AC rotating motor control device

Abstract

Provided is an AC rotating electric machine control device capable of compensating an error of an actual on-duty ratio relative to a command on-duty ratio calculated based on a command voltage with high accuracy using a simple circuit configuration. The AC rotating electric machine control device (1) detects an actual on-duty ratio (Dr) of a switching element based on a detected value of a potential of a connection point of a series connection in a series circuit, i.e., a midpoint potential, calculates an on-duty ratio error (ΔDer) based on a difference between a command on-duty ratio (Do) and the actual on-duty ratio (Dr), and corrects a voltage command or a command on-duty ratio (Do) based on the on-duty ratio error (ΔDer).

Description

Technical Field

[0001] The present application relates to an AC rotating electric machine control device. Background Art

[0002] Electric vehicles such as electric vehicles and hybrid vehicles are equipped with AC rotating electric machines as vehicle drive sources. The inverter converts the DC power supplied from the DC power supply into AC power to supply the AC rotating electric machine. The inverter includes switching elements such as MOSFET (Metal Oxide Semiconductor Field Effect Transistor).

[0003] Generally, in the switch element, since there is a delay time of on-delay and off-delay, if the on-off switching of the switch element on the positive side and the switch element on the negative side is performed at the same time, the switch elements on the positive side and the negative side may be short-circuited. In order to prevent this short circuit, a dead time for turning off both switch elements is provided between the on-period of the switch element on the positive side and the on-period of the switch element on the negative side. If there is a dead time, an error occurs in the actual on-duty ratio for the command on-duty ratio of the switch element calculated based on the voltage command.

[0004] In the technology disclosed in Patent Document 1, in order to compensate for the error in the on-duty ratio caused by the dead time, the dead time is compensated by feedforward control using a reference model circuit of a current control loop and a current command value.

[0005] In addition, in the technology disclosed in Patent Document 2, a correction instruction voltage is calculated based on a duty cycle instruction value, a detection voltage is estimated based on the terminal voltage of the winding, and a loss voltage caused by the dead time is estimated based on the difference, and dead time compensation is performed by feeding back the dead time compensation value to the dq axis current instruction value.

[0006] Prior art literature

[0007] Patent Literature

[0008] Patent Document 1: Japanese Patent No. 4681453

[0009] Patent Document 2: Japanese Patent No. 6319532 Summary of the invention

[0010] Technical problem to be solved by the invention

[0011] However, the technique of Patent Document 1 requires the determination of a threshold value of a reference model circuit. In order to improve compensation accuracy, the threshold value needs to be switched under various conditions, which requires adjustment work on an actual device.

[0012] In the technology of Patent Document 2, since the detection voltage is estimated based on the terminal voltage of the motor, a voltage detection circuit is required. In order to ensure the detection accuracy, a filter circuit for removing external noise is also required, which has the problem of increasing the circuit scale. In addition, there is a problem of increased computational processing burden in the estimation of the detection voltage and the estimation of the loss voltage.

[0013] Therefore, there is a demand for an AC rotating electric machine control device that can accurately compensate for an error in an actual on-duty ratio with respect to a command on-duty ratio calculated based on a command voltage with a simple circuit configuration.

[0014] Technical means for solving technical problems

[0015] The AC rotating electric machine control device involved in the present application is an AC rotating electric machine control device for controlling an AC rotating electric machine having an n-phase (n is a natural number greater than or equal to 2) winding, comprising:

[0016] An inverter, wherein n sets of series circuits are provided corresponding to each of the n phases, wherein the series circuits connect in series a switching element on a positive side connected to a positive side of a DC power source and a switching element on a negative side connected to a negative side of the DC power source, and a connection point of the series connection is connected to the winding of the corresponding phase;

[0017] a midpoint potential detection circuit for detecting a potential of a connection point of the series connection in the series circuit, that is, a midpoint potential;

[0018] a control circuit for controlling the switching element on and off based on a voltage command or a command on-duty ratio calculated from the voltage command,

[0019] The control circuit

[0020] Based on the detected value of the midpoint potential, the actual on-duty ratio of the switching element is detected,

[0021] Calculate the error of the on-duty ratio according to the difference between the command on-duty ratio and the actual on-duty ratio,

[0022] Based on the error of the on-duty ratio, the voltage command or the command on-duty ratio is corrected.

[0023] Effects of the Invention

[0024] According to the AC rotating electric machine control device involved in the present application, it is possible to detect the actual on-duty ratio based on the detection value of the midpoint potential and calculate the error of the on-duty ratio. Moreover, by correcting the voltage command or the command on-duty ratio according to the error of the on-duty ratio, the switching element can be turned on or off according to the command, and the control accuracy of the voltage applied to the winding can be improved. In addition, since the detection value of the midpoint potential of the midpoint potential detection circuit is used, the error of the on-duty ratio can be detected with a simple circuit structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a schematic configuration diagram of the AC rotating electric machine and the AC rotating electric machine control device according to the first embodiment.

[0026] Figure 2 This is a block diagram of a control circuit according to the first embodiment.

[0027] Figure 3 This is a hardware configuration diagram of the control circuit according to the first embodiment.

[0028] Figure 4 This is a timing chart for explaining the control operation involved in the first embodiment.

[0029] Figure 5 This is a diagram for explaining the positive and negative of the winding current and the on and off of the switching element during the dead time according to the first embodiment.

[0030] Figure 6 This is a diagram for explaining the relationship characteristics between the winding current and the on-duty ratio error according to the first embodiment.

[0031] Figure 7 This is a diagram for explaining the learning of the duty ratio error characteristic according to the first embodiment.

[0032] Figure 8 This is a diagram for explaining the correction coefficient characteristics according to the first embodiment.

[0033] Fig. 9 This is a diagram for explaining the learning permission range involved in the first embodiment.

[0034] Fig.10 This is a block diagram of a control circuit according to the second embodiment.

[0035] Fig.11 This is a diagram for explaining the setting of time allocation involved in Implementation Example 4.

[0036] Fig.12 This is a timing chart for explaining the setting of time allocation involved in Implementation Example 4. DETAILED DESCRIPTION

[0037] 1. Implementation Method 1

[0038] An AC rotating electric machine control device 1 (hereinafter simply referred to as a control device 1 ) according to Embodiment 1 will be described with reference to the drawings. Figure 1 It is a schematic configuration diagram of the AC rotating electric machine 2 and the control device 1 according to the present embodiment.

[0039] 1-1. AC rotating motor

[0040] The AC rotating motor 2 is provided with a winding of n phases (n is a natural number greater than or equal to 2). In the present embodiment, the AC rotating motor 2 is a permanent magnet type synchronous AC rotating motor having a stator provided with a winding of n phases and a rotor provided with a permanent magnet. In the present embodiment, n=3, and three phases are provided, namely, the U phase, the V phase, and the W phase. The stator is provided with three-phase windings Cu, CV, and Cw. The three-phase windings Cu, Cv, and Cw are provided in star connection. In addition, the three-phase windings may also be provided in delta connection.

[0041] The AC rotating electric machine 2 includes a rotation sensor 16 that outputs an electrical signal corresponding to the rotation angle of the rotor. The rotation sensor 16 is configured as a Hall element, an encoder, a resolver, etc. The output signal of the rotation sensor 16 is input to the control circuit 30 .

[0042] 1-2. Control device 1

[0043] The control device 1 includes an inverter 20 , a power supply voltage detection circuit 13 , a current detection circuit 17 , a midpoint potential detection circuit 18 , a gate drive circuit 12 , a redundant three-phase short-circuit circuit 19 , and a control circuit 30 .

[0044] 1-2-1. Inverter, etc.

[0045] The inverter 20 has a plurality of switching elements, and performs DC to AC conversion between the DC power supply 10 and the three-phase winding. The inverter 20 is provided with three sets of series circuits (legs: branches) corresponding to the windings of each of the three phases, and the series circuit is connected in series with a positive-side switching element 23H (upper arm) connected to the positive side of the DC power supply 10, and a negative-side switching element 23L (lower arm) connected to the negative side of the DC power supply 10. The inverter 20 includes three positive-side switching elements 23H and three negative-side switching elements 23L, a total of six switching elements. Moreover, the connection point where the positive-side switching element 23H and the negative-side switching element 23L are connected in series is connected to the winding of the corresponding phase.

[0046] Specifically, in the series circuit of each phase, the collector terminal of the switching element 23H on the positive side is connected to the positive side wire 14, the emitter terminal of the switching element 23H on the positive side is connected to the collector terminal of the switching element 23L on the negative side, and the emitter terminal of the switching element 23L on the negative side is connected to the negative side wire 15. The connection point between the switching element 23H on the positive side and the switching element 23L on the negative side is connected to the winding of the corresponding phase.

[0047] The switching element uses an element having the function of a diode connected in reverse parallel, for example, an IGBT (Insulated Gate Bipolar Transistor) having a diode 22 connected in reverse parallel, or a MOSFET (Metal Oxide Semiconductor Field Effect Transistor) having a parasitic diode connected in reverse parallel.

[0048] <Gate drive circuit 12>

[0049] The gate drive circuit 12 drives the switch element on and off. The gate drive circuit 12 is connected to the gate terminal of each switch element via the redundant three-phase short-circuit circuit 19. The gate drive circuit 12 outputs an on signal or an off signal to each switch element according to a control signal transmitted from the control circuit 30 via an optical coupler or the like, thereby driving each switch element on or off.

[0050] <Redundant three-phase short-circuit circuit 19>

[0051] The redundant three-phase short-circuit circuit 19 forcibly short-circuits or opens (turns on or off) each switching element according to a command signal from the control circuit 30. The short-circuited or opened switching elements may be the switching elements on the positive side of the three phases, the switching elements on the negative side of the three phases, or all the switching elements.

[0052] <Current Detection Circuit 17>

[0053] The current detection circuit 17 detects the winding current flowing through each phase winding. The current detection circuit 17 outputs an electrical signal corresponding to the winding current, and the output signal is input to the control circuit 30. In this embodiment, the current detection circuit 17 is a Hall element or the like provided on the wires of each phase connecting the connection point of the series circuit of the switching element and the winding. Alternatively, the current detection circuit 17 may be a shunt resistor connected in series with the series circuit of each phase.

[0054] <Midpoint Potential Detection Circuit 18>

[0055] The midpoint potential detection circuit 18 detects the potential of the connection point between the switching element on the positive side and the switching element on the negative side in the series circuit, that is, the midpoint potential. The current detection circuit 17 outputs an electrical signal corresponding to the midpoint potential, and the output signal is input to the control circuit 30. In this embodiment, the midpoint potential detection circuit 18 detects the midpoint potential of the connection point of the series circuit of each of the three phases.

[0056] In this embodiment, the midpoint potential detection circuit 18 is a circuit that turns on or off the output signal according to whether the midpoint potential is greater than or less than the potential threshold. When the midpoint potential is greater than the potential threshold, the midpoint potential detection circuit 18 turns on the output signal, and when the midpoint potential is less than the potential threshold, the midpoint potential detection circuit 18 turns off the output signal. The potential threshold is set to a voltage within the range of 0 to the power supply voltage Vdc (for example, half of the power supply voltage Vdc). Different potential thresholds can also be used at the rising edge and the falling edge of the midpoint potential. The midpoint potential detection circuit 18 can also be composed of a simplified circuit such as a comparator that compares the potential threshold and the midpoint potential.

[0057] <Power supply voltage detection circuit 13>

[0058] The power supply voltage detection circuit 13 detects the power supply voltage Vdc supplied from the DC power supply 10 to the inverter 20. The current and voltage detection circuit 13 outputs an electric signal corresponding to the power supply voltage Vdc, and the output signal is input to the control circuit 30.

[0059] <DC power supply 10>

[0060] The DC power supply 10 uses a chargeable and dischargeable power storage device (e.g., a lithium-ion battery, a nickel-metal hydride battery, or an electric double-layer capacitor). In addition, the DC power supply 10 may also be provided with a DC power converter that steps up or steps down the DC voltage, i.e., a DC-DC converter. A smoothing capacitor may also be connected between the positive side wire 14 and the negative side wire 15 of the inverter 20.

[0061] 1-2-2. Control circuit 30

[0062] The control circuit 30 controls the AC rotating electric machine 2 by controlling the inverter 20. Figure 2 As shown in FIG. 1 , the control device 30 includes an operation state detection unit 31, a basic voltage command calculation unit 32, a duty cycle error correction unit 33, and a PWM control unit 34, which will be described later. Each function of the control circuit 30 is implemented by a processing circuit provided in the control circuit 30. Specifically, the control circuit 30 is as follows. Figure 3As shown, as a processing circuit, it includes an operation processing device 90 (computer) such as a CPU (Central Processing Unit), a storage device 91 for exchanging data with the operation processing device 90, an input circuit 92 for inputting external signals to the operation processing device 90, and an output circuit 93 for outputting signals from the operation processing device 90 to the outside.

[0063] The operation processing device 90 may include an ASIC (Application Specific Integrated Circuit), an IC (Integrated Circuit), a DSP (Digital Signal Processor), an FPGA (Field Programmable Gate Array), various logic circuits, and various signal processing circuits. In addition, the operation processing device 90 may include a plurality of operation processing devices of the same type or different types to share the execution of various processes. The storage device 91 includes a RAM (Random Access Memory) configured to read data from the operation processing device 90 and write data to the operation processing device 90, and a ROM (Read Only Memory) configured to read data from the operation processing device 90. The input circuit 92 is connected to various sensors and switches such as the rotation sensor 16, the power supply voltage detection circuit 13, the current detection circuit 17, and the midpoint potential detection circuit 18, and includes an A / D converter and an input port for inputting output signals of these sensors and switches to the operation processing device 90. The output circuit 93 is connected to electric loads such as the gate drive circuit 12 and the redundant three-phase short-circuit circuit 19 , and includes an output port for outputting control signals from the arithmetic processing device 90 to these electric loads, a drive circuit, and the like.

[0064] The control circuit 30 has Figure 2 Each function of each control unit 31 to 34 is realized by executing the software (program) stored in a storage device 91 such as a ROM by the arithmetic processing unit 90, and cooperating with other hardware of the control unit 30 such as the storage device 91, the input circuit 92, and the output circuit 93. In addition, setting data such as threshold values ​​used by each control unit 31 to 34 is stored in the storage device 91 such as a ROM as part of the software (program). The functions of the control circuit 30 are described in detail below.

[0065] 1-2-2-1. Operation status detection unit 31

[0066] The operation state detection unit 31 detects the power supply voltage Vdc based on the output signal of the power supply voltage detection circuit 13. The operation state detection unit 31 detects the winding currents Iud, Ivd and Iwd flowing through the three-phase windings based on the output signal of the current detection circuit 17. In addition, the operation state detection unit 31 detects the rotation angle θ (magnetic pole position θ) and the rotation angular velocity ω of the rotor based on the output signal of the rotation sensor 16. The operation state detection unit 31 detects the midpoint potential based on the output signal of the midpoint potential detection circuit 18. The operation state detection unit 31 detects the winding current, the magnetic pole position θ, and the rotation angular velocity ω at the timing (starting time of the PWM operation cycle) of one or both of the peaks and valleys of the carrier CA (in this example, the peaks of the carrier CA).

[0067] 1-2-2-2. Basic voltage command calculation unit 32

[0068] The basic voltage command calculation unit 32 calculates the three-phase basic voltage commands Vuob, Vvob, and Vwob, which are the three-phase voltage commands before correction by the duty ratio error correction unit 33 described later. For example, the basic voltage command calculation unit 32 calculates the three-phase basic voltage commands Vuob, Vvob, and Vwob using a known vector control.

[0069] In the present embodiment, the basic voltage command calculation unit 32 calculates the basic voltage commands Vuob, Vvob, and Vwob of the three phases by current feedback control for controlling the current flowing through the winding in the dq axis rotating coordinate system. The dq axis rotating coordinate system is a two-axis rotating coordinate system consisting of a d axis determined in the N pole direction (magnetic pole position) of a permanent magnet provided on the rotor and a q axis determined in a direction 90° (π_ / 2) ahead of the d axis in electrical angle, and rotates synchronously with the rotation of the rotor in electrical angle.

[0070] The following is a detailed description. The basic voltage command calculation unit 32 calculates the d-axis current command Ido and the q-axis current command Iqo based on the target torque, power supply voltage, speed, etc., according to the current vector control method such as maximum torque current control, weak magnetic control, and Id=0 control. The target torque can be transmitted from an external device or calculated in the basic voltage command calculation unit 32.

[0071] The basic voltage command calculation unit 32 performs three-phase to two-phase conversion and rotation coordinate conversion based on the magnetic pole position θ, thereby converting the three-phase winding currents Iud, Ivd, and Iwd into d-axis current detection values ​​Idd and q-axis current detection values ​​Iqd represented by a dq-axis rotation coordinate system.

[0072] The basic voltage command calculation unit 32 performs feedback control to change the d-axis voltage command Vdo and the q-axis voltage command Vqo by PI control or the like so as to bring the dq-axis current detection values ​​Idd and Iqd close to the dq-axis current commands Ido and Iqo.

[0073] The basic voltage command calculation unit 32 performs fixed coordinate conversion and two-phase to three-phase conversion on the dq axis voltage commands Vdo and Vqo based on the magnetic pole position θ, and converts them into three-phase basic voltage commands Vuob, Vvob, and Vwob. Commands modulated by applying third harmonic injection or the like to the three-phase basic voltage commands Vuob, Vvob, and Vwob may also be used.

[0074] 1-2-2-3. PWM control unit 34

[0075] The PWM control unit 34 performs on-off control of the switch elements on the positive and negative sides of each phase by PWM control (Pulse Width Modulation). In the present embodiment, the PWM control unit 34 performs on-off control of each switch element based on the voltage instructions Vuo, Vvo, and Vwo of each phase of the three phases after the basic voltage instructions Vuob, Vvob, and Vwob of the three phases are corrected by the duty cycle error correction unit 33 described later. In addition, the voltage instructions Vuo, Vvo, and Vwo of each phase of the three phases are calculated during the period between the timing of one or both of the peaks and valleys of the carrier CA (in this example, the peaks of the carrier CA), that is, the PWM operation cycle, and are updated at the end of the PWM operation cycle.

[0076] The PWM control unit 34 generates PWM signals for each of the three phases that are switched on and off with the PWM period Tcc and the commanded on-duty ratios Do_u, Do_v, and Do_w of each of the three phases based on the voltage commands Vuo, Vvo, and Vwo of each of the three phases. As shown in the formula (1) described later, the value obtained by adding 0.5 to the value obtained by dividing the voltage commands Vuo, Vvo, and Vwo of each of the three phases by the power supply voltage Vdc becomes the commanded on-duty ratios Do_u, Do_v, and Do_w of each of the three phases. The commanded on-duty ratio Do corresponds to the on-duty ratio of the switching element on the positive side.

[0077] In the present embodiment, the PWM control unit 34 generates a PWM signal for turning each switching element on and off based on the comparison result of the voltage commands Vuo, Vvo, and Vwo of each of the three phases and the carrier CA. Figure 4As shown in FIG. 1 , it is assumed that the carrier CA is a triangular wave that oscillates with an amplitude of the power supply voltage Vdc / 2 with 0 as the center in the PWM period Tcc, and the PWM control unit 34 turns on the PWM signal when the voltage command exceeds the carrier CA, and turns off the PWM signal when the voltage command is lower than the triangular wave. If the dead time described later is ignored, for each phase, when the PWM signal is turned on, the control signal of the switch element on the positive side becomes turned on, and the control signal of the switch element on the negative side becomes turned off, and when the PWM signal is turned off, the control signal of the switch element on the positive side becomes turned off, and the control signal of the switch element on the negative side becomes turned on.

[0078] <Dead Time>

[0079] In the present embodiment, the PWM control unit 34 is configured to alternately turn on and drive the positive-side switching element and the negative-side switching element with a dead time for each phase. That is, for each phase, the PWM control unit 34 provides a dead time for turning off the driving of both the positive-side switching element and the negative-side switching element between the on-driving period of the positive-side switching element and the on-driving period of the negative-side switching element.

[0080] In this embodiment, for each phase, the PWM control unit 34 shortens the dead time of the on-drive period of the switching element on the positive side compared to the on-drive period corresponding to the command on-duty cycle Do, and extends the dead time of the off-drive period of the switching element on the negative side compared to the on-drive period corresponding to the duty cycle obtained by subtracting the command on-duty cycle Do from 1.

[0081] like Figure 4 As shown in FIG. 1 , the PWM control unit 34 generates a PWM signal for each phase that is turned on and off by the command on-duty ratio Do as described above. Then, for each phase, the PWM control unit 34 drives the switching element on the negative side to be turned off when the PWM signal is turned on, drives the switching element on the positive side to be turned on when the dead time has passed after the PWM signal is turned on, drives the switching element on the positive side to be turned off when the PWM signal is turned off, and drives the switching element on the negative side to be turned on when the dead time has passed after the PWM signal is turned off.

[0082] 1-2-2-4. Duty ratio error correction unit 33

[0083] The error between the commanded on-duty ratio Do and the actual on-duty ratio Dr is caused by various main reasons. Therefore, the duty ratio error correction unit 33 detects the on-duty ratio error ΔDer, and corrects the voltage command of each of the three phases based on the on-duty ratio error ΔDer. In the present embodiment, the duty ratio error correction unit 33 corrects the basic voltage commands Vuob, Vvob, and Vwob of the three phases based on the on-duty ratio error ΔDer, and calculates the voltage commands Vuo, Vvo, and Vwo of each of the three phases. First, the main reasons for the on-duty ratio error are described.

[0084] <Changes in the actual on-duty ratio Dr caused by the positive and negative winding current>

[0085] During the dead time, current flows through the diode provided on the positive or negative switching element and connected in reverse parallel, depending on whether the winding current is positive or negative. Therefore, the actual on-duty ratio Dr of the positive switching element varies.

[0086] like Figure 5 As shown, during the dead time, when the winding current is positive, the current flows through the diode 22 of the switch element 23L on the negative side, so the switch element 23L on the negative side is actually turned on, and the actual on period of the negative side is extended by twice the dead time compared to the on-drive period. On the other hand, since the switch element 23H on the positive side is actually turned off during the dead time, the actual on period of the positive side is maintained as the on-drive period. During the dead time, the midpoint potential of the connection point between the switch element 23H on the positive side and the switch element 23L on the negative side is turned on with the negative side of the DC power supply 10, and therefore becomes the same potential as the negative side of the DC power supply 10 (0, off in this example).

[0087] Therefore, when the winding current is positive, the actual on-duty ratio Dr of the switching element 23H on the positive side maintains the driving on-duty ratio, which is reduced by the on-duty ratio amount corresponding to the dead time compared to the command on-duty ratio Do. Figure 4 As shown, in this case, while the switching element 23H on the positive electrode side is actually turned on, it can be detected that the midpoint potential has become the same potential as the positive electrode side of the DC power supply 10 (in this example, the power supply voltage Vdc, on).

[0088] On the other hand, Figure 5As shown, during the dead time, when the winding current is negative, the current flows through the diode 22 of the switch element 23H on the positive side, so the switch element 23H on the positive side is actually turned on, and the actual on period of the positive side is extended by twice the dead time than the on drive period. On the other hand, since the switch element 23L on the negative side is actually kept off during the dead time, the actual on period of the negative side is kept as the on drive period. During the dead time, the midpoint potential is the same potential as the positive side of the DC power supply 10 because it is turned on. (In this example, the power supply voltage Vdc, on)

[0089] Therefore, when the winding current is negative, the actual on-duty ratio Dr of the switch element 23H on the positive side increases by an on-duty ratio corresponding to twice the dead time compared to the drive on-duty ratio, and increases by an on-duty ratio corresponding to the dead time compared to the command on-duty ratio Do. Figure 4 As shown, in this case, while the switching element 23H on the positive electrode side is actually turned on, it can be detected that the midpoint potential has become the same potential as the positive electrode side of the DC power supply 10 (in this example, the power supply voltage Vdc, on).

[0090] <Changes in the actual on-duty ratio Dr caused by turn-off delay and turn-on delay>

[0091] Even after the PWM signal is driven to be turned off, there is a time when the switching element continues to flow so that the current flows. Therefore, there is a time delay before the switching element is actually turned off (hereinafter referred to as the turn-on delay). In addition, after the PWM signal is driven to be turned on, there is also a time delay before the switching element is actually turned on (hereinafter referred to as the turn-on delay). The turn-off delay and the turn-on delay vary according to the magnitude of the winding current. In this case, during the period when the switching element on the positive side is actually turned on, it can be detected that the midpoint potential becomes the same potential as the positive side of the DC power supply (in this example, the power supply voltage Vdc, turn-on).

[0092] <Correction of Detected On-Duty Error ΔDer>

[0093] Therefore, the duty cycle error correction unit 33 detects the actual on-duty cycle Dr of the switching element based on the detected value of the midpoint potential, and calculates the on-duty cycle error ΔDer based on the difference between the command on-duty cycle Do calculated from the voltage command and the actual on-duty cycle Dr. However, the duty cycle error correction unit 33 corrects the voltage command based on the on-duty cycle error ΔDer.

[0094] According to this structure, the actual on-duty ratio Dr can be detected based on the detection value of the midpoint potential, and the on-duty ratio error ΔDer can be calculated. Moreover, by correcting the voltage command according to the on-duty ratio error ΔDer, the switching element can be turned on or off according to the command, and the control accuracy of the voltage applied to the winding can be improved. In addition, since the detection value of the midpoint potential of the midpoint potential detection circuit 18 is used, the on-duty ratio error ΔDer can be detected with a simple circuit structure.

[0095] As shown in the following equation, the duty ratio error correction unit 33 adds 0.5 to the value obtained by dividing the voltage commands Vuo, Vvo, Vwo of each of the three phases by the power supply voltage Vdc, and calculates the command on-duty ratios Do_u, Do_v, Do_w of each of the three phases.

[0096]

[0097] The duty ratio error correction unit 33 detects the actual on-duty ratio Dr based on the timing when the midpoint potential is larger than the potential threshold and the timing when the midpoint potential is smaller than the potential threshold.

[0098] In this embodiment, if Figure 4 As shown in the operation of the U phase in the middle, the duty cycle error correction unit 33 uses a timer to measure the cut-off period GPR1_u, which is the period from the moment when the midpoint potential of the series circuit of the U phase is less than the potential threshold and the midpoint potential becomes cut-off to the moment when the midpoint potential becomes cut-off. In addition, the duty cycle error correction unit 33 uses a timer to measure the cut-on period GPR0_u, which is the period from the moment when the midpoint potential becomes cut-off to the moment when the midpoint potential is greater than the potential threshold and the midpoint potential becomes conductive. Moreover, as shown in the following formula, the duty cycle error correction unit 33 subtracts the cut-on period GPR0_u from the cut-off period GPR1_u, calculates the on period of the switching element on the positive side, and detects the value obtained by dividing the on period on the positive side by the PWM cycle Tcc as the actual on duty cycle Dr_u of the U phase.

[0099] Dr_u=(GPR1_u-GPR0_u) / Tcc

[0100] Dr_v=(GPR1_v-GPR0_v) / Tcc

[0101] Dr_w=(GPR1_w-GPR0_w) / Tcc…(2)

[0102] As shown in formula (2), for the V phase, the duty cycle error correction unit 33 measures the off-on period GPR0_v and the off-off period GPR1_v based on the midpoint potential of the series circuit of the V phase, and calculates the actual on-duty cycle Dr_v of the V phase. For the W phase, the duty cycle error correction unit 33 measures the off-on period GPR0_w and the off-off period GPR1_w based on the midpoint potential of the series circuit of the W phase, and calculates the actual on-duty cycle Dr_w of the W phase.

[0103] The duty ratio error correction unit 33 may measure the period from the time when the midpoint potential becomes on to the time when the midpoint potential becomes off using a timer, or may directly measure the on period of the switching element on the positive electrode side.

[0104] The duty ratio error correction unit 33 is based on the last PWM operation cycle ( Figure 4 The cut-off period GPR1 and the cut-off period GPR0 measured in (n-1)) are in this PWM operation cycle ( Figure 4 The actual on-duty ratio Dr is calculated in (n))).

[0105] This PWM operation cycle ( Figure 4 The actual on-duty ratio Dr calculated in (n) corresponds to the PWM operation cycle ( Figure 4 The end time of the PWM operation cycle two cycles ago calculated in (n-2)) Figure 4 The command voltage set by (a) in FIG.

[0106] Therefore, as shown in the following equation, for the U phase, the duty ratio error correction unit 33 calculates the duty ratio error from the PWM operation cycle ( Figure 4 The command conduction duty cycle Do_u(n-2) calculated in (n-2) is subtracted from the current PWM operation cycle ( Figure 4 The actual on-duty ratio Dr_u(n) calculated in (n)) is used to calculate the on-duty ratio error ΔDer_u(n) of this PWM operation cycle.

[0107] ΔDer_u(n)=Do_u(n-2)-Dr_u(n)

[0108] ΔDer_v(n)=Do_v(n-2)-Dr_v(n)

[0109] ΔDer_w(n)=Do_w(n-2)-Dr_w(n)…(3)

[0110] For the V phase and the W phase, the duty ratio error correction unit 33 also calculates the on-duty ratio errors ΔDer_v and ΔDer_w based on the actual on-duty ratios Dr_v and Dr_w and the command on-duty ratios Do_v and Do_w.

[0111] <Learning about On-Duty Error>

[0112] The duty cycle error correction unit 33 can reflect the detected on-duty cycle errors ΔDer_u, ΔDer_v, and ΔDer_w of each of the three phases on the voltage command of each of the three phases set next. However, the detected on-duty cycle error ΔDer corresponds to the control state of the PWM operation cycle two cycles ago, so the information is delayed. In addition, in this method, it is necessary to always detect the midpoint voltage of all three phases, detect the actual on-duty cycle Dr, and detect the on-duty cycle error.

[0113] In this embodiment, the duty cycle error correction unit 33 is configured to learn the conduction duty cycle error characteristics, calculate the conduction duty cycle error ΔDerc for correction using the learned error characteristics, and correct the voltage command of each of the three phases using the conduction duty cycle error ΔDerc for correction.

[0114] As described above, the on-duty ratio error ΔDer varies depending on the positive or negative sign of the winding current and the magnitude of the winding current. Figure 6 An example of the on-duty error ΔDer corresponding to the winding current is shown in FIG. As described above, when the winding current is positive, since the current flows through the diode of the switching element on the negative side during the dead time, the actual on-duty Dr is reduced by the on-duty amount corresponding to the dead time compared to the command on-duty Do, and the on-duty error ΔDer is increased by the on-duty amount corresponding to the dead time compared to the command on-duty Do. However, since the turn-off delay and the turn-on delay vary according to the magnitude of the winding current, the on-duty error ΔDer varies according to the operating point of the winding current.

[0115] When the winding current is negative, since the current flows through the diode of the switching element on the positive side during the dead time, the actual on-duty ratio Dr increases by the on-duty ratio corresponding to the dead time compared to the command on-duty ratio Do, and the on-duty ratio error ΔDer decreases by the on-duty ratio corresponding to the dead time compared to the command on-duty ratio Do. However, since the turn-off delay and the turn-on delay vary according to the winding current, the on-duty ratio error ΔDer varies according to the operating point of the winding current.

[0116] When the winding current is 0, the on-duty error ΔDer is 0. Before and after the winding current is 0, the nonlinearity becomes stronger. As the winding current increases from 0, the on-duty error ΔDer increases sharply, and as the winding current decreases from 0, the on-duty error ΔDer decreases rapidly. When the coil current is positive and negative, the on-duty error ΔDer basically becomes a symmetrical characteristic at 0.

[0117] Therefore, the duty cycle error correction unit 33 learns the duty cycle error characteristic indicating the relationship between the winding current and the conduction duty cycle error ΔDer based on the detection value of the winding current and the conduction duty cycle error ΔDer. Furthermore, the duty cycle error correction unit 33 refers to the learned duty cycle error characteristic, calculates the conduction duty cycle error ΔDerc for correction corresponding to the detection value of the winding current, and corrects the voltage command of each of the three phases based on the conduction duty cycle error ΔDerc for correction.

[0118] <Winding current detection value for learning>

[0119] As mentioned above, the current PWM operation cycle ( Figure 4 The actual on-duty ratio Dr calculated in (n) corresponds to the PWM operation cycle ( Figure 4 The end time of the PWM operation cycle two cycles ago calculated in (n-2)) Figure 4 The command on-duty ratio Do set by (a) in the previous PWM operation cycle ( Figure 4 The value varies with the winding current (n-1)).

[0120] Therefore, the duty ratio error correction unit 33 uses the detection value of the winding current corresponding to the timing of detecting the actual conduction duty ratio Dr as the detection value of the winding current for learning. Figure 4 The winding current Iud(n-1) detected in (a) and the start time of this PWM operation cycle ( Figure 4 The winding current Iud(n) detected by (b) in (b) is used to calculate the detection value IL_ud(n) of the winding current for learning, which corresponds to the conduction duty cycle error ΔDer_u(n) calculated in this PWM operation cycle. For example, the duty cycle error correction unit 33 is calculated using the interpolation coefficient α as shown in the following formula. The interpolation coefficient α is set to a value between 0 and 1, for example, set to 0.5. In addition, for the V phase and the W phase, the detection values ​​IL_v and IL_w of the winding current for learning are calculated in the same way.

[0121] IL_u(n)=α×Iud(n-1)+(1-α)×Iud(n)

[0122] IL_v(n)=α×Ivd(n-1)+(1-α)×Ivd(n)

[0123] IL_w(n)=α×Iwd(n-1)+(1-α)×Iwd(n)…(4)

[0124] Alternatively, as shown in the following equation, as the detection values ​​IL_u, IL_v, IL_w of the winding current for learning, the duty cycle error correction unit 33 may use the current values ​​IL_ue, IL_ve, IL_we, which are obtained by performing fixed coordinate conversion and two-phase to three-phase conversion on the dq axis current instructions Ido, Iqo on the dq axis rotating coordinate system based on the magnetic pole position θest corresponding to the timing of detecting the actual conduction duty cycle Dr. In this embodiment, the duty cycle error correction unit 33 estimates the current value using the magnetic pole position θest after 1 to 2 cycles (1.5 cycles in this example) later than the start time of the PWM operation cycle in each PWM operation cycle. Then, the duty cycle error correction unit 33 sets the current values ​​IL_ue(n-2), IL_ve(n-2), IL_we(n-2) estimated in the PWM operation cycle two cycles ago as the detection values ​​IL_u(n), IL_v(n), IL_w(n) of the winding current for learning corresponding to the conduction duty cycle error ΔDer_u(n) calculated in the current PWM operation cycle.

[0125] [Mathematical formula 1]

[0126]

[0127] Here, θ is the magnetic pole position detected at the start of each PWM operation cycle, ω is the rotational angular velocity of the magnetic pole position detected at the start of each PWM operation cycle, and β is a predetermined period coefficient, for example, set to 1.5, and the magnetic pole position θest is estimated to be 1.5 times later than the start of each PWM operation cycle.

[0128] <Learning about Duty Cycle Error Characteristics>

[0129] In this embodiment, the duty ratio error correction unit 33 is configured to use Figure 7 The duty cycle error characteristics when the winding current is positive and the duty cycle error characteristics when the winding current is negative are approximated by a linear function as shown in the following equation, and learning is performed by changing the coefficients A and B of the linear function. Here, A is the slope, B is the intercept, and I is the winding current.

[0130] |ΔDer|=A×|I|+B…(6)

[0131] Since the duty cycle error characteristics when the winding current is positive and the duty cycle error characteristics when the winding current is negative are symmetrical at 0, a linear function is used to collectively learn the positive winding current and the negative winding current, and learn the relationship between the absolute value of the winding current |I| and the absolute value of the conduction duty cycle error |ΔDer|. In addition, since the duty cycle error characteristics between phases are the same, the duty cycle error characteristics of each phase are collectively learned using a linear function.

[0132] For example, the duty cycle error correction unit 33 updates the coefficients A and B of the linear function using the successive least square method with a forgetting coefficient based on the detection value IL of the winding current for learning and the conduction duty cycle error ΔDer calculated in the current PWM operation cycle. For example, the duty cycle error correction unit 33 may update the coefficients A and B of the linear function using the least square method based on the detection value IL of the winding current for learning and the conduction duty cycle error ΔDer calculated in the current and past multiple PWM operation cycles. In addition, a higher-order polynomial may be used to learn the duty cycle error characteristic.

[0133] <Learning is prohibited in non-exportable areas>

[0134] Fig. 9 A diagram showing a non-output area based on dead time. It is shown that by setting the dead time Td, a non-output area is generated for the command on-duty ratio Do during the on-drive period of the switch elements on the positive side and the negative side. In the non-output area, the on-drive period of the switch elements on the positive side and the negative side relative to the command on-duty ratio Do is different from the characteristic that the difference between the command on-duty ratio and the actual on-duty ratio becomes the dead time. Therefore, when the command on-duty ratio Do is in the non-output area, it is best not to learn the duty ratio error characteristic. Therefore, for each phase, the duty ratio error correction unit 33 updates the learning value of the duty ratio error characteristic when the command on-duty ratio Do is within the preset learning allowable range, and does not update the learning value of the duty ratio error characteristic when the command on-duty ratio is outside the learning allowable range.

[0135] <When the winding current is close to 0, learning is prohibited>

[0136] If used Figure 6As described above, the nonlinearity becomes stronger before and after the winding current reaches 0, and cannot be approximated by a linear function. Therefore, as shown in the following formula, the duty cycle error correction unit 33 updates the learning value of the duty cycle error characteristic (the coefficients A and B of the linear function in this example) when the absolute value of the detection value IL of the learning winding current is greater than the preset current threshold THi, and does not update the learning value of the duty cycle error characteristic when the absolute value of the detection value IL of the learning winding current is less than the current threshold THi.

[0137] <Calculation of on-duty error ΔDerc for correction>

[0138] As shown in the following formula, for each phase, the duty cycle error correction unit 33 calculates the conduction duty cycle error ΔDerc for correction corresponding to the detected value of the winding current by referring to the duty cycle error characteristic when the absolute value of the detected value of the winding current is greater than the current threshold value THi. Here, SIGN(X) is a function of the positive and negative signs of the output input value X.

[0139] |Iud|≥THi

[0140] ΔDerc_u=SIGN(Iud)×A×|Iud|+B

[0141] |Ivd|≥THi

[0142] ΔDerc_v=SIGN(Ivd)×A×|Ivd|+B

[0143] ·|Iwd|≥≥THi

[0144] ΔDerc_w=SIGN(Iwd)×A×|Iwd|+B

[0145] When X≥0, SIGN(X)=+1

[0146] When X<0, SIGN(X)=-1…(7)

[0147] On the other hand, as shown in the following formula, for each phase, when the absolute value of the detected value of the winding current is less than the current threshold THi, the duty cycle error correction unit 33 calculates the correction coefficient K0 corresponding to the detected value of the winding current with reference to the correction coefficient characteristic Fcmp which pre-sets the relationship between the winding current and the correction coefficient K0, and calculates the conduction duty cycle error corresponding to the detected value of the winding current with reference to the duty cycle error characteristic, multiplies the correction coefficient K0 by the conduction duty cycle error, and thus calculates the final conduction duty cycle error ΔDerc.

[0148] ·|Iud|<THi

[0149] K0_u=Fcmp(|Iud|)

[0150] ΔDerc_u=SGN(Iud)×K0_u×(A×|Iud|+B)

[0151] ·|IvdI<THi,

[0152] K0_v=Fcmp(|Ivd|)

[0153] ΔDerc_v=SIGN(Ivd)×K0_v×(A×|Ivd|+B)

[0154] ·|Iwd|<THi,

[0155] K0_w=Fcmp(|Iwd|)

[0156] ΔDerc_w=SIGN(Ivd)×K0_w×(A×|Iwd|+B)…(8)

[0157] Here, Fcmp() is a function of the correction coefficient characteristic that takes the absolute value of the winding current as an input variable, such as Figure 8 As shown, when the absolute value of the winding current is the current threshold value THi, 1 is output, and when the absolute value of the winding current is 0, 0 is output, and interpolation is performed between the two points. In this embodiment, the correction coefficient characteristic Fcmp is set to a linear function with a slope of 1 / THi and an intercept of 0. In addition, the correction coefficient characteristic can be a high-order function or table data.

[0158] In addition, as the detection values ​​Iud, Ivd, and Iwd of the winding current of each phase used in equations (7) and (8), the estimated value of the winding current at the timing of actually turning on and off the switch element based on the corrected voltage command can also be used. Specifically, as shown in equations 1 and 2 of equation (5), as the detection values ​​Iud, Ivd, and Iwd of the winding current of each phase, the duty cycle error correction unit 33 can use the current values ​​IL_ue, IL_ve, and IL_we, which are obtained by performing fixed coordinate transformation and two-phase to three-phase transformation on the dq axis current instructions Ido and Iqo on the rotating coordinate system of the dq axis based on the magnetic pole position θest. The magnetic pole position θest is set at the magnetic pole position corresponding to the timing of actually turning on and off the switch element based on the corrected voltage command, and β is set to a value between 1 and 2 (1.5 in this example). The duty ratio error correction unit 33 uses the current values ​​IL_ue, IL_ve, IL_we estimated in the current PWM operation cycle as the detection values ​​Iud, Ivd, Iwd of the winding currents of the respective phases.

[0159] The duty cycle error correction unit 33 calculates the voltage commands Vuo, Vvo, Vwo of the three phases as shown in the following equation by adding the value obtained by multiplying the on-duty cycle errors ΔDerc_u, ΔDerc_v, ΔDerc_w for correction of the three phases by the power supply voltage Vdc to the basic voltage commands Vuob, Vvob, Vwob of the three phases. Then, the PWM control unit 34 performs on-off control of the switching elements based on the corrected voltage commands Vuo, Vvo, Vwo of the three phases.

[0160]

[0161] 2. Implementation Method 2

[0162] Next, the control device 1 involved in the second embodiment is described. The description of the same structural parts as those in the first embodiment is omitted. The basic structure of the control device 1 involved in this embodiment is the same as that in the first embodiment, but the difference from the first embodiment is that the PWM control unit 34 generates a PWM signal based on the comparison result of the command on-duty Do calculated from the command voltage of each of the three phases and the carrier, and the duty error correction unit 33 corrects the command on-duty Do based on the on-duty error.

[0163] Fig.10 The block diagram of the control device 1 involved in this embodiment is shown. As shown in the following formula, the basic voltage command calculation unit 32 calculates the basic command conduction duty ratio Dob_u, Dob_v, Dob_w of each of the three phases by adding 0.5 to the value obtained by dividing the basic voltage commands Vuob, Vvob, Vwob of each of the three phases calculated in the same way as in the above-mentioned embodiment 1 by the power supply voltage Vdc.

[0164]

[0165] The duty ratio error correction unit 33 calculates the on-duty ratio errors ΔDerc_u, ΔDerc_v, and ΔDerc_w for correction of each of the three phases by the same method as in Embodiment 1. The duty ratio error correction unit 33 adds the on-duty ratios ΔDerc_u, ΔDerc_v, and ΔDerc_w for correction of each of the three phases to the basic command on-duty ratios Dob_u, Dob_v, and Dob_w of each of the three phases, as shown in the following equation, to calculate the command on-duty ratios D_u, Do_v, and Do_w of each of the three phases.

[0166]

[0167] The PWM control unit 34 performs on-off control of each switching element based on the corrected command on-duty ratios Do_u, Do_v, and Do_w of each of the three phases. The PWM control unit 34 generates a PWM signal for turning each switching element on and off based on the comparison result between the command on-duty ratios Do_u, Do_v, and Do_w of each of the three phases and the carrier CA. In the present embodiment, the carrier CA is set to a triangular wave that oscillates between 0 and 1 at a PWM period Tcc. When the command on-duty ratio Do exceeds the carrier CA, the PWM signal is turned on, and when the command on-duty ratio Do is lower than the triangular wave, the PWM signal is turned off. As in the first embodiment, for each phase, the PWM control unit 34 alternately turns on and drives the switching element on the positive side and the switching element on the negative side with a dead time.

[0168] 3. Implementation Method 3

[0169] Next, the control device 1 involved in the third embodiment is described. The description of the same structural parts as those in the first embodiment is omitted. The basic structure of the control device 1 involved in this embodiment is the same as that in the first embodiment, but the difference from the first embodiment is that the duty ratio error correction unit 33 learns the duty ratio error characteristics independently when the winding current is positive and when it is negative.

[0170] The driving signal of the gate driving circuit 12 is connected to the gate terminal of the switching element via the redundant three-phase short-circuit circuit 19. Therefore, the cut-off delay and the conduction delay caused by the intervention of the redundant three-phase short-circuit circuit 19 are generated. In particular, when the redundant three-phase short-circuit circuit 19 forcibly drives the driving signal of the switching element on the positive side of the three phases, or when the driving signal of the switching element on the negative side of the three phases is forcibly driven, the cut-off delay and the conduction delay are different between the switching element intervened by the redundant three-phase short-circuit circuit 19 and the switching element not intervened by the redundant three-phase short-circuit circuit 19. In addition, regardless of whether the redundant three-phase short-circuit circuit 19 exists, the cut-off delay and the conduction delay of the switching element on the positive side and the switching element on the negative side are sometimes different depending on the characteristic difference. Therefore, in the case where the winding current is positive and the winding current is negative, since the duty cycle error characteristic changes, the correction accuracy can be improved by independent learning.

[0171] In this embodiment, as shown in formula (12), the duty ratio error correction unit 33 learns the duty ratio error characteristics for positive use when the winding current I is positive, and learns the duty ratio error characteristics for negative use when the winding current I is negative. The duty ratio error correction unit 33 performs learning by changing the coefficients Ap and Bp of the linear function for positive use and the coefficients An and Bn of the linear function for negative use.

[0172] When I>0,

[0173] |ΔDer|=Ap×|I|+Bp

[0174] When I<0,

[0175] |ΔDer|=An×|I|+Bn

[0176] For example, when the detection value IL of the winding current for learning is positive, the duty cycle error correction unit 33 uses the successive least squares method with a forgetting coefficient to update the coefficients Ap and Bp of the positive linear function based on the detection value IL of the winding current for learning and the conduction duty cycle error ΔDer. When the detection value IL of the winding current for learning is negative, the duty cycle error correction unit 33 uses the successive least squares method with a forgetting coefficient to update the coefficients An and Bn of the negative linear function based on the detection value IL of the winding current for learning and the conduction duty cycle error ΔDer.

[0177] As shown in the following formula, for each phase, when the absolute value of the detected value of the winding current is greater than the current threshold THi, the duty cycle error correction unit 33 refers to the positive or negative duty cycle error characteristics according to the positive or negative sign of the detected value of the winding current to calculate the correction duty cycle error ΔDerc corresponding to the detected value of the winding current.

[0178] When |Iud|≥THi,

[0179] When Iud≥0, ΔDerc_u=Ap×|Iud|+Bp

[0180] When Iud<0, ΔDerc_u=-(An×|Iud|+Bn)

[0181] When |Ivd|≥THi,

[0182] When Ivd≥0, ΔDerc_v=Ap×|Ivd|+Bp

[0183] When Ivd<0, ΔDerc_v=-(An×|Ivd|+Bn)

[0184] ·|Iwd|≥THi,

[0185] When Iwd≥0, ΔDerc__w=Ap×|Iwd|+Bp

[0186] When Iwd<0, ΔDerc_w=-(An×|Iwd|+Bn)…(13)

[0187] On the other hand, as shown in the following formula, for each phase, when the absolute value of the detected value of the winding current is less than the current threshold THi, the duty cycle error correction unit 33 calculates the correction coefficient K0 corresponding to the detected value of the winding current with reference to the correction coefficient characteristic Fcmp which pre-sets the relationship between the winding current and the correction coefficient K0, and calculates the conduction duty cycle error corresponding to the detected value of the winding current with reference to the positive or negative duty cycle error characteristic according to the positive or negative of the detected value of the winding current, and multiplies the correction coefficient K0 by the conduction duty cycle error to calculate the final conduction duty cycle error ΔDerc.

[0188] When |Iud|<THi,

[0189] K0_u=Fcmp(|Iud|)

[0190] When Iud≥0,

[0191] ΔDerc_u=K0_u×Ap×|Iud|+Bp

[0192] When Iud<0,

[0193] ΔDerc_u=-K0_u×(An×|Iud|+Bn)

[0194] When |Ivd|<THi,

[0195] K0_v=Fcmp(|Ivd|)

[0196] When Ivd≥0,

[0197] ΔDerc_v=K0_v×Ap×|Ivd|+Bp

[0198] When Ivd<0,

[0199] ΔDerc_v=-K0_v×(An×|Ivd|+Bn)

[0200] ·|Iwd|<THi,

[0201] K0_w=Fcmp(|Iwd|)

[0202] When Iwd≥0,

[0203] ΔDerc_w=K0_w×Ap×|Iwd|+Bp

[0204] When Iwd<0,

[0205] ΔDercw=-K0_w×(An×|Iwd|+Bn)…(14)

[0206] 4. Implementation Method 4

[0207] Next, the control device 1 according to Embodiment 4 is described. The description of the same components as those in Embodiment 1 is omitted. The basic structure of the control device 1 according to this embodiment is the same as that of Embodiment 3, but the difference from Embodiment 1 is that the duty ratio error correction unit 33 divides time into a plurality of periods and obtains learning data of the phase set in each period.

[0208] In this embodiment, the duty cycle error correction unit 33 divides time into multiple periods, sets a phase in which the winding current is positive and a phase in which the winding current is negative in each period, and for the two phases set in each period, detects the actual on-duty cycle Dr based on the detection value of the midpoint potential, calculates the on-duty cycle error ΔDer based on the difference between the command on-duty cycle Do and the actual on-duty cycle Dr, and updates the learning value of the duty cycle error characteristic based on the detection value of the winding current and the on-duty cycle error ΔDer.

[0209] In this embodiment, if Fig.11 and Fig.12 As shown, the time for the magnetic pole position to rotate 720deg is divided into 12 periods of 60deg each, and in each period, a phase in which the detection value IL of the learning winding current is positive and a phase in which the detection value IL of the learning winding current is negative are set. In addition, the phase of each period is set so that there is no deviation in the number of periods between the phases. In this embodiment, when the detection values ​​IL_u and IL_ue of the learning winding current of the U phase become 0, the period is divided based on the adjusted magnetic pole position θestη (=θest+η+π) [rad] obtained after adjusting the offset value η, so that the magnetic pole position becomes 0. The offset value η is calculated based on the dq axis current instructions Ido and Iqo using the following formula. In addition, for the sake of convenience of explanation, [rad] and [deg] are used, but the units of the two are only converted.

[0210] [Mathematical formula 2]

[0211]

[0212] According to this configuration, the number of phases for which learning data is calculated at the same time can be reduced to two phases, one phase for which the winding current is positive and one phase for which the winding current is negative, thereby reducing the computational load. In addition, since the learning data for the phase for which the winding current is positive and the phase for which the winding current is negative can be calculated at the same time, the learning value of the positive duty cycle error characteristic and the learning value of the negative duty cycle error characteristic can be updated without deviation.

[0213] [Other embodiments]

[0214] Finally, other embodiments of the present disclosure will be described. In addition, the structures of the embodiments described below are not limited to being applied individually, and can be applied in combination with the structures of other embodiments as long as no contradiction occurs.

[0215] (1) In the above-mentioned embodiments, the three-phase case where n=3 is described as an example. However, n may be set to any natural number greater than 2, such as n=2 or 4.

[0216] (2) In the above-mentioned embodiments, the case where the duty cycle error correction unit 33 learns the common duty cycle error characteristics between phases is described as an example. However, the duty cycle error correction unit 33 may be configured to learn the duty cycle error characteristics of each phase separately. According to this structure, the individual differences of the switching elements between the phases can be learned.

[0217] (3) In the above-mentioned embodiments, the case where the midpoint potential detection circuit 18 detects the midpoint voltage of all the series circuits of the three phases, and the duty ratio error correction unit 33 detects the actual conduction duty ratio of each of the three phases and calculates the learning data is described as an example. However, the midpoint potential detection circuit 18 may be configured as a circuit that detects the midpoint voltage of only one phase of the series circuit that can detect the midpoint voltage, and the duty ratio error correction unit 33 detects the actual conduction duty ratio of only one phase that can detect the midpoint voltage, calculates the learning data, learns the duty ratio error characteristics based on the learning data of only one phase that can detect the midpoint voltage, and uses the learned duty ratio error characteristics to calculate the conduction duty ratio error ΔDerc for correction of each of the three phases. According to this structure, when the individual differences of the switching elements between the phases are small, the circuit structure of the midpoint potential detection circuit 18 can be further simplified, and the calculation burden of the learning data can be reduced.

[0218] Although the present application describes various exemplary embodiments and examples, the various features, methods and functions described in one or more embodiments are not limited to the application of specific embodiments, and can be applied to the embodiments individually or in various combinations. Therefore, it can be considered that countless variations that are not illustrated are also included in the technical scope disclosed in the present application specification. For example, it is set to include the case where at least one constituent element is deformed, added or omitted, and the case where at least one constituent element is extracted and combined with constituent elements of other embodiments.

[0219] Description of symbols

[0220] 1 AC rotating motor control device

[0221] 2 AC rotating motor

[0222] 10 DC power supply

[0223] 17 Current detection circuit

[0224] 18 Midpoint potential detection circuit

[0225] 19 Redundant three-phase short circuit

[0226] 20 Inverter

[0227] 22 Diode

[0228] 23H Switching element on the positive side

[0229] 23L Switching element on the negative side

[0230] 30 Control circuit

[0231] 31 Operation status detection unit

[0232] 32 Basic voltage command calculation unit

[0233] 33 Duty cycle error correction unit

[0234] 34 PWM control unit

[0235] Do instruction conduction duty cycle

[0236] Dr Actual on-duty cycle

[0237] Fcmp correction factor characteristics

[0238] IL Detection value of winding current

[0239] K0 correction factor

[0240] THi Current Threshold

[0241] Δder is the on-duty cycle error.

Claims

1. An AC rotating motor control device, The AC rotating electric machine control device controls an AC rotating electric machine having n-phase windings, wherein: n is a natural number greater than 2, and the characteristics include: An inverter, wherein n sets of series circuits are provided corresponding to each of the n phases, wherein the series circuits connect in series a switching element on a positive side connected to a positive side of a DC power source and a switching element on a negative side connected to a negative side of the DC power source, and a connection point of the series connection is connected to the winding of the corresponding phase; a midpoint potential detection circuit for detecting a potential of a connection point of the series connection in the series circuit, that is, a midpoint potential; and a control circuit for controlling the switching element on and off based on a voltage command or a command on-duty ratio calculated from the voltage command, The control circuit Based on the detected value of the midpoint potential, the actual on-duty ratio of the switching element is detected, Calculate the on-duty ratio error according to the difference between the command on-duty ratio and the actual on-duty ratio, Based on the on-duty ratio error, the voltage command or the command on-duty ratio is corrected, It also includes a current detection circuit, which detects the winding current flowing through the winding of each phase. The control circuit learns a duty ratio error characteristic indicating a relationship between the winding current and the conduction duty ratio error based on the detected value of the winding current and the conduction duty ratio error, and referring to the learned duty cycle error characteristic, calculating the conduction duty cycle error corresponding to the detection value of the winding current, and correcting the voltage command or the command conduction duty cycle based on the calculated conduction duty cycle error, The control circuit approximates the duty ratio error characteristic when the winding current is positive and the duty ratio error characteristic when the winding current is negative using a linear function, and changes the coefficient of the linear function to perform learning.

2. The AC rotating electrical machine control device according to claim 1, characterized in that: The control circuit updates an intercept and a slope of coefficients of the linear function using a successive least square method based on the detection value of the winding current and the on-duty ratio error.

3. The AC rotating electrical machine control device according to claim 1 or 2, characterized in that: The control circuit uses the detection value of the winding current corresponding to the timing at which the actual on-duty ratio is detected when learning the duty ratio error characteristic.

4. The AC rotating electrical machine control device according to claim 1 or 2, characterized in that: The control circuit uses the current value as the detection value of the winding current for learning the duty cycle error characteristic, and the current value is obtained by performing two-phase to three-phase conversion on the dq-axis current instruction on the rotating coordinate system of the dq axis based on the magnetic pole position corresponding to the timing of detecting the actual conduction duty cycle.

5. The AC rotating electrical machine control device according to claim 1 or 2, characterized in that: The control circuit updates the learning value of the duty cycle error characteristic when the absolute value of the detection value of the winding current is greater than a preset current threshold, and does not update the learning value of the duty cycle error characteristic when the absolute value of the detection value of the winding current is less than the current threshold.

6. The AC rotating electrical machine control device according to claim 5, characterized in that: For each phase, when the absolute value of the detected value of the winding current is greater than the current threshold, the control circuit refers to the duty cycle error characteristic to calculate the duty cycle error corresponding to the detected value of the winding current. For each phase, when the absolute value of the detected value of the winding current is less than the current threshold, the control circuit calculates the correction coefficient corresponding to the detected value of the winding current with reference to the correction coefficient characteristic that pre-sets the relationship between the winding current and the correction coefficient, and calculates the on-duty cycle error corresponding to the detected value of the winding current with reference to the duty cycle error characteristic, multiplies the on-duty cycle error by the correction coefficient, and thus calculates the final on-duty cycle error.

7. The AC rotating electrical machine control device according to claim 1 or 2, characterized in that: The control circuit uses the current value as the detection value of the winding current used when calculating the conduction duty cycle error with reference to the duty cycle error characteristic. The current value is obtained by performing two-phase to three-phase conversion on the dq-axis current command on the rotating coordinate system of the dq axes based on the magnetic pole position. The magnetic pole position corresponds to the timing of actually turning on and off the switching element based on the corrected voltage command or the command conduction duty cycle.

8. The AC rotating electrical machine control device according to claim 1 or 2, characterized in that: The control circuit updates the learning value of the duty cycle error characteristic when the command on-duty cycle is within a preset learning allowable range, and does not update the learning value of the duty cycle error characteristic when the command on-duty cycle is outside the learning allowable range.

9. The AC rotating electrical machine control device according to claim 1 or 2, characterized in that: The control circuit divides time into multiple periods, sets a phase in which the winding current is positive and a phase in which the winding current is negative in each period, detects the actual on-duty cycle based on the detection value of the midpoint potential for the two phases set in each period, calculates the on-duty cycle error based on the difference between the command on-duty cycle and the actual on-duty cycle, and updates the learning value of the duty cycle error characteristic based on the detection value of the winding current and the on-duty cycle error.

10. The AC rotating electrical machine control device according to claim 1 or 2, characterized in that: The midpoint potential detection circuit is a circuit that turns on or off an output signal according to whether the midpoint potential is greater than or less than a potential threshold.

11. The AC rotating electrical machine control device according to claim 1 or 2, characterized in that: The control circuit detects the actual on-duty ratio based on a time when the midpoint potential is greater than a potential threshold and a time when the midpoint potential is less than the potential threshold.

12. The AC rotating electrical machine control device according to claim 1 or 2, characterized in that: For each phase, the control circuit drives the switching element on the positive electrode side and the switching element on the negative electrode side alternately with a dead time in between. For each phase, the control circuit shortens the dead time during the on-drive period of the switching element on the positive side compared to the on-drive period corresponding to the command on-duty cycle, and extends the dead time during the off-drive period of the switching element on the negative side compared to the on-drive period corresponding to the command on-duty cycle.

13. The AC rotating electrical machine control device according to claim 12, characterized in that: For each phase, the control circuit generates a PWM signal that is turned on and off using the command on-duty ratio, and when the PWM signal becomes on, the switching element on the negative side is driven to be off, and after the PWM signal becomes on, when the dead time has passed, the switching element on the positive side is driven to be on, and when the PWM signal becomes off, the switching element on the positive side is driven to be off, and after the PWM signal is off, when the dead time has passed, the switching element on the negative side is driven to be on.

Images