Control device for a rotating electric machine

Abstract

Control device for a rotating electric machine (1) which: a rotating electric machine (80) which has a rotor in which a permanent magnet is arranged and which is driven by an inverter (7) which performs a power conversion between alternating current power and direct current power, a magnetic pole position (θ) of the rotor is detected by a sensorless control system, a current control using the magnetic pole position (θ) based on a deviation between a current command (Id*, Iq*) and a feedback current from the rotating electric machine (80) in a dq-axis vector coordinate system defined by a d-axis extending in a direction of a magnetic field generated by the permanent magnet and a q-axis orthogonal to the d-axis, performs dead-time compensation, in which a start point and an end point of a control pulse for switching elements (3) that make up the inverter (7) are adjusted to reduce an error in an AC output due to a dead-time period (Td), during which both an upper switching element (3) and a lower switching element (3) of the switching elements that make up a branch of a phase of the inverter (7) are controlled to an OFF state, compared to that for a case in which the control pulse is a reference pulse, and a switching control on the switching elements (3) for driving the rotating electric machine (80) is carried out, wherein In the case where the rotating electric machine (80) is controlled by setting the current command (Id*, Iq*) in the dq-axis vector coordinate system such that the magnitude of an armature current (Ia) corresponding to an output torque of the rotating electric machine (80) is minimized, the rotating electric machine is controlled such that the magnitude of the armature current (Ia) becomes equal to or greater than a lower limiting current (Ia_min) that is predetermined.

Description

Technical area

[0001] The present invention relates to an electric machine control device that controls a rotating electric machine having a rotor in which a permanent magnet is arranged and driven via an inverter that performs power conversion between direct current power and alternating current power. State of the art

[0002] A method called vector control is known as a control method for a rotating electrical machine, such as an internal permanent magnet synchronous motor (IPMSM). In vector control, control is performed by subjecting motor currents flowing through the stator coils of the motor to coordinate conversion into vector components for a d-axis extending in the direction of a magnetic field generated by a permanent magnet arranged in a rotor, and a q-axis orthogonal to the d-axis. For coordinate conversion, it is necessary to accurately detect the position of the rotor (magnetic pole position). A rotary sensor such as a resolver is often used to detect the magnetic pole position. However, for IPMSM with a magnetic characteristic, a variety of sensorless magnetic pole position detection techniques have been proposed that do not use such a rotary sensor.For example, a method is known in which a magnetic pole position is detected by utilizing an induced electromotive force generated by the rotation of the rotor. In the case where the motor rotates at a significantly low speed, such as when the motor is stationary, no induced electromotive force is generated at all, or only a small induced electromotive force is generated. Therefore, a method has also been proposed in which a magnetic pole position is estimated according to a response obtained when an observation signal with a high-frequency current or high-frequency voltage is applied to the motor.

[0003] The drive of a rotating electrical machine is often controlled by an inverter that performs power conversion between DC power and AC power. During switching control for the inverter, a so-called dead time, during which both upper switching elements and lower switching elements are turned off simultaneously, is provided to prevent short-circuiting between the positive and negative electrodes of the inverter with both upper switching elements and lower switching elements turned on. Due to the dead time, the actual effective pulse width may be reduced with respect to the effective pulse width of a switching pulse based on a modulation command, and an error from a command value may be caused in the output voltage obtained after conversion from DC power to AC power.The error can interfere with electrical detection of the magnetic pole position, causing a detection error. Therefore, dead-time compensation is sometimes performed, which reduces the error in the output voltage due to dead time by shifting the start point and end point of the effective pulse by the same amount of time.

[0004] Dead time compensation works effectively when the phase currents of the AC power are close to the center of amplitude (close to zero). However, when the proportion of the period in which the phase currents have a value close to the center of amplitude (zero) is large compared to the phase current cycle, such as when the amplitude of the phase currents is small, the magnetic pole position cannot be accurately detected for a longer period, and as a result, the accuracy of detecting the magnetic pole position may be lowered. Thus, Japanese Patent Application Publication No. JP 2012-178950 A (Patent Document 1) proposes a technique for improving detection accuracy by processing the phase of a current for controlling the control quantity of a motor (seventh to ninth paragraphs, fifty-third to fifty-seventh paragraphs, etc.).However, with this method, torque ripple tends to increase because the current is adjusted to an advanced phase and a delayed phase. In addition, the computational burden for adjusting the current phase increases. The influence of dead time depends on the motor torque and the voltage on the DC side of the inverter (DC link voltage). Even in cases where the influence of dead time is relatively small, the efficiency can be reduced by adjusting the current phase or the like. State-of-the-art documents.Patent documents

[0005] Patent Document 1: JP 2012 - 178 950 A, which corresponds to US 2012 / 0 217 849 A.

[0006] US 2010 / 0 156 330 A1, EP 0 710 582 A1 and JP 2005 - 110 343 A show further examples of control devices for rotating electrical machines. Summary of the inventionProblem to be solved by the invention

[0007] In view of the above-described background, it is desirable to provide a sensorless magnetic pole position detection technique that enables the magnetic pole position of a rotating electric machine to be electrically accurately derived by reducing a voltage error due to a dead time while suppressing the influence on the operating efficiency of the rotating electric machine. Means of solving the problem

[0008] In view of the above, according to the present invention, there is provided a rotating electric machine control device (or control device for a rotating electric machine) which: controls a rotating electrical machine having a rotor in which a permanent magnet is arranged and which is driven via an inverter that performs power conversion between AC power and DC power, a magnetic pole position of the rotor is detected by a sensorless control, performs current control using the magnetic pole position based on a deviation between a current command and a feedback current from the rotating electric machine in a dq-axis vector coordinate system defined by a d-axis extending in a direction of a magnetic field generated by the permanent magnet and a q-axis orthogonal to the d-axis, performs dead-time compensation in which a start point and an end point of a control pulse for switching elements composing the inverter are adjusted to reduce an error in an AC output due to a dead-time period during which both an upper switching element and a lower switching element of the switching elements composing one branch of a phase of the inverter are controlled to an OFF state, compared with that for a case where the control pulse is a reference pulse, and carries out a switching control on the switching elements for driving the rotating electrical machine, whereby in the case where the rotating electric machine is controlled by setting the current command in the dq-axis vector coordinate system such that an amount of armature current corresponding to an output torque of the rotating electric machine is minimized, the rotating electric machine is controlled such that the amount of armature current becomes equal to or greater than a lower limit current prescribed in advance.

[0009] For accurate detection of the magnetic pole position by sensorless control, it is useful to perform dead-time compensation. However, to perform dead-time compensation, it is generally necessary to determine the polarity of the phase current of the AC power. In this case, in the case where the amplitude of the phase current is small and the wave height of the phase current is close to the amplitude center of the phase current, as is the case when the rotating electrical machine is driven with low torque, high accuracy may be required for determining the polarity of the phase current, or the reliability of the determination may be lowered depending on the determination accuracy. According to the present configuration, the current command is set such that the armature current becomes equal to or greater than the lower limit current.The magnitude of the phase current amplitude can be ensured by limiting a reduction in the armature current. This means that the amplitude of the phase current can be ensured in such a way that the polarity of the phase current can be sufficiently determined. If the magnitude of the armature current is equal to or greater than the lower limit current, the armature current is not limited. Thus, the operating efficiency of the rotating electrical machine is only affected if the magnitude of the armature current is less than the lower limit current.Thus, with the present configuration, it is possible to provide a sensorless magnetic pole position detection technique that allows the magnetic pole position of a rotating electric machine to be electrically accurately derived by reducing a voltage error due to a dead time while suppressing the influence on the operating efficiency of the rotating electric machine. Short description of the drawings Fig. 1 is a block diagram schematically illustrating an example of the system configuration of a rotating electric machine control device. Fig. 2 is a block diagram schematically illustrating an example of the functional configuration of the rotating electric machine control device. Fig. Figure 3 illustrates the relationship between a dq-axis vector coordinate system and a δ-γ-axis vector coordinate system. Fig. Figure 4 illustrates the relationship between an α-β-axis vector coordinate system and the dq-axis vector coordinate system. Fig. 5 illustrates a dead time. Fig. Figure 6 illustrates a voltage error caused by dead time. Fig. Figure 7 illustrates an example of dead time compensation. Fig. Figure 8 schematically illustrates the relationship between an operating point and a current command in a current vector space. Fig. 9 illustrates waveforms schematically showing an example of a current command for a case where control is performed with a constant current control. Fig. Figure 10 schematically illustrates the relationship between a lower limit current and a DC link voltage. Fig. Figure 11 schematically illustrates the relationship between the lower limit current and a rotational speed. Ways of carrying out the invention

[0010] An embodiment of the present invention will now be described with reference to the drawings. As shown in the Fig. 1 and Fig. As illustrated in FIG. 2, a rotating electric machine control device (rotating electric machine control device) 1 is a control device provided with a function of detecting the rotation state (magnetic pole position and rotational speed) of an AC rotating electric machine 80 in a so-called sensorless manner without using a rotation sensor such as a resolver. According to the embodiment, the rotating electric machine 80 is an internal permanent magnet synchronous motor (IPMSM) and has a characteristic (including an inverse characteristic), that is, the magnetic properties in the N-pole direction of the permanent magnets of a rotor are different from the magnetic pole properties in a direction electrically perpendicular thereto (a direction shifted by an electrical angle of 90° therefrom).According to the embodiment, as described in detail later, a motor control device utilizes the saliency to determine the rotational state such as the magnetic pole position, magnetic pole direction, and rotational speed through sensorless control, even when the rotating electrical machine 80 is stationary or rotating at a low speed. Thus, the present invention can also be applied to other types of rotating electrical machines with saliency (for example, synchronous reluctance motors). Of course, the rotating electrical machine 80 includes a motor (electric motor), a generator (electric generator), and a motor generator that functions as both a motor and a generator as needed.

[0011] As described above, the electric machine 80 to be controlled by the rotating electric machine control device 1 has a rotor in which a permanent magnet is arranged and is driven via an inverter that performs power conversion between AC power and DC power. In other words, as shown in Fig. As illustrated in FIG. 1, the rotating electric machine control device 1 includes a rotating electric machine drive device (rotating electric machine drive device) 70 including an inverter 7 and a DC link capacitor 6, and controls the drive of the rotating electric machine 80 via the rotating electric machine drive device 70. The inverter 7 is a power conversion device connected to the AC rotating electric machine 80 to perform power conversion between DC power and multi-phase AC power (here, three-phase AC power). An arm corresponding to one phase of the AC power is constructed of a series circuit including an upper switching element and a lower switching element.The DC link capacitor 6 smoothes an intermediate circuit DC voltage Vdc, which is the DC voltage on the DC side of the inverter 7.

[0012] According to the embodiment, the rotating electric machine 80 is used, for example, as a driving power source for a vehicle such as a hybrid vehicle and an electric vehicle. Vehicles such as automobiles, which cannot receive power supplied from an overhead line like railway vehicles, are provided with a DC power source 4 such as a secondary battery (battery) such as a nickel-hydrogen battery or a lithium-ion battery, an electric double-layer capacitor, or the like, as a power source for driving the rotating electric machine 80. According to the embodiment, the DC power supply 4 is provided with a power source voltage of, for example, 200 to 400 V as a high-voltage, high-capacity DC power source for supplying power to the rotating electric machine 80.Since the rotating electric machine 80 is an AC rotating electric machine, the inverter 7, which performs power conversion between DC power and AC power (here, three-phase AC power), is provided between the DC power source 4 and the rotating electric machine 80. Hereinafter, the voltage between a positive power source line P and a negative power source line N on the DC side of the inverter 7 is referred to as "DC link voltage Vdc." The DC power source 4 can supply power to the rotating electric machine 80 via the inverter 7 and can store power obtained by power generation performed by the rotating electric machine 80.A smoothing capacitor (DC link capacitor 6), which smoothes the voltage (DC link voltage Vdc) between the positive and negative electrodes on the DC side of the inverter 7, is provided between the inverter 7 and the DC power source 4. The DC link capacitor 6 stabilizes a DC voltage (DC link voltage Vdc) that fluctuates according to fluctuations in the power consumption of the rotating electric machine 80.

[0013] The inverter 7 converts DC power with the intermediate circuit DC voltage Vdc into AC power with a plurality of phases (n phases, where n is a natural number; here, three phases) to supply the DC power to the rotating electric machine 80, and converts AC power generated by the rotating electric machine 80 into DC power to supply the DC power to the DC power source. The inverter 7 is configured to include a plurality of switching elements. The switching elements are preferably power semiconductor elements that can operate at a high frequency, such as IGBTs (insulated gate bipolar transistors), power MOSFETs (metal oxide semiconductor field effect transistors), SiC MOSFETs (silicon carbide metal oxide semiconductor FETs), SiC SITs (SiC static induction transistors), and GaN MOSFETs (gallium nitride MOSFETs).According to the embodiment, as shown in . Fig. 1, IGBTs 3 are used as the switching elements.

[0014] The inverter 7, which converts power between DC power and multi-phase AC power (here three-phase AC power), is constructed, for example, from a bridge circuit that is constructed from a number of branches, where the number corresponds to the number of phases (here three phases), as is generally known. That is, as it is in Fig. As illustrated in Figure 1, two IGBTs 3 are connected in series between the positive DC electrode side (the positive power source line P on the positive electrode side of the DC power source) and the negative DC electrode side (the negative power source line N on the negative electrode side of the DC power source) of the inverter 7 to constitute one arm. In the case of three-phase AC power, the series circuit (one arm) is connected in parallel for three lines (three phases). That is, a bridge circuit is constructed in which one series circuit set (one arm) corresponds to each stator coil 8 of the rotating electric machine 80 corresponding to one of a U-phase, a V-phase, and a W-phase.

[0015] The midpoint between the series circuits (arms) formed by a pair of IGBTs 3 for each phase, that is, the connection point between the IGBT 3 on the positive power source line P side (upper IGBT (upper switching element)) and the IGBT 3 on the negative power source line N side (the lower IGBT (lower switching element)), is respectively connected to a stator coil 8 of the rotating electric machine 80. A freewheeling diode (FWD) 5 is provided in parallel with each IGBT 3, wherein the direction from the negative electrode (negative power source line N) to the positive electrode (positive power source line P) (the direction from the lower side to the upper side) is defined as the forward direction.

[0016] As it is in Fig. As shown in Figure 1, the inverter 7 is controlled by the rotating electric machine control device 1. The rotating electric machine control device 1 is configured as an ECU (electronic control unit) composed of a logic circuit such as a microcomputer as its core element. As shown in Fig. As illustrated in FIG. 2, the rotating electric machine control device 1 is configured to include an inverter control section 10 and a rotation state information calculation section 20. The inverter control section 10 and the rotation state information calculation section 20 are each configured to include a plurality of functional sections implemented by cooperation between hardware such as a microcomputer and software (a program).For example, the rotating electric machine control device 1 sets a torque command T* based on a target torque and a target speed for the rotating electric machine 80, which are provided as a request signal from another control device or the like such as a vehicle ECU 90 via a CAN (Controller Area Network) or the like, and controls the rotating electric machine 80 via the inverter 7.

[0017] The rotating electric machine control device 1 (inverter control section 10) performs, for example, pulse width modulation (PWM) control as the pattern switching mode (voltage waveform control mode) for the IGBTs 3 composing the inverter 7.In addition, the rotating electric machine control device 1 includes, as the modes of field control for a stator, a normal field control (basic control) such as a maximum torque control in which a maximum torque is output for an armature current, and a maximum efficiency control in which the motor is driven with a maximum efficiency for an armature current, and a field regulation control such as a field weakening control in which a field current that does not contribute to a torque is supplied to weaken a field magnetic flux, and a field intensification control (field amplification control) in which, conversely, a field magnetic flux is amplified.According to the embodiment, the rotating electric machine control device 1 controls the rotating electric machine 80 by performing current control using a current vector control method in a two-axis orthogonal vector space (coordinate system) that rotates synchronously with the rotation of the rotating electric machine 80. In the current vector control method, for example, the current control is performed in the two-axis orthogonal vector space defined by a d-axis extending along the direction of a field magnetic flux generated by a permanent magnet and a q-axis that electrically leads by n / 2 with respect to the d-axis.

[0018] The gate terminal, which is a control terminal, of each of the IGBTs 3 constituting the inverter 7 is connected to the rotating electric machine control device 1 via a drive circuit 2 to be individually subjected to switching control. A high-voltage system circuit for driving the rotating electric machine 80 and a low-voltage system circuit such as the rotating electric machine control device 1, which has a microcomputer or the like as its core, have significantly different operating voltages (power source voltages for the circuits).Therefore, the drive circuit 2 (control signal drive circuit) is provided, which transmits a gate drive signal (switching control signal) to each of the IGBTs 3 with improved drive capability of the gate drive signal (capability to cause the subsequent circuit to operate, such as voltage amplitude or output current). A gate drive signal for the IGBTs 3 generated by the rotating electric machine control device 1, which is a low-voltage system voltage, is supplied to the inverter 7 as a gate drive signal for a high-voltage system circuit via the drive circuit 2. The drive circuit 2 is constructed using an insulating element such as a photocoupler or a transformer, and a drive IC, for example.

[0019] Thus, the rotating electric machine 80 is driven at a predetermined output torque and a predetermined speed via the inverter 7, which is subjected to switching control performed by the inverter control section 10. In this case, the value of a current flowing through each stator coil of the rotating electric machine 80 is fed back to the inverter control section 10. The inverter control section 10 performs PI (Proportional Integral Control) or PID (Proportional Integral Derivative Control) control of the deviation of current commands (Id*, Iq*) set according to the torque command T* to control the drive of the rotating electric machine 80.To achieve such control, a current (Iu, Iv, Iw) flowing through a conductor such as a busbar provided between one arm of the inverter 7 for each phase and one stator coil of the rotating electrical machine 80 for each phase is detected by a current sensor 30.

[0020] The vector control performed by the inverter control section 10 will be briefly described below. In such vector control, the vector space (coordinate system) is a dq-axis vector coordinate system (dq-axis vector space) defined by a d-axis extending in the direction of a magnetic field generated by a permanent magnet disposed in the rotor of the rotating electrical machine 80 and a q-axis electrically orthogonal to the d-axis. According to the embodiment, the inverter control section 10 is configured to include a torque command calculation section 11, a torque control section 12 (current command calculation section), a current control section 13 (voltage command calculation section), a modulation control section 14, and a three-phase / two-phase coordinate conversion section 15.

[0021] According to the embodiment, the torque command calculation section 11 calculates a torque command T* (target torque) based on a speed command ω* (target speed) from a higher-level ECU or the like, such as the vehicle ECU 90, and the actual speed. According to the embodiment, the rotation of the rotating electrical machine 80 is detected in a sensorless manner without using a rotation sensor such as a resolver. Thus, the actual speed is an estimated speed estimated by the rotation state information calculation section 20 and is indicated by ω^ (roof), as shown in Fig. 2 (indicated by ω^ for convenience in the description). For example, the torque control section 12 accesses a map corresponding to the torque command T* to set (determine) current commands Id*, Iq* for vector control. The current commands Id*, Iq* are set according to the dq-axis vector coordinate system described above.

[0022] The current control section 13, for example, performs PI control on the deviation between the current commands Id*, Iq* in the dq-axis vector coordinate system and the feedback currents Id, Iq to calculate voltage commands Vd*, Vq* in the dq-axis vector coordinate system. The feedback currents Id, Iq are the detected values ​​of the three-phase currents (Iu, Iv, Iw) flowing through the stator coils of the rotating electrical machine 80, and the detected values ​​have undergone coordinate conversion performed by the three-phase / two-phase coordinate conversion section 15 into the dq-axis vector coordinate system with two phases and have been fed back. The voltage commands Vd*, Vq* are subjected to coordinate conversion, which is performed by the modulation control section 14, into voltage commands for three phases.The modulation control section 14 generates a switching control signal for switching control of the inverter 7 based on the voltage commands for three phases by, for example, pulse width modulation (PWM).

[0023] The coordinate conversions performed by the modulation control section 14 and the three-phase / two-phase coordinate conversion section 15 are performed based on a magnetic pole position θ of the rotor. That is, in order to perform vector control on the rotating electrical machine 80, it is necessary to perform coordinate conversion between the real three-phase space and the two-phase dq-axis vector coordinate system. Therefore, it is necessary to accurately detect the magnetic pole position θ of the rotor. The embodiment adopts sensorless control in which the magnetic pole position θ of the rotor is estimated without providing a rotation detection device such as a resolver. Thus, the magnetic pole position θ is an estimated magnetic pole position and is indicated by θ^ as shown in Fig. 1 (indicated as θ^ for convenience in the description).

[0024] While the rotating electric machine 80 is rotating, magnetic pole position information is included in the induced electromotive force. Therefore, an induced electromotive force is estimated from an inverter output voltage and the feedback currents (Id, Iq) to estimate the rotational speed ω (estimated rotational speed ω^) and the magnetic pole position θ (estimated magnetic pole position θ^). In contrast, when the rotating electric machine 80 is stationary, no induced electromotive force is naturally generated. In contrast, when the rotating electric machine 80 is rotating at a low speed, the induced electromotive force is correspondingly small. Therefore, it is necessary to use a different method for calculating the rotational speed ω (w^) and the magnetic pole position θ (θ^).For example, in the case where the rotating electric machine 80 is stationary or rotating at a low speed, a high-frequency observation signal (observation current or observation voltage) serving as an electrical stimulus is applied to the rotating electric machine 80, and the rotational speed ω (ω^) and the magnetic pole position θ (θ^) are calculated from a response to the observation signal.

[0025] According to the embodiment, as shown in Fig. 2, two position calculation sections, namely a high-speed range position calculation section 21 and a low-speed range position calculation section 22, are provided. The high-speed range position calculation section 21 mainly calculates a rotational speed w (ω^ H ) and a magnetic pole position θ (θ^ H) in a high-speed rotation range in which an induced electromotive force (induced voltage) can be utilized. The low-speed range position calculation section 22 mainly calculates a rotation speed ω (ω^ L ) and a magnetic pole position θ (θ^ L ) in a low-speed rotation range using a high-frequency observation signal. The result (ω^ H and θ^ H ) of the calculation by the high-speed range position calculation section 21 and the result (ω^ L and θ^ L) of the calculation by the low-speed range position calculation section 22 are selected or superimposed by the switching section 23, as described below, to be used by the torque command calculation section 11, the modulation control section 14, and the three-phase / two-phase coordinate conversion section 15. For example, the switching section 23 can set the rotational speed w (ω^) and the magnetic pole position θ (θ^) by selecting one of the following modes: a low-speed calculation mode performed by the low-speed range position calculation section 22, a high-speed calculation mode performed by the high-speed range position calculation section 21, and a superimposition mode in which a weighted average consistent with the rotational speed of the result (ω^ L and θ^ L) of the calculation by the low-speed range position calculation section 22 and the result (ω^ H and θ^ H ) of the calculation by the high-speed range position calculation section 21. In addition, the switching section 23 also controls a switch 24 to switch whether a high-frequency observation signal (here "Vd h *”) should be created or not.

[0026] The method of calculating the rotational state information (θ^, ω^) by the high-speed range position calculation section 21 and the low-speed range position calculation section 22 is briefly described below. The high-speed range position calculation section 21 calculates the magnetic pole position θ (θ^) of the rotor using an "extended induced voltage model" that uses an extended induced voltage obtained by summing an induced voltage generated by the rotation of the magnetic flux generated in the rotor by a d-axis current, an induced voltage generated by variations in the current flowing through the d-axis inductance on the stator side, and an induced voltage generated by the rotation of the magnetic flux of the permanent magnet. The "extended induced voltage model" is described in detail in a technical paper by Shinji Ichikawa et al.“Sensorless Controls of Salient-Pole Permanent Magnet Synchronous Motors Using Extended Electromotive Force Models” (T. IEE Japan, vol. No. 12, 2002) and is therefore not described in detail here.

[0027] A general circuit equation (voltage equation) in the rotating coordinate system (dq-axis vector coordinate system) of a rotating electrical machine exhibiting magnetic characteristics is represented by the following equation (1). In the equation, p represents a differential operator, Ld and Lq represent the d-axis inductance and the q-axis inductance, respectively, and K represents the Ean induced voltage constant. The voltage equation by the "extended induced voltage model" is given by equation (2). The second term of equation (2) is defined as the "extended induced voltage" (which follows equation (3)). In equations (2) and (3), the symbol "·" added to "Iq" indicates a time differential of "Iq" and is used differently from the differential operator p in equation (1) because the derivative is applied only to a dotted variable. Equation (1) [VdVq]=[R+pLD−ωLqωLdR+pLq]+[0ωKE] Equation (2) [VdVq]=[R+pLd−ωLqωLqR+pLd][IdIq]+[0(Ld−Lq)(ωId−I˙q)+ωKE] Equation (3) [EdEq]={(Ld−Lq)(ωId−I˙q)+ωKE}

[01]

[0028] The term "(Ld-Lq)wId" on the right side of equation (3) indicates the induced voltage generated by the rotation of the magnetic flux in the rotor caused by a d-axis current. The term "(Ld-Lq)Iq" (with the dot omitted to indicate a time differential) on the right side of equation (3) indicates the induced voltage generated by variations in the current flowing through the q-axis inductance on the stator side. The term "ωK E" on the right side of equation (3) indicates the induced voltage generated by the rotation of a magnetic flux of the permanent magnet attached to the rotor. That is, the position information in the permanent magnet of the rotating electric machine and the inductance are all collected in the "extended induced voltage." Although a detailed description can be found in the article by Ichikawa et al. and thus is not presented here, when equation (2) is converted to a fixed coordinate system (for example, an α-β axis vector coordinate system) set on the stator of the rotating electric machine, there is no term containing a value (corresponding to the technical contribution "2θ") that is difficult to process when estimating the magnetic pole position, which makes calculation for estimation simple.When calculating the magnetic pole position using a common induced voltage model, it is necessary to use an approximation, which may reduce the estimation accuracy. However, in the case where an extended induced voltage model is used, approximation is not required, which allows the magnetic pole position θ (θ^) and the rotational speed w (ω^) to be accurately estimated.

[0029] The low-speed position calculation section 22 applies a high-frequency observation signal to the rotating electric machine 80 and calculates the magnetic pole position θ (θ^) of the rotor based on a high-frequency component contained in a feedback current as a response component to the observation signal. According to the embodiment, a high-frequency observation signal (Vd h *: V ah cos(ω^ ht)) is generated, which is to be superimposed on a d-axis voltage command Vd. As shown in Fig. 3, a phase difference of “θ h “, as a command value, a phase command “θ h *") exists between the coordinate system (γ-δ-axis vector coordinate system) that matches the observation signal and the dq-axis vector coordinate system. A response component to the phase difference superimposed on a voltage command as a command value is included in the feedback current from the rotating electric machine 80. For example, a response component to the high-frequency observation signal is extracted from the q-axis feedback current (δ-axis feedback current Iδ) using a high-pass filter, and the response component is demodulated via a high-pass filter, a superposition circuit (mixer), a low-pass filter, etc.

[0030] When the relationship between the fixed coordinate system (α-β-axis vector coordinate system) set on the stator of the rotating electrical machine 80 and the dq-axis vector coordinate system is considered, the dq-axis vector coordinate system rotates with respect to the α-β-axis vector coordinate system, and the magnetic pole position θ can be defined as a phase angle “θ” based on the α and β axes, as shown in Fig. 4. In addition, the rotational speed w of the rotor can be defined as the rotational speed w of the dq-axis vector coordinate system with respect to the α-β-axis vector coordinate system. In the case where the magnetic pole position θ is estimated by calculation as in the embodiment, the actual dq-axis vector coordinate system cannot be directly detected. Thus, as represented by the d and q axes with ^ (roof) in Fig. 4, an estimated dq-axis vector coordinate system is set based on the magnetic pole position θ^ estimated by calculation. The magnetic pole position of the rotor based on the α and β axes is defined as "θ^" with ^ as shown in Fig. 4, and the rotational speed of the estimated dq-axis vector coordinate system with respect to the α-β-axis vector coordinate system is defined as “ω^” with ^.

[0031] As it is in Fig. As illustrated in Figure 4, by converging Δθ, which corresponds to an error between the actual dq-axis vector coordinate system and the estimated dq-axis vector coordinate system, to zero, the estimated dq-axis vector coordinate system is made to coincide with the actual dq-axis vector coordinate system. That is, by converging Δθ to zero, the estimated d- and q-axes are made to coincide with the actual d- and q-axes, and therefore the magnetic pole position is accurately detected. The rotation state information calculation section 20 calculates the magnetic pole position using this principle. For example, PI control is performed on the demodulated response component by a phase synchronization section (PLL: Phase Locked Loop) such that "Δθ" becomes "0". An estimated rotation speed "ω^ L" is calculated as a result of the PI control. A distance, i.e., an angle, is obtained by integrating a rotational speed (angular velocity). Thus, an estimated magnetic pole position "θ^ L “ by integrating the estimated speed “ω^ L “ calculated using an integrator.

[0032] During the switching control for the inverter, a so-called dead time Td (see Fig. 5), which is a period in which both the upper switching elements and the lower switching elements are simultaneously turned off, is provided to prevent a short circuit between the positive electrode side and the negative electrode side in which both the upper switching elements and the lower switching elements are turned on. For example, the dead time Td is provided by delaying a rise (dashed line) of an ideal switching pulse (control pulse) based on a modulation command by a predetermined period to generate a reference pulse.Due to the dead time Td, the effective pulse width of the actual pulse (basic pulse) may be reduced with respect to the effective pulse width of a switching pulse (ideal pulse) based on a modulation command, and an error ΔV may be caused in the output voltage obtained after converting the DC power to the AC power (see . Fig. 6). The error ΔV in the voltage caused by the dead time Td can be approximately considered as a square wave voltage, as shown by the dotted line in Fig. 6. The amplitude (error ΔV) of the square wave voltage is represented by “ΔV ≈ Ed·Td·fc”, where the intermediate circuit DC voltage Vdc is given by “Ed” and the carrier frequency during modulation is given by “fc”.

[0033] As described above, the high speed range position calculating section 21 calculates the magnetic pole position (θ^ H ) based on the voltage equation (equation (2)), and the error ΔV affects the calculation result. On the other hand, the low-speed range position calculation section 22 calculates the magnetic pole position (θ^ L ) based on the response component to the observation signal, assuming that a voltage consistent with the voltage commands (Vd*, Vq*) is output. Thus, if the error ΔV is caused in the output voltage, the result of calculating the magnetic pole position (θ^ L ) is impaired.

[0034] Therefore, as stated in Fig. As shown in Figure 7 as an example, dead-time compensation is performed, in which the error ΔV in the output voltage due to the dead time Td is reduced by shifting the start point and end point of the effective pulse by the same amount of time. For example, dead-time compensation is performed according to the polarity of the three-phase currents (Iu, Iv, Iw). Fig. Figure 7 illustrates the dead time compensation procedure for a case where the phase current is positive. The upper part of Fig. Figure 7 illustrates a theoretical switching pulse based on a modulation command. In the case where the phase current is positive, the pulse width "Tp" of the theoretical switching pulse is extended by "Td / 2" on both sides (see the middle part of Fig. 7). Specifically, the ON period of the upper switching element on both sides is lengthened by “Td / 2” respectively, and the corresponding OFF period of the lower switching element on both sides is lengthened by “Td / 2” respectively (ON periods adjacent to the OFF period preceding and following the OFF period are shortened by “Td / 2” respectively). Next, as described above with reference to Fig. 5, the rise of the switching pulse is delayed by a predetermined period corresponding to the dead time Td (see the lower part of Fig. 7). The ON period of the upper switching pulse is lengthened by "Td = (Td / 2) x 2" and then shortened by an amount equal to the dead time Td, and thus the theoretical switching pulse "Tp" is maintained, with the amount of lengthening and the amount of shortening canceling each other out.

[0035] In dead time compensation for a case where the phase current is negative, the ON period of the upper switching element on both sides is shortened by "Td / 2" each, and the corresponding OFF period of the lower switching element on both sides is shortened by "Td / 2" each (ON periods adjacent to, preceding, and following the OFF period are each extended by "Td / 2"). This means that dead time compensation for a case where the phase current is negative is performed in the opposite manner to that for a case where the phase current is positive, and can be easily justified by analogy, and thus will not be illustrated or described in detail.

[0036] To perform such dead-time compensation, it is necessary to determine the polarity of the phase currents (Iu, Iv, Iw). However, when the phase current is close to zero (near the amplitude center), the reliability of the polarity determination may be reduced due to the resolution of the current sensor 30, the influence of a current ripple, etc. Therefore, the dead-time compensation may not function sufficiently, and the accuracy of detecting the magnetic pole position may also be reduced. In contrast, when the amplitude of the phase current is small, that is, when the magnitude of an armature current Ia is small, even at the wave height (peak and trough) of the phase current, the current value is close to zero, which reduces the reliability of the polarity determination.Therefore, according to the embodiment, adjustment is performed so that the amplitude of the phase current becomes equal to or greater than a certain level so that dead time compensation can be properly performed. Specifically, when setting the current commands (Id*, Iq*) in the dq-axis vector coordinate system according to the output torque of the rotating electric machine 80, the rotating electric machine control device 1 sets the current commands (Id*, Iq*) such that the magnitude of the armature current Ia becomes equal to or greater than a lower limit current (Ia_min) prescribed in advance.

[0037] Fig. 8 illustrates a current vector space (coordinate system). In Fig. 8, curves 100 (each of curves 101 to 103) represent "equal-torque curves." The equal-torque curves 100 (101 to 103) each indicate a vector locus of the armature current Ia on which the rotating electrical machine 80 outputs a certain torque. The equal-torque curve 102 corresponds to a torque lower than that for the equal-torque curve 101, and the equal-torque curve 103 corresponds to a torque lower than that for the equal-torque curve 102. The magnitude of the armature current Ia is the magnitude of a vector composed of a d-axis current (Id) and a q-axis current (Iq) (Ia 2 = ID 2 + IQ 2). A vector composed of a u-phase current Iu, a v-phase current Iv, and a w-phase current Iw in a three-phase coordinate system with u, v, and w phases, and the vector composed of the d-axis current (Id) and the q-axis current (Iq) are the same as each other and are the armature current Ia. Curves 200 (each of curves 201 to 203) indicate "constant current circuits." The constant current circuit is a vector locus on which the magnitude of the armature current Ia has a constant value. The d-axis and q-axis current commands (Id*, Iq*) are set as current values ​​at an operating point (for example, P1, P2, and P3) in such a current vector space.

[0038] In Fig. 8 is a curve 400, a vector locus (maximum torque control curve) of an operating point during the execution of maximum torque control (basic control). As described above, the rotating electric machine control device 1 executes, as the basic control, maximum torque control in which a maximum torque is output for the armature current Ia, maximum efficiency control in which the motor is driven with a maximum efficiency for the armature current Ia, or the like. Here, the maximum torque control is executed as the basic control. During the execution of the basic control, the Fig. 2 illustrated torque control section 12 to a map like that shown in Fig. 8 as an example, to set an intersection point between the equal torque curve 100, which is selected based on the torque command T*, and the maximum torque control curve 400 as an operating point, and to set the d-axis current and the q-axis current at the operating point as the current commands (Id*, Iq*). When the magnitude of the armature current Ia at the operating point is equal to or greater than the lower limit current Ia_min prescribed in advance, the current values ​​at the operating point on the maximum torque control curve 400 are set as the current commands.

[0039] For example, assume that the magnitude of the armature current Ia at the Fig. 8 is equal to or greater than the lower limit current Ia_min, and the magnitude of the armature current Ia at the operating point P2 is smaller than the lower limit current Ia_min. Since the magnitude of the armature current Ia at the operating point P1 is equal to or greater than the lower limit current Ia_min, the d-axis and q-axis current values ​​at the operating point P1 are set as the current commands (Id*, Iq*). Conversely, since the magnitude of the armature current Ia at the operating point P2 is smaller than the lower limit current Ia_min, the operating point is moved on the equal torque curve 103 until the magnitude of the armature current Ia becomes equal to or greater than the lower limit current Ia_min. Specifically, the operating point is moved to an intersection point between the constant current circuit 200 indicating the lower limit current Ia_min and the equal torque curve 103.According to the embodiment, the magnitude of the armature current Ia at the operating point P1 is set to the lower limit current Ia_min. Thus, an intersection point between the constant current loop 203 and the equal-torque curve 103 is set as the post-motion operating point P3. Then, the d-axis and q-axis current values ​​at the operating point P3 are set as the current commands (Id*, Iq*). The constant current loop 203 is also a vector locus of a constant armature current (Ia') in the constant current control, which will be described below.

[0040] In other words, the basic control in which the operating point is set to the maximum torque control curve 400 is a control in which the current commands (Id*, Iq*) are set such that the magnitude of the current Ia corresponding to the output torque of the rotating electric machine 80 is minimized. That is, the basic control is a control in which the contact point between the constant current circuit 200 and the equal torque curve 100 is used as the operating point. In contrast, the control in which the current commands are set by moving the operating point from the operating point P2 to the operating point P3 on the equal torque curve 100 is a control in which the operating point is set to the constant current circuit 200. Thus, according to the embodiment, such a control is referred to as constant current control.As described above, the constant current loop 200 is also a vector locus of a constant armature current (Ia') in the constant current control. In the constant current control, the operating point is set to an intersection point between the constant current loop 200 and the equal torque curve 100.

[0041] In the case where the magnitude of the armature current Ia falls below the lower limit current Ia_min in the basic control, the rotating electric machine control device 1 executes the constant current control in which the current commands (Id*, Iq*) are set such that the magnitude of the armature current Ia coincides with the lower limit current Ia_min. That is, the rotating electric machine control device 1 (the torque control section 12) can execute the basic control and the constant current control, and switchably executes the basic control and the constant current control according to the magnitude of the armature current Ia at an operating point based on the torque command T*. In the case where control including such constant current control is performed, the operating point follows a locus such as that indicated by the arrow Y in Fig. 8 is specified.

[0042] In the case where control is performed with the constant current control, the rotating electric machine control device 1 preferably increases the armature current Ia gradually from zero to at least the lower limit current Ia_min when the rotating electric machine 80 starts to rotate, and decreases the armature current Ia gradually from at least the lower limit current Ia_min to zero when the rotating electric machine 80 becomes stationary. Fig. Figure 9 illustrates waveforms that schematically illustrate an example of a current command for a case where control is performed with constant current control. The top part of Fig. 9 indicates a speed command. Fig. Figure 9 schematically illustrates a case in which a so-called speed sweep is performed, in which the speed is increased from zero at a certain speed, maintained at a certain speed, and then allowed to decrease to zero over time. The second part from the top in Fig. 9 indicates the actual speed of the rotating electrical machine 80. It is indicated that the actual speed follows the speed command well. The third and fourth parts from the top in Fig. 9 indicate torque commands. The third part from the top indicates a torque command for normal maximum torque control. The fourth part from the top indicates a torque command for a case where control is performed using constant current control. As can be seen from the waveforms of the two torque commands, the torque command for maximum torque control (basic control) and the torque command for constant current control are identical to each other.

[0043] The second part from below of Fig. 9 indicates a current command (an armature current Ia command) for maximum torque control. The lowest part of Fig. Figure 9 shows a current command for a case where control is performed using constant current control. The current command for maximum torque control is substantially proportional to the torque command. In contrast, in the case where control is performed using constant current control, a current command equal to or greater than the lower limit current Ia_min is set regardless of the torque command when the torque command value is in a low-torque region. That is, when the torque command value is in the low-torque region, the current command is set such that the magnitude of the armature current Ia is not less than the lower limit current Ia_min. In the case where speed sweep is performed as shown in Fig. 9, such a low torque region occurs when the rotating electric machine 80 starts to rotate and when the rotating electric machine 80 stops. Thus, as shown in the lower part of Fig. As illustrated in Figure 9, when the rotating electric machine 80 starts rotating, the rotating electric machine control device 1 gradually reduces the armature current Ia from zero to at least the lower limit current Ia_min. In contrast, when the rotating electric machine 80 stops rotating, the rotating electric machine control device 1 gradually reduces the armature current Ia from at least the lower limit current Ia_min to zero.

[0044] Specifically, the rotating electric machine control device 1 gradually increases the current command to the lower limit current Ia_min at a control start time "t1" and maintains the current value. The current command is maintained at the lower limit current Ia_min until time "t3" at which the current command, which has been increased proportionally to the torque command since time "t2", reaches the lower limit current Ia_min in the maximum torque control. The transition of the current command from time t1 to time t3 corresponds to the vector locus along the arrow Y1 in Fig. 8. The transition of the current command after time "t3" is similar to that for a case of maximum torque control. That is, the rotating electrical machine 80 is controlled according to a current command proportional to the torque command. The transition of the current command after time "t3" corresponds to the vector locus along the arrow Y2 in Fig. 8.

[0045] The same control is also performed in the case where the rotating electric machine 80 is decelerated to stop. The rotating electric machine control device 1 decreases the current command proportional to the torque command from time "t5." When the decreasing current command reaches the lower limit current Ia_min at time "t6," the rotating electric machine control device 1 maintains the current command at the lower limit current Ia_min. At a control end time "t8" that is after time "t7" at which the current command proportional to the torque command becomes zero, the rotating electric machine control device 1 gradually decreases the armature current Ia to zero. As in the case of maximum torque control, the transition of the current command from time "t5" to time "t6" corresponds to the vector locus along the arrow Y2 in Fig. 8 in the opposite direction. In addition, the transition of the current command after time “t6” corresponds to the vector locus along the arrow Y1 in Fig. 8 in the opposite direction.

[0046] As described above, the error ΔV in the AC voltage is represented by “ΔV ≈ Ed·Td·fc”, where the intermediate circuit DC voltage Vdc is given by “Ed” and the carrier frequency during modulation is given by “fc” (see Fig. 6). This means that the error ΔV is proportional to the DC link voltage Vdc. In the case where the DC link voltage Vdc is low, the error ΔV is small, and the influence on the derivation of the magnetic pole position θ is also small. If the armature current Ia is larger than necessary (larger than the current applied in the basic control), even in the case where the influence on the derivation of the magnetic pole position θ is small, a loss will be increased. Therefore, the value of the lower limit value Ia_min is preferably set according to the DC link voltage Vdc.For example, in the case where the DC link voltage Vdc is less than a predetermined lower limit DC link voltage Vdc_min, it can be determined that the influence of the error ΔV on the calculation of the magnetic pole position θ is small, so basic control is always performed with the lower limit current Ia_min set to zero. Alternatively, the lower limit current Ia_min can be varied according to the DC link voltage Vdc.

[0047] Fig. 10 illustrates an example of the lower limit current Ia_min set according to the intermediate circuit DC voltage Vdc as described above. Fig. 10 illustrates an example in which the lower limit current Ia_min is set to become larger as the intermediate circuit DC voltage Vdc, which is a voltage on the DC side of the inverter 7, becomes higher in the case where the intermediate circuit DC voltage Vdc becomes equal to or larger than the lower limit intermediate circuit DC voltage Vdc_min prescribed in advance. Although Fig. 10 illustrates an example in which the lower limit current Ia_min is set to continuously increase, the lower limit current Ia_min may be set to gradually increase.

[0048] In addition, in the case where the magnetic pole position θ is derived by the high-speed range position calculation section 21, the expanded induced voltage given by equation (3) is used as described above. It is clear from equation (3) that as the speed w becomes higher, "wId" and "ωK E “ will also increase, which also increases the expanded induced voltage (equation (3) is repeated below). Equation (4) [EdEq]={(Ld−Lq)(ωId−I˙q)+ωKE}

[01]

[0049] If the induced voltage (extended induced voltage) is sufficiently large with respect to the error ΔV in the AC voltage, the error ΔV can be ignored during the derivation of the magnetic pole position θ. As understood from equation (3), as the rotational speed w of the rotor of the rotating electric machine 80 increases, the induced voltage becomes higher, and thus the value of the lower limit current Ia_min is preferably set according to the rotational speed ω. For example, in the case where the rotational speed ω is lower than a lower limit speed w_min prescribed in advance, the magnetic pole position θ is mainly calculated by the low-speed range position calculation section 22, and thus the lower limit current Ia_min is set to a constant value with respect to the rotational speed ω.In the case where the speed w is equal to or greater than the lower limit speed ω_min, the value of the lower limit current Ia_min is preferably varied according to the speed w.

[0050] Fig. 11 illustrates an example of the lower limit current Ia_min set according to the speed ω as described above. Fig. 11 illustrates an example in which the lower limit current Ia_min is set to become smaller as the rotational speed w of the rotor of the rotating electrical machine 80 becomes higher in the case where the rotational speed ω is equal to or greater than the lower limit speed ω_min that has been prescribed in advance. Although Fig. 11 illustrates an example in which the lower limit current Ia_min is set to continuously decrease, the lower limit current Ia_min may be set to gradually decrease.

[0051] As described above, the rotating electric machine control device 1 reduces the influence of the dead time Td on the derivation of the magnetic pole position θ by reliably performing dead time compensation. Thus, the lower limit current Ia_min is preferably set to be equal to or less than a minimum current value that enables dead time compensation. It is known that the effect of dead time compensation differs according to the detection accuracy of the three-phase currents (Iu, Iv, Iw), the stability of the three-phase currents (as well as the size of a noise component being small), and the electrical characteristics of the switching elements such as IGBTs, although this is not described in detail. Thus, the minimum current value that enables dead time compensation is preferably set with such elements in mind.For example, the minimum current value is preferably set according to the resolution of the current sensor 30, the amplitude of a ripple component of the three-phase current detected by the current sensor 30, the element characteristics (such as stray capacitance and on-resistance value) of the switching elements constituting the inverter 7, etc. Other embodiments

[0052] Other embodiments of the present invention are described below. The configuration according to each embodiment described below is not limited to independent application, but can be applied in combination with the configuration of other embodiments as long as no contradiction occurs. (1) Above, an embodiment of the present invention was illustrated by way of example, according to which the rotating electric machine serving as a driving power source for an automobile is to be controlled. However, the rotating electric machine to be controlled by the rotating electric machine control device according to the present invention is not limited to a rotating electric machine serving as a driving power source for an automobile. For example, the rotating electric machine to be controlled may be a rotating electric machine that drives an electric oil pump, a compressor for an air conditioner, or the like, even in the case where the rotating electric machine is mounted on an automobile.Alternatively, the rotating electrical machine to be controlled may be a rotating electrical machine for a railway vehicle or the like that receives power from an overhead line. Alternatively, the rotating electrical machine to be controlled may be a rotating electrical machine that drives an air conditioner, a washing machine, or the like in a workplace or at home. (2) In the above, a configuration in which the expanded induced voltage is used as the induced voltage was exemplified. However, such a configuration does not prevent the use of the induced voltage based on a general voltage equation such as the one exemplified in equation (1) in calculating the magnetic pole position. For example, even if the induced voltage is a general induced voltage, it becomes higher as the rotational speed w of the rotating electrical machine increases, and if the induced voltage is sufficiently large with respect to the error ΔV in the AC voltage due to the dead time Td, the error ΔV can be ignored during the derivation of the magnetic pole position θ. (3) As described above, the low-speed range position calculation section 22 superimposes an observation signal on a voltage command. However, the configuration of the low-speed range position calculation section 22 is not limited to this. It is possible to adopt a variety of configurations in which a high-frequency observation signal is supplied to the rotating electric machine, and in which the magnetic pole position is estimated according to a response to the observation signal. For example, an observation signal may be superimposed on a current command. (4) A variety of techniques other than those described above can be applied to detecting the magnetic pole position using sensorless control. Overview of the implementation example

[0053] An overview of the rotating electric machine control device (1) according to the embodiment will be briefly described below.

[0054] According to one embodiment, the rotating electrical machine control device (1) controls: a rotating electrical machine (80) having a rotor in which a permanent magnet is arranged and which is driven via an inverter (7) which performs a power conversion between alternating current power and direct current power, a magnetic pole position (θ) of the rotor is detected by a sensorless control, performs current control using the magnetic pole position (θ) based on a deviation between a current command (Id*, Iq*) and a feedback current from the rotating electric machine (80) in a dq-axis vector coordinate system defined by a d-axis extending in a direction of a magnetic field generated by the permanent magnet and a q-axis orthogonal to the d-axis, performs dead time compensation in which a start point and an end point of a control pulse for switching elements (3) composing the inverter (7) are adjusted to reduce an error in an AC output due to a dead time period (Td) during which both an upper switching element (3) and a lower switching element (3) of the switching elements composing one branch of a phase of the inverter (7) are controlled to an OFF state, compared to that for a case where the control pulse is a reference pulse, and carries out a switching control on the switching elements (3) for driving the rotating electrical machine (80), wherein in the case where the rotating electric machine (80) is controlled by setting the current command (Id*, Iq*) in the dq-axis vector coordinate system such that an amount of an armature current (Ia) corresponding to an output torque of the rotating electric machine (80) is minimized, the rotating electric machine is controlled such that the amount of the armature current (Ia) becomes equal to or greater than a lower limit current (Ia_min) prescribed in advance.

[0055] For reliable detection of the magnetic pole position (θ) by a sensorless controller, it is useful to perform dead time compensation. However, to perform dead time compensation, it is generally necessary to determine the polarity of the phase current of the AC power. Among these, in the case where the amplitude of the phase current is small and the wave height of the phase current is close to the amplitude center of the phase current, such as in the case where the rotating electrical machine (80) is driven with a low torque, high accuracy may be required for determining the polarity of the phase current, or the reliability of the determination may be reduced depending on the determination accuracy. In the present configuration, the current command (Id*, Iq*) is set such that the armature current (Ia) becomes equal to or greater than the lower limit current (Ia_min).The magnitude of the amplitude of the phase current can be ensured by limiting a reduction in the armature current (Ia). That is, the amplitude of the phase current can be ensured such that the polarity of the phase current can be sufficiently determined. In the case where the magnitude of the armature current (Ia) is equal to or greater than the lower limit current (Ia_min), the armature current (Ia) is not limited. Thus, the operating efficiency of the rotating electrical machine (80) is only affected in the case where the magnitude of the armature current (Ia) is less than the lower limit current (Ia_min).Thus, with the present configuration, it is possible to provide a sensorless magnetic pole position detection technique that allows the magnetic pole position (θ) of the rotating electrical machine (80) to be electrically accurately derived by reducing the voltage error due to the dead time while suppressing the influence on the operating efficiency of the rotating electrical machine (80).

[0056] To control the rotating electric machine (80), generally, maximum torque control in which maximum torque is output for the armature current (Ia), maximum efficiency control in which the rotating electric machine is driven at maximum efficiency for the armature current (Ia), or the like are performed as the basic control. In this case, for example, in the dq-axis vector coordinate system, the operating point is set on a line such as a vector locus (e.g., the maximum torque control line (400)) of the operating point at which the maximum torque is output and a vector locus (e.g., a maximum efficiency control line) of the operating point at which the rotating electric machine (80) is driven at maximum efficiency, and the current value at the operating point is set as the current command (Id*, Iq*).On the other hand, in order to set the current command (Id*, Iq*) such that the magnitude of the armature current (Ia) becomes equal to or greater than the predetermined lower limit current (Ia_min), it is necessary to set the operating point to a line of a vector locus (for example, the constant current circuit (201 to 203)) of the operating point at which the armature current (Ia) coincides with the lower limit current (Ia_min). Thus, the rotating electric machine control device (1) is preferably configured to perform control based on at least two such types of operating points. Although the vector locus and the operating point have been illustrated here as an example for ease of understanding, this does not mean that it is essential to set the operating point in actual control.According to a preferred embodiment, the rotating electric machine control device (1) can be capable of executing basic control in which the current command (Id*, Iq*) is set such that the magnitude of the armature current (Ia) is minimized according to the output torque of the rotating electric machine (80), and constant current control in which the current command (Id*, Iq*) is set such that the magnitude of the armature current (Ia) coincides with the lower limit current (Ia_min) in the case where the magnitude of the armature current (Ia) drops below the lower limit current (Ia_min) in the basic control, and the rotating electric machine control device (1) can switchably execute the basic control and the constant current control according to the magnitude of the armature current (Ia).

[0057] Preferably, in the rotating electric machine control device (1), the armature current (Ia) is gradually increased from zero to at least the lower limit current (Ia_min) when the rotating electric machine (80) starts rotating, and the armature current (Ia) is gradually decreased from at least the lower limit current (Ia_min) to zero when the rotating electric machine (80) stops. With this configuration, the current command can be appropriately set to the lower limit current (Ia_min) or greater even in a low-torque region.

[0058] The error in the AC voltage due to the dead time (Td) is proportional to the voltage (DC link voltage (Vdc)) on the DC side of the inverter (7), the frequency of the control pulse during switching control at the switching elements (3) (modulation frequency during modulation), and the period of the dead time (Td). Thus, when the DC link voltage (Vdc) is low, the error in the AC voltage due to the dead time (Td) is small, and the influence on the derivation and detection of the magnetic pole position (θ) by the sensorless controller is also small. If the armature current (Ia) larger than necessary is supplied, even when the influence on the derivation of the magnetic pole position (θ) is small, a loss will be increased. Therefore, the value of the lower limit current (Ia_min) is preferably set according to the DC link voltage (Vdc).That is, according to a configuration, in the rotating electric machine control device (1), the lower limit current (Ia_min) is preferably set to become continuously or stepwise larger as a DC link voltage (Vdc), which is a voltage on a DC side of the inverter (7), becomes higher in the case where the DC link voltage (Vdc) is equal to or larger than a lower limit DC link voltage (Vdc_min) prescribed in advance.

[0059] An induced voltage generated by the rotation of the rotor of the rotating electric machine (80) increases in proportion to the rotational speed (w) of the motor. In the case where the magnetic pole position (θ) is calculated and detected based on the induced voltage generated by the rotation of the rotor, and if the induced voltage is sufficiently large with respect to the error in the AC voltage, the error during the derivation of the magnetic pole position (θ) can be ignored. If the armature current (Ia) larger than necessary is supplied even in the case where the influence on the derivation of the magnetic pole position (θ) is small, a loss is increased. The induced voltage becomes higher as the rotational speed (ω) of the rotor becomes higher, and thus the value of the lower limit current (Ia_min) is preferably set according to the rotational speed (w).That is, according to a configuration, in the rotating electric machine control device (1), the lower limit current (Ia_min) is preferably set to continuously or stepwise become smaller as the rotational speed (w) of the rotor becomes higher in the case where the rotational speed (w) is equal to or higher than a lower limit speed (ω_min) prescribed in advance.

[0060] In the rotating electric machine control device (1), the lower limit current (Ia_min) is preferably set to be equal to or greater than a minimum current value that enables dead time compensation. In order to reduce the influence of the dead time (Td) on the derivation of the magnetic pole position (θ), dead time compensation can preferably be reliably performed. When the lower limit current (Ia_min) is set to be equal to or greater than the minimum current value that enables dead time compensation, the magnetic pole position (θ) of the rotating electric machine (80) can be accurately electrically derived by performing dead time compensation, and the voltage error due to the dead time (Td) can be reduced. Industrial applicability

[0061] The present invention can be applied to a rotating electric machine control device that controls a rotating electric machine having a rotor in which a permanent magnet is arranged and driven via an inverter that performs power conversion between AC power and DC power. List of reference symbols 1 Rotating electrical machine control device 3 IGBT (switching element) 7 inverters 80 Rotating electrical machine 90 vehicle ECU Ia armature current Ia_min Lower limit current Id* d-axis current command (current command) Iq* q-axis current command (current command) Td dead time (dead time period) Vd d-axis voltage command Vdc intermediate circuit DC voltage Vdc_min lower limit DC link voltage ΔV error θ magnetic pole position θ^ estimated magnetic pole position (magnetic pole position) ω speed ω^ estimated speed (RPM) ω_min lower limit speed

Claims

[1] Control device for a rotating electrical machine (1), which: controls a rotating electrical machine (80) having a rotor in which a permanent magnet is arranged and which is driven via an inverter (7) which performs a power conversion between alternating current power and direct current power, a magnetic pole position (θ) of the rotor is detected by a sensorless control, performs current control using the magnetic pole position (θ) based on a deviation between a current command (Id*, Iq*) and a feedback current from the rotating electric machine (80) in a dq-axis vector coordinate system defined by a d-axis extending in a direction of a magnetic field generated by the permanent magnet and a q-axis orthogonal to the d-axis, performs dead time compensation in which a start point and an end point of a control pulse for switching elements (3) composing the inverter (7) are adjusted to reduce an error in an AC output due to a dead time period (Td) during which both an upper switching element (3) and a lower switching element (3) of the switching elements composing one branch of a phase of the inverter (7) are controlled to an OFF state, compared to that for a case where the control pulse is a reference pulse, and carries out a switching control on the switching elements (3) for driving the rotating electrical machine (80), wherein in the case where the rotating electric machine (80) is controlled by setting the current command (Id*, Iq*) in the dq-axis vector coordinate system such that an amount of an armature current (Ia) corresponding to an output torque of the rotating electric machine (80) is minimized, the rotating electric machine is controlled such that the amount of the armature current (Ia) becomes equal to or greater than a lower limit current (Ia_min) prescribed in advance. [2] A control device for a rotating electrical machine (1) according to claim 1, wherein the control device for a rotating electrical machine is capable of to carry out a basic control in which the current command (Id*, Iq*) is set such that the magnitude of the armature current (Ia) is minimized in accordance with the output torque of the rotating electrical machine (80), and to carry out a constant current control in which the current command (Id*, Iq*) is set such that the magnitude of the armature current (Ia) matches the lower limit current (Ia_min) in the case where the magnitude of the armature current (Ia) in the basic control falls below the lower limit current (Ia_min), and the control device for a rotating electrical machine (1) switchably executes the basic control and the constant current control according to the amount of the armature current (Ia). [3] A control device for a rotating electrical machine (1) according to claim 1 or 2, wherein the armature current (Ia) is gradually increased from zero to at least the lower limit current (Ia_min) when the rotating electrical machine (80) starts to rotate, and the armature current (Ia) is gradually decreased from at least the lower limit current (Ia_min) to zero when the rotating electrical machine (80) becomes stationary. [4] A control device for a rotating electric machine (1) according to any one of claims 1 to 3, wherein the lower limit current (Ia_min) is set to become continuously or stepwise larger as a DC link voltage (Vdc), which is a voltage on a DC side of the inverter (7), becomes higher in the case where the DC link voltage (Vdc) is equal to or larger than a lower limit DC link voltage (Vdc_min) prescribed in advance. [5] A control device for a rotating electrical machine (1) according to any one of claims 1 to 4, wherein the lower limit current (Ia_min) is set to continuously or stepwise become smaller as the rotational speed (ω) of the rotor becomes higher in the case where the rotational speed (ω) is equal to or higher than a lower limit speed (ω_min) prescribed in advance. [6] A control device for a rotating electrical machine (1) according to any one of claims 1 to 5, wherein the lower limit current (Ia_min) is set to be equal to or greater than a minimum current value enabling dead time compensation.

Images