* which have been obtained at this point in time may be employed. In this case, the delay unit 105a may be omitted.
[0079] In the present embodiment, it is not necessary to perform direct detection of the primary magnetic flux. Further, the primary magnetic flux can be estimated, regardless of whether or not there is saliency, and the primary magnetic flux control is performed while correcting the deviation of the phase difference ϕc.
[0080] Thus, by performing estimation of the primary magnetic flux with the phase difference ϕc which is a parameter having a strong correlation with an output torque, it is possible to estimate the primary magnetic flux with high accuracy even in the area where the output torque is large. This makes a drive of the rotary electric motor 3 stable in the area where the output torque is large, in other words, an area where the rotary electric motor 3 can be driven stably is extended. Further, even in the area where the output torque is large, the rotary electric motor 3 can be driven at a high-efficient operating point.
<The Third Embodiment>
[0081] In the present embodiment, shown is a technique to achieve the effect shown in the second embodiment when the electric motor controller 1 obtains the estimation value or a measured value of the primary magnetic flux [λ1].
[0082] As shown in Fig. 9, the electric motor controller 1 of the present embodiment has a constitution in which the second calculation unit 103A is replaced with a second calculation unit 103B and a primary magnetic flux command correction unit 107 is further included in the constitution of the electric motor controller 1 of the second embodiment.
[0083] Now, it is assumed that the primary magnetic flux [λ1] = [λ1γc λ1δc]
^{t} is estimated by a method other than that shown in the second embodiment. A correction value [Λγ** Λδ**]
^{t} of the primary magnetic flux command (hereinafter, also referred to as a primary magnetic flux command correction value [Λ1**]), which satisfies the following Eq. (12) together with the primary magnetic flux [λ1], is obtained by Eq. (13). In this equation, introduced is the estimation value [λ1^] of the primary magnetic flux which is described in the second embodiment. Further, for easy understanding, a component Λγ* of the primary magnetic flux command value is also clearly specified (actually, Λγ* = 0).
[Mathematical 6]
[0084] By performing the primary magnetic flux control in the second embodiment, the right side of Eq. (12) becomes zero. Therefore, when the primary magnetic flux control is performed on the primary magnetic flux [λ1] on the basis of the primary magnetic flux command correction value [Λ1**] obtained by Eq. (13), the effect can be achieved the same in the second preferred embodiment. In other words, it is natural that it is not necessary to perform the direct detection of the primary magnetic flux, and further, it is possible to perform the primary magnetic flux control while correcting the deviation of the phase difference ϕc, not depending on the method of measuring or estimating the primary magnetic flux [λ1], regardless of whether or not there is saliency.
[0085] In this case, it is not necessary to replace the primary magnetic flux command value [Λ1*] in the feedforward term [F] with the primary magnetic flux command correction value [Λ1**]. As can be understood from Fig. 2, this is because the inductive voltage ω • Λδ* appearing on the γc axis is determined, regardless of whether or not there is the deviation [ΔΛ].
[0086] On the other hand, the feedback term [B] is determined on the basis of the deviation between the primary magnetic flux [λ1] and the primary magnetic flux command correction value [Λ1**]. Therefore, introducing the deviation of the primary magnetic flux, [ΔΛ'] = [Λγ**-λ1γc Λδ**-λ1δc]
^{t}, the feedback term [B] is obtained by the following equations.
[Mathematical 7]
[0087] Fig. 10 is a block diagram showing a configuration of the first calculation unit 102 and a second calculation unit 103B. As described above, since the feedforward term [F] uses the primary magnetic flux command value [Λ1*], instead of the primary magnetic flux command correction value [Λ1**], the first calculation unit 102 is employed also in the present embodiment, like in the first and second preferred embodiments.
[0088] On the other hand, since the calculation for obtaining the feedback term [B] uses the primary magnetic flux command correction value [Λ1**], the second calculation unit 103B has a configuration which is slightly different from that of the second calculation unit 103A. Specifically, since Λγ* = 0 in the second calculation unit 103A, this is not inputted but is prepared in the second calculation unit 103A. On the other hand, in the second calculation unit 103B, the γc-axis component Λγ** of the primary command correction value [Λ1**] is inputted. Further, though the command value Λδ* is inputted in the second calculation unit 103A, the δc-axis component Λδ** of the primary command correction value [Λ1**] is inputted in the second calculation unit 103B. In the configuration shown in Fig. 10, other configurations are the same as the configuration shown in Fig. 5.
[0089] The primary magnetic flux command correction unit 107 inputs thereto the primary magnetic flux command value [Λ1*], the estimation value [λ1^] of the primary magnetic flux (calculated by the primary magnetic flux estimation unit 105 as described in the second embodiment), and the primary magnetic flux [λ1] which is estimated by another method. Then, by performing the calculation of Eq. (13), the primary command correction value [Λ1**] is outputted.
<Variations>
[0090] Estimations of the primary magnetic flux [λ1] by other methods other than the method shown in the second embodiment will be exemplified below.
[0091] With reference to Fig. 7, in consideration of the γc-axis component (vγc-R • iγc) and the δc-axis component (vδc-R • iδc) of an internal inductive voltage ω • Λδ, the estimation values λ1γc^ and λ1δc^ of the primary magnetic flux Λδ are obtained as -(vδc-R • iδc)/ ω and (vγc-R • iγc)/ ω, respectively.
[0092] Further, when the estimation value Λδ^ of the primary magnetic flux Λδ is obtained, with reference to Fig. 7, setting χ = ϕ - ϕc, an estimation value χ^ of the angle χ is obtained by Eq. (16).
[Mathematical 8]
[0093] Therefore, the estimation values λ1γc^ and λ1δc^ are obtained as -sin(χ^) • Λδ^ and cos(χ^) • Λδ^, respectively.
[0094] Now, the estimation value Λδ^ of the primary magnetic flux Λδ can be calculated by using, for example, the estimation value of the primary magnetic flux in the α-β fixed coordinate system of the rotary electric motor 3. Herein, the α-β fixed coordinate system has the α axis and the β axis, and employs the β axis in 90-degree phase advance with the α axis. As described earlier, for example, the α axis is adopted in phase with the U phase.
[0095] Introducing an α-axis component λ1α^ and a β-axis component λ1β^ of the estimation value Λδ^ of the primary magnetic flux Λδ, the estimation value λδ^ of the primary magnetic flux Λδ is obtained by Eq. (17).
[Mathematical 9]
[0096] Now, as shown in Eq. (18), the α-axis component λ1α^ and the β-axis component λ1β^ can be obtained by integration of the α-axis component V0α and the β-axis component V0β of the internal inductive voltage ω • Λδ with respect to the time. The α-axis component V0α can be calculated as Vα-R • iα from an α-axis component Vα of an applied voltage V observed outside and an α-axis component iα of the current [I] flowing in the rotary electric motor 3. Similarly, the β-axis component V0β can be calculated as Vβ-R • iβ from a β-axis component Vβ of the applied voltage V observed outside and a β-axis component iβ of the current [I] flowing in the rotary electric motor 3. The applied voltage V is obtained from, for example, the three-phase voltage supplied from the voltage supply source 2 to the rotary electric motor 3 in conformity with Fig. 4.
[Mathematical 10]
[0097] Further, when the α-axis component λ1α^ and the β-axis component λ1β^ are obtained, the estimation values λ1γc^ and λ1δc^ can be also obtained by another method. In other words, the estimation values λ1γc^ and λ1δc^ can be obtained by Eq. (19) by using the phase θ of the δc axis with respect to the α axis.
[Mathematical 11]
[0098] Further, the α-axis component λ1α^ and the β-axis component λ1β^ can be obtained by another method. As described above, since the applied voltage V can be obtained from the three-phase voltage supplied from the voltage supply source 2 to the rotary electric motor 3, the U-phase component Vu, the V-phase component Vv, and the W-phase component Vw can be measured. As described above, the three-phase current Iu, Iv, and Iw flowing in the rotary electric motor 3 can be measured. Therefore, the U-phase component λ1u^, the V-phase component λ1v^, and the W-phase component λ1w^ of the estimation value Λδ^ of the primary magnetic flux Λδ can be obtained by Eq. (20), like by Eq. (18).
[Mathematical 12]
[0099] By performing the coordinate conversion of the UVW-phases and the α-β fixed coordinate system, the α-axis component λ1α^ and the β-axis component λ1β^ can be obtained by Eq. (21). Therefore, by further using Eq. (19), the estimation values λ1γc^ and λ1δc^ can be obtained.
[Mathematical 13]
[0100] When complete integration is performed in the integral calculation of Eqs. (18) and (20), the DC component is superimposed and the error in the estimation of the magnetic flux thereby becomes larger. Therefore, it is preferable to perform the well-known incomplete integration.
[0101] Further, instead of Eq. (11), the phase difference ϕc can be estimated as follows. Though Fig. 11 corresponds to Fig. 7, a q' axis is newly employed. Herein, the q' axis is adopted in phase with a voltage V'. The voltage V' is a synthesis of the inductive voltage ω • Λδ caused by the primary magnetic flux and a voltage having a δc-axis component ω • Ld • iγc and a γc-axis component (-ω • Ld • iδc).
[0102] Introducing a leading phase angle ϕc' of the γc axis viewed from the q' axis and a leading phase angle ζ of the q' axis viewed from the q axis, an estimation value of the phase difference ϕc can be obtained as a sum of the angles ϕc' and ζ. Then, the angles ϕc' and ζ can be obtained by Eqs. (22) and (23), respectively.
[Mathematical 14]
[0103] In any one of the above-described preferred embodiments, the electric motor controller 1 includes a microcomputer and a memory device. The microcomputer executes each process step (in other words, each procedure) described in a program. The above memory device can be constituted of one or a plurality of memory devices such as a ROM (Read Only Memory), a RAM (Random Access Memory), a rewritable nonvolatile memory (EPROM (Erasable Programmable ROM) or the like), a hard disk unit, and the like. The memory device stores therein various information and data and the like, also stores therein a program to be executed by the microcomputer, and provides a work area for execution of the program.
[0104] It can be grasped that the microcomputer functions as various means corresponding to each of the process steps described in the program, or that the microcomputer implements various functions corresponding to each of the process steps. Further, the electric motor controller 1 is not limited thereto, and some or all of the various procedures executed by the electric motor controller 1, or some or all of the various means or various functions implemented by the electric motor controller 1 may be achieved by hardware.